r

El ia s M. A w a d & Ha ssa n Gh a z ir i

ALWAYS LEA R N IN G PEARSON

KNOWLEDGE
MANAGEMENT
This page is intentionally left blank.
 KNOWLEDGE
MANAGEMENT
Elias M. Awad
Virginia Bankers Association Professor o f Bank Management
University o f Virginia

Hassan M. Ghaziri
Associate Professor
School o f Business
American University o f Beirut

PEARSON
Credits borrowed from other sources and reproduced, with permission, in this textbook
appear on appropriate page within text.

Copyright © 2007 Dorling Kindersley (India) Pvt. Ltd

Licensees of Pearson Education in South Asia

No part of this eBook may be used or reproduced in any manner whatsoever without the
publisher’s prior written consent.

This eBook may or may not include all assets that were part of the print version. The publisher
reserves the right to remove any material present in this eBook at any time.

ISBN 9788131714034
eISBN 9789332506190

Head Office: A-8(A), Sector 62, Knowledge Boulevard, 7th Floor, NOIDA 201 309, India
Registered Office: 11 Local Shopping Centre, Panchsheel Park, New Delhi 110 017, India
“The great end of life is not know ledge but action.”
—Thomas H enry Huxley

This text is dedicated to the m em ory of

Firefighter D ana H annon and his 342 FD N Y brothers
who died in the line of duty,
W orld Trade C enter,
Septem ber 11,2001.
—Elias Awad

To my wife Blandine,
who showed me that seeking the truth entails
sharpening the taste of beauty, and
who constantly rem inds me that technology
w ithout hum anity leads to chaos.
—H assan G haziri
This page is intentionally left blank.
 I I I I I I I I I I I I I I

Contents
Preface 17
About the Authors 23

PART I: THE BASICS 25

CHAPTER 1: Working Smarter, Not Harder 25
In a N utshell 25
W hat Is K now ledge M anagem ent? 26
The Knowledge Organization 27
W hy K now ledge M anagem ent? 34
H ow It Cam e A bout 39
KM M yths 45
KM Life Cycle 48
Im plications for K now ledge M anagem ent 50
Sum m ary 51
Terms to Know 52
Test Your U nderstanding 52
K now ledge Exercises 53
R eferences 54

CHAPTER 2: Understanding Knowledge 55

In a N utshell 56
D efinitions 57
Knowledge 57
Intelligence 57
Experience 58
Common Sense 58
C ognition and Knowledge M anagem ent 59
D ata, Inform ation, and K nowledge 60
Data 60
Information 60
Knowledge 61
Types of Knowledge 66
Shallow and Deep Knowledge 66
Knowledge as Know-How 66

7
Contents

Reasoning and Heuristics 66

Common Sense as Knowledge 68
From Procedural to Episodic Knowledge 68
Explicit and Tacit Knowledge 71
E xpert Knowledge 72
Chunking Knowledge 72
Knowledge as an Attribute o f Expertise 72
H um an Thinking and Learning 75
Human Learning 15
Learning by Experience 75
Learning by Example 76
Learning by Discovery 76
Im plications for Knowledge M anagem ent 76
Sum m ary 77
Terms to Know 79
Test Your U nderstanding 79
Knowledge Exercises 80
R eferences 80

CHAPTER 3: Knowledge Management Systems Life

Cycle 82
In a Nutshell 83
C hallenges in Building KM Systems 84
C onventional Versus KM System Life Cycle 86
Knowledge M anagem ent System Life Cycle
(KM SLC) 89
System Justification 92
Role o f Rapid Prototyping 101
Selecting an Expert 101
Role o f the Knowledge Developer 102
User Training 106
Im plications for Knowledge M anagem ent 109
Sum m ary 110
Terms to Know 111
Test Your U nderstanding 111
Knowledge Exercises 112
R eferences 113

PART II: KNOWLEDGE CREATION AND CAPTURE

CHAPTER 4: Knowledge Creation and Knowledge
Architecture 115
In a Nutshell 116
Knowledge C reation 116
N o n ak a’s M odel of Knowledge C reation and
Transform ation 119
 Contents i a i i

Know ledge A rchitecture 121

The People Core 121
Identifying Knowledge Centers 122
The Technical Core 124
Build In-Home, Buy, or Outsource 135
Im plications for Knowledge M anagem ent 138
Sum m ary 139
Terms to Know 140
Test Your U nderstanding 141
K now ledge Exercises 142
R eferences 143

CHAPTER 5: Capturing Tacit Knowledge 144

In a N utshell 145
W hat Is Know ledge C apture? 146
E valuating the E xpert 150
Levels o f Expertise 152
Capturing Single Versus Multiple Experts' Tacit
Knowledge 152
Advantages and Drawbacks o f Using a Single Expert
Pros and Cons o f Using Multiple Experts 154
D eveloping a R elationship with Experts 156
Creating the Right Impression 156
Understanding the Expert's Style 157
Preparing fo r the Session 157
Deciding Where to Hold the Sessions 158
Approaching Multiple Experts 158
Fuzzy R easoning and the Q uality of Knowledge
C apture 159
Analogies and Uncertainties in Information 160
Understanding Experience 160
The Language Problem 161
The Interview As a Tool 162
Types o f Interviews 163
G uide to a Successful Interview 165
Setting the Stage and Establishing Rapport 165
Phrasing the Questions 166
Question Construction 166
Thing to Avoid 167
Reliability o f Information from Experts 168
Errors Made by the Knowledge Developer 169
Problems Encountered During the Interview 170
Ending the Interview 171
Issues to Assess 172
R apid Prototyping in Interviews 173
Benefits and Drawbacks 173
Im plications for Knowledge M anagem ent 174
Contents

Summary 175
Terms to Know 176
Test Your Understanding 176
Knowledge Exercises 177
References 179

CHAPTER 6: Other Knowledge Capture

Techniques 180
In a Nutshell 181
Overview 181
On-Site Observation 182
Brainstorming 183
Electronic Brainstorming 184
Protocol Analysis 185
What Is a Scenario? 188
Protocol Procedure o f the Diabetic Foot KM System
Consensus Decision Making 190
The Repertory Grid 191
Nominal-Group Technique (N G T) 193
The Delphi Method 194
Concept Mapping 195
Procedure 196
Concept Mapping and Semantic Nets 197
Blackboarding 198
Implications for Knowledge Management 200
Summary 200
Terms to Know 202
Test Your Understanding 202
Knowledge Exercises 203
References 205

PART III: KNOWLEDGE CODIFICATION AND SYSTEM

IMPLEMENTATION 209
CHAPTER 7: Knowledge Codification 209
In a Nutshell 210
What Is Knowledge Codification? 210
Why Codify? 212
Diagnosis 212
Instruction/Training 212
Interpretation 213
Planning/Scheduling 2 13
Prediction 2 13
Things to Remember 2 13
 Contents i i i i

M odes of Knowledge Conversion 214

How to Codify Knowledge 215
Codifying Tacit Knowledge 216
Codification Tools and Procedures 217
Knowledge Maps 217
Decision Tables 221
Decision Trees 221
Frames 223
Production Rules 224
Case-Based Reasoning 228
Knowledge-Based Agents 230
The Knowledge D eveloper’s Skill Set 231
Knowledge Requirements 231
Skills Requirements 232
Implications for Knowledge Management 234
Summary 235
Terms to Know 236
Test Your Understanding 237
Knowledge Exercises 237
References 239

CHAPTER 8: System Testing and Deployment

In a Nutshell 241
Quality and Quality Assurance 242
Knowledge Testing 243
Key Definitions 243
Issues to Consider 244
Approaches to Logical Testing 246
Circular Errors 246
Redundancy Errors 248
Unusable Knowledge 248
Subsumption Errors 248
Inconsistent Knowledge 249
Approaches to User Acceptance Testing 249
Test Team and Plan 249
User A cceptance Test Criteria 251
User Acceptance Test Techniques 251
Managing the Testing Phase 253
KM System Deploym ent 254
Issues Related to D eploym ent 254
Selection o f the Knowledge Base Problem 254
Ease o f Understanding the K M System 255
Knowledge Transfer 256
Integration Alternatives 256
The Issue o f Maintenance 25 8
Organizational Factors 259
12 7 7 7 7 Contents
Other Factors 260
Role o f the Champion 260
U ser Training and D eploym ent 261
Preparing fo r K M System Training 261
Combating Resistance to Change 262
Postim plem entation Review 263
Security Considerations 264
Im plications for K now ledge M anagem ent 264
Sum m ary 265
Terms to Know 267
Test Your U nderstanding 267
Knowledge Exercises 268
R eferences 270

CHAPTER 9: Knowledge Transfer and Knowledge

Sharing 272
In a Nutshell 273
Knowledge Transfer As a Step in a Process 273
The Knowing-Doing Gap 276
Prerequisites fo r Transfer 277
Transfer M ethods 284
Nature o f the Problem 284
Transfer Strategies 285
Inhibitors o f Knowledge Transfer 287
How Knowledge Is Transferred 288
R ole of the Internet in K nowledge Transfer 293
Internet Service Providers 294
Stability and Reliability o f the Web 295
Unique Benefits o f the Internet 296
Limitations 297
Im plications for K nowledge M anagem ent 298
Sum m ary 298
Terms to Know 300
Test Your U nderstanding 300
K nowledge Exercises 300
R eferences 301

CHAPTER 10: Knowledge Transfer in

the E-World 302
In a N utshell 302
The E-W orld 303
Intranets 303
Extranets and Knowledge Exchange 306
Groupware 308
Groupware Applications 312
 Contents a i a a 13

E-Business 313
The Value Chain 315
Supply Chain Management (SCM) and Knowledge
Exchange 315
Implications for Knowledge Management 322
Summary 324
Terms to Know 325
Test Your Understanding 326
Knowledge Exercises 326
References 327

PART IV: KM SYSTEM TOOLS AND PORTALS 329

CHAPTER 11: Learning from Data 329
In a Nutshell 330
The “Learning” Concept 330
Data Visualization 331
Neural Networks As a Learning Model 331
The Basics 334
Supervised and Unsupervised Learning 335
Business Applications 336
Relative Fit with Knowledge Management 337
Association Rules 338
Market Basket Analysis: PETC O 340
Classification Trees 341
Definition 341
Decision fo r Granting a Loan 342
Tree Construction 342
Implications for Knowledge Management 342
Summary 343
Terms to Know 344
Test Your Understanding 345
Knowledge Exercises 345
References 346
Web Sites 346

CHAPTER 12: Data Mining—Knowing the

Unknown 348
In a Nutshell 349
What Is Data Mining? 349
Definitions 351
Data Mining and Business Intelligence 351
Business Drivers 352
Technical Drivers 354
 Role o f Statistics 354
Machine Learning 354
Data Warehouses 354
OLAP 355
Evolution o f the Decision-Making Architecture 356
D M V irtuous Cycle 356
Business Understanding 357
Develop the D M Application 358
D ata M anagem ent 360
Data Sources 361
Taxonomy o f Data 361
Data Preparation 361
Model Building 362
Parameter Settings and Tuning 362
Model Testing and Analysis o f Results 362
Taking Action'and Deployment 363
Postdeployment Phase 363
DM in Practice 363
Role of DM in C ustom er R elationship
M anagem ent 367
Customer Acquisition 367
Campaign Optimization 368
Customer Scoring 368
Direct Marketing 368
Integrating DM , CRM , and E-Business 368
Implication for Knowledge M anagem ent 369
Sum m ary 372
Terms to Know 372
Test Your U nderstanding 373
Knowledge Exercises 374
R eferences 376
Selected W eb Sites 377

CHAPTER 13: Knowledge Management Tools and

Knowledge Portals 379
In a Nutshell 380
Portals:The Basics 380
What Is a Portal? 380
Evolution o f Portals 383
Key Characteristics 384
Illustration 385
The Business Challenge 386
Portals and the Business Transformation 387
Market Potential 389
K nowledge Portal Technologies 392
Key Functionality 392
 Contents i a a a 15
Collaboration 392
Content Management 394
Collaboration versus Categorization: The Case o f the World
Bank 396
Intelligent Agents 398
Implications for Knowledge Management 399
Who Is Building Enterprise Portals? 399
Who Sponsors Enterprise Portals? 399
Implementation Issues 399
Bandwidth 400
Portal Product Selection 401
Summary 402
Terms to Know 403
Test Your Understanding 404
Knowledge Exercises 404
References 404

PART V: ETHICAL, LEGAL, AND MANAGERIAL ISSUES 409

CHAPTER 14: Who Owns Knowledge? Ethical
and Legal Issues 409
In a Nutshell 410
Knowledge Owners 411
Knowledge fo r Sale 4 11
Releasing Knowledge Gained on the Job 411
Becoming an Expert via a Corporate Knowledge-Based
System 413
Legal Issues 414
The Liability Question 4 14
The Basis o f Liability 415
Copyrights, Trademarks, and Trade Names 4 16
Warranties 4 18
Strict Liability 420
Legal Disputes in Knowledge Management 420
Web Linking and Domain Name Disputes 421
The Malpractice Factor 423
The Ethics Factor 424
What Is Ethics? 424
Ethical Decision Cycle 427
Major Threats to Ethics 421
Improving the Climate 430
Where to Start? 430
Code o f Ethics 431
The Privacy Factor 431
Implications for Knowledge Management 432
Summary 433
Terms to Know 434
16 i ■i . Contents

Test Your U nderstanding 434

Knowledge Exercises 435
R eferences 436

CHAPTER 15: Managing Knowledge Workers 437

In a N utshell 438
W hat Is a Knowledge W orker? 438
Personality and Professional Attributes 439
Business Roles in the L earning O rganization 441
Management and Leadership 442
Work Management Tasks 443
Work A djustm ent and the K nowledge W orker 444
Profile o f Vocational Needs and Reinforcers
o f Knowledge Workers 446
Smart Leadership Requirements 446
Technology and the Knowledge W orker 447
Knowledge Worker's Skills 450
Role of Ergonom ics 450
Role of the CKO 452
M anagerial C onsiderations 455
Becoming a Change Leader 455
M anaging Knowledge Projects 456
The Soft Side Always Wins 456
Im plications for Knowledge M anagem ent 460
Sum mary 461
Terms to Know 462
Test Your U nderstanding 463
Knowledge Exercises 463
R eferences 463

EPILOGUE What More Do We Need

to Know? 465
The Shareability Factor 466
The H um an E lem ent 467
There Is M ore to Know 468
The Social Factor 469
O ne Final Note 470
R eferences 471

Index 473
 ■

Preface
We live in a world that changes by the minute. Change often moves organizations and
advances p eo p le’s intellect. For change to be effective, organizations as well as people
m ust change. H ow ever, even the most intelligent individuals can becom e handicapped
in ineffective organizations. People, being creatures of habit, are so busy working the
job that they fail to see change around them. They get carried away with behavior that
did well in the past, not realizing that it no longer produces effective results in a fast-
changing business.
The key to change and growth is awareness, sharing ideas, and coming up with new
and innovative ways of staying ahead of the com petition. It involves learning, innovat-
ing, and adopting behavior designed to improve quality and perform ance. As M cLuhan
said, “ E verybody ex periences far m ore than he understands. Yet, it is experience,
rath er than understanding, that influences behavior.”
So, sm art people need sm art organizations. The nam e of the game is integration
and co o p eratio n for com petitive advantage. For com panies to m ake use of hum an
ex p erience and intelligence, they m ust provide a sharing environm ent, em pow er
people with tools, and create a climate for learning and testing new ways of doing busi-
ness. As Philip Kotler, professor of m arketing. N orthw estern University, wrote, “Every
com pany should work hard to obsolete its own line .. . before its com petitors do.”

i « « « Why This fiook?

K now ledge m anagem ent is grow ing rapidly. M ore and m ore com panies have built
knowledge repositories, supporting such knowledge varieties as custom er knowledge,
product developm ent knowledge, custom er service knowledge, hum an resource m an-
agem ent know ledge, and the like. Even new job titles have appeared —from know l-
edge developer, to knowledge facilitator, to corporate knowledge officer. O ther signs
of the growing knowledge m anagem ent em phasis are in the many newspapers, m aga-
zines, journals, conferences, and sem inars that appear daily worldwide. Even vendors
on the Internet and in established locations tout specialized knowledge m anagem ent
applications of all kinds and aim ed at all levels of the organization. By all accounts,
knowledge m anagem ent is becom ing a necessary feature of today’s business culture.
In our opinion, people change the world; books pave the way. This text is about
knowledge, how to capture it, how to transfer it, how to share it, and how to m anage it.
In this volume, we introduce the many facets of knowledge m anagem ent—from con-
cepts, to people, to tools, to procedures. In term s of tools, we cover the use of the com -
puter for capturing and sharing tacit knowledge and the netw ork that transfers tacit

17
18 a a i i Preface

and explicit know ledge. C on sid er the follow ing questions regarding one aspect of
knowledge and com puters:
• D o you know a m ore effective way to preserve the knowledge of an expert who
might be retiring soon than by preserving it through a com puter?
• How could tod ay ’s corporations afford brain drain and loss of knowledge and
expertise w ithout a way to capture it so that less-experienced employees could
use it to advantage?
• If knowledge is to d ay ’s best corporate asset, would it not m ake sense to find a
way to preserve, nurture, share, and protect such capital?
The actual carrier of the knowledge, w hether it is e-mail, groupw are, or peer-to-
peer meetings, will involve people with an attitude of sharing com m on knowledge in
the interest of the project or the organization. Technology is only a tool em ployed to
expedite processes.
A no th er reason for writing this book is dissatisfaction with existing books for the
academ ic sector. T here are many good books on the m arket w here each book gives a
specialized picture of know ledge rath er than a com prehensive view of what know l-
edge is and how it is em bedded in today’s organization. For example, one book focuses
on com m on know ledge based on studying a n um ber of established enterprises.
A n o th er book brings up the concept of learning in action as a guide to putting the
learning org an izatio n to w ork. A n o th e r know n book discusses how organizations
m anage what they know. O th er books are either extrem ely behavioral focused or lean
heavily tow ard know ledge autom ation. This text is neither. It is process oriented. It
strikes a balance betw een the behavioral aspects of knowledge and knowledge m an-
agem ent and refers to technology as a m edium for knowledge transfer, especially in
the e-world.

« « « « How the Text Is Organized

This text is about knowledge m anagem ent—what it is and how to use it for com petitive
advantage. Figure 1 is a graphic layout of how the text is organized. Part I presents the
basics of knowledge managem ent and focuses on the concept of knowledge and know l-
edge m anagem ent system life cycle. C hapter 1 introduces the m eaning of knowledge,
how knowledge came about, the myths, and a sum m ary of the knowledge m anagem ent
life cycle. C hapter 2 addresses cognition and knowledge m anagem ent, types of knowl-
edge, expert knowledge, and hum an thinking and learning. C hapter 3 elaborates on the
challenges in building know ledge m anagem ent systems and the knowledge m anage-
m ent system life cycle.
Part II addresses the im portant area of knowledge creation and knowledge capture.
The focus is on the various knowledge tools to capture tacit know ledge as a step to
storing it for know ledge sharing. C hapter 4 brings up N o naka’s m odel of knowledge
creation and transform ation and knowledge architecture centered on people and tech-
nical cores. C hapter 5 lays out in detail the tacit knowledge capture process using the
interview as a tool. It also covers how to evaluate and develop a relationship with the
expert. C hapter 6 discusses other tacit knowledge capture tools such as brainstorm ing,
protocol analysis, D elphi m ethod, and concept mapping.
Part III exam ines how tacit knowledge is codified and how the resulting knowledge
base is im plem ented. In C hapter 7, we deal with knowledge codification tools such as
knowledge maps, decision tables, decision trees, and production rules. C hapter 8 elabo-
 Preface i i i i 19

Parti

FIGURE 1: Organization of the Text

rates on various approaches to logical and user-acceptance testing, how to m anage

the testing phase, and issues related to deploym ent. C h apter 9 expounds on know l-
edge tran sfer and know ledge sharing. Specifically, we focus on the prerequisites for
know ledge tran sfer, know ledge tra n sfe r m ethods, and th e role of the In te rn e t in
know ledge transfer. Finally, C hapter 10 focuses on the e-world in term s of knowledge
transfer within a com pany’s intranet, extranet, groupw are, and groupw are applications.
E -business is discussed in term s of know ledge transfer betw een businesses, supply-
chain m anagem ent, and custom er relationship m anagem ent.
P art IV addresses the m ore technical aspects o f knowledge management. The focus is
on data m ining and know ledge m anagem ent tools and portals. In C hapter 11, the focus is
on learning from d ata thro u g h neural nets, association rules, and classification trees.
C hapter 12 deals with data-m ining concepts, tasks, processes, and practice. C hapter 13
addresses portals, know ledge p ortal technologies, and key m anagerial issues.
20 ■ i ■ ■ Preface

P art V brings up the ethical, legal, and managerial issues in know ledge m anage-
m ent. The question is who owns know ledge —the know ledge w orker or the firm? In
C h ap ter 14, we discuss the liability of the know ledge developer, the expert, and the
user. We also highlight the m eaning of copyrights, tradem arks, and trade names. M ajor
threats, to ethics in know ledge sharing are also covered. C hapter 15 concludes the text
w ith a focus on th e know ledge w o rk er and how to m anage know ledge work.
Specifically, we elaborate on the skills set of the knowledge w orker and how technol-
ogy can be a critical tool for advancing knowledge work. The Epilogue is a futuristic
view of know ledge m anagem ent and the likely direction it will take to help the learn-
ing organization succeed in an increasingly com petitive global environm ent.

■■■■ Key Features

This text incorporates several key features for easy learning:
• Learning by exam ple is evident throughout the text. Concepts, principles, or p ro -
cedures th at are either technical or new are followed by examples or illustrations
for easy learning.
• Boxed vignettes throughout each chapter are brought in from the field through
journals or W eb sites.
• Illustrations are incorporated w here necessary for clearer understanding of the
m aterial. Each graph, figure, or table has been carefully sketched to ensure that
the key concept being represented stands out clearly. W hen actual adaptation is
used, references are provided so the student can locate the knowledge source.
• Im plications fo r knowledge m anagem ent relates chapter m aterial to knowledge
m anagem ent or m anagem ent decision making.
• Knowledge exercises available at the end of each chapter offer a variety of both
straightforw ard and m ore thought-provoking applications. The solutions are
available in the instru cto r’s m anual on the Web site: www.prenhall.com /awad.
• A sum m ary at the end of each chapter brings into focus the essence of the chap-
ter. G ood sum m aries can be a useful guide for chapter coverage. D epending on
learning preference, reading the sum m ary before reading the chapter can be an
effective learning approach.
• Terms to K now are selected key term s from each chapter. A glossary of these
term s is available on the W eb site: w ww.prenhall.com /awad.

■■«■ Supplements
COMPANION WEBSITE (www.prenhall.com/awad)
The te x t’s C om panion W ebsite includes the In stru c to r’s M anual, Test Item File, and
M icrosoft Pow erPoint slides.
A ccess to the in stru cto r’s section of the site, w here the In stru cto r’s M anual and
Test Item File are housed, requires a valid user ID and password to enter. You simply
need to register yourself as the instructor of the course by going to the Web site and
com pleting the initial instructor registration process. U pon com pletion of the process,
your registration request will be forw arded to your sales representative for validation.
 Preface ■aa■ 21

If you have any problem with your authorization, please contact your Prentice Hall
sales representative.

INSTRUCTOR’S MANUAL
A specially p rep ared m anual is available for the instructor. The m anual provides the
following support m aterial:

• A syllabus based on a q u arter or a sem ester plan as a guide to teaching the course
• L ecture assistance , including the objectives of each chapter, teaching notes, and
chapter summ ary
• Solutions to know led g e exercises at the end of each chapter

TEST ITEM FILE

The Test Item File contains m ultiple choice, true-false, and essay questions. The ques-
tions are rated by the level of difficulty and answers are referenced by page number.

POWERPOINT PRESENTATIONS
The slides illum inate and build upon key concepts in the text. They are available for
both students and instructors for dow nload from www.prenhall.com /awad.

Acknowledgments
In a nutshell, every effort has been m ade to m ake K n o w led g e M a n a g em en t truly
understandable, high on lucidity, and practicable. In preparing the m anuscript, we kept
in mind people, not com puters, as the final decision makers. A lthough the underlying
technology in autom ating know ledge work ensures the availability of tacit knowledge
in knowledge bases and other repositories, hum ans have the final say in the way orga-
nizations and society m ust perform .
B efore the m anuscript found its way to the publisher, various versions were tested
in the classroom over a 2-year period in the U nited States and abroad. T here were
successive revisions and updates resulting from student feedback and feedback from
professional review ers and attendees to our m any sem inars on the subject. We would
specifically like to thank the following people:

M ark A ddleson —G eorge M ason U niversity

Jim A lb e rs—Pacific L utheran U niversity
D orine A n d rew s—G eorgetow n University
Frada B urstein —M onash University
D avid F earon —C entral C onnecticut State U niversity
Mike G o d frey —California State University, Long Beach
B arbara G rabow ski—B enedictine U niversity
Paul G ra y —C larem ont G rad u ate U niversity
William H a fn e r—Nova Southeastern U niversity
Paul Johnson —U niversity of Miami
Ravi R ao P atn ay ak u n i—Temple U niversity
Shan Ling P in —N ational U niversity of Singapore
Steven R uth —G eorge M ason U niversity
D oug Tuggle —A m erican U niversity
Joanne Twining—University of D enver
Because no know ledge m anagem ent book is ever com plete, future revisions are
inevitable. A fter having gone through this m aterial, you are invited to share your expe-
rience, ideas, or thoughts. Feel free to e-m ail the authors at em a3z@ virginia.edu or
ghaziri@ aub.edu.lb, respectively.

Elias M. Awad H assan M. Ghaziri

U niversity of Virginia School of Business
C harlottesville, Virginia A m erican University of B eirut
Voice: 434-924-3423 Beirut, Lebanon
Voice: 961 1 352700
About the Authors
Dr. Elias M. Awad is the Virginia B ankers A ssociation Professor of B ank M anagem ent
at the University of Virginia. H e has over 40 years IT experience in the academic, p ub-
lishing, and consulting areas. H e is one of the w orld’s leading IT instructors and sem i-
n ar p re se n te rs in the banking industry. Dr. A wad is the C E O of In te rn atio n a l
Technology G roup, LTD., an IT consulting group with offices in Chicago, New York,
B eirut, Damascus, and Bratislava. His book publication record goes back to the early
1960s with best sellers across the IT discipline in systems analysis and design, database
m anagem ent, know ledge m anagem ent, hum an resources m anagem ent, and building
know ledge autom ation systems. His publications have been translated into G erm an,
Spanish, Portuguese, C hinese, A rabic, R ussian, and Braille. Dr. A w ad’s publications
have earn ed in tern atio n al recognition for lucidity, logical flow, and presen tatio n of
m aterial based on experien ce in the field. H e has delivered professional sem inars
internationally to m ajor corporations and governm ent agencies in 26 countries, includ-
ing K orea, R ussia, C am bodia, C anada, M exico, K azikhstan, M oldova, U zbekistan,
Saudi A rabia, L ebanon, Jordan, and Egypt.
Dr. Hassan M. Ghaziri is an associate professor in o p e ratio n and inform ation
m an ag em en t system s at the School of Business, A m erican U niversity of B eirut.
Dr. G haziri has 12 years of experience in IT teaching, consulting, and research. H e per-
form ed his activities for the academ ic and business com m unities in A m erican,
E uropean, and M iddle E astern cultures. Dr. G haziri is the founder of R esearch and
C om puter A ided M anagem ent-SA L , an IT com pany providing enterprise integrated
systems and focusing on helping com panies bridge the gap betw een business practices
and know ledge technologies. His research and developm ent interests include knowl-
edge discovery and building enterprise know ledge systems. He has authored and co-
authored over 20 scientific publications in prem ier journals.

23
This page is intentionally left blank.
 PARTI: T H E BASICS

TER I I I I I I I

Working Smarter,
Not Harder
Contents

In a Nutshell
W hat Is Knowledge M anagem ent?
The K nowledge O rganization
W hat KM Is N ot A bout
Why Knowledge M anagem ent?
The D rivers
H ow It Cam e A bout
Key Challenges
H istoric Overview
KM Myths
KM Life Cycle
Role of Trust in the KM Life Cycle
Im plications for K nowledge M anagem ent
Sum mary
Terms to Know
Test Your U nderstanding
Knowledge Exercises
R eferences

K n o w in g ignorance is strength
Ignoring kn ow ledge is sickness
—L a o T s u

In a Nutshell
Welcome to the twenty-first century and the knowledge society. The business landscape
is changing rapidly. The com petitive environm ent is no longer linear or predictable.
Survival and success depend entirely on the organization’s ability to adjust to the dy-
namics of the business environm ent. Changes in inform ation technology (IT) have gen-
erated gaps in access and control of inform ation and knowledge. Even when these gaps
are bridged, several fundam ental challenges rem ain. H ow do we apply knowledge for
26 ■ ■ ■ ■ PART I The Basics

value-added and competitive advantage? How do we convert inform ation into knowl-
edge? How do we use technology to convert challenges into opportunities? Knowledge
m anagem ent is the solution for realigning the firm ’s technical capabilities to create the
knowledge that drives the firm forward.
There is obvious room for change in the way we work and communicate and in rela-
tionships and processes am ong people within and across organizations. To be em pow -
ered to face these challenges m eans not only accessing technology, but also developing
the ability to m anage knowledge. In the final analysis, the key questions an organization
m ust consider are “D oes your com pany know w hat you know ?” “D o you know what
you know ?” “H ow do you m ake best use of the knowledge you have?” It also means
thinking “out-of-the-box,” where “the box” is what represents all the tried-and-true pro-
cedures that have w orked in the past. There is less room for “packaged solutions” to
solve m ost of a firm ’s problems. Knowledge m anagem ent m eans thinking outside the
b ou n d aries of cu rren t practices, products, services, and organizations. The new and
unpredictable business environm ent puts a prem ium on innovation and creativity much
m ore so than it has in the past. This explains our chapter title —working sm arter, not
harder. It is “obsoleting what you know before others obsolete it and profit by creating
the challenges and opportunities others haven’t even thought ab o u t” (M alhotra 2000).
We have progressed from the data processing age of the 1960s to the information
age of the 1980s to the knowledge age of the 1990s. The latest transform ation represents
the most fundam ental change since the introduction of the digital com puter 4 decades
ago. Knowledge and intellectual capital (viewed here as accrued knowledge) represent
our corporate and national wealth. Knowledge w orkers are found in every organiza-
tion, and they are the backbone of every successful business. Knowledge w orkers use
technology to reason through problem s and reach successful solutions. Com puter-aided
software gives them an edge over w orkers using conventional methods. In this chapter,
we discuss know ledge m anagem ent, the general concepts and m yths surrounding
knowledge m anagem ent, the relationship betw een knowledge m anagem ent and m an-
agem ent inform ation systems, and how an innovative organization looks at knowledge.
In the past decade, th ere has been m uch discussion ab o u t the im portance of
know ledge and know ledge m anagem ent. Roy Vagelos, the chief executive officer of
M erck & Co., told Fortune magazine that devoting time and resources to the proper
m anagem ent of knowledge is slowly, but surely, gaining support in many organizations.
For a com pany to m anage knowledge, it must first inventory its people, systems, and
decisions. Professional knowledge w orkers within the com pany must be identified, and
th eir functions m ust be defined. K now ledge technologies m ust be in co rp o rated to
reen g in eer the en tire business process. M ajor decisions should be review ed, and a
know ledge system for m aking each decision should be developed. The com pany’s
inform ation system should also be exam ined to determ ine how to benefit from em erg-
ing knowledge technologies. This self-assessment makes a com pany m ore cognizant of
its strengths and weaknesses. It should also lead to changes that are m ore in tune with
the com petitive nature of the business environm ent.

■ What Is Knowledge Management?

K nowledge m anagem ent (KM ) is a newly emerging, interdisciplinary business model
that has knowledge within the fram ew ork of an organization as its focus. It is rooted in
m any disciplines, including business, economics, psychology, and inform ation m anage-
m ent. It is the ultim ate com petitive advantage for today’s firm. Knowledge m anage-
m ent involves people, technology, and processes in overlapping parts (see Figure 1.1).
 CHAPTER 1 Working Smarter; Not Harder ■ ■ ■ i 27

FIGURE 1.1: Overlapping

Human, Organizational, and
Technological Factors of KM

As can be deduced from the select definitions in Table 1.1, researchers as well as prac-
titioners have yet to agree on a definition. How ever, each definition of KM contains
several integral parts:
• U sing accessible know ledge from outside sources
• E m bedding and storing knowledge in business processes, products, and services
• R epresenting know ledge in databases and docum ents
• Prom oting know ledge grow th through the organization’s culture and incentives
• Transferring and sharing know ledge throughout the organization
• Assessing the value of know ledge assets and im pact on a regular basis
In some ways, KM is about survival in a new business w orld—a world of competition
that increases in complexity and uncertainly each day (see Box 1.1). It is a world that
challenges the traditional ways of doing things. The focus is not only on finding the right
answers, but also on asking the right questions. W hat w orked yesterday may or may not
work tomorrow. The focus is on “doing the right thing” rather than “doing things right”
so that core com petencies do not becom e core rigidities in the future (M alhotra 2000).
KM is the process of capturing and m aking use of a firm ’s collective expertise any-
w here in the business —on paper, in docum ents, in databases (called explicit kn o w l-
edge), or in p eo p le’s heads (called tacit knowledge). Figure 1.2 implies that up to 95 per-
cent of inform ation is preserved as tacit knowledge. It is the fuel or raw m aterial for
in n o v atio n —the only com petitive advantage that can sustain a com pany in an unpre-
dictable business environm ent. It is not intended to favor expert systems of the early
1990s, w hen co m p u ters w ere pro g ram m ed to em u late hum an e x p e rts’ th ought
processes. The goal is to present a balanced view of how com puter technology captures,
distributes, and shares know ledge in the organization by linking hum an experts and
docum ented know ledge in an integrated KM system.
The goal is for an organization to view all its processes as know ledge processes.
This includes know ledge creatio n , dissem ination, upgrade, and application tow ard
organizational survival. Today’s know ledge organization has a renew ed responsibility
to hire know ledgeable em ployees and specialists to m anage know ledge as an intangi-
ble asset in the sam e way th at one calls on an investor to m anage a financial portfolio.
A firm seeks to add value by identifying, applying, and in teg ratin g know ledge in
u n p re c e d e n ted ways, m uch like an in v estor adds value by unique com binations of
stocks and bonds. The process is p art science, part art, and part luck.

THE KNOW LEDGE ORGANIZATIO N

A conceptual structure of the know ledge organization is shown in Figure 1.3. The m id-
dle layer addresses the KM life cycle—know ledge creation, know ledge collection or
capture, know ledge organization, knowledge refinem ent, and knowledge dissemination.
28 . . . . PART I The Basics

TABLE 1.1 Alternative Definitions of Knowledge Management

• Knowledge management is the process of gathering a firm’s collective expertise wher-
ever it resides—in databases, on paper, or in people’s heads—and distributing it to where
it can help produce the biggest payoff (Hibbard 1997).
• KM is a newly emerging, interdisciplinary business model dealing with all aspects of
knowledge within the context of the firm, including knowledge creation, codification,
sharing, and how these activities promote learning and innovation (encompassing tech-
nology tools and organizational routines in overlapping parts) (Berkeley 2001).
• KM caters to the critical issues of organizational adaptation, survival, and competence in
the face of increasingly discontinuous environmental change. Essentially, it embodies
organizational processes that seek synergistic combinations of data and information pro-
cessing capacity of information technology, and the creative and innovative capacityof
human beings (Malhotra 1999).
• Knowledge management is the art of creating value from an organization’s intangible
assets (Sveiby 2000).
• Knowledge management is the classification, dissemination, and categorization of infor-
mation and people throughout an organization (Taft 2000).
• Knowledge management is the discipline of capturing knowledge-based competencies
and then storing and diffusing that knowledge into business. It is also the systematic and
organized attem pt to use knowledge within an organization to improve performance
(KPMG 2000).
• KM is really about recognizing that regardless of what business you are in, you are com-
peting based on the knowledge of your employees (Johnson 2001).
• KM is a conscious strategy of getting the right knowledge to the right people at the right
time; it is also helping people share and put information into action in ways that strive to
improve organizational performance (O ’Dell et al. 2000).
• Knowledge management is a framework, a management mind-set, that includes building
on past experiences (libraries, data banks, smart people) and creating new vehicles for
exchanging knowledge (knowledge-enabled intranet sites, communities of practice, net-
works) (O ’Dell et al. 2000).
• KM is accumulating knowledge assets and using them effectively to gain a competitive
advantage (Brooking 1996).
• KM is a framework within which the organization views all its processes as knowledge
processing, where all business processes involve creation, dissemination, renewal, and
application of knowledge toward organizational sustenance and survival (Malhotra 2000).
• Knowledge management in ades a combination of software products and business prac-
tices that help organizations capture, analyze, and distill information (Craig 2000).
• KM is not about technology; it is about mapping processes and exploiting the knowledge
database. It is applying technology to people’s minds (Deveau 2000).
• Knowledge management is the sharing of information throughout a company or even
between business partners. It creates an environment in which the company leverages all
its knowledge assets (Trepper 2000).
• KM can automate the classification of documents while using machine logic that comes
as close as possible to human logic (Hersey 2000).
• Knowledge management is a discipline of identifying, capturing, retrieving, sharing, and
evaluating an enterprise’s information assets (Bair 2001).
a ■ i a ■ ■
 CHAPTER 1 Working Smarter, Not Harder aaa1 29

1 1 ■ 1 ■ ■ BOX 1.1 1 1 ■ i 1 1

AN EXAMPLE OF KNOW LEDGE MANAGEMENT.

The p redaw n silence at the U.S. em bassy in bu t such stories are an im p o rta n t in g red ien t in
Indonesia is shattered by autom atic weapons fire, w hat observers describe as the largest knowledge
and a small force of U.S. M arines barely prevents m an ag em en t effort in the w orld. T he m ilitary ’s
the attackers from entering the com pound. W ithin know ledge m an ag em en t program s are so com -
the next few hours, the embassy sends inform ation prehensive, in fact, th a t th e p riv ate sector can
ab o u t the attack to the C om bat D ev elo p m en t learn m uch from them , from m ore effective ways
C om m and in Q uantico, V irginia. T here, a team to apply inform ation technology to new ways of
analyzes the d ata and sees th at it m atches a p a t- teaching.
tern of activity of a fu n d am en talist reb el group Army Knowledge Online will become mission-
operating in the region. critical, literally, says pro g ram m anager Maj.
A n o th e r team in Q u an tico taps in to a Charles Wells. “Yesterday, we had a very straight-
C om m unity of Practice know ledge d ata b a se to forw ard threat: We knew who the enem y was, and
review new tactical m aneuvers being developed we had a lot of detailed plans to stop an attack,”
in a m odeling gam e room . E x p erts from across says Wells. “Today, we are faced with a variety of
the U.S. M arine Corps join an online discussion, challenges—regional instabilities, econom ic dan -
aided by the latest intelligence inform ation and gers, and the p ro life ra tio n of w eapons of mass
satellite p h o to g rap h s m erged into an electronic destruction. H ow are we going to accomplish all
work space. W ithin a few hours, a new urban tactic th ese m issions w ith a m uch sm aller force? The
is developed, tested in the m odeling gam e room , senior A rm y leadership sees know ledge m anage-
and sent to the M arine com m and in Indonesia. m ent as the key.”
The com m ander coordinates plans with the As advice to o ther know ledge m anagers, Lt.
USS Winston Churchill, a guided-missile destroyer C m dr. Ju dith G odw in, Pacific F le et know ledge
some 500 kilom eters to the east. L ater that day, he m anager, says, “D o n ’t build som ething and expect
calls for fire and, 90 seconds later, the ship launches peo p le to com e. They w o n ’t.” In stead , she says,
a dual salvo of Tom ahawk missiles at a key rebel “Find com m unities th a t are trying to get a task
com m unications cen ter 10 kilom eters from the done. E nable them with a specific tool so they can
embassy. see the value it can bring to w hat they are doing
“I t’s just a story,” says A lex B ennet, the U.S. right now. Then it will grow from there.”
N avy’s d ep u ty C IO for e n te rp rise in teg ratio n ,

SOURCE: Excerpted from Anthes, Gary H. “Charting a Knowledge Management Course,” Computerworld, August 21,2000, p. 38ff.

I I fl I I I

The final step is the m aintenance phase, which ensures th at the know ledge dissem i-
nated is accurate, reliable, and based on com pany standards set in advance. The outer
layer is the im m ediate environm ent of the organization—technology, culture, supplier
and custom er intelligence, com petition, and leadership. Such an environm ent has a lot
to do with how well an organization goes about developing and im plem enting its KM
life cycle —also called the KM process.
The ideal know ledge o rg an izatio n is one w here peo p le exchange know ledge
across the functional areas of the business by using technology and established p ro -
cesses. A s depicted in Figure 1.4, people exchange ideas and knowledge for policy for-
m ulation and strategy. K now ledge is also internalized and adopted within the culture
30 ■ a■a PART I The Basics

1 1 1 9 1 1 FIGURE 1.2: Uncaptured Tacit Knowledge

SOURCE: Bair, Jim “Knowledge Management Technology: Beyond Document Management.”

www.strategy-partners.com/spweb/pages.nsf/all+pages/White%20Page%20%20%20Knowledge%20
Management%20Technology:%20Beyond%20Document%20Management?opendocument September 19,
2001, pp. 1-6. Date accessed August 2002.

of the organization. All know ledge w orkers (people) are in an environm ent w here they
can freely exchange and produce know ledge assets by using various technologies. This
process influences the com pany as a whole in a positive way.
A know ledge organization derives knowledge from several sources:
• C ustom er know ledge—their needs, who to contact, custom er buying power, etc.
• Product know ledge—the products in the m arket place, who is buying them , what
prices they are selling at, and how much m oney is spent on such products
• Financial know ledge—capital resources, w here to acquire capital and at what
cost, and the integrating in financial practices
• Personnel practices know ledge—the expertise available, the quality service they
provide, and how to go about finding experts, especially in custom er service

FIGURE 1.3: The Knowledge Organization

 CHAPTER 1 Working Smarter, Not Harder ■ ■ ■ ■ 31

FIGURE 1.4: Ideal Knowledge Management

SOURCE: Gravallese, Julie. “Knowledge Management,” The MITRE Advanced Technology Newsletter,
April 2000, vol. 4, no. 1. www.mitre.org/pubs/edge/april_00/4, Date accessed August 2002.

The idea of “m anaging” know ledge is abstract. K now ledge is not som ething we
typically th in k of as being m anaged, b u t ra th e r som ething th a t is individually con-
trolled, personal, and autonom ous. To be able to m anage knowledge, one m ust first be
able to elicit an individual’s know ledge from th at individual. The hum an aspect of both
know ledge and m anaging are integral. T here is also the issue of m easuring knowledge:
If you cannot m easure it, you cannot m anage it.
O n e u n iq u e in d ic a to r of KM in action is seeing peo p le think actively, n o t
passively—thinking ahead, not behind. It is an environm ent w here custom er service
is im proved through b etter problem -solving, w here new products are available to the
m ark et m o re quickly, and w here the o rganizational processes th a t deliver the new
products continue to im prove through innovation and creativity of the people behind
the product and the production process. This is w here technology, netw orking, and data
com m unication infrastructure play an im portant role. Technology has m ade knowledge
sharing and innovation m ore feasible (see Figure 1.5).

What KM Is Not About

• K nowledge m anagem ent is not reengineering. R eengineering implies one-shot,
drastic “electrical shock” change in organizational processes to im prove efficiency.
It is a m echanical shift from one stage of operation to a m ore efficient stage, and it
usually Involves radical changes of business processes and the people involved. In
contrast, KM implies continuous change and addresses future threats and unique
opportunities. T here is continuous learning, unlearning, and relearning to ensure
sm ooth change from top to bottom . The focus is on change that will generate grad-
ual but solid gains in the com petitive environm ent. Know ledge m anagem ent is
engrained in the day-to-day operations of the business and directed by people who
are directly connected with the changing world of their com pany’s business.
• K nowledge m anagem ent is not a discipline. It is an other way of im proving quality,
profitability, and growth.
• K now ledge m anagem ent is n ot a philosophic calling. KM goes to the core of an
organization’s intangible asset (know ledge), revisits the knowledge, and taps into it.
32 a«aa PART I The Basics

• New products
• New markets
• Smarter problem-solving
• Value-added innovation
• Better customer service
• More efficient processes
• More experienced staff

Organizational Benefits

Codified Technology

I 9 9 I ■ I FIGURE 1.5: Knowledge Management and Innovation

• K nowledge m anagem ent is not intellectual capital, per se. Intellectual capital (IC)
represents the value of a com pany’s tradem arks, patents, or brand names.
Intellectual capital is a com pany’s collective brainpow er, or a com posite of expe-
rience, knowledge, inform ation, and intellectual p ro p e rty —all the property of the
organization. A lthough treated in the literature the sam e as knowledge, know l-
edge, p er se, is the consequence of actions and interactions of people with infor-
m ation and know ledge exchange based on experience over tim e (see Box 1.2).
• Knowledge m anagem ent is not based on information. Inform ation can becom e
know ledge after people use it in ways that create value. Knowledge has been
viewed as inform ation in action. A s we shall explain in C hapter 2, inform ation is
context-sensitive; know ledge is consensus-oriented.
• K nowledge m anagem ent is not about data. D ata (facts w ithout context) or infor-
m ation (interp retatio n or p atterns of data) is not knowledge. As we shall find in
C hapter 2, data by itself is not actionable knowledge (O ’D ell et al. 1998,17).
• Knowledge value chain is not inform ation value chain. In inform ation value
chains, the key com ponent is a technological system guiding the com pany’s busi-
ness processes, viewing hum ans as passive processors. In contrast, knowledge
value chains view hum ans as the key com ponents assessing and reassessing infor-
m ation stored in a technological system. Best practices into organizational busi-
ness processes are carried out after active hum an inquiry, and such processes are
continuously upd ated in line with the changing external environm ent.
• K nowledge m anagem ent is not limited to gathering inform ation fro m the com -
pany's dom ain experts or retiring employees and creating databases accessible by
intranets. KM is a collective concept of the organization’s entire core knowledge.
• K now ledge m anagem ent is not digital networks. KM is about im proving business
processes with people and technology in mind. Effective technology is the
enabler of KM , and people m ust be in the equation from the start to use technol-
ogy effectively.
 CHAPTER 1 Working Smarter, Not Harder i aa a 33

■ ■ i 1 ■ i BOX 1.2 I a I « a ■

INTELLECTUAL CAPITAL
Tracking a com pany’s physical assets is straig h t- N ader, senior vice president and chief operating
forw ard enough, as long as y ou’re counting com - officer at the Pittsburgh-based bank.
puters, adding salaries, and estim ating h eatin g R on Griffin, C IO at The H om e D epot, Inc.,
bills. But m anaging intellectual capital is a differ- says th at A tlan ta-b a sed hom e im provem ent re -
ent ball game, and one in which few com panies tailer has tried -an d -tru e stru ctu res in place for
consistently hit hom e runs. m easuring, m aintaining, and growing intellectual
Intellectual capital involves a com pany’s em - capital. The com pany uses a nine-box grid system
ployee expertise, unique organizational systems, and to m easure each e m p lo y ee ’s p erform ance and
intellectual property. For example, if a com pany’s potential, and it offers developm ental courses to
book value is $10 per share and its stock is selling for b ring em ployees up to speed on certain issues.
$40 per share, the difference is often attributed to The categories m easured include leadership abil-
intellectual capital. “When you subtract book value ity, how an em ployee fits into the H om e D epot
from m arket value, the remaining is all the intellec- culture, financial acum en, and project m an ag e-
tual and knowledge and m arket capital. It includes m ent capabilities
all the patents they might have and all other intangi- For its part, H om e D epot posts a bulletin on
bles,” says Vish Krishna, associate professor of m an- its intranet with quick references on topics such as
agement at the McCombs School of Business at the how to repair a leaky toilet or build a deck. That
University of Texas at Austin. way, know ledge is available for em ployees to
O nce a com pany identifies its intellectu al rem ain up to speed and to pass such inform ation
capital, the next step is to m aintain it. O ne of the along to custom ers. “ I t’s not just about selling
techniques th at D ollar B ank uses to m anage product in our business; it's a lot of the knowledge,
intellectual capital is to keep em ployees involved and we train on that extensively,” Griffin says.
in decision m aking and planning, says A braham

SOURCE: Excerpted from Taylor, Christie. “Intellectual Capital” Com puters orld, March 12,2001, p. 51.

1 I I I I 1

• Knowledge m anagem ent is not about “knowledge capture, ” per se. Knowledge can-
not be captured in its entirety. Problem s involving collaboration, cooperation, and
organizational culture must be addressed before one can be sure of reliable
knowledge capture.
R egardless of the business, a com pany com petes based on the know ledge of its
employees. A com pany also has a m anagem ent m ind-set that relies on past experience
(such as sm art people, docum ents, or databases) and creates a new way for exchanging
knowledge by using intranets, the Internet, local area networks, and the like. C onsider
the case of a British superm arket chain that used a custom er data-m ining application
to assess buying behavior. A fter running correlation analyses am ong several variables,
it quickly discovered a clear association betw een the purchases of diapers and beer by
m ale custom ers on Friday afternoons. A rm ed with this knowledge, the store began to
stack diapers and beer together.
It should be clear by now that people are the determ inants of KM success. The best
softw are is insufficient if you do not have people willing and ready to cooperate and
m a PART I The Basics

collaborate. Sharing knowledge based on m utual trust is the critical com ponent in the
entire KM process. To illustrate, Lotus describes its R aven as resem bling the “collabo-
rative capabilities of N otes” but going further by “defining the relationships betw een
people and content.” Raven is a key elem ent of L otus’ overall knowledge m anagem ent
strategy, which includes services, methodologies, and solutions from Lotus and IBM, all
captured under the unifying them e of “people, places, and things.” The com pany can
create an em ployee profile detailing the em ployee’s specialty and his or her projects,
and then store the profile in a database that can be edited and accessed by users and
cow orkers via e-m ail, instant messaging, and conferencing setups. The req uirem ent
with R aven is for em ployees to update their respective profiles (databases) consis-
tently and regularly to m aintain the “knowledge portal.” If knowledge is not reinvested
in this m anner, it cannot be reliably created at a later time. Software update alone is
not the answer.
Raven provides a single portal that will allow end users and communities to find
and discover useful inform ation and applications on a given subject, m ake the user
aware of other know ledgeable people in the company, and organize all related tasks,
teams, and projects. With its unique architecture, it is capable of generating a content
map, creating expertise profiles, creating and hosting com m unities, clustering docu-
m ents, m ining skills, locating experts, and searching and browsing in a user-friendly
m anner. Such a KM tool can satisfy many of the user’s needs that might not be fulfilled
currently in the firm. O ne can plan to use such a product for the KM im plem entation.
H ow ever, to fit a firm ’s needs perfectly, an inform ation technology consultant should
be contacted for p roper fit.

Why Knowledge Management?

A new word for the consum er in today’s m arket, “prosum er,” refers to the consum er
who is no longer in the passive m arket where goods are offered at the exact face value.
Prosum ers are m ore educated consumers, and they dem and more. They provide feed-
back to m anufacturers regarding the design of products and services from a consum er
perspective (B exar 2000). This has initiated new and radical changes in the business
world. Even with recent technology developm ents such as networking, e-mail, and the
Web, business has yet to fully respond to the societal, cultural, and technical challenges.
H ow ever, a positive response to know ledge sharing and know ledge m anagem ent
am ong grow th-oriented firms is beginning to appear.
KM has already dem onstrated a num ber of benefits and has offered justification
for further im plem entation. The Internet facilitated its developm ent and growth via
fast and timely sharing of knowledge. By sharing knowledge, an organization creates
exponential benefits from the knowledge as people learn from it. This makes business
processes faster and m ore effective and em pow ers em ployees in a unique way. For
exam ple, M icrosoft’s H otm ail service advanced the wide use of e-mail that allowed
users to exchange inform ation through any Web browser. Today's W eb-based interface
is the norm for most Internet service providers.
Based on a num ber of published studies, KM has had a positive impact on business
processes. The goal is to capture the tacit knowledge required by a business process
and encourage knowledge w orkers to share and com m unicate knowledge with peers.
W ith such knowledge, it is easier to determ ine which processes are m ore effective or
less effective than others. The main constraint in KM, however, is initially capturing it.
H ow ever, if an organization can succeed in capturing and dispersing knowledge, the
 CHAPTER 1 Working Smarter; Not Harder a■aa 35

benefits are endless. A com pany can leverage and m ore fully utilize intellectual assets.
It can also position itself in responding quickly to custom ers, creating new m arkets,
rapidly developing new products, and dom inating em ergent technologies.
A n o th er benefit of KM is the intangible return on knowledge sharing rather than
k now ledge h oarding. Too often, em ployees in one p a rt of a business start from
“sc ra tc h ” on a p ro je c t because the know ledge n e ed e d is som ew here else but not
know n to them . To illustrate one docum ented case, a departm ent of A T& T spent close
to $80,000 for inform ation that was available in a technical inform ation docum ent from
its associate company, Bell R esearch C orporation, for $13 (Skyrm e 1999,2).
As a result of KM, systems have been developed to gather, organize, refine, and
distribute know ledge throughout the business. In his study of Sm art Business, B otkin
(1999) suggests six top attributes of know ledge products and services:
• Learn.The m ore you use them , the sm arter they get and the sm arter you get, too.
• Im prove with use.These products and services are enhanced rather than depleted
when used, and they grow up instead of being used up.
• A nticipate.Knowing w hat you want, they recom m end w hat you might want next.
• Interactive.T here is two-way com m unication betw een you and them.
• Remember. They record and recall past actions to develop a profile.
• C ustom ize.They offer unique configuration to your individual specifications in
real tim e at no additional cost.
D uring the 1960s and 1970s, technology was focused on autom ating high-volume
static processes such as claims processing, m ortgage loan updating, airline reservation
systems, and the like. The em ergence of e-com m erce in the late 1980s and 1990s showed
how inform ation technology could im plem ent a new way of doing business effectively.
Ever-increasing processing power, high bandw idth data transm ission, and netw orking
m ade it possible to reenvision how business gets done. It has also changed the business
environm ent and introduced new com petitive imperatives. A m ong them are:
• Reacting instantly to new business opportunities, which led to decentralized deci-
sion m aking (and com petency) at the fro n t lines, where the action is. W ith that
cam e the desire to build m utual trust betw een know ledge w orkers and m anage-
m ent and to cooperate in handling tim e-sensitive tasks.
• B uilding better sensitivity to ubrain drain. ” It has been said that “expertise gravi-
tates tow ard the highest b id d er” (A pplehans et al. 1999,17). M ore and m ore com -
panies realize the im portance of m anaging and preserving expertise turnover. For
the hum an resources d epartm ent, the key question is “H ow does the firm replace
expertise w hen it retires, resigns, or simply leaves?”
• Ensuring successful partnering and core competencies with suppliers, vendors, cus-
tomers, and other constituents. Today’s technology has enabled com panies to
reengineer the ways to do business. G etting partners up to your speed requires
m ore than fast technology. K nowledge w orkers and others within the com pany
should ensure th at cooperation and coordination of workfare practiced for the
good of the firm.
W ith the expected depletion of highly know ledgeable w orkers due to the retire-
m ent of baby boom ers, KM will be im portant in the upcom ing years. It is essential that
organizations capture and preserve the know ledge of senior colleagues so that younger
em ployees can m ake im m ediate use of it and im prove upon it to m ake the business run
even m ore sm oothly and m ore efficiently. M errill Lynch has w orked to incorporate KM
to sh o rten the learning curve of its 18,000 brokers worldwide. The newly developed
36 • ■ ■ ■ PART I The Basics

know ledge-based system facilitates sharing of know ledge and quickly enables less-
trained brokers to achieve perform ance levels ordinarily associated with m uch m ore
experienced professionals. The system captures and disperses know ledge about trans-
actions, trading rules, yields, securities features, availability, tax considerations, and new
offerings. Sharing of this know ledge allows brokers to serve millions of clients w orld-
wide with sophisticated investm ent advice and detailed, up-to-date inform ation and
data. It acts as an instant training vehicle.
A ndersen Consulting (now Accenture) provides another example of a well-developed
know ledge-sharing system , called A N et. This electronic system connects em ployees
and encourages the sharing of knowledge. A N et allows an em ployee to use the total
know ledge of A ccenture (form erly A rth u r A ndersen) to solve a custom er problem
anyw here in the world through electronic bulletin boards and to follow up with visual
and data contacts. In theory, A N et expands the capabilities and knowledge available
to any custom er to th at of the entire organization. It further enhances em ployee p ro b -
lem-solving capacity by providing access to com piled subject, custom er-reference, and
resource files available either directly through the system or from C D -R O M s avail-
able to all offices. B ased on experience, A cc en tu re re p o rte d th a t technological
changes alone could not m ake A N et successfully used by employees. M ajor changes
w ithin the organization, such as changes in incentives and culture, w ere needed to cre-
ate participation. A sum m ary of KM justification is shown in Table 1.2.
Com panies that fail to em bed a viable KM operation probably suffer from several
oversights or pitfalls:

• Failing to m odify the com pensation system to reward people w orking as a team.
The traditional m ethod of com pensating people based on the old-fashioned
“inform ation-hoarding” practice does not work in a know ledge-sharing environ-
m ent. M erit increases and bonuses should be based on team contribution and
team perform ance rath er than quantity or volume.
• Building a huge database that is supposed to cater to the entire company.
G eneralized systems do not usually work well, because inform ation and know l-
edge are not stratified to address specialized areas of expertise. Ideally, the

TABLE 1.2 KM Justification

• Creates exponential benefits from the knowledge as people learn from it
• Has a positive impact on business processes
• Enables the organization to position itself for responding quickly to customers, creating
new markets, developing new products, and dominating emergent technologies
• Builds mutual trust between knowledge workers and management and facilitates cooper-
ation in handling time-sensitive tasks
• Builds better sensitivity to “brain drain”
• Ensures successful partnering and core competencies with suppliers, vendors, customers,
and other constituents
• Shortens the learning curve, facilitates sharing of knowledge, and quickly enables less-
trained brokers to achieve higher performance levels
• Enhances employee problem-solving capacity by providing access to compiled subject,
customer-reference, and resource files available either directly through the system or
from CD-ROMs available to all offices.
■ i i i a ■
 CHAPTER 1 Working Smarter, Not Harder i i i i 37

hum an resources d epartm ent should first determ ine who works' best with whom
based on com m onality of job type or job experience and then discover the knowl-
edge th at can be shared for each em ployee to be m ore successful.
• Viewing K M as a technology or a hum an resources area. This oversight relates to
the earlier o n e —w here hum an resources and inform ation technology efforts are
poorly c o o rd in ated —and defeats the purpose behind em bedding KM into the
fabric of the organization. The two departm ents should w ork jointly at introduc-
ing KM as p art of the organizational processes.
• Placing too m uch emphasis on technology. A lthough intranets, know ledge-based
tools, data warehouses, and other com puter-based softw are are part of the way
to d ay ’s organization m ust adopt, technology is only the enabler of knowledge
m anagem ent. The know ledge it m akes available m ust be organized and dissemi-
n ated to hum an decision m akers to be of any use.
• Introducing K M into the organization via a sim ple project to m inim ize possible
losses. This is the wrong way to start KM. A com pany should start with a strategy
and a cham pion, with a focus on a worthwhile, high profile project that can set the
tone for the rest of the organization. It is a high risk approach, but one that is
m ost likely to pay dividends in the long run.
• Pursuing K M without being ready. Spurred by the paradigm shift in our economy,
many corporations pursue KM without evaluating w hether they are organizationally
ready (Stewart et al. 2000,45). In other words, corporations that have been operating
under classical managem ent principles cannot be successful in adopting KM without
m ajor changes in culture, m anagem ent attitudes, and communication skills.
• H aving p oor leadership. Like any high priority project, KM is best implemented with
determ ined cham pions and top m anagem ent com m itm ent. For example, G eneral
E lectric (G E ) recognizes an organizational culture open to ideas from all levels of
the company. By encouraging best-practice sharing, the com pany can grasp the
know ledge within the em ployees and innovate the organization’s processes. Jack
Welch, form er CEO , has established a knowledge m anagem ent university and fre-
quently teaches the classes himself. Only 10 percent of the 96 com panies surveyed
by the C onference B oard sponsored by Pricew aterhouseC oopers identified the
C E O as a com ponent of a KM initiative. By integrating the C E O of the company
into the KM system, KM acquires a level of im portance and respect that would
otherw ise be lacking. G E has incorporated all levels of the business and is well
designed to share knowledge. The com pany is successfully able to use em ployee
input and know ledge to produce a strategic advantage (Jones 1999,3-18).

The Drivers
W ith these ju stifications to consider, several key KM drivers are w orth noting.
Each driver m akes a com pelling case for KM.

Technology Drivers. The proliferation of technology, data communications, networking,

and wireless transmission has revolutionized the way employees store, communicate, and
exchange data at high speed. The W orld W ide Web has changed KM from a fad to an
e-business reality. With a personal com puter costing less than $600, anyone can access
people and inform ation at any time and from anywhere. Today, the same technological
infrastructure m akes it possible to store, com m unicate, and exchange knowledge. This
makes technology a core capability leveler, leaving knowledge as a competitive differentia-
tor. It means that although technology can move information or knowledge from Chicago
to Bombay at lightening speed, it is people who turn knowledge into timely and creative
38 a a a a PART I The Basics

decisions. Tomorrow’s successful companies are ones that use information technology to
leverage their employees’ knowledge in ways that make knowledge immediately available
and useful. It also implies quality maximization and cost minimization over the long term.
Process Drivers. O ne of the m ost critical sets of KM drivers is designed to improve
work processes. Im plied in this area is the elim ination of duplicate mistakes by learning
from the past and by transferring the best experiential knowledge from one location or
project in the firm to another. Starting from scratch with each project m akes no sense
in term s of efficiency, productivity, and value-added contribution to the com pany’s b o t-
tom line. The value of knowledge sharing is shown in Box 1.3.
A nother area w here KM can im prove process is the way com panies react to m ar-
ket changes. “Just in tim e” is one approach to minimizing investm ent in inventory and
m ore expeditiously m eeting the dem ands of the consum er. R esponsiveness th at
exceeds the com petition becom es the key co n trib u to r to differentiation. It requires
know ledge of co n tro l processes. KM m eans allow ing com panies to apply unique
know ledge that m akes them m ore responsive to m arket changes by the hour.
Personnel-Specific Drivers. This area of KM drivers focuses on the need to create
cross-functional team s of know ledge w orkers to serve anyw here in the organization
and minimize personnel turnover as a threat to collective knowledge. M ore and m ore of
what was once viewed as independent firms are now closely coupled. Products and ser-
vices are jointly handled from diverse disciplinary areas (such as packaging, m anufac-
turing, engineering, and technical skills), w here creative cooperation is essential for
innovation. B rainstorm ing, co m petitive response, and proactive p o sitio n in g —all
require collaboration and coordination of various tasks within and am ong corporations.
A n o th er personnel-specific driver is minimizing knowledge walkouts. Highly m ar-
k etab le em ployees with unique know ledge can spell d isaster for th eir em ployer.
C om petence drain that goes to the com petition is probably the worst that can happen
to a com pany struggling through the new knowledge economy.
K now ledge-R elated Drivers. Several KM drivers relate to the very concept of
know ledge sharing and know ledge transfer within the firm. They include revisiting
overlooked em ployee knowledge, making critical knowledge available at the time it is
needed, and finding a m echanism to expedite available knowledge for im m ediate use.
C om panies often know what they know but have difficulty locating it. Take the case of
a custom er who w anted to return a product that was initially purchased from the same
outfit in a different city to a local chain store. A code had to be entered into the com -
puter to debit the initial store by the price of the product and then credit the local store
by the same am ount. There was only one em ployee in the local store who knew the
code. She happened to be on vacation. The custom er service em ployee could not find
critical existing know ledge in time. So, she had to contact the other store for instruc-
tions on how to handle the returned item. Counting wait time and learning the proce-
dure took close to 1 hour, while the custom er was waiting.
F inancial D in e r s . As an asset, knowledge defies econom ic theory, where assets are
subject to diminishing returns over the long run. Knowledge assets increase in value as
m ore and m ore people use them . W ith this in m ind, know ledge follows the law of
increasing returns —the m ore knowledge is used, the m ore value it provides. KM pro -
vides a w orthw hile opportunity to integrate knowledge in a way that enriches the qual-
ity of decision m aking throughout the organization.
In the final analysis, the goal of KM is to produce a positive return on investm ent
in people, processes, and technology. It m eans m easurable efficiencies in production,
 CHAPTER 1 Working Smarter; Not Harder ■■■i 39

AN EXAMPLE OF THE VALUE OF KNOW LEDGE SHARIN G

LEARNING FROM MISTAKES that the retailer that does this the fewest num ber
of times wins. As a m atter of fact, no one at Best
M istakes happen. Best Buy Co., the M inneapolis-
Buy has m ade this particular m istake even once.
based retailer of consum er electronics, personal
“O ne of our technicians found the sim ilarity of
com puters, and hom e appliances, does sp eak er
w ires for the airbag and the sp eakers and dis-
installation and sim ilar tasks. For B est Buy, p re -
seminated that information,” M eyer says. “The haz-
venting m istakes is a focus for know ledge m an -
ard was prom oted to a top tip and sent to every
agem ent. KM is ab o u t creating a co rp o rate cul-
m obile installer across our en terprise th at day.”
ture that learns from experience, so if m istakes do
O f course, that glitch is only one of many possible
happen, they will never be repeated.
mistakes. O ther installations, for example, require
To see how quickly the bottom -line benefits of
drilling close to a car’s gas tank.
knowledge sharing prove themselves at Best Buy,
Today, before an installer ever touches a car,
let’s look at the results of one of its original three
he or she goes to the com puter to print out any
KM pilots, in the auto service and installation bays.
tips for the specific vehicle and product. If anyone
“We have over 400 stores, two car bays per store,
learns som ething new, a tip is input to be validated
1,400 service technicians, and 400 different audio
by the gatekeeper and dissem inated through the
products,” says A llen M eyer, KM program direc-
system. In these ways, knowledge sharing becomes
tor. “How many different ways can you put a radio
a process em bedded in standard operating proce-
in a car? O ne: the right way. The em phasis on
dures. “W ith a new m odel w here this is always
knowledge m anagem ent and knowledge sharing is
going to be a problem , it may be inevitable that [a
to be able to m odel best practices behavior to
mistake] happens once,” M eyer admits. “The real
avoid cost and improve custom er satisfaction.”
tragedy is if it happens even a second time because
If, for example, a technician was to set off the
the inform ation d id n ’t get out. If we avoid that
Porsche B o x ste r’s airbag, B est Buy w ould lose
scenario, how much m oney can we save?”
revenue and custom er confidence. You m ight say

SOURCE: Excerpted from Barth, Steve. “Learning from Mistakes,” Knowledge Management, April 2001, pp. 41-47.

sales, and services on a daily basis; im proving the quality of decision m aking at the
front lines; bringing your business partners up to speed; improving em ployee m orale to
ensure low turnover and effective decision making; and im proving custom er-em ployee
relations through b etter trust in know ledge w orkers’ expertise.

■■i ■ How It Came About

Traditionally, there has been a predefined “recipe for success,” where com panies have
been run by com pany m ottos and age-old business practices th at were accepted as
m ethods for success. Doing the same stuff harder and harder has dem onstrated to many
leading firms that they are losing touch with the quickly changing business environm ent
and losing m arket share. The successes of yesterday no longer translate into successes
of tom orrow. In to d ay ’s business world, the heartbeat of the firm depends on the con-
stant revam ping of systems to rem ain com petitive. There is trem endous emphasis on
quick response instead of planning. Business can no longer rely on preset rules with
40 a a a i PART I The Basics

which to operate, because the business environm ent is constantly evolving at an alarm -
ing speed. To be successful, business firms must redefine and question their current
knowledge stored in corporate databases, while creating new practices to fit the busi-
ness environm ent. As a reaction to the questionable benefits from downsizing, business
process reengineering, and oth er cost-cutting m easures in the 1980s and 1990s, knowl-
edge m anagem ent surfaced as the best next step to addressing the com petition in a
hard-to-predict environm ent.
The inform ation rev o lu tio n has placed em phasis on sharing huge am ounts of
inform ation that is now accessible on the Internet. In a time of “e-everything,” infor-
m ation is accessible from business-to-business, business-to-consum er, and consumer-
to-consum er (M alhotra 2000). C orporate A m erica has begun to use this inform ation
availability to their advantage. E xternal relationships such as supply chain m anage-
m ent have been successfully used to im prove productivity and flexibility based on
sharing betw een suppliers and customers. C om panies have taken this idea of inform a-
tion sharing through KM to work within the firm. A ided by technology, employees now
can share knowledge internally, in an effort to m ake the corporation a m ore productive
enterprise.
KM is slowly gaining acceptance across industries. Several factors triggered inter-
est in KM:

• The pace o f change has accelerated dramatically during the past decade. C om pa-
nies are looking at innovative ways of taking on the com petition. Innovation is
the one core com petency needed by all organizations (D rucker 1969).
• Globalization and geographic dispersion changed the organization’s scope. M ore
and m ore organizations are trying to lean on years of experience to m anage their
global com m itm ent in a timely and profitable fashion.
• D ownsizing and reengineering resulted in sta ff attrition and knowledge drain. This
prom pted organizations to assess their knowledge core and m ake m ore effective
use of it. R eengineering assumed a one-tim e fix to a situation. This created a
vicious cycle, where solutions becam e new problems. It failed to recognize rapid
changes in tod ay ’s m arket.
• Netw orking and data com m unications m ade it easier and faster to share know l-
edge. Knowledge sharing is becoming the best way to distribute expertise across
and around the firm via technology. Technology alone is insufficient.
• The increasing dom inance o f knowledge as a basis fo r im proving efficiency and
effectiveness triggered m any companies to fin d the means fo r utilizing the know l-
edge they have gained fro m previous experience.

With these factors, it is easy to see how knowledge m anagem ent works for the sur-
vival of the firm. Knowledge is the key. It is the core com petence of any business. It is
a function that can and should be em bedded into every business process —new pro d -
ucts and services, new channels of distribution, new m arketing strategies, and new
industry definitions. Technology is the backbone, and hum an com ponents are neces-
sary to utilize it.
In a 2000 survey of 243 dom estic and international organizations by a leading con-
sulting firm, the goal was to assess the current status of KM in business. O ver two-
thirds of the respondents claimed they had a KM strategy in place. A m ong the high-
lights of the report are the following:

• KM provides real benefits.

• C om panies with KM are b etter than those w ithout it.
 CHAPTER 1 Working Smarter; Not Harder ■i ■■ 41

• O rganizations are failing to tackle K M ’s challenges. Failure to integrate KM

into everyday work, lack of time to learn and share knowledge, lack of training,
little personal benefit, and failure to use knowledge effectively continue to be
a problem .
• O rganizations are struggling with understanding the cultural im plications of KM.
They are finding that it com plicates their job description.
• C om panies still see KM as purely a technological solution. The shift in focus is
occurring now, but hum an im portance has yet to be realized.
• There is m uch to be accom plished in adopting and integrating KM, but the future
looks bright (K PM G 2000).
H um an resources is com m only identified as having a leading role in knowledge
m anagem ent. W hen P ricew aterhouseC oopers sponsored a survey of 90 com panies,
they found that 32 percent of em ployees believed that hum an resources were involved
in know ledge m anagem ent, with a 25 percent response for inform ation technology.
A lthough inform ation technology and technology in general are a large part of knowl-
edge m anagem ent, people continue to be the driving force behind knowledge m anage-
m ent (N ew m an 1999).
Key Challenges
The biggest challenge that KM vendors face is explaining what KM is and how it
can benefit a corporate environm ent. To get their foot in the door, vendors must inte-
grate KM into the corporate culture —a shift in both organizational and individual phi-
losophy. The culture comes from the top of the organization. G ood or bad, it defines
the m ajor aspects of em ployee behavior on a daily basis. If the culture is not one that
encourages co o p eratio n and trust am ong em ployees, em ployees will not cooperate.
M any experts suggest that the first objective is to m ake know ledge sharing profitable
for the em ployee and the firm as well. W hen one tries to change another person’s con-
ditioned behavior, it is best to tie a form of com pensation to changes in behavior, per-
form ance, or cooperation or to offer some kind of rew ard as a m otivator. C ultural dif-
ference and their im pact on knowledge sharing are shown in Box 1.4.
A fter the organizational culture is m odified, the next challenge is to evaluate the
firm ’s core knowledge, by em ployee, by departm ent, and by division. O ne source com -
pares assessing returns from a KM initiative to “calculating a payback from providing
em ployees with telephones, p a p e r and pens, etc.” (G lasser 1999, 17), In evaluating
these initiatives, one must look at the hum an as well as the technological investment.
A third challenge is learning how knowledge can be captured, processed, and acted
on. KM m ust allow an organization not only to “stockpile” and access inform ation, but
to get at the history of how decisions were reached as well. Turning knowledge into
action requires gleaning the inform ation that has m eaning and relevance for a particu-
lar organization. Because many executives resist the idea of m anaging knowledge like
o th er assets, a com pany has to incorporate KM into the cultural fabric —a shift that
requires changes in both organizational and personal philosophy.
A fourth challenge in KM is addressing the still neglected area of collaboration.
Sound collaboration capabilities m ean helping com pany em ployees share docum ents
and needed inform ation for all kinds of projects. Beyond efficiency and productivity,
the real benefit of collaboration is innovation. It is a natural outcom e of people w ork-
ing together for a com m on cause in a group or an organizational setting. Prom oting
innovation is the essence of KM. Technology can m ake this possible via online m eet-
ings, discussion databases, and other techniques m ade available through the IT depart-
m ent. To illustrate, imagine you missed your flight on the way to your hom e office to
42 ■ aaa PART I The Basics

----------- « i' i i « i BOX 1.4 1 1 1 1 1 1 -----------

CULTURAL DIFFERENCES AND THEIR IMPACT ON KNOW LEDGE SHARING

W EAVING TWO INTO ONE K now ledge m anagem ent techniques and p rin -
ciples are vital to a sm ooth transition. KM n atu -
Making a merger work is not easy. When faced with
rally was at the forefront when Sopheon PLC, a
a different corporate culture, it is not uncom m on
L ondon-based developer of KM tools, m erged
for the acquiring entity to try to impose its own sys-
with the M inneapolis-based KM consultancy
tems, values, and strategies on the acquired. This
Teltech R esource N etw ork Corp. In this case,
approach can produce a defection of talent and a
m anagem ent had to harm onize several core dif-
decline in stock price. O ne exam ple is the unifi-
ferences th at could have underm ined the deal.
cation of D aim ler-Benz AG and Chrysler Corp.
Both sides had to learn the finer points of trans-
Touted beforehand as a “merger of equals,” in real-
atlantic culture differences. “Europeans like to get
ity it was an acquisition. Daimler-Benz, the dom i-
to know you a little first and then do business,
nant partner, has gradually been imposing its will.
w hereas in A m erica it is often the o th er way
O bservers say that the cultural conflicts betw een
round,” says Paul Corney, a business and strategic
these companies have been particularly damaging
advisor to Sopheon in E ngland, who acted as a
when it comes to sharing knowledge. “G erm an engi-
facilitator for the merger.
neers are leery of having their ‘superior’ technology
These examples suggest that a smooth merger
incorporated into ‘low bred’ A m erican cars,” says
transition depends on speed of execution and effec-
Kay H am m er, an en trep ren eu r and president of
tive communication to resolve cultural barriers. “If
Evolutionary Technologies International, Inc.
the cultures are not meshed quickly, knowledge will
“Similarly, G erm an m arketers are having trouble
exit before it can be harvested,” says Barry Calogero,
figuring out what the U.S. consumer wants. If they
executive vice president at Robbins Gioia, Inc., a
are too arrogant to listen to their American counter-
m anagem ent consultancy in Alexandria, Virginia.
parts, there’s going to be even more trouble ahead.”
Following management efforts to assure employees
C ultural issues and o th e r intangibles can
of the wisdom and express the strategic advantages
make or break a merger. The decision about buy-
of the mergers, assessing and sharing knowledge
ing or selling a com pany should include analysis of
management help to cement the deal.
the cultures in the form of a know ledge audit.

SOURCE: Excerpted from Robb, Drew. “Weaving Two into One,” Knowledge Management, April 2001, pp. 33-38.

participate in an im portant m eeting regarding the launching of a new product. Instead

of missing the debates and participation in strategic discussions, a KM system with col-
laborative features can allow you to log on from your hotel room and jum p into an
online debate about the product in question. You could participate in the m eeting as it
is taking place. You could even vote on the final decision as if you were there in person.
A fifth challenge is to continue researching KM to im prove and expand its current
capabilities. The research should discover ways of gathering, storing, processing, and
distributing know ledge custom ized to the unique structure and operation of the firm.
M eantim e, it should be recognized th at the w icked en v ironm ent of discontinuous
change is n o t tem porary. The KM process m eans co o p eratio n and com m itm ent to
change. O rg an izatio n s m ust act on w hat they learn and continually ad ap t to the
dynam ic environm ent to reap the benefits of KM initiatives. O nce initiated, there is no
 C HAPTER 1 Working Smarter; Not Harder i ■ i i 43

looking back at the classical ways of doing business. A sum m ary of the key challenges
is shown in Table 1.3.
A final challenge is how to deal w ith tacit know ledge. A s we shall discuss in
C hapter 2, organizational learning helps us deal with tacit knowledge. U nfortunately,
m ost of the tim e is spent dealing with explicit knowledge. Tacit know ledge is difficult to
express and, therefore, difficult to codify, transfer, or share.
Like any other initiative, with challenges come KM tools or m ethods that take vari-
ous forms. The m ost com m on tools include the Internet and intranets, data warehousing,
docum ent repositories, best-p ractice repositories, datab ase m ining tools, work-flow
tools, work-flow applications, and online application-processing tools (Taft 2000,14).
T hese tools connect people to people and people to in form ation on a global basis.
Briefly, two key technologies are of special im portance:
• Installing an intranet is the first K M tool that allows com pany employees to access
a wealth o f inform ation fro m anywhere at anytim e to conduct business. Active
know ledge m anagem ent assures accuracy, currency, reliability, and integrity of
stored know ledge round the clock.
• D eveloping a videoconferencing system m akes it practicable fo r remote knowledge
workers to initiate a face-to-face discussion over a telecommunication netw ork
Such a setup is ideal for sales staff in the field or w orkers encountering problem s
on an offshore oil rig. L ater in the text, we discuss the kind of technology
required to m ake this interface possible.
A n im portant question that is often asked is “How would one know that KM ini-
tiatives are taking p lace?” Several indicators are w orth watching in a business:
• Em ployees throughout the organization sharing the best practices—through data-
bases as well as through personal interactions
• A know ledge center th at prom otes know ledge skills and facilitates knowledge
flow for all em ployees to use in know ledge sharing
• A fine-tuned intran et and groupw are facilities to expedite inform ation and
know ledge flow at all times, regardless of time, distance, or location.
• W orking know ledge team s from all departm ents or divisions, whose focus is iden-
tifying, developing, and prom oting ways to apply KM throughout the organization
R elated to KM initiatives are KM user issues w orth considering. W ith the user in
m ind, we need to question the processes required for KM success, organizationally and
technologically, and how a KM system is m easured to transform individual and team
know -how into w hat the firm should “know ” O ne solution is to develop taxonom ies of

TABLE 1.3 Key Challenges in KM

Key Challenges in KM ____________________________________________________
• Explaining what KM is and how it can benefit a corporate environment
• Evaluating the firm’s core knowledge by departm ent and by division
• Learning how knowledge can be captured, processed, and acted upon
• Addressing the neglected area of collaboration
• Continuing research into KM to improve and expand its current capabilities
• Learning to deal with tacit knowledge
44 i i i a PART I The Basics

organizational know ledge for easy knowledge transfer and distribution. Success can be
assured with a proven im plem entation plan based on experience.

Historical Overview
K now ledge has been the staple source of com petitive advantage for m any com pa-
nies for hundreds of years. For example, the idea of passing knowledge to an apprentice
from a m aster was used extensively during m edieval times. Passing the “family recipe”
th at m akes a certain product unique from one generation to another also attests to the
no tio n of know ledge tra n sfe r and know ledge sharing. A lth o u g h such tra n sfe r was
extrem ely slow, it opened the door to m odern m ethods of knowledge m anagem ent that
can exploit faster m edia of know ledge exchange, such as the Internet.
The recorded history of knowledge dates back to Plato and Aristotle, but its m odern-
day u n d erstan d in g is cred ited to scholars like D aniel Bell (1973), M ichael Polanyi
(1958,1974), Alvin Toffler (1980), and the Japanese guru, Ikujiro N onaka (1995). O ther
writers like Sveiby (1997) and Stew art (2000) prom oted the concept knowledge as the
core asset of an organization. A s we shall find in C hapter 2, Polyani and N onaka iden-
tify two kinds of knowledge: Tacit know ledge is highly experiential and is found in the
heads of em ployees, customers, and vendors; explicit knowledge can be found in books,
docum ents, d ata banks, co rp o rate policy m anuals, and the like. The learning process
involves the in tersectio n of b o th kinds of know ledge and the resulting know ledge
transform ation process.
In the 1960s, D ru c k e r coined the term s “know ledge w o rk ” and “know ledge
w o rk er” w hen he was discussing the role of know ledge in organizations. H e was the
first to suggest that the U.S. econom y has shifted from an econom y of production to a
know ledge economy, w here the basic resource is knowledge, not capital. This m eans a
shift to m arket-driven as opposed to product-driven orientation (D rucker 1969).
In the early 1970s, researchers at M IT and Stanford were analyzing ways in which
com panies produced, used, and diffused knowledge. This was the first essential step in
the evolution of know ledge m anagem ent, as we know it today. The idea of knowledge
being a corporate asset had not yet caught on, and it was not until the 1980s that com -
panies truly began to value knowledge. They also began to find ways to m anage know l-
edge through such technologies as artificial intelligence and expert systems. E xpert sys-
tem s had their heyday in the 1980s and early 1990s, w hen developers focused on trying
to develop “thinking m achines” to em ulate hum an experts rath er than using m achines
to im prove hum an thin k in g —the essence of KM.
D uring the 1990s, the onset of the Internet, the inform ation superhighway, allowed
KM to take off. The In tern et facilitated access to publications about the concept of KM
and how to im plem ent it. W ith the help of the Internet, KM becam e a feasible concept
for m any com panies. It p ro v id ed m ore o p p o rtu n itie s for know ledge sharing and
know ledge transfer than there had been in the past. In term s of methodology, KM was
briefly presented in total quality m anagem ent (TQ M ) and team w ork. Business process
reengineering (B PR ), downsizing, and outsourcing w ere also attem pts to im prove the
perform ance of the firm, although they had lim ited success. They resolved the produc-
tivity factor, but drained know ledge from the organization.
These attem pts w ere m ore like round one, w here com panies m anaged their know l-
edge assets in the sam e way they m anaged physical assets. Physical goods were stored
in the w arehouse, but for the intellectual equivalent, it was in the know ledge reposi-
tory. W hen d atab ases and “w areh o u ses” w ere full (to o m any physical assets), they
began thinking about supply chain m anagem ent (SCM ), trying to m atch the supply of
goods with dem and and reduce inventories to w hat was actually ordered for produc-
 CHAPTER 1 Working Smarter; Not Harder a a 1 i 45

tion. It was m ore like rew arding the efficiency-driven prediction of the future based on
past tre n d s—doing things right. In contrast, in round two of KM, com panies began to
realize th at to fit the supply of knowledge to the dem and for it in products and services,
they needed to toy with how knowledge w orkers did their jobs. To be effective, KM has
to be “baked in to ” the job and be part of the fabric of the work to bring in knowledge
w hen n ee d e d and ex p o rt it anyw here in the o rg anization w hen it is acquired
(D avenport 1999,2). This is w here we began to see a shift from “doing things right” to
“doing the right thing” —working sm arter, not harder.
G iven the progress m ade in autom ating procedures in the 1970s and com m unica-
tions and netw orking (mostly through e-mail) in the 1980s, the focus of technology in
the 1990s was on cognitive com puting to augm ent the knowledge work of humans. O f
these, the In tern et and intranets have had the most profound im pact on spreading the
know-how. From a know ledge perspective, the Internet and accom panying technolo-
gies have dem onstrated several characteristics in knowledge m anagem ent:
• The Internet is an incredible in fo rm a tio n source. Internet access is available
worldwide. It m eans a com pany’s know ledge w orkers can access inform ation and
share know ledge anywhere, anyplace, anytime, w ithout delay
• W ith the W orld W ide Web, every user can share a n d update in fo rm a tio n at will. This
is especially attractive with the decreasing cost of com munications.
• The Internet uses a universal co m m u n ica tio n standard protocol. This protocol,
TCP/IP, m akes inform ation access and exchange accessible from anywhere there
is a com puter and an In tern et service provider.
• T he Internet p ro v id es qu icker interaction a n d co m m u n ica tio n with fe llo w k n o w l-
edge workers. This interaction can be one-on-one or as a group.

■■■« KM Myths
In contrast to what KM is, we have tried to clarify w hat KM is not. KM is not a separate
area or function in a business, represented by a KM departm ent or a KM process. KM
is interw oven into all of an organization’s processes. A lthough effective utilization of
technology is essential, KM is not constrained by collecting know ledge from dom ain
experts and building netw orked databases or databases supported by the com pany’s
intranet. Finally, KM is not defined in term s of the specific knowledge needs of every
employee, the relevant know ledge needed, or the knowledge to be shared.
There are several myths as well:
Myth 1: Knowledge m anagem ent is a fad. As m entioned earlier, there are m any pes-
simists in industry who doubt the “good fit” potential of KM. Being at a crossroads,
vendors push older softw are products under the KM label. B PR and artificial intelli-
gence had their positive turn, although they suffered from raised expectations. Unlike
earlier trends, however, true KM becom es em bedded in the way people work in busi-
ness. So, knowing w hat you know or w hat you need to know cannot be a fad.
Myth 2: Knowledge m anagem ent and data w arehousing are essentially the sam e.The
term data w arehousing implies a repository of data, not knowledge. Knowledge, p er se,
is how you take inform ation and transform it into action. D ata w arehousing is critical
for KM. It is w here data, critical docum ents, e-mail, and other forms of inform ation are
available for eliciting know ledge at the tim e w hen it is needed. For exam ple, Sears,
R oebuck & Co. has a custom er data w arehouse with dem ographic inform ation on over
100 million households to help the sales force im prove m arketing and sales quality. For
46 ■ a ■ a PART I The Basics

example, a repairperson working on a custom er’s refrigerator notices through the KM

system that such a custom er is a likely prospect for a new freezer. D ata mining serves a
sim ilar purpose, in th at p attern s within a mass of data allow m anagem ent to b etter
understand trends and directions in a product or consum er p reference—a necessary
dim ension of KM.

Myth 3: Knowledge management is a new concept. As a concept, KM has been practiced

by successful firms as far back as the early 1980s. Companies like Ford and General Motors
Corporation have been exchanging design information and collaborating on design proj-
ects worldwide using technology all along. Today’s version of KM goes under customer
profiling, where a superm arket clerk scans a store-generated customer card to determine
patterns of purchases and consumer preferences by date, by product, and by location. The
idea is the same as it has been for the past 2 decades, except that in today’s KM, technology
has taken on a special role in the way knowledge is shared and disseminated.

Myth 4: K nowledge m anagem ent is m ere technology. This is a serious misconception.

KM is really about people, relationships, and a new way of w orking together as an
entity in an organizational setting. It is a unique way of thinking about work and about
working. Imagine a knowledge com m unity of em ployees with com m on interests shar-
ing inform ation on best practices that help everyone do a m ore efficient job. KM will
work only if there is trust and confidence am ong coworkers. O ver 80 percent of all
technology-centered KM efforts have been known to fail because of a lack of attention
to people (W hiting 1999,44).

Myth 5: Technology can store and distribute human intelligence. D ata may be stored in a
centralized database for employee access, but that does not ensure that employees will use
the information. In a turbulent competitive environment, one cannot assume that compa-
nies can predict the right information for the right employee. So, it is hardly the case that
technology distributes human intelligence. It is impossible to build a KM system that pre-
dicts who the right person is and at what time he or she needs specific information for deci-
sion making. Tacit knowledge exists within a person’s brain; information or “knowledge”
stored within the database can be viewed as a valuable exchange between people to make
sense of a situation but should not be interpreted as human intelligence. In other words,
knowledge repositories stored in computers do not allow for renewal of existing knowl-
edge and creation of new knowledge. KM should be considered as a system to be used with
concentration on the human aspect aided by technology for decision making.

Myth (>; Knowledge m anagem ent is another form of reengineering. R eengineering is

efficiency-driven —a one-tim e attem pt at introducing radical change in organizational
processes to im prove efficiency. The emphasis is on cost reduction and making better
use of existing operations. Jum p-starting such a business, however, often results in fail-
ure. KM is an ongoing renew al of organizational processes to learn in advance about
the com pany’s future opportunities and contingencies. The concentration is on value-
added activities that dem and innovation and creativity. This is ingrained in the day-to-
day processes of the business. Technology plays a critical role in the way inform ation
becom es available at electronic speed.

Myth 7: Company employees have difficulty sharing knowledge. The answer is yes and
no, depending on a num ber of factors: attitude of the knower, who the requester is, com-
pany culture, sensitivity of the knowledge requested, availability of attractive motivators,
and trust level among company personnel. U nder the traditional business model, employ-
ees with unique knowledge accumulated over years of experience tended to protect “tu r f ’
 CHAPTER 1 Working Smarter; N ot Hardei 47

by not sharing such knowledge. In a KM environm ent, where knowledge sharing means
great potential for everyone including the organization, knowledge workers need to be
sold on how knowledge sharing will bring them m utual benefits. The term sharing means
“willingly giving away a p a rt” and “holding in common.” It is a “give to get” attitude, and
because “knowing” is personal, asking som eone to share is to ask him or her to give some-
thing of themselves. M ature or secure people in a stable work environm ent tend to share
knowledge m ore than others whose experience is to the contrary. Also, knowledge sharing
can improve bonds between people, provided the act of sharing is reciprocal.

Myth 8: K now ledge m anagem ent works only within an organization. O n the surface,
this may be true, but some of the m ost valuable know ledge comes from the outside —
suppliers, brokers, governm ent agencies, and customers. The problem with extending
KM initiatives to outside sources is incom patible technology, security issues, and com -
plexity of the design.

Myth 9: Technology is a b etter alternative than face-to-face. We have seen over the
years that when it comes to real-life experience and use of hum an knowledge, technol-
ogy does not hold all the answers. The emerging mind-set within today’s forward-looking,
creative organizations is th at KM m ust entail cultural and organizational change as
well as technology-based innovations. D ata w arehousing and data mining are all con-
trib u to rs to ex tractin g and sharing know ledge, b u t the best know ledge resides in
h um an minds. This m akes a face-to-face ap p ro ach to know ledge acquisition and
know ledge sharing a b etter alternative.

Myth 10: It is a “no brainer” to share what you know. In general, secure and m ature peo-
ple are less reluctant to share what they know with others. U nfortunately in traditional
business, people with years of experience tended to hoard knowledge rather than share it,
because it gave them leverage, control, and assurance of a job. Furtherm ore, “knowing” is
personal. To ask people to share know ledge is tan tam o u n t to expecting them to give
something of themselves. Sharing knowledge often depends on who the requester is* how
sensitive is the knowledge requested, the attitude of the “knower,” and the motivational
forces at play. To share knowledge, the business has to undergo special employee training,
instill trust within the business, and give employees and m anagem ent a chance to cement
relationships based on trust. A summary of the KM myths is shown in Table 1.4.

TABLE 1.4 Myths About Knowledge Management

The Myths o f Knowledge Management___________________________________
1. Knowledge management is a fad.
2. Knowledge management and data warehousing are essentially the same.
3. Knowledge management is a new concept.
4. Knowledge management is mere technology.
5. Technology distributes human intelligence.
6. Knowledge management is another form of reeingineering.
7. Company employees have difficulty sharing knowledge.
8. Knowledge management works only within an organization.
9. Technology is a better alternative than face-to-face.
10. It is “no brainer” to share what you know.
a a a a a a
48 ■ aa« PART I The Basics

■« » « KM Life Cycle
KM goes th ro u g h a series of steps, m aking up an ongoing life cycle. The four-step
process, sum m arized in Table 1.5, includes gathering, organizing, refining, and dissem -
inating. These and alternative approaches to the KM life cycle are explained in detail
in C hapter 3.
The capturing phase deals with knowledge capture and includes e-mail, audio files,
digital files, and the like. In this phase, it is im portant to go to all the sources available
and never judge the usefulness of the captured know ledge until after it is subjected to
exhaustive testing. In this phase, KM systems are an ideal approach to eliciting and rep -
resen tin g know ledge in to a form th a t can be available to m any users —a key KM
process. This process is discussed in detail in later chapters.
A fter the capturing phase, captured data or inform ation should be organized in a
way that can be retrieved and used to generate useful knowledge. O ne can use index-
ing, clustering, cataloging, filtering, codifying, and other m ethods to do the organizing.
Speed, user-friendliness, efficiency of access, and accuracy are im portant elem ents to
consider throughout the organizing phase.
A fter organizing the inform ation, it should be refined. D ata m ining can be applied
in this phase. D ata m ining takes explicit know ledge found in databases and transform s
it into tacit knowledge. D ata-m ining software is used to find patterns in data, predict
behavior, and w arn against future problem s based on the data supplied in data w are-
houses. For exam ple, the sales records for a particular brand of tennis racket might, if
sufficiently analyzed and related to o th er m arket data, reveal a seasonal correlation
with the purchase by the sam e parties of golf equipm ent.

TABLE 1.5 Four-Process View of

Knowledge Management
Four-Process View o f KM
• Capturing —D ata entry
—Scanning
—Voice input
—Interviewing
—Brainstorming
• Organizing —Cataloging
—Indexing
—Filtering
—Linking
—Codifying
• Refining —Contexualizing
—Collaborating
—Compacting
—Projecting
—Mining
• Transfer —Flow
—Sharing
—A lert
-P u sh
 C HAPTER 1 Working Smarter, Not Harder ■ ■ a ■ 49

A fter the refining phase, know ledge should be dissem inated or transferred. This
includes m aking know ledge available to em ployees via tutorials or guidelines for effec-
tive use. Predictive m odels can be designed to alert users to consequences of certain
projects or hum an resource activities. The key point is not to let stored or available
know ledge sit idle in a repository like a database. It should be available to authorized
users to contribute to the corporate com petitive advantage.
Figure 1.6 shows the relationship betw een the KM life cycle and four key areas in
th e org an izatio n . T aken to g e th e r, one can u n d e rsta n d a viable set of relationships
betw een KM and m anagem ent decision making, organizational culture, organizational
personnel, and inform ation technology. A s we shall see in C hapter 3, each area could
have an im pact on the way KM is installed and m aintained in the organization.

Role of Trust in the KM Life Cycle

K now ledge-based organizations are learning hum an systems. O ne of the require-
m ents th a t en courage the flow of know ledge is establishing an environm ent w here
em ployees feel free to share insights, experiences, and know-how. This boils down to
the issue of trust. Trust supports the KM process by giving em ployees clear impressions
that reciprocity, free exchange, and proposing innovations will be recognized and fairly
com pensated. In contrast, lack of tru st encourages em ployees at all levels to hoard
know ledge and build suspicion in people and organizational processes.
Trust m eans integrity, consistent com m unication, and proven willingness of the
organization to integrate em ployees into the decision-m aking process. For trust to be
effective, an organization m ust m ake an overt effort to em bed trust in the culture of
the business. This m eans tackling the environm ent of trust from' several angles: hum an
resources, m an ag em en t initiatives, top m anagem ent su pport, and d em onstrating to
know ledge em ployees th at they belong. It becom es the cem ent th at glues an organiza-
tion to g eth er in its effort to address an unpredictable com petitive environm ent. The
bottom line is th at one cannot have an open, candid dialogue with som eone he or she
does not trust. If one trusts his or her peers, then there should be no concerns about
w here they stand.
W ith trust in m ind, the key question is how does a com pany develop trust? First,
trust should be built at all levels of the organization. A sum m ary view is to break down
50 ■ i ■ * PART I The Basics

the barriers created by traditional organizational structures. R ather than managing by

controlling and closely supervising employees, the alternative is to em pow er them and
open lines of com m unication betw een m anagers and workers. D uring the annual p er-
form ance evaluation session, each em ployee should be given a chance to explain what
he or she did to build a sense of com m unity in their departm ent, division, or the com -
pany as a whole. Some of the suggestions offered include:

• D ecentralizing or flattening the organization structure to allow decision making

by team w ork
• R educing control-based m anagem ent and encouraging m anagem ent by results so
that em ployees and team s can get their rightful credit for contribution to the
com pany’s profitability
• Revisiting the com pany’s mission statem ent and ethics policy to dem onstrate the
com pany’s new views about organizational values and opening lines of com m uni-
cations through periodic em ployee satisfaction surveys, brainstorm ing sessions,
and the like
• Assessing and im proving em ployee responsibilities and accountability on a com -
pany-wide basis; elim inating unnecessary directives or barriers
• Recognizing em ployees through bonus program s, knowledge em ployee of the
m onth, on the em ployee bulletin board, or in the new spaper
• Installing an em ployee training program , especially geared for improving
em ployee com m itm ent to knowledge sharing

It can be concluded, then, that trust is critical to a successful KM program . D espite

its value, it continues to be a most difficult intangible to create and m aintain, especially
in large corporations like Ford, G eneral M otors, IBM, and others. The potential is in its
lasting effect once it is p ro p erly em b ed d ed into the o rganization and m aintained
within the corporate policies that m ake sense for KM.

■ Implications for Knowledge Management

A ny discussion on know ledge m anagem ent is bound to produce both optim ists and
doubters. H ow ever, we know now that the source of organizational w ealth is som e-
thing specifically hum an —knowledge. Applying knowledge to what we already do con-
tributes to productivity. In contrast, applying know ledge to what is new is innovation.
K now ledge is the key to both goals. This will be the focus of C hapter 2.
The rap id ad v an cem en t of know ledge exchange tools (such as the In tern et,
intranets, and groupw are) is setting the stage for knowledge m anagem ent. The great
potential of this em erging concept lies in how well we share the “know -how ” to aug-
m ent the decision-m aking quality and capacity of the organization. For those whose
product is knowledge, know ledge w orkers are distinguished from inform ation handlers
by their ability to solve unique problem s with renew ed effectiveness and add value in
an unpredictable com petitive environm ent. The em phasis is on augm enting the abili-
ties of the hum an u se r—working sm arter, not harder. They use their knowledge rather
than data or inform ation.
In m ost corporations, the m anagem ent of know ledge is still uncharted territory.
Few executives u n d erstan d how to m ake the m ost effective use of their com pany’s
know ledge. They may know about tangible assets like the physical plant or existing
inventory, but they have trouble putting a price tag on the value of the talent or exper-
tise in the sales force.
 CHAPTER 1 Working Smarter; Not Harder a a a a 51

To p rom ote a know ledge m anagem ent environm ent, com panies m ust consider cul-
tural, social, and organizational changes as well as technological support. They m ust
also add systems and applications in their IT infrastructure in the interest of integra-
tion and know ledge sharing. For this reason, m ore and m ore com panies are beginning
to recognize the im portance of know ledge m anagem ent systems. This area is covered
later in the text.
O ne area th at a com pany m ust address is how to m anage its intellectual capital—
the sum of know ledge of the hum an resources of the organization. As we shall find in
C hapter 2, the know ledge found in policy m anuals, case histories, training m aterials,
and em ployees’ heads is the m ost valuable asset of the com pany in term s of replace-
m en t costs. To d evelop and m ain tain this know ledge is extrem ely costly. U nless a
com pany review s its core know ledge and takes steps to m anage it, it is subject to
p o ten tial disaster resulting from resignations, turnover, loss of its com petitive edge,
and the like.
Finally, in the grand schem e of things, the trend tow ard know ledge sharing is good
for cow orkers, for the com pany, and for society as a whole. The key converter from
hoarding know ledge to sharing know ledge is trust. To gain a com petitive advantage in
an unpredictable environm ent, we m ust learn to m aster the art of knowledge sharing
and know ledge m anagem ent. E fforts can be realized only if there is follow-through.
C o rp o rate culture m ust change first. K now ledge is not a technology, but an activity
enabled by inform ation technology and produced by people. Com panies must commit
to changing their corporate structure by assigning know ledge workers, whose responsi-
bilities include m otivating and organizing the corporate, to adapt to a new way of busi-
ness that transcends the entire landscape.

Summary 1 1 1 1
• Today’s know ledge age brings with it an increasingly complex business environ-
m ent worldwide. C om panies need to address know ledge sharing, knowledge
collaboration, and know ledge dissem ination to be able to com pete in an un-
predictable m arketplace.
• KM is the process of gathering and m aking use of a firm ’s collective expertise
anyw here in the business. A firm seeks to add value by creatively identifying,
applying, and integrating know ledge in unprecedented ways.
• KM is not about reengineering, discipline, or one based on inform ation only. It is
not a fad or a way of appeasing em ployees around an em ployee-oriented concept.
It is about change, and tom orrow ’s business is the nam e of the game. People are
the determ inants of KM success.
• T here is plenty of justification for KM. K now ledge sharing m eans faster process-
ing, no duplication of effort, and reacting m ore responsively to new business
opportunities. It also m eans ensuring successful partnering and core com petencies
with suppliers, vendors, custom ers, and other constituents.
• Com panies that fail to install a KM environm ent could be guilty of using the old
rew ard system, having a database that lacks focus, having poor coordination of
the IT or hum an resources effort, choosing the w rong project, and having poor
leadership.
• There are several drivers that provide a com pelling case for adopting KM: tech-
nology, process, personnel-specific, know ledge-related, and financial drivers. In
the final analysis, the goal of KM is to produce a positive return on investm ent in
people, processes, and technology.
52 m i PART I The Basics

• Several factors triggered interest in KM: need for innovation, globalization, easier
navigation with telecom m unications, and the increasing dom inance of knowledge
as a way of im proving efficiency and effectiveness of the firm.
• KM dates back to Plato and A ristotle, although its m odern-day understanding
began in the 1960s. D uring the 1990s, the onset of the Internet, the inform ation
superhighway, allowed KM to take off. Q uicker interaction and com m unications
with fellow know ledge workers, w hether one-on-one or as a group, m ade it all
possible.
• Several KM m yths have been cited in an effort to clarify what KM is not. Once
the concept becom es widely accepted and the KM process becom es standardized,
m ost of these myths should disappear.
• The KM cycle essentially begins with knowledge gathering, followed by know l-
edge organization, know ledge refinem ent, and know ledge transfer. The role of
trust is critical throughout the entire life cycle. In C hapter 3, the KM life cycle,
including know ledge m aintenance and update, will be discussed in detail.

Terms t o K n o w 1 1 1 1
Downsizing: Reducing the physical, personnel, and func- knowledge workers to serve anywhere in the organiza-
tional processes of an organization in an effort to tion and minimize personnel turnover as a threat to
improve its efficiency through reduced costs and collective knowledge.
improved performance. Process driver: A driver or a motivator designed to
Financial driver: A driver or a motivator that views improve work processes through KM.
knowledge assets as something increasing in value as Reengineering: One-shot drastic “electrical shock”
more and more people make use of it. change in organizational processes to improve effi-
Intellectual capital: The value of a company’s trademarks, ciency; a mechanical shift from one stage of operation
patents, brand names, and the like; a company’s collec- to a more efficient stage.
tive brainpower or a composite of experience, knowl- Supply chain management: The integration of all activi-
edge, information, and intellectual property. ties associated with the flow and transformation of
Internet: A system of interconnected data communication goods from the raw materials to the customer as well as
or computer networks on a global basis. the related information flows.
Knowledge management: The process of gathering and Technology driver: A driver or a motivator designed to
making use of a firm’s collective expertise wherever it use technology for storing and transferring knowledge
resides—on paper, in databases, or in people’s heads. throughout the organization.
Knowledge-related driver: A driver or a motivator that Trust: A feeling of confidence in another person;
relates to the concept of knowledge sharing and knowl- having confidence or faith in another person or in
edge transfer within the firm. a relationship.
Knowledge-sharing: A process of transferring human Videoconferencing: A computer-based system designed
knowledge about a process or a procedure to others in to simulate face-to-face meetings with two-way full-
the organization; ability and willingness of people to motion video along with two-way audio. Participants
exchange specialized experience with others for the can view one another during long-distance conferences
common good of the organization. or meetings.
Personnel-specific driver: A driver or a motivator that
focuses on the need to create cross-functional teams of

Test Y o u r U n d e rsta n d in g * * 1 1
1. Select one definition of KM in Figure 1.2 and explain the reason(s) for your
choice.
2. KM involves people, technology, and processes in overlapping parts. Explain
the KM concept.
3. O ne u n iq u e in d icato r of KM in action is seeing peo p le think ahead, not
behind. D o you agree with this statem ent? Explain why you agree or disagree.
 CH A PTER 1 Working Smarter, Not Harder a a a ■ 53

4. In w hat way is KM n o t about:

a. reengineering
b. a discipline
c. data
d. know ledge capture
5. D istinguish betw een:
a. KM and intellectual capital
b. a cham pion and a m anager
c. the In tern et and an intranet
d. trust and know ledge sharing
6. W hat factors justify the adoption of KM into an organization?
7. W hat oversights or pitfalls impede the introduction of KM into an organization?
8. Briefly explain the factors that triggered interest in KM.
9. The chapter explains several challenges in KM. Select three challenges and
elaborate on each.
10. H ow w ould one know that KM initiatives are taking place?
11. Sum m arize in your own words the history leading to today’s KM.
12. H ow does the In tern et contribute to the use of KM in organizations?
13. Explain five KM m yths and the reasons for viewing them as myths.
14. List and briefly explain the key steps in the KM life cycle.
15. H ow does a com pany develop trust? Be specific.

K n ow led ge Exercises * a 1 *
1. Suppose you w ere asked to do a 15-minute presentation before the m anagers
of a small retailer about the pros and cons of knowledge m anagem ent. W hat
would you say? O utline the content of your talk.
2. H ow easy do you think it is to understand know ledge m anagem ent? Why?
3. S earch th e In te rn e t and c u rren t jo u rn a ls for surveys th a t show how well
com panies are adopting (or struggling w ith) KM. R ep o rt your findings in
class.
4. H ow should a com pany assess its know ledge core? D etail a procedure that
can be sold to top m anagem ent.
5. W hat do you think distinguishes to d a y ’s know ledge age from th a t of the
inform ation age of the 1980s and 1990s?
6. Search the literature, including the Internet, and prepare a three-page report
on som e of the problem s in developing and im plem enting a know ledge m an-
agem ent system.
7. In your own words, describe the im plications th at a know ledge m anagem ent
system has on an industry of your choice.
8. W orking w ith a classm ate, conduct an interview or a telep h o n e survey
regarding a local business’s
a. fam iliarity with KM
b. level of literacy in KM
9. A business m anager, a program m er, and a psychologist all w ant to becom e
KM designers. W hich do you feel will have the least difficulty? Why?
10. A car dealer who has just learned about know ledge m anagem ent thinks it
could be ideal for separating “tire-kickers” from serious buyers. W ould this
be a typical KM application? W hy?
11. V isit a sm all business (such as a h ard w are sto re or a bak ery ) and stay
long enough to get a feel for the info rm atio n exchanged, custom er q u es-
tions, and the way the business is run. W rite a short rep o rt about the core
know ledge of th e business. A re th e re know ledge w orkers th ere or just
clerks?
54 aaaa PART I The Basics

12. Find cases or current inform ation on trends in KM by searching the Internet.
Locate a com pany that is reported to have succeeded (or failed) in going the
KM route. W rite a three-page report sum m arizing your findings.
13. Visit a local bank (not a branch) and identify know ledge workers. W hat is
unique about know ledge w orkers? H ow do you distinguish them from regu-
lar em ployees?

R ef er en c es a a a a
Applehans, Wayne, Globe, Alden, and Laugero, Graig. Man- Newman, B. “Managing Knowledge: The HR Role,” HR
aging Knowledge. Boston, MA: Addison-Wesley, 1999. Executive Review, vol. 6, no. 4,1999, pp. 3-18.
Bair, Jim. “Knowledge Management Technology: Beyond Nonaka, Ikujiro. The Knowledge-Creating Company.
Document Management,” www.strategy-partners.com/ London: Oxford University Press, 1995.
spweb/pages.nsf/all+pages/White %20Paper%20% O ’Dell, Carla S., Essaides, Nilly, Ostro Nilly (contributor),
20Knowledge%20Management%20Technology: C. Grayson (preface). I f Only We Knew What We
% Beyone% 20Document%20M anagement?open Know: The Transfer o f Internal Knowledge and Best
document, Date accessed August 2002. Practice. New York: Free Press, 1998.
Barth, Steve. “Learning from Mistakes,” Knowledge Man- Polanyi, Michael. Knowing and Being. Chicago: The
agement, April 2001, pp. 41^17. University of Chicago Press, 1974.
Bell, Daniel. The Coming o f Post Industrial Society. Polanyi, Michael. Personal Knowledge: Toward a Post-
London: Oxford University Press, 1995. Critical Philosophy. Chicago: The University of
Bexar County. “Information Technology Industry Trend,” Chicago Press, 1958.
www.co.bexar.tx.us/bcis/Documentation/DOC_It_ Robb, Drew. “Weaving Two into One,” Knowledge
Industry_Trends/doc_it_industry_trends.htm, Date Management, April 2001, pp. 33-38.
accessed August 2002. Skyrme, David J. “Knowledge Networking,” The
Botkin, James. Smart Business: How Knowledge Intelligent Enterprise, November 1999, pp. 9-15.
Communities Can Revolutionize Your Company. Skyrme, David. “Knowledge Management: Making It
New York: Free Press, 1999. Work,” www.skyrme.com/pubs/lawlib99.htm, Date
Brooking, A. Intellectual Capital. London: Thompson accessed August 2002.
Press, 1996. Stewart, Kathy, et al. “Confronting the Assumptions
Craig, Robert. “Knowledge Management Comes of Age.” Underlying the Management of Knowledge: An
ENT, vol. 5, April 2000, pp. 36-37. Agenda for Understanding and Investigating
Davenport, Tom. “Knowledge Management, Round Two,” Knowledge Management,” The Data Base for Advances
CIO Magazine, November 1,1999, pp. 1-4. in Information Systems, Fall 2000, pp. 41-45.
Deveau, Denise. “Knowledge Is Business Power,” Com- Stewart, Thomas A. Intellectual Capital: The New Wealth
puting Canada, vol. 26, no. 8, April 14,2000, p. 14. o f Organizations. New York: Doubleday Press, 1997.
Drucker, P. F. The Age o f Discontinuity. New York: Sveiby, Karl-Erik. “Intellectual Capital and Knowledge
H arper Publishing, 1969. Management,” www.sveiby.com/articles/Intellectual
Glasser, Perry. “The Knowledge Factor,” CIO Magazine, Capital.html, Date accessed August 2002.
January 1999, pp. 12-18. Sveiby, Karl-Erik. The New Organizational Wealth:
Gravallese, Julie. “Knowledge Management,” The M IT R E Managing and Measuring Knowledge-Based Assets.
Advanced Technology Newsletter, vol. 4, no. 1, April New York: Berrett-Kohler, 1997.
2000, pp. 1-3. www.mitre.org/pubs/edge/april_00, Date Taft, Darryl. “Stopping Knowledge Overflow —
accessed August 2002. Knowledge Management Tools Still in Developmental
Hersey, Ian. “Need: Net-Speed Knowledge Tools,” Phase,” Computer Reseller News, February 2000, p. 14.
Electronic Engineering Times, September 4,2000, p. 90. Taylor, Christie. “Intellectual Capital,” Computerworld,
Hibbard, Justin. “Knowing What We Know,” March 12,2001, p. 51.
InformationW eek, O ctober 20,1997, p. 46. Toffler, Alvin. The Third Wave. New York: Bantam
Johnson, Cindy. Director of Collaboration and Knowledge Books, 1980.
Sharing at Texas Instruments. Interview February 2,2002. Trepper, Charles. “Keep Your Knowledge In-House,”
KPMG Consulting. Knowledge Report 2000, November 1999. InformationW eek, September 4,2000, p. 55.
Malhotra, Yogesh. “Knowledge M anagement for the New Whiting, Rick. “Myths and Realities,” InformationW eek,
World of Business,” www.brint.com/km/whatis.htm, November 22,1999, p. 42ff.
D ate accessed August 2002.
 Understanding
Knowledge
Contents

In a N utshell
D efinitions
Knowledge
Intelligence
E xperience
C om m on Sense
C ognition and Know ledge M anagem ent
D ata, Inform ation, and K now ledge
D ata
Inform ation
K nowledge
Types of Knowledge
Shallow and D eep Knowledge
Know ledge as K now -H ow
Reasoning and H euristics
C om m on Sense as K now ledge
From Procedural to Episodic K nowledge
Explicit and Tacit Know ledge
E xpert Knowledge
C hunking Knowledge
K now ledge as an A ttrib u te of Expertise
H um an Thinking and L earning
H um an L earning
L earning by E xperience
Learning by Exam ple
Learning by D iscovery
Im plications for K now ledge M anagem ent
Sum m ary
Terms to Know
Test Your U nderstanding

55
i ■ PART I The Basics

Knowledge Exercises
R eferences

I think, therefore I am.

—R e ne D esc a r t es

In a Nutshell
The most critical w ord in the KM area is knowledge. U nfortunately, there is a continu-
ing m yth that know ledge resides only in books, reports, or documents. This cannot be
true, any m ore than viewing musical notes on a page constitutes music. W hat we have
h ere are rep resen tatio n s of inform ation and music, respectively. W hat do we m ean
when we talk about know ledge? Two key issues are distinguishing betw een knowledge
and inform ation and determ ining how they are interrelated. Knowledge is neither data
nor inform ation, although it is related to both. The term s are not interchangeable, and
knowing w hat is needed often determ ines organizational success or failure.
This chapter exam ines the concept of knowledge as the heart of an organization’s
p roductivity and grow th and discusses the classifications of know ledge. In tellig en t
behavior implies the ability to u nderstand and use language and to store and access
relevant experience at will. H um ans acquire expertise—that is, they learn via experi-
ence. E x p ertise in c o rp o ra te s the ability to reaso n and to m ake deductions; it also
includes the concept of com m on sense. This m akes hum an intelligence and the know l-
edge that hum ans amass over time the prim ary organizational asset. A n organization’s
technology or telecom m unications netw ork is only a vehicle for know ledge transfer
and know ledge exchange; it cannot replace hum an knowledge.
A n essential criterion of knowledge and learning is memory. Learning by discov-
ery is less understood than learning by experience or by example. Several approaches
to learning are covered later in the chapter. A knowledge base is a critical com ponent
of know ledge m anagem ent. K nowledge developers need 10 understand the theory and
m eaning of know ledge early in the knowledge capture phase and becom e fam iliar with
the unique kinds of know ledge available in the corporation under study. It should be
pointed out th at an ex p ert’s know ledge is not lim ited to inform ation or complex pro -
cedures. K now ledge em braces a w ider sphere than inform ation. Likewise, a knowledge
base is not the sam e as a database. A database has a predeterm ined structure; a know l-
edge base is a set of facts and inference rules for determ ining new inform ation and
“sm arter” know ledge for decision making.
K now ledge as know -how may be either shallow or deep knowledge. It may also be
p rocedural, declarative, sem antic, or episodic know ledge. M ore recently, know ledge
has been classified as explicit or tacit knowledge. All these classifications are covered
in the chapter.
A n o th e r goal of this ch ap ter is to show th at hum an know ledge and logical p ro -
cesses can be captured and rep resented in a valid way to solve certain kinds of p ro b -
lems. (The w ords valid and validity as used in the text do not have the same m eaning
as truth. A conclusion, for exam ple, follows logically from facts and rules using infer-
ence procedures; it can be a false, or not correct, conclusion, even though it may still be
a valid one.)
 C HAPTER 2 Understanding Knowledge a i i ■ 57

m i Definitions
B efo re discussing know ledge and its m any ram ifications, kn o w le d g e needs to be
defined in relation to intelligence, experience, and c o m m o n sense.

KNOW LEDGE
We define k n o w le d g e as “u n d erstan d in g gained th ro u g h experience or study.” It is
“know -how ” or a fam iliarity with how to do som ething th at enables a person to p er-
form a specialized task. It may also be an accum ulation of facts, procedural rules, or
heuristics. These elem ents are defined as follows:

• A fact is a statem ent of some elem ent of truth about a subject m atter or a domain.
For example, milk is white and the sun rises in the east and sets in the west are facts.
• A procedural rule is a rule th at describes a sequence of relations relative to the
main. For exam ple, always check the traffic when entering a freeway; if the gas
gauge indicates less than a q u arter of a tank of gas, look for a gasoline station.
• A heuristic is a rule of thum b based on years of experience. For example, if a p er-
son drives no m ore than 5 miles above the speed limit, then that person is not
likely to be stopped for speeding.

A beneficial aspect of knowledge is that it can com pensate for some search time. A
hum an expert who knows a set of solutions can get a job done without much searching for
information. Conversely, a hum an novice in a video game searches a vast num ber of alter-
native moves at each juncture because he lacks experiential knowledge. Unfortunately,
without the aid of knowledge that allows the novice to immediately eliminate inappropri-
ate approaches, this m ethod encompasses too many approaches to evaluate.
A n o th er aspect of know ledge is specificity; it cannot be transferred from one pro b -
lem dom ain to another. Therefore, one m ust have the surgeon’s know -how to repair a
heart valve, the auto transm ission specialist’s know -how to replace a reverse gear, and
the p a in te r’s know -how to create an accom plished portrait. These kinds of extensive
know ledge are referred to as tacit kn o w le d g e and often take m any years to acquire.
Finally, values, beliefs, and integrity are related to knowledge. This has a lot to do
with w hat the know er perceives, accepts, and concludes from the environm ent. People
g enerally o rganize and synthesize th e ir know ledge by th e ir values. N o n ak a and
Takeuchi suggest th at “know ledge, unlike inform ation, is about beliefs and com m it-
m e n t” (N onaka and Takeuchi 1995). M ore recently, we began to attach integrity to the
w hole process of know ledge capture, know ledge sharing, and know ledge m aintenance.
Integrity m eans reliability, trustw orthiness, privacy, and confidentiality. Integrity cuts
across the discipline, regardless of com pany size or resources.

INTELLIGENCE
Intelligence refers to the capacity to acquire and apply knowledge. It is the ability to build
or improve upon knowledge, to transform as much of on e’s knowledge as possible into
knowledge that can be used to m ake good decisions. A n intelligent person is one who has
the ability to think and reason. This distinction separates the novice from the m aster in a
game like chess. Knowledge conversion is directly responsible for much of the expert’s
efficiency in applying knowledge and for the difficulty of m aking it explicit.
C onsider this exam ple: R ecent research into the true m eaning of intelligence illus-
trates very well the difficulty of defining the term . The organization doing this research
decided to get to the bottom of the question once and for all and, given its im portance,
58 a a a i PART I The Basics

assigned its m ost senior scientist to it. The esteem ed scholar spent several m onths con-
ducting this research. A t the end of that period, the scientist gathered a num ber of col-
leagues together, held up in front of them the artificially intelligent artifact chosen as
the subject of the research, and said, “Ladies and gentlem en, this is a therm os bottle. It
keeps hot stuff hot, and it keeps cold stuff cold. My question is, how does it know ?”
A bility to u n d e rsta n d and use language is a n o th e r a ttrib u te of intelligence.
Language understanding is not easy to acquire, especially for the existing technology.
For exam ple, consider the statem ent, T h e city o f F a irm o u n t is u n d er 6 fe e t o f water:
D oes this m ean th at the city is com pletely underw ater, with the tallest building below
the w ater level? A n o th er exam ple is the statem ent, T he su n b ro k e through the clouds.
H ow literally should one in terp ret this statem ent? O f course, both the m eaning of the
words and the context of the statem ents determ ine how a read er should understand
the messages. Prior know ledge and com m on sense also en ter the picture.
Memory, or the ability to store and retrieve relevant experience at will, is part of
intelligence. H ow th e b rain sto res and retriev es in fo rm atio n or know ledge is still
unclear. Later, the text includes a discussion of knowledge organization and how it is
exploited in the KM building life cycle.
Learning is know ledge or skill th at is acquired by instruction or study. It is the
inevitable consequence of intelligent problem -solving. Intelligent people learn quickly
and m ake effective use of w hat they learn. Inasm uch as problem -solving and knowl-
edge organization have been successfully dem onstrated in the business enterprise, the
sam e success has yet to be shown in technology or com puter programs. People learn
from experience; to date, com puters have not.

EXPERIENCE
Experience relates to w hat we have done and what has historically happened in a spe-
cific area of work. In Latin, the w ord experience m eans “to put to the test.” People with
deep know ledge in a given subject have been tested by experience. Experience also
leads to expertise. Think of Sherlock H olm es investigating a m urder. The goal is to find
the m u rderer. H o lm es’s reasoning and deductions rely on all evidence collected; he
w orks backw ard from the goal until the suspect is caught. E xpertise is also intuition
and the ability to access o n e ’s knowledge rapidly to achieve an efficient and successful
outcom e.
E x p erien ce is closely re la te d to know ledge. K now ledge develops over tim e
through successful experience, and experience leads to expertise. A n expert is som eone
who knows w hat he or she does not know and is the first one to tell you so. Firms hire
experts to benefit from th eir experience and proven know ledge in solving com plex
problem s. P eo p le use ex p erien ce to change facts into know ledge, which se p arates
novices from experts. E xceptions do occur, how ever. Bach, for exam ple, was expert
m usician at 5 years of age. In general, w ithout experience, one would hardly be consid-
ered an expert. E xperience in using know ledge allows people to refine their reasoning
processes in a know ledge m anagem ent environm ent.

CO M M O N SENSE
Common sense refers to the unreflective opinions of ordinary humans, which comes
naturally to a child as young as 3 or 4 years old. For example, m ost youngsters know
th at if they touch a hot stove, they will get burned. In contrast, a com puter could be
told all kinds of things about hot stoves and the effect of heat on the hum an skin, and it
still would not perceive w hat would happen if it “to uched” a hot stove: M achines lack
com m on sense. C om m on sense is not easily learned or acquired
 C H A PTER 2 Understanding Knowledge ■■■i 59

TABLE 2.1 Basic Knowledge-Related Definitions

Artificial Emulating or imitating something natural or real
Common sense Innate ability to sense, judge, or perceive situations; grows stronger
over time
Fact A statem ent that relates a certain element of truth about a subject
m atter or a domain
Heuristic A rule of thumb based on years of experience
Intelligence The capacity to acquire and apply knowledge; ability to understand and
use language; ability to store and retrieve relevant experience at will;
learning from experience
Knowledge Understanding gained through experience; familiarity with the way to
do something to perform a task; an accumulation of facts, procedural
rules, or heuristics
Procedural rule A rule that describes a sequence of relations relative to the domain
■ a i a ■ a

Lack of com m on sense m akes technology “b rittle ”; th at is, com puters rarely go
beyond the scope of their data w arehouse or know ledge base. M any im portant projects
assum ed by hum ans in business today require com m on sense, which is only partially
understood by to d ay ’s com puter. Table 2.1 sum m arizes these concepts.

a a l ■ Cognition and Knowledge Management

Cognitive psychology provides an essential background for understanding knowledge
and expertise. The goal of cognitive psychology is to identify the cognitive structures
and processes th at relate to skilled perform ance within an area of operation. Cognitive
science in general is the interdisciplinary study of hum an intelligence. Its two m ain
com ponents are experim ental psychology, which studies the cognitive processes that
constitute hum an intelligence, and artificial intelligence, which studies the cognition of
com puter-based intelligent systems. O ur focus in this text is on the first com p o n en t—
how cognition contributes to the perform ance of intelligent workers.
W ith these relationships in m ind, one can see cognitive psychology’s contribution
to KM. U n d erstan d in g the lim itations and biases provided by cognitive psychology
helps in u nderstanding expertise. H um an lim itations —such as m em ory capacity and
the physical limits im posed by hum an sensory and m otor system s—m ust be considered
w hen attem pting to understand how the hum an ex p ert‘carries out a task.
The process of eliciting and rep resen tin g expert know ledge typically involves a
knowledge developer and one or m ore hum an experts. To capture hum an knowledge, the
developer interviews the expert(s) and asks for inform ation regarding a specific area of
expertise that the expert is adept at solving. The expert may be asked to “think aloud,” to
verbalize his or her thought processes, while solving the problem . People cannot always
give com plete, accurate reports of their m ental processes. Experts may have greater dif-
ficulty in conveying some kinds of knowledge, such as procedural knowledge (explained
later in the chapter). Psychologists have long been aware of problem s related to verbal
reports, and through research, they have developed m ethods for circumventing them.
Cognitive psychology research contributes to a b ette r understanding of w hat con-
stitutes know ledge, how know ledge is elicited, and how it should be represented in a
co rp o rate know ledge base for others to tap. B ecause know ledge developers should
60 m i PART I The Basics

take knowledge elicitation (also called knowledge capture) seriously, they should have
a strong educational and practical background in cognitive psychology and cognitive
processes. Knowledge capture techniques are covered in Chapter 4.

■■«« Data, Information, and Knowledge

The main benefit of discussing knowledge this early in the text will become obvious
later, when the analysis turns to knowledge workers and how knowledge developers
capture and codify knowledge from human experts. Knowledgeable experts tend to be
adept at explaining how they arrive at decisions or solutions because they have years of
experience. They have become experts by adding and refining knowledge, not just by
capturing and storing information. Likewise, knowledge developers, whose job is to
acquire and represent experts’ knowledge, need to understand the many ramifications
of knowledge early in order to decide whether a particular expert has the requisite
knowledge for building or updating a reliable knowledge base.

DATA
D ata are unorganized and unprocessed facts. They are static; they just sit there. For
example, John is 6 feet ta ll This is data; it does not necessarily lead one anywhere..
However, the meaning one brings to the evaluation of this data could be important.
Such an evaluation may indicate that John’s height would make him an asset to the
basketball team. This becomes information.
D ata is a set of discrete facts about events —structured records of transactions.
When a customer goes to the store and buys merchandise, the number of socks and the
price he or she paid are all data. The data tells nothing about the motivation behind the
purchase, the quality of the socks, or the reputation of the store. Quantitatively, stores
evaluate patterns of purchases, num ber of customers purchasing specific items, and
other items those custom ers purchased. Evaluations such as these can be used to
derive information about customer behavior, the price-sensitivity of certain merchan-
dise, and the like. This means that data is a prerequisite to information.
All organizations need data, and some companies depend more heavily on data
than others. For example, insurance companies, banks, the internal revenue service, and
the social security administration are heavy number crunchers. Millions of transactions
are processed daily. The problem with too much data is that it offers no judgment and
no basis for action. This means that an organization must decide on the nature and vol-
ume of data needed to create information.

INFORMATION
The word information is derived from the word inform , which means “to give shape to”;
information means shaping the data to arrive at a meaning in the eyes of the perceiver.
Information is an aggregation of data that makes decision making easier. It is also facts and
figures based on reformatted or processed data. For example, a profit and loss statement
provides information. It is an assembling of facts into a form that shows an organization’s
state of health over a specific time period. Here is another example of information:

Five farmers o f northern Beirut, who had switched crops from watermelon to
sugarcane with the high hope o f a quick profit , could not bear the anguish o f
crop failures fo r two consecutive seasons. They committed suicide after having
to sell the farm to pay the bank loan.
 CHAPTER 2 Understanding Knowledge aaaa 61

Unlike data, information is understanding relations. It has meaning, purpose, and

relevance. It has a shape, because it is organized for a purpose. The data may havet>een
reorganized, statistically analyzed, or have had errors rem oved—all performed to add
meaning to a message, a report, or a docum ent. The m edium is not the message,
although it could affect the message. An analogy: Having a telephone does not ensure
worthwhile conversation, although certain telephones make the message clear and
more easily understood. Today, having more information technology is not a guarantee
for improving the state of information.
Information is accessible to employees and managers through the company’s local
area networks, intranet, e-mail, Internet, satellite infrastructure, snail mail, or hand
delivery. Unlike data that emphasizes quantity and efficiency of processing, the focus
of information is qualitative: Does the report tell me something I don’t know? Is there
new meaning in the semiannual report? The implication is that data becomes informa-
tion when meaning or value is added to improve the quality of decision making.

KNOWLEDGE
Knowledge has always been an essential component of all human progress. Our ances-
tors must have em ployed an enorm ous am ount of knowledge to form an axe-like
object. From know-how to use seeds for planting to the invention of machinery, to
travel to the m oon—all required an accumulation of special knowledge to achieve the
task. When it comes to basics, people use their intelligence and creativity to come up
with the value-added products and services that take on the competition. Knowledge
capital is essentially a reflection of how well an organization leverages the knowledge
of its workforce, the needs of its customers, and the knowledge of the suppliers to
ensure value-added outcome. Knowledge capital is the way an organization derives
wealth from its information resources on a regular basis.
Knowledge is the most cherished remedy for complexity and uncertainty. It is a
higher level of abstraction that resides in people’s minds. It is broader, richer, and much
harder to capture than data or information. People seek knowledge, because it helps
them succeed in their work.Tiwana views knowledge as actionable (relevant) informa-
tion available in the right format, at the right time, and at the right place for decision
making (Tiwana 2000).
Knowledge has different meanings, depending on the discipline where it is used. In
this text, knowledge is “human understanding of a specialized field of interest that has
been acquired through study and experience.” It is based on learning, thinking, and
familiarity with the problem area in a departm ent, a division, or in the company as a
whole. The focus is on sustainable competitive advantage. Knowledge is not informa-
tion, and information is not data. Davenport and Prusak (2000) define knowledge as “a
fluid mix of framed experience, values, contextual information, and expert insight that
provides a framework for evaluating and incorporating new experiences and informa-
tion.” See Figure 2.1 for alternative definitions of knowledge, data, and information.
O ther characteristics of knowledge are summarized in Box 2.1.
Knowledge is derived from information in the same way information is derived
from data. It may be viewed as an understanding of information based on its perceived
im portance or relevance to a problem area. It can also be thought of as a person’s
range of information. Embracing a wider sphere than information, knowledge includes
perception, skills, training, common sense, and experience. It is the sum total of our per-
ceptive processes that helps us to draw m eaningful conclusions. For example, an
investor requires knowledge to evaluate two companies’ profit and loss statements in
order to determine which one is the healthier company.
PART I The Basics

Data Information Knowledge

Statements about reality Organized, systematized Human interaction with reality
(Acharya 2001) data (Acharya 2001) (Acharya 2001)
Unsorted bits of fact Data that has been sorted, Meaningful links people make in
(D ixon 2000) analyzed, and displayed their minds between information
A representation o f a (D ixon 2000) and its application in action in a
fact, number, word, Data that has been assigned specific setting (Dixon 2000)
image, picture, or a meaning (Liebowitz and The whole set of insights, experi-
sound Wilcox 1999) ences, and procedures that are
M easurements Data that is meaningful or considered correct and true
(Applehans et al. useful to som eone and that, therefore, guide the
1999) (Dickerson 1998) thoughts, behavior, and com -
A discrete, objective Potential for action; resides in munication of people
fact about events the user (Malhotra 1998) (Liebowitz and Wilcox 1999)
(Davenport and A statement of fact about An ideational (i.e., conceptual
Prusak 2000) measurements (Applehans rather than physical) construct
et al. 1999) generated via the agency of
the human mind (Housel and
Descriptive knowledge
Bell 1999)
(characterizing the state
of some past, present, Ability to turn information and
future, or hypothetical data into effective action
solution) (Holsapple and (Applehans et al. 1999)
Whinston 1996) A n organizational resource con-
Data that makes a differ- sisting of the sum of what is
ence (Davenport and known (H olseapple 1996)
Prusak 2000) A fluid mix of framed experience,
values, contextual information,
and expert insight that pro-
vides a framework for evaluat-
ing and incorporating new
experiences and information
(Davenport and Prusak 2000)
Systematizing and structuring
information for a specific pur-
pose (Johannessen et al. 1994)
Information whose validity has
been established through tests
of proof (Libeskind 1996)

FIGURE 2.1: Alternative Definitions of Data, Information, and Knowledge

In addition to evaluation, inform ation becomes knowledge with questions like

“What implications does this information have for my final decision?” “How does this
inform ation relate to other inform ation I received yesterday?” “W hat value can I
derive from this information that adds value to the finished product?” Regardless of
the process that organizes knowledge in the individual, it prom otes the sought-after
expertise for value-added decision making in the business. See Table 2.2 for a summary
definition of each of the three terms.
As can be seen, information is all around, but only a fraction of it is useful in problem-
solving. Knowledge has to be built and requires regular interaction with others in the
know in the organization. It is social, time critical, interactive, evolving, and created for a
 CHAPTER 2 Understanding Knowledge aaaa 63

a ■ i i v ■ b o x 2.1 a a a i i i

SELECT CHARACTERISTICS OF KNOWLEDGE

Knowledge involves a human interaction with • Knowing takes place in relation to existing
reality (or with information about reality, or infor- know ledge—it is placing things in context, in
mation about other knowledge or information), relation to existing constructions of reality,
where the human is the subject and acts as the content, and concepts.
active, creative elem ent, and modifies the latter by • Knowledge involves a judgment, a subsumption of
way of reconstructing it. Knowledge involves the particular under the universal. It involves a
attribution of meaning and significance by the certain amount of synthesis and integration of dis-
knower as a person. In fact, every reconstruction creet information under a category, a construc-
is a reinterpretation as well. tion, or an attribution of a causality or justifiabil-
When I know something, it is relative to me. There ity, relative to the knower’s frame of reference.
can be no knowledge without me. It is always in • Knowledge has a moment of categorical im pera-
relation to my existence and my knowing it. With tive and can induce a cognitive dissonance
my death dies my world, and with it my knowl- between belief and practice, between the past and
edge. In knowing something, I individualize, sub- the present, between the present and the future,
jectify, and appropriate it and make it my own. between what is and what ought to be, and so on,
What I know, in the process, becomes m y own. and therefore, can form a springboard for poten-
Knowledge is essentially social in nature. We tial action. In other words, knowledge by definition
need universal categories for generation, expres- is driven into practice.
sion, representation, storage, retrieval, and • Knowledge is always a part of a dynamic system.
reappropriation of knowledge. The categories Knowledge has the tendency to go for more of
are universal in the sense that (a) they are capable itself, to bypass itself, and to constantly develop
of holding the same meanings for all humans itself. It is only limited by mental and environ-
belonging to the same community and (b) the mental constraints.
categories can be socialized in terms of being
• Knowledge is gregarious by nature and has a ten-
shared, reconstructed, and applied by other
dency to socialize itself. Socialization is the means
humans belonging to the concerned universe
by which individual knowledge gets reinforced,
of discourse.
challenged, modified, improved, and validated.
In knowing something, I believe it to be true.
• Knowledge processes are always a part of an open
Without this belief, it could just be some infor-
system. It is like a game where the goalpost keeps
mation, without that stamp of individualized
shifting itself. The meanings, the dictionaries, and
identity marked on it. This belief is a part of a
even the rules o f the language are always in flux —
system of beliefs, values, and rationality, and
as volatile as the turns in modern life. Knowledge
hence constitutes a responsibility and potential
creation, by definition, is a process of innovation.
commitment.

SOURCE: Excerpted from Acharya, Jagabandhu. “What Is Knowledge?” kmx.totalkm.com/whatisk.html, April 27,2001, pp. 2-5,
Date accessed August 2002.

purpose but drawing on experience from other times and domains. Cooperation and pro-
ductivity are expected as people work to achieve, not to control.Teamwork is a prerequisite
for people to talk, compare, and exchange thoughts, leading to a culture that makes it clear
that “What is my job?” is less important than “What is the purpose of what I am doing?”
Knowledge, not information, can lead to a competitive advantage in business.
Information is thus closer to the decision-making process than is data. For example,
64 a ■ ■ ■ PART I The Basics

TABLE 2.2 Progression from Data to Information to Knowledge

Data Unprocessed or raw facts; a static set of transactional elements, such as
211102345.
Information Processed data; an aggregation of data that have meaning. For example, a
financial analysis report that begins with account number 211-10-2345.
Knowledge A person’s range of information, embracing a wider sphere than informa-
tion. It includes perception, skills, training, common sense, and experi-
ence. For example, a financial analyst might say, “A ll indicators tell me
that I’d better pull out o f X Y Z stock before year end.”

inventory tracking involves storing, retrieving, and reporting information in a structured

format. Decision-making tools that structure information are called decision support sys­
tems (DSS) (spreadsheets, financial planning, resource allocation, and production sched-
uling) in which the decision approach is advisory and the problems addressed are semi-
structured. These systems are used in situations in which the end user makes decisions.
DSS performs the kind of information processing necessary to speed the decision-
making process and thereby saves the user from sifting through the information unas-
sisted. The user might query a sales-forecasting software (or DSS) package, “What if I
lowered the price of product X by 8 percent?” The software would evaluate the impact
of the price cut and might respond, “Your sales should increase by 32 percent.” The
user would then decide what to do with this information.
Figure 2.2 approaches data, information, and knowledge in support of the definitions
in Table 2.2. Data is the raw material for the so-called “number crunching” or data pro-
cessing. The emphasis is on quantitative capture of facts—the backbone of accounting and
payroll applications. For example, an alphabetic list of company employees, the hours
worked, and the hourly rates are all data. When we organize data in a format that adds
meaning, it becomes information. In this respect, information is more qualitative' than
quantitative (or algorithmic) than data. Using the same example, when the employee list
shows employee earnings by departm ent and titles, it becomes useful information.
Information has been defined as data “in formation” (Dixon 2000). It is data that is classi-
fied, sorted, displayed, communicated verbally or graphically, or in the form of tables.
In contrast to inform ation, knowledge is viewed as a body of inform ation,
processes, and experience that centers on a particular subject. It is actionable informa-
tion. Knowledge is links that people make between information and how it is applied
in action in a specific domain. For example, the employee list (by departm ent and title)
may be evaluated by an experienced personnel director for inequity in pay. Inequity in
pay is a specific domain of concern to every organization.
Wisdom is the highest level of abstraction, with vision, foresight, and the ability to
see beyond the horizon. It is the summation of one’s career experience in a specialized
area of work. The power of vision is illustrated in Box 2.2.
One can see that knowledge-rich organizations have a reason for adopting knowl-
edge management into their infrastructure. Knowledge management moves a firm to a
new level of quality, creativity, and sustainable competitive advantage. Unlike technology,
which disappears as a medium of competitive advantage (everyone has it), knowledge
advantage is sustainable because over the long haul it generates continuing advantages. To
ensure continuity, a corporation with a knowledge management environment must main-
tain and enhance its knowledge base and knowledge activities on a regular basis.
 CHAPTER 2 Understanding Knowledge a ■a a 65

Nonalgorithmic Nonprogrammable
(heuristic)
Wisdom

Knowledge
(actionable information)

Algorithmic Programmable

Data, Information, and Knowledge

FIGURE 2.2 : Data, Information, and Knowledge

-- i 1 a 1 1 1 BOX 2.2 1 1 1 i ■ 1 -----

THE YO UN G M A N AND THE STARFISH— AN EXAMPLE OF VISION

A wise man was taking a sunrise walk along the past the breaking waves. “It made a difference for
beach. In the distance, he caught sight of a young that one,” he replied.
man who seemed to be dancing along the waves. That young m an’s actions represent som e-
As he got closer, he saw that the young man was thing special in each of us. We are all gifted with
picking up starfish from the sand and tossing the ability to make a difference. Each of us can
them gently back into the ocean. shape our own future. Each of us has the power to
“What are you doing?” the wise man asked. help our organizations reach their goals.
“The sun is coming up, and the tide is going Vision without action is merely a dream.
out; if I don’t throw them in, they’ll die.” Action without vision just passes the time.
“But young man, there are miles and miles of Vision with action can change the world.
beach with starfish all along it—you can’t possibly —Joel Arthur Barker
make a difference.”
The young man bent down, picked up another
starfish, and threw it lovingly back into the ocean,

SOURCE: Anonymous.
a i i i PART I The Basics

Types of Knowledge
K n o w led g e is cla ssified in to a variety o f types. W h en co n sid erin g k n o w led g e m a n a g e -
m en t, the k n o w le d g e d e v e lo p e r sh o u ld be fam iliar w ith each ty p e and k n ow h o w to tap
into it during k n o w le d g e capture.

SHALLOW AND DEEP KNOWLEDGE

O n e w ay to classify k n o w led g e is to d eterm in e w h eth er it is sh a llo w or d eep . Shallow,
or read ily reca lled surface, knowledge in d ica tes m in im al u n d ersta n d in g o f the p ro b lem
area. For ex a m p le, ap p roval o f lo a n a p p lica tio n s for secu red lo a n s o f less than $1,000
d e p e n d in g on a ssets and salary w o u ld b e b a sed e sse n tia lly on a fe w b a sic ru les that
hardly req u ire hum an co n su lta tio n . In con trast, a lo a n ap p roval sch em e that em p lo y s
14 v ariab les (in clu d in g c u sto m e r ’s cred it rating, n et w orth , pattern o f p ayin g o n tim e,
and c o lle g e e d u c a tio n ) w o u ld b e m o r e c o m p le x an d m o r e risky. Deep knowledge
acq u ired th rou gh years o f e x p e r ie n c e w o u ld b e req u ired to d e c id e on such a loan .

KNOWLEDGE AS KNOW -HOW

K n o w le d g e b a sed on read in g and training is m uch d ifferen t from k n o w le d g e b a sed on
practical e x p e rie n c e that sp an s m an y years. K n o w led g e b a sed on k n ow -h ow , or a ccu -
m u la ted lesso n s o f practical ex p er ie n c e , is w hat is n e e d e d for b u ild in g ex p ert system s.
T h e p ro b lem w ith practical e x p e r ie n c e is that it is rarely d o c u m e n te d . C apturing such
e x p e r ie n c e req u ires sp ecia l to o ls. K n o w le d g e ca p tu re p ro c ed u r es and to o ls are c o v -
ered in C h ap ter 4.
K n o w -h o w d istin g u ish es an ex p ert from a n o v ice. If yo u w ere told y o u n e e d e d an
o p e r a tio n to sa v e y o u r e y e s ig h t, y o u p r o b a b ly w o u ld n o t w a n t an in tern to d o it.
In stead , you w o u ld w ant an ex p ert e y e su r g e o n ’s se a so n e d e x p e r ie n c e b a sed on h u n -
dred s o f sim ilar op era tio n s.
E xp erts rep resen t their k n o w -h o w in term s o f heuristics, ru les o f th um b b a sed on
their e x p e rie n c e —em p irical k n o w led g e. In b u ild in g a k n o w le d g e b ase, h eu ristics g e n -
erally o p era te in the form o f if/th en statem en ts: “I f such and such c o n d itio n s exist, then
such and such a c tio n s result." L en a t has d e fin e d h e u ristic s as “c o m p ile d h in d sig h t”
(L en a t 1982). K n o w -h o w is not b o o k k n o w le d g e ; it is practical ex p e rie n c e. It can be
e x p r e sse d as ru les o f th u m b or h eu ristics if o n e w ere to b u ild a k n o w le d g e b ase for
k n o w le d g e sharing w ithin the orga n iza tio n .

REASONING AND HEURISTICS

H u m a n s reason in a variety o f ways:

1. R e a so n in g by analogy: relatin g o n e c o n cep t to an oth er. For e x a m p le, a battery is

a n a lo g o u s to a reservoir. B oth are u sed to store and p ro v id e en erg y for pow er.
2. Form al reason in g: u sin g d e d u c tiv e or in d u ctiv e m eth o d s. D e d u c tiv e re a so n in g
(a lso ca lled ex a ct rea so n in g b eca u se it d ea ls w ith e x a ct facts and e x a ct c o n c lu -
sio n s) tak es k n ow n p rin cip les and a p p lies th em to in sta n ces to infer so m e sort o f
a c o n clu sio n . For exam p le:
IF Fred is taller than N an cy A N D N a n cy is taller than Sarah,
T H E N Fred is taller than Sarah.
A n o th e r exam p le:
IF m o th ers are w o m en A N D Sarah is a m oth er,
T H E N Sarah is a w om an .
 CHAPTER 2 Understanding Knowledge ■i ai 67

Deduction makes use of major and minor premises. Almost any argument
can be formed using this type of reasoning process. Here are two examples:
a. Major premise: Each titanium coil leaving the mill must be 100 percent
quality tested.
M inor premise: One part of the coil has not passed the test.
Conclusion: The entire coil should be rejected.
b. M ajor premise: In Turkey, all citizens get a pension from age 60 on.
M inor prem ise: Kamal Elberlik, age 60, is a citizen of Turkey.
Conclusion: Elberlik is receiving a pension.
As is evident from the examples, the idea behind deductive reasoning is to
generate new knowledge from previously specified knowledge. If the original
rule is true, then the deduction will be valid. If a knowledge base, for example,
uses only deductive inference and the information assimilated is true, then one
can depend on all the inferred conclusions to be valid. Unfortunately, much of
common-sense reasoning (explained later) is nondeductive by nature.
In terms of inductive reasoning, it works the other way around. Inductive rea-
soning is reasoning from a set of facts or individual cases to a general conclusion—
from specific examples to general rules. Inductive reasoning is the basis of scien-
tific discovery. Consider the following examples:
a. Premise: Chronic unemployment causes social problems.
Premise: Illiteracy causes social problems.
Premise: Recession causes social problems.
Premise: Drug trafficking causes social problems.
Conclusion: Chronic unemployment, illiteracy, recession, and drug traffick-
ing cause social problems.
b. Premise: He is an avid fisherman.
Premise: He is an avid hunter.
Premise: He is an avid mountain climber.
Conclusion: He likes outdoor sports
c. Premise: Admission depends on status.
Premise: Admission depends on race.
Premise: Admission depends on appearance.
Premise: Admission depends on financial wealth.
Conclusion: Admission is subjective.
In each example, the inference is an induction: It goes from a finite number
of instances to a conclusion about every member of a class.
3. Case-based reasoning (CBR). Suppose a person was diagnosed with a condition
that required major surgery. How would that person choose a surgeon? People
tend to prefer an older surgeon with years of practical experience over someone
fresh out of residency. The younger surgeon might be well versed in book
knowledge, but experience is a better predictor of success among surgeons.
Their experience is judged based on the number of cases they handled and the
success of each case.

The idea of reasoning from relevant past cases is attractive because it is so similar
to the process human experts often use to solve problems successfully. The process of
choosing a surgeon indicates the perceived importance of case experience in expert
problem-solving. Manipulating past problem-solving examples is critical in problem
areas such as law, medical diagnosis, prediction, design, and process control. Because
68 i a « a PART I The Basics

experts tend to forget over time, capturing these cases in a knowledge base means
reaping future benefits from past successes (and failures) of experts’ work.
For years, knowledge developers have claimed that human experts reason about a
problem by recalling similar cases encountered in the past. In fact, they reason by anal-
ogy: The expert tries to figure out how one case is similar (or dissimilar) to other cases
solved in the past. A case is knowledge at an operational level. It is an episodic descrip-
tion of a problem and its associated solution. Box 2.3 illustrates an instance of human
CBR, in which the reasoner (the father, who is also an attorney) uses previous cases
similar to the current case as precedent to get the judge to dismiss the case.

COMMON SENSE AS KNOWLEDGE

Common sense is another type of knowledge that all human beings possess in varying
forms and in varying amounts. It is a collection of personal experiences and facts
acquired over time and the type of knowledge that humans tend to take for granted.
For example, if som eone asked you to look up S hakespeare’s phone number, you
would know that such a task is impossible. Common sense tells you that Shakespeare is
dead and that the telephone was not invented until years after his death.
A human expert uses extensive common sense knowledge. Such reasoning is so
common to the expert that it is basically ignored. Unfortunately, a computer does not
possess common sense. Although computers can calculate p i to a thousand decimal
places in seconds, they “know” nothing about some of the simplest things that people
learn early in life, such as oranges are to eat, but baseballs are not, and that legs are
permanently attached to the body.

FROM PROCEDURAL TO EPISODIC KNOWLEDGE

A fourth way to classify knowledge is according to whether it is procedural, declara-
tive, semantic, or episodic (see Figure 2.3).
Procedural knowledge is an understanding of how to do a task or carry out a pro-
cedure. It is knowledge contained in the application of a procedure. Procedural knowl-
edge usually involves psychomotor skills, such as holding onto the handrail while riding
an escalator.
Some procedural knowledge is not psychomotor, however. For example, when a person
learns a language and speaks it fluently, it becomes a natural part of the person. In the case
of an expert, when the same knowledge is used over and over again in a procedure, it comes
to be used automatically. Knowledge developers must select appropriate techniques for
tapping this type of knowledge. Procedural knowledge is covered more fully in Chapter 4.
Declarative knowledge is information that experts can easily discuss. Unlike pro-
cedural knowledge, it is “awareness knowledge,” or routine knowledge of which the
expert is conscious. It is shallow knowledge that is readily recalled, because it is simple,
uncomplicated information. This type often resides in short-term memory, the part of
the brain that retains information for brief periods of time. It kicks in when you are at
an airport and decide to call a friend in the area, for example. You look up the friend’s
phone number jn the phone book and memorize it well enough to dial it. Chances are
by the time you have boarded the flight, you will have forgotten the number. You
remembered it only long enough to dial it.
Shallow, declarative knowledge is also used in diagnosing the electrical system of a
car. The rule of thumb is that dim headlights probably indicate a faulty battery. The
simple rule that if the headlights are dim, then the battery is faulty is obvious. It does
not explain how the electrical system works; it just states a causal relationship between
the headlights and the battery. As we shall find in Chapter 4, knowledge developers
 CHAPTER 2 Understanding Knowledge i aai 69

1 I 1 i I 1 BOX 2.3 i ■ ■ 1 1 i

REASONING WITH THE JUDGE

An attorney’s 17-year-old son was issued a speed- Son: Your honor, it is impossible to go exactly
ing ticket for driving 72 miles per hour in a 65-mph 65 miles per hour while driving. I kept looking at
zone. The young man vehemently denied going that the speedometer. I know I was not driving at 72
fast, but the radar report unmistakably indicated a miles per hour the way this cop drummed it up.
speeding violation. Father and son were scheduled Judge: (looks at the father) Sir, are you the
to appear in juvenile court later that month. father?
The week before the scheduled court date, Father: I am, your honor.
the father went to court. He talked to parents Judge: Do you have anything to say before I
waiting outside the courtroom. He sat in on sev- give my decision?
eral cases to size up the judge and plot strategy Father: Your honor, we have no quarrel with
for defending his son’s case. Every parent he the radar as an electronic system. It is known for
talked to —case after case —indicated that the measuring speed but not the driver’s attitude or
judge almost never let the juvenile off the hook. experience. I would like to submit as evidence this
Penalties ranged from a $60 fine to suspension of young m an’s academic record (all As) and extra-
the driver’s license. The judge dismissed only one ordinary accom plishm ents (Eagle Scout, presi-
case out of 40 that day. The father decided to use dent of his class, no prior speeding tickets). A
it as the basis for his son’s defense. num ber of cases have been brought before this
On the scheduled court date, father and son court to mitigate the severity of the charge. For
appeared before the judge: example, in the case of Jane M a lo n e y v. C o m m o n -
Judge: Young man, how do you plead, guilty w ealth o f M assach u setts and B ill C ro ll v. C o m m o n -
or not guilty? w ealth o f M assach u setts.
Son: No contest, your honor. Judge: (quickly scans the record as he remem-
Judge: It sounds like you’ve been coached. bered both cases cited). Bailiff! Hand the driver’s
Officer, go ahead with your report. (The officer goes license to this boy’s father. (To the father) Sir, you
through the ritual of showing the radar report.) are to keep your son’s license for 10 days before
Judge: Young man, tell this court your side of he’s allowed to drive. Next time, bring your check-
the case. book with you. Case dismissed.

find this type of knowledge useful in the early phase of knowledge capture, because it
promotes familiarity with the type of expertise of the human expert. The best way to
acquire declarative knowledge is with a structured interview.
Semantic knowledge is a deeper kind of knowledge. It is highly organized, “chun-
ked’' knowledge that resides in long-term memory. Such knowledge may have been
there for years and may have been used so often that the information seems like sec-
ond nature. Semantic knowledge includes major concepts, vocabulary, facts, and relation-
ships. Returning to the headlight and battery example, semantic knowledge about the
system would consist of understandings about the battery, battery cables, lights, the
ignition system, and so forth, and their interrelationships. On the basis of this knowl-
edge, one can build rules about causal relationships among those things. In the case of
the headlights, one might know that dim headlights can be caused by a loose battery
cable, a bad alternator, or a drain on the electrical system. At this point, a real expert
(in this case, a certified mechanic) enters the picture (see Figure 2.4).
a a a a PART I The Basics

Shallow Procedural Knowledge

Knowledge Knowledge of how to do a task that is essentially motor in
nature; the same knowledge is used over and over again.

Declarative Knowledge
Surface-type information that is available in short-term
memory and easily verbalized; useful in early stages of
knowledge capture but less so in later stages.

Semantic Knowledge
Hierarchically organized knowledge of concepts, facts, and
relationships among facts.

Episodic Knowledge
Knowledge that is organized by temporal spatial means,
not by concepts of relations; experiential information that
Deep is chunked by episodes. This knowledge is highly compiled
Knowledge and autobiographical and is not easy to extract or capture.

I 1 1 1 1 1 FIGURE 2.3: From Procedural to Episodic Knowledge

FIGURE 2.4: A Comparison of Declarative and Semantic Knowledge

Declarative Knowledge
Therefore
Headlights Battery
Are Dim Is Faulty

Semantic Knowledge
 CHAPTER 2 Understanding Knowledge i aaa 71

Episodic knowledge is knowledge based on experiential information, or episodes.

Each episode is chunked in long-term memory. In general, the longer a human expert
takes to explain or verbalize his or her knowledge, the more semantic or episodic it is.
In contrast, an expert can recall procedural or declarative knowledge in a relatively
short time, because it is readily available in short-term memory.
An interesting aspect about episodic knowledge is that its use is automated. For
example, have you ever driven from point A to point B and yet not remembered many
details of how you got there? This is a common experience. Driving information is so
chunked that most people have trouble remembering and explaining it. In the process
of conveying the expert’s knowledge, the expert explains by examples, or scenarios.
Special tools and techniques are required for capturing such knowledge.

EXPLICIT AND TACIT KNOWLEDGE

A nother approach to categories of knowledge is whether the knowledge is tacit or
explicit (Nonaka and Takeuchi 1995). Tacit knowledge is knowledge embedded in the
hum an mind through experience and jobs. Coined by H ungarian medical scientist
Michael Polanyi (1891-1976), it includes intuitions, values, and beliefs that stem from
years of experience. It is the knowledge used to create explicit knowledge and is best
com m unicated personally through dialogue and scenarios, with use of m etaphors.
Therefore, knowledge is not private, but social. Socially relayed knowledge becomes
part of the real-life experience of the knower.
In contrast, explicit knowledge is knowledge codified and digitized in books, docu-
ments, reports, white papers, spreadsheets, memos, training courses, and the like.
Explicit knowledge can be retrieved and transm itted more easily than tacit knowl-
edge. Because it is knowledge learned directly from experience, tacit knowledge (also
referred to as whispers in the ears) is difficult to share across space and time. The per-
son with tacit knowledge has no conscious experience of what it is that is causing his or
her activity. Chapter 4 elaborates on methods and tools for tapping tacit knowledge.
Tacit and explicit knowledge have been expressed in terms of knowing-how and
knowing-that, respectively. They have also been viewed as embodied knowledge and
theoretical knowledge, respectively (Sveiby 1997). Knowing-how, or embodied knowl-
edge, is characteristic of the expert who makes judgment or decisions without explicitly
going through a set of rules. By contrast, knowing-that, or theoretical knowledge, is
when a person learning a skill articulates available knowledge using explicit instruc-
tion, rules, principles, or procedures.
On the surface, explicit or documented knowledge is easier to identify, because it is
a physical entity that can be measured and distributed. It can be stored as a written
procedure or as a process in a computer. With that in mind, it is reusable for decision-
making purposes. By contrast, tacit knowledge is personal and hard to formalize and
communicate. It is primarily heuristics (rules of thumb), mind-sets, and unconscious
values. The downside is that it is occasionally wrong and hard to change. Because the
human mind is the storage medium, tacit knowledge is vulnerable to loss. When it is
shared, vulnerability is reduced, and it is easier to reuse (Foster 2002).
Understanding what knowledge is makes it easier to understand that knowledge
hoarding is basic to hum an nature. As long as job security is based on the special
knowledge that we have, it continues to be a challenge to support knowledge sharing.
So, when it comes to building a KM system, the company that assumes “if we build,
they will rush” has a surprise coming. A strong incentive and proper change in the cul-
ture of the business are part of the planning that must take place before going KM.
72 a a i ■ PART I The Basics

m i Expert Knowledge
Most people realize that knowledge cannot be directly observed. What can be
observed is the expertise that relies on knowledge to produce solutions. A person with
expert knowledge can solve a complex problem more quickly and more effectively
than someone else can. When the expert’s advanced skills and years of experience are
enhanced by attention to detail, quality results. Not surprisingly, most experts are per-
fectionists. They want the solution to be exactly what is called for by the problem.'
Expert knowledge is not just a head full of facts or a repository of information for
the intellect. It is information woven into the mind of the expert for solving complex
problems quickly and accurately. Take the case of chess experts. They not merely rec-
ognize thousands of chessboard moves, but they consider them in ways relevant to
playing the game. These are all finely tuned to the requirements of the game. The pat-
terns experts learn to recognize are also unique, and the possibilities they consider all
tend to be good ones. It is obvious that a good problem-solving strategy depends on
how much you know.

CHUNKING KNOWLEDGE
Knowledge is stored in an expert’s long-range memory as chunks. Knowledge compila­
tion, or chunking, enables experts to optimize their memory capacity and process
information quickly. Chunks are groups of ideas or details that are stored and recalled
together as a unit. For example, an auto mechanic who had been rebuilding Porsche
transmissions for 18 years was able to rem em ber 140 different steps flawlessly.
Although the sequences and combinations were difficult for him to describe, he consis-
tently completed each job according to specifications.
Chunking prom otes expert performance. The more chunking a person does, the
more efficient is his or her recall. The drawback is that chunking makes it difficult for
experts to be aware of their own knowledge so that they can describe it to others. For this
reason, decompiling chunked knowledge and putting it into words is not an easy task.

KNOWLEDGE AS AN ATTRIBUTE OF EXPERTISE

Experts appear to differ from novices by the greater knowledge of their specialties.
Our society has always recognized exceptional persons, whose performance is vastly
superior to that of the average individual. Speculation about how people acquire extra-
ordinary abilities has been a subject of discussion for many years. Based on current lit-
erature, it seems that in virtually all areas of specialization, insight and knowledge accu-
mulate quickly, and the criteria for expert performance undergo continuous change.
To become an expert in a specialized area, one is expected to master the requisite
knowledge, surpass the achievements of already recognized eminent people, and make
unique contributions to the specialized field. This means that if individuals are innately
talented, they cannot easily or quickly achieve an exceptional level of performance
without preparation time and years of successful practice. In their classic study of
expertise in chess, Simon and Chase (1973) learned that no one had attained the level
of international chess master (grandmaster) with less than 10 years of intense prepara-
tion with the game. Bobby Fischer and Salo Flohr were only a year shy of the 10-year
preparation time (Hayes 1981).
The unique performance of a knowledgeable expert (versus a novice) is clearly
noticeable in decision-making quality. This implies that knowledgeable experts
 CHAPTER 2 Understanding Knowledge aaa■ 73

are more selective in the information they acquire, are better able to acquire infor-
mation in a less-structured situation, and agree more than novices regarding the in-
form ation they consider im portant for decision making or problem-solving (Shan-
teau 1988).
Aside from quantifying soft information, knowledgeable experts tend to catego-
rize problems on the basis of solution procedures (as opposed to surface procedures),
embedded over time in the expert’s long-range memory and readily available on recall.
Because of this strategy, experts tend to use knowledge-based decision strategies, start-
ing with known quantities to deduce unknowns. Should a first-cut solution path fail, the
expert can trace back a few steps and then proceed again —a valuable process in decid-
ing on the appropriate action(s).
In contrast, nonexperts use means-end decision strategies to approach the same
problem situation. They focus on goals instead of essential features of the task, making
the job laborious and unreliable (Sweller, et al., 1983). The net effect is that experts
seem to demonstrate better-developed procedural knowledge for problem-solving. If
procedural knowledge is appropriate for the problem, experts proceed to encode and
finalize a decision more quickly and effectively than nonexperts. As a result, experts
reason differently and solve problems more quickly than nonexperts. Experienced cab
drivers, for example, will recognize a shorter route while traveling to their destinations
and tend to take that route even though the rider m isinterprets such a m aneuver
(Awad 1999).
Taken at face value, experts are the beneficiaries of the information or knowledge
that comes inevitably from experience rather than innate talent bestowed upon the
select few who eventually become experts in their specialty. Experts work harder, rely
less fully on routines, and seem to be engaged in extending their knowledge rather than
merely exploiting it. With knowledge as an attribute of expertise, the focus is not only
on what it includes but also on how it is acquired and how it works. Within the hidden
knowledge of experts are declarative, procedural (skill), and episodic (heuristic)
knowledge (Awad 1999). Procedural knowledge manifests itself in perform ance;
episodic knowledge is the invisible knowledge lying in long-range memory. It repre-
sents years of proven experience in a domain.
In a nutshell, certain individuals consistently perform at higher levels than others
and are labeled experts. They possess attributes that account for their outstanding per-
formance. W hether these attributes are innate talent or w hether they are acquired
through knowledge, practice, and experience, there is reason to believe that experts
succeed primarily because of their proven knowledge and how well they apply and
upgrade it on a daily basis. The conceptual model in Figure 2.5 essentially suggests that
academic knowledge contributes to conceptual knowledge, which is a prerequisite for
practical (in the field) knowledge. Practical knowledge is a contributor to experience
that over time leads to expertise. Expertise means minimum or no errors in decision
making or problem-solving.
The main conclusion regarding corporate experts is for the company to find ways
to retain them by various means, such as recognition, bonuses, and giving an opportu-
nity for mentorship, knowledge sharing, creativity, independence, ability utilization,
and the like (see Box 2.4). The key phrase is knowledge sharing. Mechanisms must be
installed to encourage the sharing of expertise throughout the organization, and there
must be strong support from upper m anagem ent. W ithout such effort, brain drain
could be devastating for a forward-looking organization operating in an uncertain
environment.
74 ■ ■ ■ ■ PART I The Basics

I I I ■ I I FIGURE 2.5: A Conceptual Model of Expertise

I I I ■ I ■ BOX 2.4 ■ ■ ■ ■ ■ ■

THE IMPORTANCE OF EXPERTS

TAKING CARE OF EXPERTS ceal the expert’s identity, leaving it up to the
expert to participate or communicate directly
After the initial pride of being recognized as an
with an individual once a query is made.
expert wears off, being on call to an entire company
It m akes sense for companies to reinforce
may not seem very attractive to busy professionals.
appreciation of the value of experts’ contribu-
“Many experts’ greatest concern about Yellow
tions. Peer recognition is a good way to do this,
Pages and other locator software is that they will
according to Smith. “We implemented a recogni-
be swamped with basic questions,” says Richard
tion program where engineers vote for the best of
McDermott, a knowledge management consultant
the best practices,” he says. More tangible incen-
in Boulder, Colorado. “So, they often hide out.”
tives can help, too. Some companies, including
An effective knowledge-sharing system should
Schlumberger, offer prizes for sharing and create
give experts safe, efficient, and pleasant ways to
point systems for collaborative efforts.
help their coworkers. Knowledge leaders must
Experts also find that they like sharing what
show these experts how their participation will help
they know. “Occasionally, a worker’s job changes
their own colleagues, so they don’t see it “as some
so much that providing expertise becomes as sig-
sort of headquarters deal,” says Reid Smith, vice
nificant a part of the job as the tasks in the origi-
president of knowledge management for Schlum-
nal job description,” says Susan Hanley, director
berger Ltd. in Sugar Land, Texas. “Explain why
of content and collaboration m anagem ent at
you’re doing this and what you’re trying to achieve.”
Plural, Inc., in B ethesda, Maryland. Some even
Recognizing that experts may want to control
change careers so they can do it more.
their availability, some software applications con-

SOURCE: Excerpted from Cuthbertson, Bruce. “Prospective for Experts,” Knowledge Management, June 2001, p. 30.

a a ■ a ■ ■
 CHAPTER 2 Understanding Knowledge aai « 75

Human Thinking and Learning

Because knowledge is the focus of knowledge managem ent, knowledge developers
need to understand how humans think and learn. Scientists have long tried to under-
stand the human brain as part of their process of building computers that may someday
duplicate the hum an expert’s thought process in problem-solving. Imagine a child
using blocks to build a tower. As soon as the tower is com pleted, the child takes a
whack at the tower, destroying it. Next, the child builds a higher tower and then
destroys it as well, and so on. Eventually, the child becomes hungry, and the pattern of
B UILD and DESTROY begins to degenerate. The child gives the tower one final
swipe, destroying it on the way to the kitchen. These spontaneous activities have
proved to be difficult for computers, mainly because no one knows why people do
them and, therefore, knows how to instruct the computer to do them.
According to Marvin Minsky (1991), the human mind is a “society of minds” that is
hierarchically structured and interconnected so that the BUILD, DESTROY, and
H U N G ER agents of the child are minds that represent the self, promote intelligence,
and provide the basis for acquiring knowledge. The study of artificial intelligence has
introduced more structure into human thinking about thinking. So many activities of
the computer resemble human cognitive processes that human and machine “thinking”
are converging in many applications, despite the differences between the brain’s archi-
tecture and the com puter’s. For example, both mind and machine accept data and infor-
mation, manipulate symbols, store items in memory, and retrieve items on command.
Obviously, humans do not receive and process information in the same way that
machines do. For instance, humans receive information via sensing—seeing, smelling,
tasting, touching, and hearing. This system of receiving external stimuli promotes a kind
of thinking and learning that is unique to humans. On a macro level, computers and
humans receive inputs from a variety of sources. Computers receive information from
keyboards, speech, touch screens, and other external sensors. On a micro level, both the
central processing unit of a computer and the human brain receive all information as
electrical impulses.The difference is that computers must be programmed to do specific
tasks. Performing one job does not transcend onto other jobs as it does with human.

HUMAN LEARNING
Memory is an essential component of learning, because it accommodates learning. One
interesting aspect of healthy human memory is that it never seems to run out of space.
Also, as humans acquire more and more knowledge, they generally experience little
interference with the recall ability or the quality of the inform ation in memory. In
other words, as people learn new facts, they integrate them in some way with what they
think is relevant and organize the resulting mix to produce valuable decisions, solu-
tions, or advice. Such learning ability is the basis of accumulating knowledge, experi-
ence, and expertise.
For humans, learning occurs in one of three ways: learning by experience, learning
by example, and learning by discovery. The next section explores these types in an
effort to see how they contribute to human knowledge.

LEARNING BY EXPERIENCE
The ability to learn by experience is a mark of intelligence. When an expert is selected
whose knowledge someone wants to acquire, the expert is expected to have years of
experience reworking problems and looking into different angles for solving difficult
■ ■ ■ a PART I The Basics

TABLE 2.3 Three Types of Human Learning

Learning by experience Trial and error or reworking problems is used to acquire experi-
ence in problem-solving. A n expert uses experience to explain
how a problem is solved.
Learning by example Specially constructed examples or scenarios are used to develop
the concept(s) the student is expected to learn. In knowledge
capture, the human expert uses a scenario to explain how a
problem is solved.
Learning by discovery This is an undirected approach, where humans explore a problem
area with no advance knowledge of what their object is.
------------------------------------------------ a a i a a a -------------------------------------------------

problems. One way of testing potential experts is to observe their recall ability.
Experts, who know a lot about a particular problem, have been found to remember
facts in that problem area much more easily and more quickly than nonexperts, who
presumably have fewer facts to recall. This type of information would be important for
the knowledge developer to keep in mind when understanding a human expert’s range
of knowledge.

LEARNING BY EXAMPLE
Like learning by experience, learning by example is a good contributor to accumulating
knowledge over time. In learning by example, specially constructed examples are used
instead of a broad range of experience. Much classroom instruction is composed of
teaching by example—providing examples, cases, or scenarios that develop the concepts
students are expected to learn. Because this method allows students to learn without
requiring them to accumulate experience, it is more efficient than learning by experience.

LEARNING BY DISCOVERY
Learning by discovery is less understood than learning by example or by experience. It
is an undirected approach in which humans explore a problem area without advance
knowledge of the objective. No one understands why humans are so good at this. It is
difficult to teach, and it will be years before we can benefit from this approach com-
mercially. Table 2.3 summarizes the three types of learning.

Implications for Knowledge M anagem ent

Knowledge awareness benefits entire organizations. With today’s emphases on sustain-
able competitive advantage, added value, and improved productivity, a firm’s manage-
ment needs to create, innovate, monitor, and protect its knowledge inventory. More
specifically, a KM environment means a focus on generating new knowledge; transfer-
ring existing knowledge; embedding knowledge in products, services, and processes;
developing an environment for facilitating knowledge growth; and accessing valuable
knowledge from inside and outside the firm. When this happens, it is beyond survival.
In fact, it is beyond intranets and databases—the technology that supports KM.
Some sources claim that 20 percent of an organization’s knowledgeable personnel
can operate 80 percent of the organization’s day-to-day business. The human resources
manager can play an important role in identifying the knowledge core of the organiza-
 CHAPTER 2 Understanding Knowledge i i i i 77

tion, recommending ways to preserve this critical core, and building a robust, long-
range plan to ensure top-quality operation. Without such preparation, corporate talent
could potentially erode through a brain drain that spells disaster for any business. At
the same time, professionals with expertise are naturally drawn to organizations that
recognize and reward expertise, especially when that expertise directly contributes to
the firm’s productivity. Such matches explain the stability and growth of many success-
ful “learning” companies.
Based on the discussion in the chapter, several ideas should be considered for
how a com pany should perform in order to create and m aintain sustainable com-
petitive advantage. First, there should be more emphasis on tapping, sharing, and pre-
serving tacit knowledge and the total knowledge base of the company. A com-
pany’s knowledge base includes explicit and tacit knowledge and exists internally in
the business as well as within the firm ’s external connections. Second, companies
should focus on innovation and the processes that convert innovation to new products
and services. Knowledge sharing and an emphasis on the total knowledge base pro-
mote innovation.
Finally, it is important to consider a renewed focus on organizational learning sys-
tems and systemic thinking throughout the organization. This is a realistic expectation,
because knowledge is closely related to learning, which is the outcome of regular and
continuous interactive learning. Systems thinking means understanding how the vari-
ous parts of the company work. This includes learning behavioral patterns in the sys-
tem and the culture or system environm ent in which employees and administrators
operate. In other words, systemic thinking is expected to support innovation and con-
tinuous im provem ent processes, social competence, and interactions, as well as the
total knowledge base.
What good is knowledge if it cannot be shared? If knowledge is power, sharing
it will multiply power across the business. Unfortunately, sharing knowledge is an un-
natural thing. One person’s knowledge is an added value to that person’s career path.
Knowledge management is designed to solve the problem of unrecycled knowledge.
Systems have been developed to gather, organize, refine, and distribute knowledge
throughout the business. Virtually all such systems should have six key attributes:
learning capability, improving with use, knowing what you want, two-way communica-
tion between the system and you, recalling past actions to develop a profile, and unique
configuration to your individual specifications in real time.
In the final analysis, communication and connection make knowledge sharing an
ongoing activity. Technology can only do so much to create a formal system. Success
with KM exists when the culture is ready to communicate and connect. The end result
is “community,” built around knowledge and based on vision.

S u m m a ry 1 1 , 1
• Intelligent behavior has several attributes:
• The ability to understand and use language
• The ability to store and retrieve relevant experience at will
• Learning by example, from experience, or by discovery
• Several key terms are worth noting:
• K now ledge —understanding gained through experience
• Intelligence—the capacity to apply knowledge
78 i a i a PART I The Basics

• Heuristics—rules of thumb based on years of experience

• Common sense—innate ability to sense, judge, or perceive situations that
grows stronger over time
• Experience—changing facts into knowledge to refine a reasoning process
• A distinguishing feature of human learning is that as people learn new facts, they
integrate them in some way and use the resulting mix to generate value-added
decisions, solutions, or advice.
• Humans learn by experience, by example, and by discovery. Learning by discov-
ery is less understood than learning by example and by experience. Learning con-
tinues to be a major concern in knowledge management.
• Knowledge developers, whose job is to capture experts’ knowledge, need to be
well prepared and to have a clear understanding of the distinctions among knowl-
edge, information, and data. They must focus on knowledge as it relates to the
problem area.
• The relative importance of data, information, and knowledge is a function of
the importance of the problem, the decision approach, the nature of the prob-
lem, and the number of persons affected. Whereas data plays a relatively
trivial role in problem-solving, knowledge occupies a major role. The decision
approach is advisory and relates to a difficult problem affecting many people
in the organization.
• Expert knowledge is clustered, or “chunked,” in long-range memory. Chunking
promotes expert performance, but can also make it difficult for experts to
be aware of their own knowledge in a way that allows them to describe it
to others.
• Humans have common-sense knowledge, a collection of personal experiences
and facts acquired over time. The fact that common-sense reasoning is so
strong in experts makes it difficult for knowledge developers to capture their
deep knowledge.
• Knowledge can be classified by procedural, declarative, semantic, or episodic means.
• Procedural knowledge is knowledge that is used over and over again.
• Declarative knowledge is knowledge that the expert is aware or conscious
of. It is shallow knowledge.
• Semantic knowledge is chunked knowledge that resides in the expert’s
long-range memory.
• E pisodic knowledge is knowledge based on experiential inform ation.
Each episode is chunked in long-range memory.
• Common sense is inferences made from knowledge about the world. Reasoning is
the process of applying knowledge to arrive at solutions. It works through the
interaction of rules and data.
• Deductive^reasoning deals with exact facts and conclusions. The idea behind
deductive reasoning is to generate new knowledge from previously specified
knowledge. In contrast, inductive reasoning is reasoning from a set of facts to gen-
eral principles. Induction usually produces results without explanation.
• Knowledge has also been classified as tacit and explicit knowledge. Tacit knowl-
edge, or “know-how,” is stored in people’s minds and is not so easy to capture or
share. By contrast, explicit knowledge is codified and digitized in the form of
records, reports, or documents and is reusable for decision making.
• Case-based reasoning is reasoning by analogy. Human experts reason about a
problem by recalling similar cases encountered in the past.
 CHAPTER 2 Understanding Knowledge a a ■■ 79

T er ms to Kn o w ■■■■
Case: Knowledge at an operational level; episodic Inferencing: Deriving a conclusion based on statements
description of a problem and its associated that only imply that conclusion.
solution. Intelligence: The capacity to acquire and apply knowl-
Case-based reasoning: A methodology that records and edge through the ability to think and reason.
documents previous cases and then searches the rele- Knowledge: Understanding, awareness, or familiarity
vant case(s) to determine their usefulness in solving a acquired through education or experience.
current problem; problem-solving a case by analogy Learning: Knowledge or skill acquired by instruction
with old ones. or study.
Chunking: Grouping ideas or details that are stored and Learning by discovery: Acquiring new ideas by exploring
recalled together as a unit. a problem area with no advance knowledge of what is
Common sense: Possessing common knowledge about being sought.
the world and making obvious inferences from this Learning by example: Acquiring new ideas based on spe-
knowledge. cially constructed examples or scenarios.
Compilation: The way a human translates instructions Learning by experience: Acquiring new ideas based on
into meaning language or response. hundreds of previously stored concepts.
Decision support systems (DSS): Computer-based infor- Logic: The scientific study of the process of reasoning
mation systems that combine m odels and data for solv- and the set of rules and procedures used in the reason-
ing complex problems with extensive user involvement. ing process.
Declarative knowledge: Surface information that experts Memory: The ability to store and retrieve relevant experi-
verbalize easily. ence at will.
Deductive reasoning: A lso called exact reasoning; takes Premise: Provides the evidence from which the conclu-
known principles (exact facts) and applies them to sion must necessarily follow; evaluates the truth or
instances to infer an exact conclusion. falsehood with some degree of certainty.
Deep knowledge: Knowledge based on the fundamental Procedural rule: A rule that describes a sequence of rela-
structure, function, and behavior of objects. tions relative to the problem area.
Episodic knowledge: Knowledge based on experiential Reasoning: The process of applying knowledge to arrive
information chunked as an entity and retrieved from at solutions based on the interactions between rules
long-term memory on recall. and data.
Experience: The factor that changes unrelated facts into Scenario: The formal description of how a problem situa-
expert knowledge. tion operates.
Expert: A person whose knowledge and skills are based Semantic knowledge: Highly organized, “chunked”
on years of specialized experience. knowledge that resides in the expert’s long-term mem-
Expertise: The skill and knowledge possessed by some ory and represents concepts, facts, and relationships
humans that result in performance that is far above among facts.
the norm. Shallow knowledge: Readily recalled knowledge that
Explicit knowledge: Knowledge codified in documents, resides in short-term memory.
books, or other repositories. Short-term memory: The part of the human brain that
Fact: A statement of a certain elem ent of truth about a retains information for a short period of time.
subject matter or a problem area. Tacit knowledge: Knowledge used to create explicit
Heuristic: A rule of thumb based on years of experience. knowledge; the mind-set of individuals that
Inductive reasoning: Reasoning from a given set of facts includes intuitions, values, and beliefs that stem
or specific examples to general principles or rules. from experience.

Test Y o u r U n d e r s ta n d in g 1 1 , 1
1. If intelligence is the capacity to acquire and apply knowledge, what is
knowledge?
2. Briefly explain the key attributes of intelligent behavior.
3. Distinguish between:
a. fact and rule
b. knowledge and common sense
c. experience and heuristics
d. learning by example and learning by discovery
80 ■* i « PART I The Basics

4. Define episodic knowledge and semantic knowledge. Give an example of each.

5. Illustrate by example the possible relationship between:
a. knowledge and information
b. knowledge and data.
6. Why is knowledge compiled? Discuss its relationship to long-range memory.
7. Review the types of knowledge discussed in the chapter. Use an illustration
of your own to show how each type differs from the other types.
8. What is the difference between tacit knowledge and explicit knowledge?
Give an example of each.
9. If shallow knowledge is declarative knowledge, what is deep knowledge? Be
as specific as you can.
10. Illustrate the differences between deductive and inductive reasoning. Under
what conditions is one preferred over the other?

K n o w le d g e E xercises « • • 1
1. “People do not think in the same way as machines, because people are bio-
logical.” Do you agree?
2. What type of knowledge is used in each of the following activities:
a. tying a shoelace
b. debugging a computer program
c. baking a pie
d. replacing a car’s flat tire
e. negotiating peace with a hostile country
f. driving in congested traffic
Explain each classification.
3. List five heuristics that you employ in everyday life. By what kind of learning
have you arrived at these rules of thumb?
4. Determine the type of reasoning in each of the following cases:
a. Liz did not deposit money in her checking account.
Liz is a customer of the bank.
Liz’s check bounced (insufficient funds).
Conclusion: Checks drawn against a negative account balance will bounce.
b. Drivers who exceed the speed limit get speeding tickets.
Donna is a licensed driver.
Donna drove her car 20 miles over speed limit.
Conclusion: Donna will get a speeding ticket.
(What is wrong with this reasoning?)
c. A customer whose account balance drops below $100 during the month
is subject to a $5 charge.
Bob is a bank customer.
Bob has a checking account.
Bob’s checking account balance dropped to $98.
Conclusion: Bob’s account will be charged $5.

R e fe re n c es 1 1 1 1
Applehans, Wayne, Globe, Alden, and Laugero, Graig. Awad, Elias M. Selected Attributes o f the Domain Expert.
Managing Knowledge. Boston, MA: Addison- Unpublished manuscript 1999, pp. 1-13.
Wesley, 1999. Chiem, Phat. “Trust Matters," K nowledge Management,
Acharya, Jagabandhu. “What Is Knowledge?’' May 2001, pp. 50-52.
kmx.totalkm.com/whatisk.html, April 27,2001, pp. 2-5, Cuthbertson, Bruce. “Prospective for Experts," K nowl-
D ate accessed August 2002. edge Management, June 2001, p. 30.
 CHAPTER 2 Understanding Knowledge ■a a a 81

Davenport,Thomas H., and Prusak, Laurence. Working July/August 1998, pp. 58-60, www.brint.com/km/
Knowledge. Boston, MA: Harvard Business School whatis.htm, Date accessed August 2002.
Press, 2000, pp. 1-24. Minsky, Marvin. “Conscious Machines." In Machinery o f
Dixon, Nancy M. Com m on Knowledge. Boston, MA: Consciousness Proceedings, National Research Council
Harvard Business School Press, 2000, pp. 12-13, of Canada, 75th Anniversary Symposium on Science in
17-32. Society, June 1991.
Foster, Allan. “Grasping the Tacit Knowledge Nettle." Nonaka, Ikujiro, and Takeuchi, Hiro. The Knowledge-
Knowledge M anagement, April 1998, p. 32. Creating Company: How Japanese Companies Create
Garvin, David A. Learning in Action. Boston, MA: the Dynam ics o f Innovation, New York: Oxford
Harvard Business School Press, 1999, pp. 47-90. University Press, 1995.
Hayes, J. R. The Complete Problem Solver. Philadelphia, Pfeffer, Jeffrey, and Sutton, Robert. The K nowing-D oing
PA: Franklin Institute Press, 1981. Gap. Boston, MA: Harvard Business School Press,
Holsapple, C. W., and Whinston, A. B. Decision Support 2000, pp. 5,12,181.
Systems. Minneapolis, MN: West Publishing Co., 1996. Roberts-Witt, Sarah. “Knowledge Everywhere,"
pp. 11-124. K nowledge M anagement, June 2001, pp. 36-41.
Horibe, Frances. Managing K nowledge Workers. New Shanteau, James, “Psychological Characteristics and
York: John Wiley & Sons, Inc., 1999, pp. 1-18. Strategies of Expert D ecision Makers.” Acta
Housel,Thomas, and Bell, Arthur. Measuring and Man- Psychologica, vol. 68,1988, pp. 203-15.
aging Knowledge. New York: McGraw-Hill, Inc., 2001. Simon, H. A., and Chase, W. G. “Skill in Chess," American
Johannessen, J. A., 01ais£n>Johan, and Olsen, B. Scientist, vol. 61,1973, pp. 3 9 4 ^ 0 3 .
Mismanagement o f Tacit K now ledge, www.program. Sveiby, Karl-Erik. “Tacit Knowledge." D ecem ber 31,1997.
forskningsradet.no/skikt/johannessen.php3, www.sveiby.com/articles/polanyi.html, Date accessed
Date accessed August 2002. August 2002.
Koulopoulos,T. M., and Frappaolo, Carl. Smart Things to Sweller, John, Mawyer, R. F., and Ward, M. R.
Know A bout Knowledge Management. Milford, CT: “D evelopm ent of Expertise in Mathematical Problem
Capstone Publishing, 2000, pp. 37-66. Solving," Journal o f Experimental Psychology: General,
Lenat, Douglas B. K nowledge-Based Systems in Artificial D ecem ber 1983, pp. 639-61.
Intelligence. New York: McGraw-Hill Book Co., 1982. Tiwana, Amrit. The K nowledge Management Toolkit.
Liebeskind, J. P. “Knowledge, Strategy, and the Theory of Upper Saddle River, NJ: Prentice Hall, 2000, pp. 55-98.
the Firm," Strategic Management Journal, Winter 1996, Witt, Roberts. “Knowledge Management: Know What
pp. 93-107. You Know," PC M agazine, July 1,2000, p. 165.
Malhotra, Yogesh. “Knowledge Management for the New
World of Business." Journal fo r Quality & Participation ,
 Knowledge
Management
Systems Life Cycle
Contents

In a Nutshell
Challenges in Building KM Systems
Conventional Versus KM System Life Cycle
Knowledge Management System Life Cycle (KMSLC)
System Justification
Role of Rapid Prototyping
Selecting An Expert
Role of the Knowledge Developer
Role of Quality Assurance
User Training
Implications for Knowledge Management
Summary
Terms to Know
Test Your Understanding
Knowledge Exercises
References

There is nothing more difficult to plan, more

doubtful of success, nor more dangerous to
manage than the creation of a new system. For
the initiator has the enmity of all who would
profit by the preservation of the old system and
merely lukewarm defenders in those who would
gain by the new one.
—M a c h ia v e l l i, 1513
 CHAPTER 3 Knowledge Management Systems Life Cycle i ■1 i 83

The building of knowledge management can be viewed as a life cycle that begins with
a master plan and justification and ends with a system structured to meet KM require-
ments for the entire company. A knowledge team representing the thinking of the firm
and a knowledge developer with expertise in knowledge capture, knowledge design,
and knowledge implementation ensure a successful system. Lack of planning, struc-
ture, and order can invite disaster.
The most critical phase of the KM systems life cycle is identifying the immediate,
intermediate, and long-term needs for the prospective system. This means reviewing
the knowledge core of existing employees; conducting a cost/benefit analysis to deter-
mine the justification for and potential benefits of the candidate system; and determ in-
ing the tools and the procedures to ensure completeness, accuracy, integrity, and opera-
tional success of the installation.
The knowledge m anagem ent systems life cycle (KMSLC) centers around three
questions:
1. What is the problem that warrants a solution by the KM system? How important
is the problem? What clues indicate that the system should be built? What will
the user and the company as a whole gain from the system?
2. W hat developm en t strategy should be considered? Who is going to build the
system?
3. What process will be used to build the system?
These questions are interrelated, and they lead to other questions when the knowl-
edge developer gets involved in the system development process. In deciding on a KM
system, it is im portant to consider top m anagem ent support early in the planning
phase. Employee support and participation through the knowledge capture phase also
make it easier to navigate through the development process. This chapter focuses on
the strategic planning and the justification for KM system developm ent and the
process of building it.
The concept of “life cycle” is not new. It can be applied to virtually every endeavor,
personal and business. Here are some examples:
College: Admission, education, graduation
Term paper: Introduction, body, conclusion, references
Air flight: Boarding, takeoff, cruising, landing, disembarking
Faculty: Assistant professor, associate professor, professor
Information system
development: Problem definition, analysis, design, implementation
These life cycles have some characteristics in common:
• Discipline, order, or segmentation into manageable activities or phases
• Good documentation for possible changes or modifications of the system in the
future
• Coordination of the project to ensure the cycle is completed on time
• Regular management review at each phase of the cycle
Structure and order are to KMSLC what chapters and paragraphs are to a book.
Can you imagine trying to read this book if the paragraphs occurred in random order?
You would be exposed to the same material, but it would lack order and, therefore,
84 a a ■ a PART I The Basics

meaning. In the same way, a well-defined life cycle is essential for successful develop-
ment and maintenance of knowledge management in a business environment.

■ Challenges in Building KM Systems

On the surface, KM seems exciting, but several challenges lay ahead that hinder KM
system development. Among the key challenges are the following:
• Culture. Changing organizational culture is not an overnight exercise. The number
one challenge is getting people to share their knowledge rather than hoarding it (see
Box 3.1). To do so means changing people’s attitudes and behaviors. The company
that develops the right combination of incentives for employees to collaborate and
share their knowledge will go a long way toward ensuring a successful KM system.
There is no question that knowledge is power, and no one wants to give it up. Tradi-
tionally, employees hoard knowledge because they understandably worry that they
will lose advantage in the organization. KM has to make knowledge sharing attractive
enough to be lasting, not just for the company but for the employees as well.
• Knowledge evaluation. The company that tackles the evaluation issue will have
both a head start on the employee incentive problem and an edge on refining
KM processes for maximum profit. Assessing the worth of information is a cru-
cial step if a company wants to refine its methods or create a reward system for
employees who have generated the “best” knowledge (see Box 3.2).

BOX 3.1 a 1 I I fl 1

THE SHARING OF KNOWLEDGE

Both inside and outside a company’s gates, great machine, or at the water cooler. Knowledge man-
things often begin at those fortunate m om ents agement proponents today believe that those
when people with knowledge and vision pool chance interactions that mix and reformulate
their dreams. A perfect example: WorldCom, Inc., knowledge can and should be systematically en-
the company that merged with MCI to form MCI couraged. It can happen in one office, but time and
WorldCom, Inc., now the fourth-largest telecom- geography are no longer constraints to collaborative
m unications com pany in the world, had its ori- work. The creation of a profitable idea—like MCI
gins at a coffee shop in Hattiesburg, Mississippi, WorldCom—always takes hard work but is never
when four executives who knew the business just dumb luck. Larry Prusak, executive director of
got together exactly 1 month after AT&T’s long- IBM ’s Institute for Knowledge Management in
distance monopoly was broken by antitrust laws. Waltham, Massachusetts, says, “All of life and busi-
Is it luck or knowledge? ness is a game of odds. Just as HR policies increase
Of course, businesspeople can meet by design the odds of employee retention, and good customer
at regular times and places, but the spontaneous service increases the odds toward repeat business,
appearance of a group that can develop a shared knowledge management is about increasing the
vision has historically been the product of a chance odds toward knowledge being transferred, utilized,
meeting on the stairs, on the elevator, by the coffee and contributing to innovation.”

SOURCE: Excerpted from Glasser, Perry. “The Knowledge Factor,” CIO Magazine, December 1 5 ,1998-January 1,1999. pp. 1-9.

a a a a a a
 CHAPTER 3 Knowledge Management Systems Life Cycle aaaa 85

i a a i i a bo x 3.2 a a a a a a

ESTIMATING THE WORTH OF KNOWLEDGE

Figuring a return for knowledge management ini- firm. Though satisfied with Kworld’s early opera-
tiatives is like trying to calculate payback from pro- tion, a nagging concern for Turillo remains finding a
viding employees with telephones, paper, and pens. proper way to evaluate the information in KPMG’s
According to international and U.S. chief knowl- knowledge repository, the Global Knowledge Ex-
edge officer Michael J. Turillo, Jr., of KPMG Peat change. Assessing the worth of inform ation is a
Marwick LLP, the consultancy invested $40 million crucial step if a company wants to refine its methods
in a knowledge managem ent initiative called or create a reward system for the employees who
K w orld during a 5-month period in 1998. The have generated the “best” knowledge. A simple hit
investment included capital outlays for hardware, count on inform ation use reveals little about its
software, and the development cost of executive worth. Instead, the interest is in quality, who is
and management time consumed by planning and using the information, and whether they are deci-
testing the system for the New York C ity-based sion makers.

SOURCE: Excerpted from Glasser, Perry. “Knowledge Factors,” CIO Magazine, December 15,1998-January 1,1999, p. 3.

a a a a a a

Unfortunately, estimating the worth of knowledge is complicated. Reliable met-

rics have yet to be made available.
• Knowledge processing. Many companies do not realize the importance of the
human element in KM. Effective KM systems must allow organizations not only
to store and access information but to document how decisions were reached as
well. Specifically, techniques should be identified that will capture, store, process,
and distribute the kind of knowledge that cannot be readily tabulated in rows or
columns (see Box 3.3).

a a a a a a bo x 3.3

THE PROCESSING OF KNOWLEDGE

The m erger of Peat Marwick International and base design respects the legacy of KPM G’s con-
Klynveld Main G oerdeler in 1987 produced a sulting practice, which remains organized around
babel of cultures in the American, British, Dutch, product, industry, and geography. KPM G users
and German partnership known today as KPMG once had to physically locate the work of a spe-
Peat Marwick LLP. “KPMG was not so much a cialist senior partner and potentially read through
global company as it was a collection of geo- dozens of documents. With Kworld, expert infor-
graphically identified franchises,” says interna- mation is available on the desktop. Extracting in-
tional and US. chief knowledge officer Michael J. house knowledge about ER P installations for the
Turillo, Jr. petroleum industry in Germany, for example, is
Kworld —K PM G ’s knowledge intranet that now only a few mouse clicks away.
started in 1997—is the company’s strategic bet to
make all its pieces work together. The knowledge (icontinued)
86 a a ■ ■ PART I The Basics

(continued) York City office.The editors capture nonquantifi-

able knowledge from feeds that zap as many as
Keeping the repository current occupies 15 8,000 published papers, speeches, books, and maga-
full-time “knowledge editors” in K PM G ’s New zine articles across their screens daily.

SOURCE: Excerpted from Glasser, Perry. “The Knowledge Factor,” CIO Magazine, December 15,1998-January 1,1999, pp. 6-7.

---------------------------------------------------- ■ ■ 1 1 1 1 -------------------------------------------------------------------------------------- :-------------

• Knowledge implementation. When it comes to KM, an organization must commit

to change, learning, and innovating if it is to realize leadership in the marketplace.
Technology has already made it possible to implement knowledge collaboration,
regardless of time and place. One of the im portant tasks in KM is extracting
meaning from information that will have an impact on special problems or mis-
sions. Lessons learned in the way of feedback are stored for others facing the
same problem in the future. That is why prisoners of war are debriefed after their
release. Their experience is stored in a repository for future behavioral adjust-
ment that could improve the survival and life style of future prisoners of war.

■■■■ Conventional Versus KM System Life Cycle

Those who have developed conventional information systems in the past and are now
involved in building KM systems need to see the relationship between the conven-
tional systems approach and KM system development. This perspective extends to the
role of the systems analyst versus that of the knowledge developer. This section exam-
ines these important distinctions.

Key Differences
Some striking differences distinguish conventional systems development from KM
system development:
1. The systems analyst deals with data and information obtained from the user. The
user is highly dependent on the analyst for the solution (a conventional infor-
mation system). The knowledge developer deals with knowledge captured from
people with known knowledge in the firm. The developer is highly dependent on
them for the solution.
2. The main interface for the systems analyst is with the novice user, who knows the
problem but not the solution. In contrast, the main interface for the knowledge
developer is the knowledgeable person who knows the problem and the solution.
There is no comparable expert in a conventional information system process.
3. Conventional system development is primarily sequential; that is, particular steps are
carried out in a particular order. Design cannot be initiated without analysis, testing
cannot be done without a design, and so on. In contrast, KMSLC is incremental and
interactive. A KM system is not built in a few large steps; rather, it evolves toward a
final form. Rapid prototyping as a knowledge capture tool plays a major role in KM
system evolution. The flowcharts in Figure 3.1 illustrate these differences.
4. In the development of conventional information systems, testing occurs toward
the end of the cycle after the system has been built. In KMSLC, the knowledge
 CHAPTER 3 Knowledge Management Systems Life Cycle a ■■■ 87

Conventional System KM System

Life Cycle Life Cycle

Recognition of Need Evaluate Existing

and Feasibility Study Infrastructure

FIGURE 3.1 1 Comparison of the Development Life Cycles of a Conventional Information System Life
Cycle and a KM System Life Cycle

developer tests (verifies and validates) the evolving system from the beginning
of the cycle.
5. The discipline of system developm ent and system m aintenance is much more
extensive for conventional information systems than it is for KMSLC. KM system
maintenance is delegated to knowledge editors, whose job is to ensure a reliable
system and to upgrade the system to standards.
6. The conventional system life cycle is process-driven and documentation-oriented,
with emphasis on the flow of the data and the resulting system. It fosters the
“specify then build” approach. The KMSLC is result-oriented. The emphasis is on
a “start slow and grow” incremental process.
7. Conventional system development does not support tools like rapid prototyping
because it follows a set sequence of steps. KMSLC utilizes rapid prototyping (to
be explain later in the text), incorporating changes on the spot, which augments
and refines the KM system until it is ready for use. Thus, the prototype evolves
I The Basics

Repeated
Cycle(s)

Repeated fl fl fl i fl fl
Cycle(s) ■— “ ™1

FIGURE 3.2: Rapid

Prototyping Process

into the final KM system. As shown in Figure 3.2, in knowledge capture, tasks are
structured, and the knowledge sought is reform ulated until the knowledge
worker or the resident expert judges that it is right. This process promotes verifi-
cation of the KM system. Verification answers the question, Is the system built
right? The final test is validation, which ensures that authorized users are satisfied
with what the KM system offers.

Key Similarities
Although obvious differences exist between the conventional information system
development life cycle and KMSLC, so, too, do certain similarities:
1. Both cycles begin with a problem and end with a solution. The problem solution
is pitched to benefit the company user and the organization as a whole.
2. After setting up the strategic plan, the early phase of the conventional system life
cycle begins with information gathering to ensure a clear understanding of the
problem and the user’s requirem ents. In KMSLC, the early phase requires
knowledge capture, which later becomes the foundation of the knowledge base.
Both information and knowledge must be represented in order for a system to
produce results.
3. Verification and validation of a KM system resemble conventional system test-
ing. Verification ensures that the KM system is clear of errors. This is similar to
alpha testing or debugging a conventional system program(s). In contrast, KM
system validation and conventional system beta (user acceptance) testing ensure
that the system meets the user’s requirements prior to deployment.
4. B oth the knowledge developer and the systems analyst need to choose the
appropriate tools for designing their respective systems. Choice of tools and
methodologies will be covered in Chapter 4.

Users Versus Knowledge Workers

Users and knowledge workers haive similarities and differences. For example, in
terms-of cooperation, the user must work with the systems analyst by providing the
information or documents needed for building a successful information system. In con-
trast, the knowledge worker or the expert “owns” the knowledge being shared or given
away. Most experts are cooperative, although no knowledge developer should take
them for granted. The main difference is that the expert does not have the same vested
interest in the system that the user has; in most cases, the expert will not be the final
user. Table 3.1 summarizes differences between the user and the expert.
 CHAPTER 3 Knowledge Management Systems Life Cycle * i ■■ 89

TABLE 3.1 Comparison of Users and Experts

Attributes User Expert
D ependence on the system High Low to nil
Cooperation Usually cooperative Cooperation not required
Tolerance for ambiguity Low High
Knowledge of problem High Average/low
Contribution to system Information Knowledge/expertise
System user Yes No
Availability for system builder Readily available N ot readily available

■■■■ Knowledge M anagem ent System Life Cycle

(KMSLC)
Due to a lack of standardization in the field, several approaches have been proposed for a
KMSLC. Some approaches borrow from the conventional system development life cycle;
others prescribe steps unique to knowledge management. Table 3.2 lists representative

TABLE 3.2 Alternative Approaches to Knowledge Management Development

Davenport
Garvin L iebow itz and and Prusak
Step Tiwana (2000) Dixon (2000) (2000) Wilcox (1997) (2000)

1. Analyze existing infrastructure Team performs a task Acquiring Build knowledge Acquisition
knowledge
2. Align knowledge management Team explores the re- Interpreting Organize and hold Dedicated
and business strategy lationship between knowledge resources
action and outcome
3. D esign the knowledge infra- Common knowledge Applying Distribute and Fusion
structure gained knowledge pool
4. Audit existing knowledge Knowledge transfer Apply knowledge Adaptation
assets and systems system selected to work object
5. Design the KM team Knowledge translated Knowledge
into a form usable networking
by others
6. Create the KM blueprint Receiving team adapts
knowledge for its
own use
7. D evelop the KM system
8. Deploy, using the results-driven
incremental m ethodology
9. Manage change, culture, and
reward structures
10. Evaluate performance, measure
ROI, and incrementally refine
the KMS
90 ■ ■ ■ ■ PART I The Basics

Stage Key Question (s) Outcome

Evaluate existing What is the problem? Statement of objectives
infrastructure Is the system justifiable? Performance criteria
Is the system feasible? Strategic plan

Form the KM team Who should be on the team? Standardized procedure for
How will the team function? system development

Knowledge capture What and whose knowledge Acquisition of knowledge core

should be captured?
How would knowledge cap-
ture proceed?

D esign KM blueprint How will knowledge be D esign of KM system

(master plan) represented? Hardware/software implemen-
tation details
Test plan
Security, audit, and operating
procedures

Test the KM system How reliable is the system? Peer reviews, walkthroughs

Implement the KM system What is the actual operation? User-friendly system

How easy is it to use? Training program

Manage change and reward D oes the system provide the Satisfied users
structure intended solutions?

Post-system evaluation Should the system be Reliable and up-to-date system

modified?

FIGURE 3.3: KM System Development Life Cycle

approaches. A cursory examination reveals many similarities and overlaps. We propose a

hybrid life cycle that makes best use of the existing ones (see Figure 3.3). A representative
knowledge management system is summarized in Box 3.4.
Note that although conventional system development and KM system develop-
ment differ in some im portant ways, the two approaches are fundamentally similar.
The conventional approach may still be used in developing KM systems, but it is being
replaced by iterative design, prototyping, early testing, and other variations.

1 1 1 « ■ 1 BOX 3.4

THE LIFE CYCLE OF A KNOWLEDGE MANAGEMENT SYSTEM

BUILDING A KNOWLEDGE MANAGEMENT edge m anagement auditing system to advise the
SYSTEM FROM THE GROUND UP bank on compliance with the federal bank exam-
iners who audit the bank on a quarterly basis. The
With the Wall Street scandals of 2002 (A rthur
office of the Comptroller of the Currency (OCC),
Andersen, Enron, WorldCom), a large com m er-
cial bank ordered the developm ent of a knowl- (continued)
 CHAPTER 3 Knowledge Management Systems Life Cycle aaaa 91

(icontinued ) how that examiner worked. Finally, the develop-

ers com bined the six m ethods into one set that
a bureau of the U.S. Treasury, is the regulator of described the examiners’ consensus.
the national banking system. To oversee this system Taping the sessions was trem endous help.
of 4,500 banks, the OCC administers 125 offices It saved having to write everything down and
across the country, staffed by 2,500 bank examin- allowed the developers to concentrate on what
ers. These examiners range from 25-year veterans was said, and they were better able to probe their
to novice assistants. thought process for inconsistencies and missing
To create a bank knowledge m anagem ent pieces. The transcripts also allowed them to see
auditor or BANKOR, the b ank’s president their responses verbatim.
formed a knowledge m anagement team consist- Three m ore sessions were held. The fourth
ing of the bank’s auditor, a representative from and last session used four cases and involved vis-
each m ajor departm ent, and a m em ber of the its to all six examiners. This round was the sign-off
board of directors to launch the project. The for testing the model. During this visit, the exam-
senior vice president of the loan departm ent iners analyzed the cases and com pared their
chaired the team. BA N K O R ’s knowledge was results with that of the system. After some mi;ior
captured through extensive interviews with six modification, the examiners approved BANKOR
federal examiners over a 15-month period. as a “fair representation” of their own analyses.
Fortunately, one knowledge developer was an A t this point, the developers were ready to pre-
examiner, and the other was an IT expert in sent the system for final user acceptance testing.
knowledge management system design. A fter knowledge capture, the developers
Interviewing the Examiners Given the seri- designed a blueprint that specified how BANKOR
ous nature of bank auditing and solvency, six sea- would be codified. They adopted CORVID —a
soned examiners were chosen for the job. With knowledge autom ation package developed by
each exam iner coming from different states EXSYS Corporation. In addition, they decided on
where the bank does business, this strategy helped the im plem entation and testing details. Testing
BANKOR address everyone’s concerns and made plays a critical role in the codification of knowl-
it easier to gain universal acceptance. Once se- edge, using rules. The package is capable of inter-
lected, the knowledge developers scheduled a facing with databases as well as documented
weeklong session with each examiner. They held knowledge in various repositories. It also is capable
the first interview not really knowing what to of im porting and exporting text information for
expect. They had selected several case studies for printout or documentation update.
the examiners to analyze. The first mistake was The bank president turned out to be the best
that the cases were too obvious. A fter the first champion of the project. He let it be known that
interview, it was apparent that it did not take an the system must be a winner. He attended sev-
experienced exam iner to analyze the cases and eral meetings of the KM team and brainstormed
reach conclusions. So, in preparation for the suc- issues dealing with auditing procedures and in-
ceeding interviews the developers selected more tegrity. He also m ade sure that all details were
complex cases that would require the examiners properly docum ented for m aintenance and fu-
to wrestle with their own decision-making process ture upgrade. The team also became the end
and apply true expertise. user when it was time to test the system. It was
Following the first interview, it took 3 months totally committed to seeing the system succeed at
to assemble and collate the captured notes. The each phase of the life cycle. The experience
developers reduced the notes for an individual turned out to be an effective knowledge-sharing
exam iner’s cases to a general conclusion about experience.

■ ■ ■ a a i
92 ■ ■ ■ ■ PART I The Basics

Evaluate Existing Infrastructure

The basis for a KM system is satisfying a need for improving the productivity and
potential of employees and the company as a whole. To do so, we need to gain familiarity
with various com ponents making up the KM strategy and supporting technology.
Identifying and evaluating the current knowledge environment makes it easier to point
out the critical missing gaps and justify the formation of a new KM environment. The
psychology behind evaluating the current knowledge infrastructure is giving the per-
ception that the current way of doing things is not conveniently abandoned in prefer-
ence for a brand new system.
As a part of this phase, we focus on system justification, scoping the evaluation,
and determining feasibility. Each step is explained here:

SYSTEM JUSTIFICATION
The goal of this step is for the knowledge developer to justify whether it is worth
undergoing the kinds of changes and investment that ensure top management support.
KM system justification involves answers to specific questions:

1. Is current knowledge going to be lost through retirement, transfer\ or departure to

other firms? Existing knowledge infrastructure becomes a problem when knowledge
workers who have been handling highly complex problems for years suddenly
decide to leave or are due to retire, resign, or simply go with the competition.
Because replacing company experts whose knowledge has been individually
hoarded is not easy to do, a KM system that provides shared expertise could
be added assurance for company survival and future growth. In the case of
BANKOR (see Box 3.4), the auditing process of existing bank auditors needed ver-
ification and possible update to a system that assured federal bank examiner com-
pliance. The need was timely and a necessary step to ensure integrity of operation.
2. Is the proposed K M system needed in several locations? If a company has only
one expert to handle a particular problem at a given time, similar problems
in other locations have to wait for the expert’s attention or someone in that loca-
tion might have to address the problem from scratch. (This is why it took more
than 18 months to put out the fires in Quwait’s 650 oil wells after the Gulf War.)
In such a case, a KM system can be used to distribute application expertise uni-
formly throughout a company’s facilities.
3. A re experts available and willing to help in building a K M system? Without reli-
able knowledge or knowledge workers, there can be no knowledge-based system.
For certain complex problems, a knowledge developer may have to work with
several knowledge workers, although dealing with several knowledge workers
presents all kinds of knowledge capture or validation challenges. With the
BANKOR project, m ultiple examiners were needed to ensure consistency of
decision making, especially with the bank doing business in several states.
4. D oes the problem in question require years o f experience and cognitive reasoning
to solve? One indicator that a problem might be a candidate for a KM system is
its nonprocedural or nontrivial nature. If the knowledge can be learned in a few
weeks and the pattern of solution is routine, then the problem is probably not
sufficiently heuristic to qualify for a KM system application. With BANKOR,
years of practical experience were required for bank auditing.
5. When undergoing knowledge capture , can the expert articulate how the problem
will be solved? The ability to explain a procedure or illustrate how a problem is to
be solved is not a skill everyone possesses. Often, the more of an expert a person
 CHAPTER 3 Knowledge Management Systems Life Cycle ■i a a 93
is, the more difficulty that person has explaining a complex procedure or a solu-
tion. One of the main obstacles is explaining heuristics. In order to do this, experts
must dig into their deep knowledge and verbalize stored heuristics. For the
process to work, the knowledge developer must also have good communication
and interpersonal skills to elicit the expert’s responses.
6. H ow critical is the knowledge to be captured? Criticality of an application has to
do with how vital it is to the company’s survival or productivity. A project that is
considered a reaction to a recent catastrophic failure in an area of the firm ’s
operation deserves close evaluation. The BANKOR project was a reaction to the
corporate scandals resulting from devious auditing corruption.
7. A re the tasks nonalgorithmic? That is, are heuristics used to solve the problem?
The problem domain should be considered a judgment m aker rather than a calcu-
lation processor. Conventional software performs calculations on data to produce
information; a KM system takes inform ation and transforms it into knowledge
and practical advice. See Figure 3.4 for a comparative view of the two systems.
8. Is there a champion in the house? For a KM system to succeed, it must be supported
from within the organization. The organization must have a champion who believes
that a KM system can be beneficial for the company and is willing to support its
development aggressively. The champion is often in a good position to coordinate
the developm ent of the system. Such a person usually has credibility with the
employees, understands their concerns, speaks their language, and knows the key
people in the organization. With BANKOR, the president was popular with bank
officers and employees, even though he had been president for less than 2 years.
His enthusiasm and honesty got the team to work overtime many days at a time.

The Scope Factor

When deciding how large the KM system should be, it is im portant to “scope” the
project first. The term scoping means “limiting the breadth and depth of the project
within the financial, human resource, and operational constraints.” The project should

FIGURE 3.4: Calculation and Judgment Views in Problem-Solving

94 « ■ * i PART I The Basics

be small enough for the knowledge developer’s capabilities, especially if it is the orga-
nization’s first attempt at developing a KM system. It must also be completed quickly
enough for the users to foresee its benefits. A complex, drawn-out project whose devel-
opment takes months and years is likely to falter on the way to completion or cause the
user to lose patience and interest, which makes training and successful use difficult.
As part of the scoping factor, several areas must be included:
• Readiness of the company’s current technology. This includes intranet, local area
networks, extranet (a company’s extended network connecting suppliers, cus-
tomers, and salespersons), decision support tools, and other information technol-
ogy tools.
• Identification of gaps and areas needing improvement in current technology to
see how well such technology will match the technical requirements of the pro-
posed KM system.
• General review and understanding of the benefits and limitations of KM tools
and components that become part of the feasibility study.

The Feasibility Question

The first phase to consider is feasibility. A feasibility study addresses the follow-
ing questions:
• Is the project doable? Can it be completed within a reasonable time?
• Is it affordable? Do the system’s potential benefits justify the cost of development?
• Is it appropriate? Just what can the firm expect to get out of it?
• Is it practicable? How frequently would the system be consulted and at what
cost?
In BANKOR’s experience, the project was clearly doable, affordable, appropriate,
and practicable, because the bank wanted to be the leader in demonstrating integrity
and reliability to customers, stockholders, and federal examiners. Actually, the project
had to be carried out, given the sensitivity of accurate auditing in the banking sector.
Including an assessment of costs against tangible and intangible benefits, these
questions evaluate the project in detail. The result of the study is a proposal that sum-
marizes what is known and what is going to be done. Once approved by top manage-
ment, it paves the way for KM system design. Although a feasibility study is not con-
sidered a formal part of KM system development, identifying the proposed system’s
many ramifications (economic, technological, and behavioral) is necessary front-end
work. This way, there should be no surprises. Attempting such a project makes no sense
unless the heart and soul of the company are committed to the effort, especially if it is
the organization’s first KM system.
Econom ic feasibility, better known as cost/benefit analysis , determ ines to what
extent a new system is cost-effective. Technical feasibility is determined by evaluating the
hardware and supportive software within the framework of the company’s IT infrastruc-
ture. Behavioral feasibility includes training management and employees in the use of
the KM system. All these factors must be viewed collectively for a successful installation.
The traditional approach to conducting a feasibility study could be useful in build-
ing a KM system. It involves several tasks:
• Form a K M team. Forming a KM team means formalizing a commitment to the
KM development process. Team members are selected from various areas in the
organization; selection is based on criteria designed to bring the best representa-
tives for the duration of the project. This is exactly what was done to form the
 CHAPTER 3 Knowledge Management Systems Life Cycle ai aa 95
KM team for BANKOR. There was fair representation of both talent and depart-
ments on a full-time basis.
• Prepare a master plan. A master plan lays out the steps to be taken throughout the
life cycle of the project. It could be a physical representation like a flowchart that
specifies the activities to be carried out and the relationships among the activities.
• Evaluate cost/performance o f the proposed K M system. In this phase, the key
question is “What are we getting in return for the cost of building this new KM
system?” The terms used are breakeven analysis and payback analysis. Breakeven
is the point at which the costs of the current and the proposed systems are equal.
Beyond that is net savings or net benefits. Payback is deciding how long it will
take a system to produce enough savings to pay for developmental costs.
Essentially, performance criteria are evaluated against tangible and intangi-
ble costs to determine the most cost-effective system. Obviously, the more tangi-
ble the costs, the easier it will be to measure the benefits. It is often the case that
the benefits are more intangible than tangible, which makes it difficult to sell to
top management (see Box 3.5).
• Quantify system criteria and costs. There are various metrics to do so. One proce-
dure for quantifying system criteria is as follows:
• Determine a weight factor for each criterion based on its perceived impact on
the system.
• Assign a relative rating to each criterion. For example, on a scale of 1 to 5,1 is
rated “poor” and 5 is rated “excellent.”
• Multiply the weight factor by the relative rating and sum the score for each
alternative system. Table 3.3 shows an example of a weighted candidate
evaluation system compared to the existing or an alternative one. Assuming
the weight and rating factors are accurate, the system with the higher score is
judged to be the better system. Additional supportive information would be
helpful in reinforcing the choice.
In summary, several factors must be considered in identifying an appropriate KM
system. Many such projects fail because of poor front-end evaluation. Also, without a
champion or someone within the firm who can support and promote the project at all
times, the project is likely to fall by the wayside.

Importance of User Support

In conventional information systems, systems analysts often build the system and
then worry about selling it to the user through demonstrations and training. Such “sell-
ing” should begin early in the development process. The same is true with KM systems.
The user should be sold on it early in the KMSLC. At this stage of the development life
cycle, a knowledge developer needs to address a number of questions:
• Does the intended user know the new KM system is being developed? How is it
perceived?
• How involved should the ultimate user be in the building process?
• What user training will be required when the KM system is up and running?
• What operational support must be provided?
These questions prompt idea generation. They help determ ine whether the pro-
posed KM system is doable and justifiable; they also allow a move to the next step,
knowledge capture. Without an honest “sell” to the user, all kinds of problems can arise
when it is time to implement the system. Upper management should also give its bless-
ing to the project early by allocating sufficient resources for its development.
96 ■* i ■ PART I The Basics

BOX 3.5

VALUATION OF KNOWLEDGE CAPITAL

CALCULATING KNOWLEDGE CAPITAL many consider to be intangible. However, with
the rising importance of knowledge assets, the time
The valuation of knowledge capital makes it pos-
has come to place the management of knowledge
sible to assess the worth of the people who pos-
on the agenda of executive managers, financial
sess the accumulated knowledge about an organi-
analysts, and shareholders. To that end, one must
zation. They are the individuals who leave the
start by developing an independently verifiable
workplace every night (and may never return),
quantification of the worth of knowledge assets.
storing in their heads the know-how acquired
The allocation of the respective contributions
while receiving full pay. Their brains are reposito-
of knowledge capital and financial capital to prof-
ries of knowledge accumulated over untold hours
its can be made if one recognizes that financial
of listening and talking while not delivering any
capital is now a commodity —readily available at
goods or services to paying customers.
a price that reflects the interest rate that a firm
The employees’ minds, and the files they man-
pays for its borrowings. However, what makes a
age, carry a share of the company’s knowledge cap-
company prosper is not financial capital—which
ital. This makes every employee a custodian of the
anyone can obtain for a price —but the effective-
most im portant assets a firm owns, even though
ness with which knowledge capital is put to use.
these assets never show up on any financial reports.
Therefore, the annual returns realized on knowl-
In contrast, the custodianship of financial assets
edge capital can be isolated after paying a
has become a well-defined discipline that depends
“ren tal” for the financial capital and then sub-
on procedures and regulations on how to account
tracting that amount from the profits.
for and report these resources. Over a period of
What remains is “economic profit," or the
many years, an elaborate framework involving
economic value-added, because it accounts for
accounts, auditors, reporting standards, and govern-
those missing elements that represent everything
ment oversight has been developed to do that. The
not shown on a conventional balance sheet. By fil-
financial assets of corporations represent only a
tering out the contributions of financial capital
small share of total corporate assets. By far, the most
from the reported profits, we are left with a resid-
important assets are now in the form of corporate
ual that is entirely attributable to what knowledge
knowledge, but the custodianship, accounting, and
capital has actually delivered. In other words,
reporting standards for these resources do not exist.
knowledge value-added is the annual yield a firm
Financial executives have shown a rem ark-
realizes from its knowledge capital assets.
able reluctance to put num bers on something

SOURCE: Excerpted from Strassmann, Paul A. “Calculating Knowledge Capital ” files.strassmann.com/pubs/km/1999-lO.php,

October 1999. Date accessed August 2002.

In summary, the first step in the KMSLC is to decide what is wanted, identify the
goals of the proposed system, assess feasibility, plan how to begin, locate a champion,
inform users and their m anagers of the project, and gain their support within the
scope of the project. Above all, every KM system must begin with ready knowledge,
knowledgeable experts ready to work with a knowledge developer in building the
infrastructure.
 CHAPTER 3 Knowledge M anagement Systems Life Cycle u mmm 97

TABLE 3.3 Weighted Evaluation Matrix

System A System B
Evaluation Criteria Weight Rating Score Rating Score
Performance
Knowledge quality 4 5 20 4 16
System reliability 2 4 08 4 08
Knowledge currency 3 4 12 3 09

Costs
Payback 2 3 06 4 08
System developm ent 4 4 16 3 12
System operation 3 4 12 3 09
User training 5 5 25 3 15
99 78

Role of Strategic Planning

As with any other application, there are concerns to ponder before launching a
new system. Issues such as planning your new KM system, selling it to the employees
and top management, providing good service and maintaining security—all take on a
new meaning when applied to KM developm ent and support. As a consequence of
evaluating the existing infrastructure, the company should devise a strategic plan with
vision aimed at advancing the short-term and long-term objectives of the business with
the KM system in mind.
What is being emphasized is the ultimate goal of a KM system—connecting users
with knowledge content. That is why it is so im portant to review knowledge needs
throughout the company and to inventory knowledge content assets that you wish to
capture. It is risky to plunge ahead before strategizing. Consider the following areas:
• Vision —What is your business trying to achieve? How will it be done? How will
the KM system fulfill the goals? How will you measure success?
• Resources —How much can your business afford to build the right KM system? Is
there talent aboard that can take the responsibility and assure that vision
becomes reality?
• Culture —Is your business politically amenable for coordinating efforts to support
this new approach to knowledge sharing? Who will ultimately control the KM
system’s content and user feedback?
Once business strategy is made final, the next step is to demonstrate a viable link
between KM and strategy. The idea is to elevate KM to the higher level of strategy and
ensure that the final system is in compliance with strategy. Related to this step is the
importance of analyzing knowledge gaps (things that need corrections or additions) and
relating them to strategic gaps. There is also the need to determine the right questions to
ask and to package knowledge as part of the proposed KM system. The upshot is to use
knowledge to create value for the product, the service, or the organizational process.
Figure 3.5 illustrates the relationship betw een and the need to match business
strategy with that of KM strategy early in the KM system life cycle. Threats and feedback
PART I The Basics

Business Strategic
Environment Plan

FIGURE 3.5: Matching Business Strategy with KM Strategy

from the environment (such as government regulations, customers, and competitive

threats) impact business strategy on a day-to-day basis. Likewise, the business environ-
ment impacts both KM strategy and KM technology, which enable KM strategy. The
strategic plan, in turn, drives KM strategy, which means that KM strategy must be
within the framework of the company’s strategic plan for it to work and be supported
by top management.

Form the KM Team

After evaluation of the company’s existing infrastructure is complete, a KM team
should be formed; this team stays with the building process until the final installation.
Forming a KM team means:
• Identifying the key units, departments, branches, or divisions as the key stakeholders
in the prospective KM system. Each stakeholder unit has certain knowledge require-
ments and one or more experts who could represent the unit on the KM team. In
forming the BANKOR KM team, a representative-from every department involved
in the federal examiners’ quarterly bank audit was included in the KM team. Each
member brought unique knowledge and cited specific audit strategies that were
tested through knowledge capture; the outside examiners acted as experts.
• Balancing the team size and competency organizationally, strategically, and tech-
nologically. In other words, the KM team should have a balanced complement of
experts in each of these areas to collaborate on the multidimensional nature of
the KM system building process.
Teamwork can be pleasurable and smooth running or it can be fraught with prob-
lems. Team success depends on a number of factors:
• Caliber o f team members in terms o f personality, communication skills, and experi-
ence. Engaging personalities with good communication skills usually do better at
 CHAPTER 3 Knowledge Management Systems Life Cycle i ■■■ 99
problem-solving and conflict resolution than laid-back personalities. Also, team
members with experience in a specific domain garner attention and respect
within the team.
• Team size. There should be representation with qualifications. In a large multina-
tional firm, where there are hundreds of departments or divisions, a representa-
tive from each area would make teamwork unwieldy and counterproductive.
Based on the author’s consulting experience, team size from 4 to 20 is manage-
able. Beyond that, a team is divided into subcommittees with more distinctive
assignments. When that happens, things tend to drag, and coordination among
subcommittees becomes a challenge.
• Complexity o f the project. The complexity of a KM system depends on the size of
the organization, the nature of the products, the level of sophistication of existing
technology, and the expertise of team members. Working against a tight deadline
can also make the project more difficult to handle. In the case of BANKOR, the
KM system was extremely complex, especially with the tax and regulatory
requirements of the states in which the bank conducted business.
• Leadership and team motivation. Effective leadership can generate motivation,
cooperation, and coordination within the team to ensure a successful outcome.
Team motivation means a volunteering spirit and willingness to devote time for
the KM system building process. This is influenced by the size of the team, the
perceived level of expertise of each team member as compared to other team
members, and how freely team members are encouraged to participate.
• Promising more than can be realistically delivered. There is nothing worse for a
team than to build high hopes for top management when the odds are stacked
against such achievement. Underestimating factors such as development costs
and outcomes beclouds credence of the team to deliver with integrity.

Knowledge Capture
In the third step in the KMSLC, knowledge capture takes on several directions and
meanings. Briefly, explicit knowledge is captured in repositories from documentation,
files, and other media. In contrast, tacit knowledge is captured from company experts
and from knowledge stored in databases for all authorized employees. D ata mining
also enters the picture in terms of using intelligent agents that analyze the data ware-
house and come up with new findings that could lead to new products or services.
The focus of Chapter 5 is on capturing tacit knowledge by using various tools and
methodologies. Knowledge capture involves eliciting, analyzing, and interpreting the
knowledge that a human expert uses to solve a particular problem. It corresponds to
systems analysis in conventional system development. In both cases, interviews with
knowledgeable people are used to capture information and knowledge. Sometimes,
interviewing the ultimate user of the knowledge management system is just as impor-
tant as interviewing the expert whose knowledge you are trying to capture.
In KM system development, the knowledge developer acquires heuristic knowl-
edge from the expert(s) in order to build the knowledge base. In Chapter 4, we discuss
problems and procedures in the knowledge capture process. For example, the user has
more to gain from the knowledge base than the expert who is helping to build it. The
expert hands over years of knowledge for a system that someone else will use. The bene-
fits are not the same, which explains some experts’ lack of cooperation and motivation
throughout the system development life cycle.
Knowledge capture and transfer are often carried out through teams, not just
through individuals. As shown in Figure 3.6, common knowledge is created by asking a
100 ■ i ■ a PART I The Basics

FIGURE 3.6: Knowledge Capture and Transfer Through Teafns

team to perform a task. A fter the outcom e is achieved, the knowledge developer
explores with the team the relationship between team action and outcome. If the team
agrees that common knowledge has been gained, it becomes available in a knowledge
base by translating it into a form usable by others in the organization. It can be in the
form of a document, a case with rules and parameters, or a scenario.
Knowledge capture plays a unique role in various phases of the KMSLC. As shown
in Table 3.4, knowledge capture includes determining feasibility, choosing the expert,
tapping the expert’s knowledge, and retapping knowledge to plug gaps in the system
and to verify and validate the knowledge base after the system is in operation. So, in
KMSLC, knowledge capture is an evolving step, not a one-time, front-end step as it is
in conventional information system development.

TABLE 3.4 Knowledge Capture Activities in the KMSLC

KM SLC Step Knowledge Capture A ctivity
1. Determ ine feasibility • Seek out a champion
• Locate a cooperative expert

2. Capture knowledge • Apply appropriate tools to capture

expert’s knowledge

3. Design KM blueprint • Design KM architecture

4. Verify and validate the KM system • Correct for knowledge integrity and
work closely with expert through rapid
prototyping

5. Implement the KM system • Work with the user to ensure system

acceptance and proper training

6. Manage change and reward structure • Reinforce change and reward compliance
through human resources

7. Evaluate postsystem • Recapture new knowledge and update

the knowledge base
 CHAPTER 3 Knowledge Management Systems Life Cycle « i i i 101

ROLE OF RAPID PROTOTYPING

Most knowledge developers use an iterative approach in knowledge capture, which means
a series of repeated actions. For example, the knowledge developer starts with a small-
scale system, a prototype, based on the limited knowledge captured from the expert dur-
ing the first few sessions. What turns the approach into rapid prototyping is the following:
• The knowledge developer shows the expert a skeletal procedure based on rudi-
mentary knowledge acquired from the expert during the past two sessions.
• The expert reacts by saying, “This is all right, but there’s more to it. Let me tell
you what you need to incorporate into this prototype . . . ” This is precisely how
the knowledge developers in the BANKOR system conducted their interviews.
• While the expert watches, the knowledge developer enters the additional knowl-
edge into a computer-based system that represents the prototype.
• The knowledge developer reruns the modified prototype and continues entering
additional knowledge or making modifications as suggested by the expert until
the expert says, “This is about right. I think you’ve got it this time.”
This spontaneous, iterative building of a knowledge base is referred to as rapid
prototyping. The process continues through several sessions and weeks of work until
the knowledge base moves out of the prototype stage and becomes a full-fledged
knowledge base that is ready for use. Rapid prototyping is essential for large systems,
because the cost of a poorly structured system that is unusable can be prohibitive.
Other benefits of rapid prototyping include the following:
• It documents for the expert and others that progress is being made on the project.
• Mistakes can be quickly corrected.
• The system is tested each time new knowledge or modifications are incorporated.
• It yields a tangible product at an early stage.
• The system “grows” in step with increasing understanding of what the user will
learn and how the expert will provide it.
• It promotes accelerated knowledge capture.
• It demonstrates the capabilities of the resulting knowledge base.
The “smash and grab” approach —in which the knowledge developer “grabs” the
attention of the expert for a few hours over a few sessions and ends up with a “work-
able” knowledge base —is not recommended. Rem em ber that knowledge bases are
based on human heuristics (rules of thumb) and human knowledge, not on quantita-
tive or syntactic data. Rapid prototyping fosters maturation and “value-added” activi-
ties. It requires patience, concentration, and knowing how to make changes on the fly
without errors.

SELECTING AN EXPERT
In expressing their differing views about the nature of expertise, some argue that the
entire breadth of human expertise must be studied before attempting to capture the
knowledge necessary to build a knowledge base. Others support a narrower focus that
limits the role of the expert to knowledge capture activities. In either case, the goal is a
knowledge base that represents expertise, rather than the expert.
In working with the expert, the knowledge developer will occasionally run into
vague expressions of thought or reasoning processes that must either be clarified by
the expert or factored into the system. For example, “fuzzy” expressions can be as-
signed certainty factors to reflect the extent of “fuzziness” of the expression. Certainty
102 i fl a i PART I The Basics

factors require careful analysis to assure proper representation in the knowledge base.
In fact, much of the knowledge that the expert provides is based on heuristics, which
tend to be qualitative, rather than quantitative.
Finally, a competent and cooperative expert is essential to the success of knowledge
capture. The expert must be able to communicate information understandably and in
sufficient detail. The questions that face every knowledge base project are as follows:
• How can one know that the so-called expert is in fact an expert? In the example
of BANKOR, the examiners were certified and each had more than 10 years’
experience.
• Will the expert stay with the project to the end?
• What backup is available in case the expert loses interest, decides to leave, or sim-
ply is no longer available?
• How is the knowledge developer to know what is and what is not within the
expert’s area of expertise?
Sometimes, either because of the nature of the problem or its importance to the
organization, more than one expert should be involved in addressing the problem
domain. Dealing with several experts is not an easy task. The knowledge developer
may have to use different tools in different ways to access the experts’ knowledge and
promote agreement among them before the captured knowledge becomes part of the
knowledge base. Dealing with single and multiple experts is discussed in Chapter 5.

ROLE OF THE KNOWLEDGE DEVELOPER

The knowledge developer is an important player in the KMSLC. As explained in detail
in Chapter 4, he or she is the architect of the system. This one person identifies the
problem domain, captures the knowledge, writes and tests the heuristics that represent
the knowledge, and coordinates the entire project from beginning to end.
Such a pivotal job requires certain qualifications. The most im portant attributes are
excellent communication skills, an understanding of knowledge capture tools, familiar-
ity with technology, tolerance for ambiguity, ability to work well with other profession-
als including experts, being a conceptual thinker, and having a personality that moti-
vates people to work together as a team. The knowledge developer’s job requires
interaction with a number of individuals throughout the KMSLC. The most frequent
interactions, shown in Figure 3.7, are with the organization’s knower(s), the knowledge
worker, and most likely, the champion. Each person can make or break the project. In
the case of BANKOR, the knowledge developers had a great rapport with the presi-
dent, whose time and effort championed the application through implementation.

Design the KM Blueprint

This phase is the beginning of designing the IT infrastructure and the knowledge
management architecture. The knowledge management team relies on such a blueprint
to proceed with the actual design and deployment of the KM system. A KM blueprint,
referred to as the K M system design, addresses several im portant issues. Once created,
the next step is to actually put together the KM system.
1. Aim for system interoperability and scalability with existing company IT infra-
structure.
2. Finalize the scope of the proposed KM system with realized net benefits in mind.
3. Decide on the required system components, such as user interface options,
knowledge directories, and mining tools.
 CHAPTER 3 Knowledge Management Systems Life Cycle i i ■■ 103

■ ■ ■ ■ ■ « FIGURE 3.7: The Central Role of the Knowledge Developer

4. Develop the key layers of the KM architecture to meet your company’s require-
ments. As shown in Figure 3.8, the key layers are as follows:
a. The user interface is what the user sees and works with in terms of accessing
and working with the knowledge base and other repositories. Without an
effective user interface, even the best KM system is bound to fail. A browser
is used to link the user to the Internet.
b. The authentication/security layer screens incoming requests for security pur-
poses. For example, firewalls are installed by the company’s IT departm ent
to ensure that certain files do not leave company premises or certain incom-
ing information does not contaminate existing repositories. Passwords and
other authorization protocols are used to ensure integrity of the knowledge
base and reliability of the overall systems at all times.
c. Collaborative agents and filtering are designed to provide a near-personal-
ized presentation of knowledge or information to m eet the user’s require-
ments. Intelligent agents (specialized software) are employed to do “intelli-
gent” searches in a database, knowledge base, or other repositories to
expedite user requests and display the right information.
d. The application layer is used to communicate with the actual application in
use. This is where the user begins to do som ething useful with the n et-
works—browsing a Web site, sending e-mail, or transferring a file between
file servers and client computers. It answers the question “What informa-
tion do I send to my partner?” It is simply two useful programs talking to
each other. It also defines request and response formats. For example, an
e-mail client browser program talks to the e-mail server program, saying
som ething like “Deliver this message to [email protected].” Rem em -
ber that each type of program (e-mail, file transfer, and so forth) has its
own protocol. The application layer protocol assumes that the next layer
down (transport layer) will take care of passing the message along to the
destination.
104 ■ ■ ■ ■ PART I The Basics

Authentication/Security Layer
(includes access identification, firewalls, and user recognition)
Internal Layer That
the Company’s IT
Infrastructure Controls

Collaborative Agents and Filtering

(intelligent robots that can disseminate news, warnings, and notifications)
Agent Technology Is
Intelligence Within a
Knowledge Management System

Application Layer
(collaborative work tools, videoconferencing, decision support tools,
group decision support tools, yellow pages, and so forth)
Upper Part of the Data
Communication Network Layer

Transport/Internet Layer
[transmission control protocol/Internet protocol (TCP/IP), e-mail, document exchange,
simple mail transport protocol (SMTP), and so forth]
Manages Transmission
of Data Between Computers

Physical Layer
(cables, physical wires, modems for transmission)

Transmits Raw Data in Bit

Format to Destination

Repositories

Data Warehouse Documents

and Other
Databases
Files
Legacy Applications

FIGURE 3.8: Key Layers of a KM System Infrastructure

SOURCE: Adapted from Tiwana, Amrit. The Knowledge Management Toolkit. Upper Saddle River, NJ:
Prentice Hall, 2000, p. 309.

Rem em ber that there are standards that specify how two application
programs should communicate at the application layer. For example, in Web
service, there is a program called “browser” that resides on the client PC and
a Web server application program on the Web server. The standard on the
application layer is the hypertext transport protocol, or HTTP. Its function is
to govern requests and responses between the browser and the Web server
application program. It is usually the beginning of a Web address in the form
 CHAPTER 3 Knowledge Management Systems Life Cycle i ■■« 105

of http://. E-mail uses other applications that use different application layer-
standards such as SMTP (simple mail transfer protocol) and POP (point of
presence) application layer standards to send and receive e-rpail, respectively.
e. The transport Internet layer is the most critical part of the KM network sys-
tem. The function of the transport layer is to manage the transmission or
the flow control of data between two com puters or across a network. It
also allows two com puters to communicate, even if the com puters were
made by different vendors. The standard for the transport layer is transmis­
sion control protocol (TC P ), which specifies how two host computers will
work together.
In contrast, the Internet layer routes messages across multiple nodes of
networks for delivery. It also handles network congestion to minimize per-
formance problems. A typical message is “Send this packet to com puter
number 190.172.63.08 via computer number 123.32.12.14, which is on a net-
work one hop away.” The standard for routing messages is Internet protocol
(IP ). TCP and IP work together, and that is the reason that in the literature
you always see them as TCP/IP.
f. The physical layer is the lowest layer in the journey of a message from source
to destination. It converts bits into signals for outgoing messages and signals
into bits for incoming messages. It answers the question “How do I use the
medium (cable, physical wiring) for transmission?” The physical layer also
uses modems (telephone network standards) to transmit the message as raw
data to the destination. Modems are used only to link a user host to the first
router (a box with intelligent software that routes messages to other routers).
g. Repositories are hard disks or storage devices that hold explicit and tacit
knowledge and the rules associated with them. O ther repositories include
legacy (old or traditional applications like accounts receivable) databases,
Web databases, e-mail databases, and UNIX databases.

Test the KM System

The testing phase involves two steps: a verification procedure and a validation pro-
cedure. A verification procedure ensures that the system is right—that the programs do
what they are designed to do. The internal makeup of the system is checked to see that
the right knowledge is available when needed in the form at needed and that any rules
will fire when they are supposed to fire. In this way, the technical performance of the
system is evaluated.
The second procedure is called validation. This test ensures that the system is the
right system —that it meets the user’s expectations, that it is user friendly, and that it will
be usable and scalable on demand. Validation provides assurance that the solutions or
advice derived from the knowledge management system comes close enough to that of
the hum an expert. In other words, the validation process checks reliability of the
knowledge m anagem ent system. In the case of BANKOR, the system was initially
tested weekly to see how well the bank met the federal regulatory audit procedures.
Several changes had to be'm ade because of oversight on em bedding old auditing
requirem ents into the current system. Also, a money laundering module had to be
rewritten to reflect the changes made after the federal government began to track Bin
Laden’s assets in the United States and abroad.
Validation of KM systems is not foolproof. System validation is a long-term item:
Human experts must monitor system performance continually. Eventually, systems will
be able to renew their certification and vouch for their own reliability.
106 a ■ ■ a PART I The Basics

Implement the KM System

Once the appropriate knowledge has been captured, encoded in the knowledge base,
verified, and validated, the next task of the knowledge developer is to implement or deploy
the proposed system on a computer or a server. Implementation means converting a new
KM system into actual operation. Conversion is a major step in implementation. The other
steps are postimplementation review and KM system maintenance. Another way of look-
ing at implementation is that it is the transformation of a precise representation of knowl-
edge into a machine-executable equivalent—a specific program or software package.

ROLE OF QUALITY ASSURANCE

Quality assurance is the development of controls to ensure a quality KM system. This
feature is actually emphasized in every stage of the system development process. The
goal is to have as error-free and reliable a system as possible. Knowledge developers
under time and budget constraints could compromise quality, especially during knowl-
edge capture. The chief knowledge officer (CKO) has the responsibility to ensure that
every step in knowledge management is error-free.
There are several kinds of errors to look for in knowledge base design:
• Reasoning errors. Errors can occur in inductive or deductive reasoning. An ex-
ample of an inductive reasoning error is This firm has never had a layoff There-
fore, it will never have a layoff in the future. The likelihood of error is higher if this
conclusion is accepted. Likewise, a deductive reasoning error is If you are preg-
nant, then you will be overweight. The deduction that pregnancy causes weight
gain does not apply in all cases. Even if it did, one must still consider the question
of how much weight is “overweight.” Such issues are illustrative of fuzzy logic,
covered in detail in expert systems courses.
• Ambiguity. A statement is ambiguous when it has more than one meaning or can
be interpreted in more than one way. For example, “Sell the bad stock” is an
ambiguous directive. Which stock is bad? What is meant by “bad”? How bad is
“bad”? The subjective nature of decision making in such a case raises questions of
validity, reliability, and integrity of the decision and, therefore, the knowledge base.
• Incompleteness. A statement that represents some but not all pieces of knowledge
is incomplete knowledge. For example, does “Adjust the air conditioner” mean to
turn it on or off, or to make it cooler or warmer?
• False representation. A statement that incorrectly represents pieces of knowledge
is false representation. Take the example, “The small key will open the door.”
Which small key? An incorrect representation presents the possibility of two
kinds of errors: false positive and false negative error. With false positive error, a
rule is accepted when it should be rejected. In the “small key” example, if the
small key will not open the door when it should, then accepting the key gives a
false positive. In contrast, when a rule is rejected when it should be accepted, it is
descriptive of false negative.

USER TRAINING
A major part of the implementation phase is training company employees on the new
KM system. The level and duration of training depend on the user’s knowledge level
and the system’s attributes. Users range from novices to experts. A novice is generally a
casual user with limited knowledge of IT. An expert is someone with prior IT experi-
ence and who keeps abreast of the technology. Users are also classified as tutor, pupil,
or customer.
 CHAPTER 3 Knowledge Management Systems Life Cycle aaaa 107

The tutor user acquires a working knowledge of an existing knowledge base in

order to keep the system current. This end user is responsible for system maintenance.
The knowledge developer normally trains the tutor user on procedures for incorporat-
ing new knowledge into the knowledge base.
The responsibilities and interest of pupil users and customer users fall within the
actual use of the KM system. The knowledge developer’s role is to familiarize the end
users with the ways in which the system will be best used. The pupil user is an unskilled
worker trying to gain some understanding of the captured knowledge. The purpose of
training pupil users is to gain their acceptance of the KM system. The customer user is
one with interest in knowing how to use the KM system on a regular basis.
Neither the pupil user nor the customer user needs in-depth training during the
early stages of KM system development. Training in the early phases uses development
funds inefficiently. During the time between prototyping and a working system, the
pupil users and customer users may forget most details.
User training is also influenced by the requirem ents of the KM system. System
requirements range from simple and user friendly to advanced and less user friendly. A
user friendly KM system is almost self-instructional with menu-driven features and easy-
to-use manuals. The duration of training is usually measured in hours to days. Advanced
KM systems require knowledge of queries in high-level programming or artificial intelli-
gence (A l) languages and take days to weeks to learn, even for the experienced user.
Training must be geared to the specific user based on capabilities, experience, and
system complexity. The most effective training is supported by a well-written user man-
ual, easy-to-use explanatory facility, and job aids. User manuals can be invaluable train-
ing documents, especially if the user is geographically isolated. A well-written user
manual is highly illustrated and contains an index for reference. Graphics, photographs,
and templates provide quick reference and are invaluable teaching aids. Comparable
information should also be available in the software via a help key on the keyboard.

Manage Change and Reward Structure

Regardless of what is being converted or how well deploym ent is carried out,
knowledge sharing does not come easy to many people. Implementation means change,
and people in general resist change. Implementation of a KM system is the initiation of
a new order. People become anxious when they do not know what the system will offer
and how it will affect their current jobs and decision-making quality. The result is stress
and further resistance to change. The resistors include the following:

• Experts. Some domain experts have anxiety about the potential impact of sharing
knowledge of their jobs in the organization. Experts on the way to retirement
worry less, but they may still lack motivation unless they are properly compen-
sated for their efforts.
• Regular employees (users). Participants in general resent lack of recognition
(sometimes compensation), especially when they have put time in building the
KM system.
• Troublemakers. Those left out of building the KM system or chronic complainers
tend to obstruct the installation, cause delays, and may even prompt cancellation
of the installation.
• N arrow-minded “superstars. ” Technical people in the organization’s IT depart-
ment sometimes resist any change that they did not initiate or approve in
advance. Others veto a project not in their area of interest. Without management
support, such resistance can spell doom for a new installation.
108 • a « i PART I The Basics

Resistance is displayed in three personal reactions:

• Projection: hostility toward peers.

• Avoidance: withdrawal from the scene, such as calling in sick.
• Aggression: killing the system, because of uncertainty of its operation or use.

A psychological element that explains resistance to change is the value that users
place on knowledge and decision making. In most organizations, knowledge means
power. While knowledge developers build KM systems that promote knowledge shar-
ing through knowledge bases, domain experts who stand to lose monopoly on knowl-
edge may resist a new system installation. Most resistance relates to the perceived
impact of the system on one’s job or status in the organization. Of course, a few people
resist change of any kind.
Resistance also has much to do with the individual personality, the organizational
structure, and the group relations within the area where the KM system will be
installed. User education, training, and participation in the building process can help
reduce resistance to change.
Because a m ajor user concern in KM system im plem entation is how to work
the system, users frequently ask such questions as W hat functions are available?
How do I access each function? How do I know if the system has answered my
questions correctly? How current is the knowledge base? A nother user concern is how
the KM system reaches conclusions or lines up with the real-world problem as seen
by the user. The knowledge developer must dem onstrate the system and provide
detailed training.

Postsystem Evaluation
A fter the KM system has been deployed and the operation is “up and run-
ning,” the effect of the new system on the organization should be carefully evaluated.
System impact must be assessed in terms of its effects on the people, procedures, and
performance of the business. More specifically, the main areas of concern are quality
of decision making, attitude of end users, and costs of knowledge processing and
update.
Several key questions are asked in the postimplementation stage:

• How has the KM system changed the accuracy and timeliness of decision making?
• Has the new system caused organizational changes? How constructive have the
changes been?
• How has the new KM system affected the attitude of the end users? In what way?
Was it worth it?
• How has the new KM system changed the cost of operating the business? How
significant was it?
• In what way has the new system affected relationships between end users in the
organization?
• Do the solutions derived from the new system justify the cost of investment?

The objective is to evaluate the KM system against standards and determine how
well it meets the goals set in advance. This process is actually related to validation. The
user initiates the review, which prompts a procedure for maintenance or enhancement.
Enhancement means upgrading the system to meet a new set of requirements; mainte-
nance means making corrections to meet the initial system requirements.
 CHAPTER 3 Knowledge Management Systems Life Cycle i ■ a ■ 109

Implications for Knowledge M anagem ent

Assuming the technology is available and ready to use, several managerial factors
should be considered:

• An organization considering a KM system as part of its information systems envi-

ronment must make a commitment to user education and training prior to building
the system. Knowledge sharing is not that straightforward in many organizations.
• Top management should be approached with facts about the costs and benefits of
the proposed KM system. Being sold on the project means assurance of financial
and technical support.
• If a KM system is anywhere on the organization’s horizon, human resources or
the IT departm ent should begin training knowledge developers and others who
have the potential to do knowledge engineering.
• Domain experts must be recognized and rewarded in ways that make them feel it
is worth their time to cooperate. Assigning an expert to the project without such
rewards can jeopardize the whole process.
• Finally, for an organization to anticipate its future technology needs, it is
extremely important to do long-range strategic planning. Such planning can help
the firm to attain its desired outcomes. Introducing leading-edge technology such
as KM systems can help the organization to achieve competitive advantage.

One of the critical issues raised during implementation is knowledge system main-
tenance. Maintenance of KM systems continues to be a nebulous area. Some of the
questions to be addressed by management include the following:

• Who will be in charge of maintenance?

• What skills should the maintenance specialist have? What is the best way to train
him or her?
• What incentives should be provided to ensure quality maintenance?
• What kinds of support and funding are needed?
• What relationship should be established between the maintenance of a KM sys-
tem and the IT staff of the organization?

The enhancement function is closely related to maintenance. Because experts con-

stantly upgrade their knowledge, it makes sense to do the same with the KM system
to keep it current. One question is “How quickly does the problem change?” The
answer then determines how one would justify the enhancement, given the nature of
the solution(s).
A nother managerial issue to consider is how one would know whether the imple-
m entation would be a success. In addition to technical considerations, one must be
aware of the people factor. People rallying behind technology can increase the success
of implementation. Additionally, a KM system cannot succeed if the issues of cost and
how that cost is going to be absorbed have not been resolved. Any of these issues can
inhibit successful implementation.
Other managerial issues affect the direction and outcome of KM system projects.
For example, the depth of the knowledge required to solve the problem is often an
unknown commodity due to the nature of the expertise and the limitations of the
knowledge developer. Forcing clarification too early in the building phase can result in
a less-than-satisfactory system. Therefore, managing expectations and scoping the sys-
tem are crucial early tasks.
110 ■ ■ ■ i PARTI The Basics

F in ally, th e r o le o f th e e x p e r t in sy s te m im p le m e n ta tio n is a lso im p o rta n t.

M a n a g em en t sh o u ld p rop erly c o m p e n sa te th e d om ain ex p ert for his or her effort in
d e v e lo p in g a K M sy stem . T he C K O m ay also p lay a p ro a ctiv e ro le in the im p le m e n ta -
tion p h a se by sim p ly sharing in terest in th e K M sy stem . A lack o f this kind o f en th u si-
asm and o v ert su p p ort p u ts a d a m p er o n th e w h o le id ea o f k n o w le d g e sharing and
k n o w le d g e m a n a g em en t.

Su m m a r y a a a a
• Building a KM system can be viewed as a life cycle. The life cycle begins when a
knowledge management system is determined to be doable, affordable, and prac-
ticable, with value added for company profitability and growth. Knowledge devel-
opers interview users and work with experts to develop a system. Discipline, good
documentation, coordination, and regular management review characterize the
development of KM system life cycle.
• Conventional and KM systems’ development life cycles differ.
• A systems analyst deals with data and information obtained from the user;
the knowledge developer deals with tacit knowledge acquired from
human experts.
• The main interface for the systems analyst is the user, who knows the
problem but not the solution. The main interface with the knowledge
developer is the human expert, who knows the problem and the solution.
• Conventional system developm ent is primarily sequential; KMSLC is
incremental and interactive.
• Testing is done at the end of conventional information system develop-
ment; verification and validation are perform ed throughout the KM
development life cycle.
• Conventional and KM systems’ development life cycles are also similar.
• Both begin with a problem and end with a solution.
• Both begin with information gathering to ensure a clear understanding of
the users’ requirements or the problem at hand.
• Both involve testing the system and ensuring that the system is satisfactory.
• Particular development tools are used to build each system.
• The first step in building a KM system is identifying the problem domain, fol-
lowed by a feasibility study for evaluating the problem in detail. This includes
weighing the total costs against the potential tangible and intangible benefits.
Early scoping is important. The project should be of a manageable size, and the
organization should be able to foresee its benefits.
• Knowledge capture involves elicitation, analysis, and interpretation of the knowl-
edge that a human expert uses to solve a particular problem.
• Most KM systems begin as small-scale prototypes based on the limited knowl-
edge acquired during the first few sessions with the human expert. The system
grows gradually as the knowledge developer gains new insights from the expert
and adds them to the prototype.
• Once the KM blueprint is made final, verification and validation ensure that the
system is right and that we have the right system, respectively.
• System implementation is the process of organizing the knowledge and integrat-
ing it with the testing strategy of verification and validation. A system must be
modified or updated as new knowledge is captured.
 CHAPTER 3 Knowledge Management Systems Life Cycle i i i ■ 111

• Most barriers to the development of a KM system are nontechnical. Lack of sup-

port from top management, knowledge developers’ limited interpersonal skills,
experts’ poor communications skills, and users’ resistance are all barriers that
must be addressed.
• The ultimate goal of every KM system is successful implementation and deploy-
ment. Implementation means change, and people in general resist change. The
resistors include experts, nonexperts, troublemakers, and narrow-minded techni-
cal “superstars.” Resistance is displayed in the form of projection, avoidance, and
aggression and has much to do with the individual personality, the organizational
structure in which the user works, and the group relations in the area where the
system will be installed. User education, training, and participation can help
reduce or control resistance to change.

Terms t o K n o w 11,1
Aggression: Resistance to KM systems through em ployee Projection: Resistance to expert systems through
sabotage of the system. em ployee display of hostility toward peers.
Avoidance: Resistance to KM systems through em ployee Pupil user: An unskilled em ployee trying to learn or gain
withdrawal from the job or scene. some understanding of the captured knowledge.
Champion: Individual within the organization who Rapid prototyping: Spontaneous, on-the-spot, iterative
believes the project will benefit the company and is approach to building KM systems; an iterative process
willing to take risks in supporting its development; has by which the knowledge developer shows the domain
credibility with management, experts, and users; and expert what the KM system looks like based on the
has access to key persons in the business. knowledge captured to date.
Customer user: A user interested in knowing how to use Systems analyst: A specialist who gathers information
the system for problem solving on a regular basis. from the user or the user’s staff in order to define a
Deployment: Physical transfer of the technology to the problem and determine alternative solutions and their
organization's operating unit. consequences within conventional information system
Enhancement: Upgrading the system to meet a new set of developm ent processes.
requirements. Tutor user: A user with a working knowledge of the KM
Implementation: The process of organizing the knowl- system or knowledge base and the responsibility for
edge and integrating it with the processing strategy for system maintenance.
final deployment. Validation: A system test to ensure the right system from
KM system life cycle (KMSLC): The steps through which a technical view; a system that meets the expert’s
a knowledge management system project goes before it expectations.
becomes operational. Verification: A system test to ensure the proper function-
Maintenance: Making necessary corrections so that the ing of the system; addresses the intrinsic properties o f
KM system continues to meet the initial system the KM system.
requirements.

Test Y o u r U n d e r s ta n d in g a 1 a a
1. Why is it helpful to view the building of a KM system as a life cycle?
2. In what ways do conventional and KM systems’ development life cycles dif-
fer? How are they similar?
3. Distinguish between:
a. verification and validation
b. knowledge developer and systems analyst
c. pupil user and tutor user
d. projection and avoidance
4. Successful KM system implementation depends on several factors. Briefly
explain each factor.
5. How important are organizational factors in system implementation?
112 * « i • PART I Tlw Basics

6. Why is the place of training important? Why is duration of training important?

7. What do you think determines the success of user training?
8. How do users differ from experts?
9. Of the steps making up the KMSLC, which one do you consider the most
critical? Why? Which would be the most time-consuming?”
10. What is rapid prototyping? How is it useful in building KM systems?
11. Elaborate on the main steps of a feasibility study. When should a feasibility
study be conducted? Why?
12. What does it mean to “scope” a project? Give an example.
13. What is your understanding of a human expert? How would you select one?
14. Write an essay in which you describe the role of the knowledge developer and
his or her relationship with the key persons in the knowledge capture process.
15. Briefly highlight the functions and attributes of the chief knowledge offi-
cer (CKO).

K n o w l e d g e Exercises 1 1 1 1
1. Crozet Country Club's use of a knowledge management system can be justi-
fied on several fronts. The system captures the human resources manager’s
knowledge of how she makes her hiring decisions. If she leaves the club, the
system will help to ensure that her expertise and know-how are not lost.
The “ Hiring KM System” saves the club time and money by making
accurate decisions more rapidly than the normal way of deciding on an appli-
cant. Because the club receives a large number of applications to fill a rela-
tively small number of positions, efficient applicant evaluation is of critical
importance. However, handling and sorting the plethora of applications was
formerly time-consuming and haphazard. This system reduces the time taken
to fill vacant positions. More importantly, it helps to ensure that employees
have the necessary qualifications.
The hiring KM system promotes a higher level of consistency and quality
in hiring decisions. Prior to implementing the system, the manager did not
have a “wage matrix" to use in determining appropriate wages for employ-
ees, and the process was quite random. This led to some conflict within the
organization. Furtherm ore, the m anager hired some applicants without
focusing enough on their ability, and they turned out to be poor employees.
In essence, the KM system helps guarantee that all of the necessary variables
are fully considered in each decision.
a. Is this sufficient justification for the KM system? Why or why not?
b. Is the system likely to replace the human resources manager, who was
instrumental in sharing her knowledge, which is captured in the knowl-
edge base?
2. The goal of this project is to develop a consumer-lending knowledge base to
guide the junior bank officer through the decision of whether an auto loan
should be approved. In the loan departm ent, there are experienced senior
loan officers who are willing to share their knowledge with others.
You have been assigned the job of building a KM system for the bank.
Explain in detail the life cycle of this project.
3. A KM system for a large retailer was designed to help human resources
develop a qualified candidate pool from the many applications they receive
on a daily basis. The knowledge on hiring available in the KM knowledge
base suggests the wage each accepted applicant should receive, based on his
or her skills and relevant experience.
The three positions addressed are (a) floor salesperson, (b) customer
service representative, and (c) gift wrap employee. The system encodes the
 CHAPTER 3 Knowledge Management Systems Life Cycle 113

knowledge of the expert in each area and the knowledge of the person who
makes the hiring decisions into rules that are used to perform the task. The
store’s human resources manager played a key role throughout knowledge
capture; her experience is stored in a specialized knowledge base. During her
absence, a junior human resources person can query the knowledge base for
information based on the expert’s (human resources manager) opinion in
various combinations of circumstances and constraints. This made certain
decisions easy to make, which made the hiring process faster and improved
the overall efficiency of the human resources department.
a. Is anything missing from the development life cycle? What made you
think so?
b. Do you think more than one person developed the KM system? Explain.
4. A high-tech firm designed a KM system for the student housing office of
a major university in March 2001 and made the system available through
the university’s intranet; three campuses were involved. A fter the sys-
tem passed verification and validation, the knowledge developer held
a wrap-up m eeting with representative end users from each branch. She
spent 2 hours going through the operations-manual and running examples
through the newly installed system to dem onstrate ease of use and ease
of access.
The attendees were quite impressed with the many features—a color
screen, easy-to-follow menu, and display of the reasons to justify all kinds of
answers. The system also links legacy databases such as tuition adjustment
and meal plans, depending on the housing arrangements.
One month after the 2-hour training session, the knowledge engineer
sent a questionnaire to all attendees, who, by then, were end users of the KM
system. Some of the questions were as follows:
• Did the operations manual help you understand the system?
• Did the knowledge developer give you a working understanding of the
system?
• Did you understand the questions asked by the system?
• Does the system meet your expectations? If not, please offer any sugges-
tions you might have.
• Did the system provide adequate and correct answers?
• Do you think that the system is usable in your environment with immedi-
ate change? Be specific.
a. Based on the information provided, evaluate the training approach fol-
lowed by the knowledge developer.
b. Could you deduce a training plan in this case? Explain.
c. Critique the questionnaire used to follow up on the installation.

R eferences 1 1 1 1
Angus, Jeff. “Harnessing Corporate Knowledge,” Delio, Michelle. “Stepping Up, Moving On, Getting Smarter,”
Inform ationweek, August 7,2000, pp. 1-3, www. Knowledge Management, June 2001, pp. 51-54.
informationweek.com/798/prknowledge.htm, Date Dixon, Nancy M. Com m on Knowledge. Boston, MA:
accessed August 2002. Harvard Business School Press, 2000.
Date, Shruti. “Agencies Create CKO Posts to Get in the Flash, Cynthia. “Who Is the CKO?” Knowledge
Know,” Governm ent Computer N ews, November 8, M anagement, May 2001, pp. 37-41.
1999, pp. 1-2. Foster, Allan. “Grasping the Tacit Knowledge Nettle,”
Davenport, Thomas H., and Prusak, Laurence. Working Knowledge M anagement, no. 1, April 1998, p. 32.
Knowledge. Boston, MA: Harvard Business School Frank, Diane. “Feds Ponder CKO R o)&?~Federal
Press, 2000, pp. 1-24. Com puter Week, June 28,1999, pp. 16-18.
114 i a a i PART I The Basics

Freeman, Miller. “Applying Knowledge to Profitability,” Skyrme, David. “D o You Need a CKO?” www.skyrme.com/
Intelligent Enterprise, March 1,2000, p. 15. insights/27cko.htm, Date accessed August 2002.
Garvin, James. A Guide to Project Management B ody o f Strassmann, Paul A. “Calculating Knowledge Capital,”
Knowledge. Newton Square, PA: Project Management files.strassmann.com/pubs/km/1999-10.php, October
Institute 2000. 1999, Date accessed August 2002.
Glasser, Perry. “The Knowledge Factor,” CIO M agazine, Tiwana, Amrit. The Knowledge Management Toolkit.
D ecem ber 1 5 ,1998-January 1,1999, pp. 1-9. Upper Saddle River, NJ: Prentice Hall, 2000.
Housel, Thomas, and Bell, Arthur. Measuring and M an- Trepper, Charles. “Keep Your Knowledge In-House,”
aging Knowledge. New York: McGraw-Hill, Inc., 2001. Informationweek, September 4,2000, pp. 1-3.
Liebowitz, Jay, and Wilcox, Lyle C. Knowledge Manage- www.informationweek.com/802/prknowledge.htm.
ment. Ft. Lauderdale, FL: CRC Press, 1997. D ate accessed November 22,2002.
Paquette, Pat. “CKO:Trendiest Job in Business,”
www.hightechcareers.com/doc498e/trendiest498e.html,
Date accessed August 2002.
PART II: K N O W L E D G E CREATION AND CAPTURE

a a a i i i E R a a a a a a a

Knowledge
Creation and
Knowledge
Architecture
Contents

In a Nutshell
Knowledge Creation
N onaka’s Model of Knowledge Creation and Transformation
Knowledge Architecture
The People Core
Identifying Knowledge Centers
The Technical Core
Build In-House, Buy, or Outsource?
Implications for Knowledge Management
Summary
Terms to Know
Test Your Understanding
Knowledge Exercises
References

Where success is concerned, people are not

measured in inches, or pounds, or college degrees,
or fam ily background; they are measured by the
size o f their thinking.
—J. M. CUPELLO

115
116 7777 PART II Knowledge Creation and Capture

■ In a Nutshell
In the previous chapters, we discussed the concept of knowledge management, what
knowledge is, and the development process that deploys a KM system for a business
organization. In this chapter, the key questions are: Where does knowledge reside?
How would one create knowledge? What does it take to design a KM system? The lit-
erature differs on how to address these questions, because much depends on the type
of knowledge, the culture of the organization, and how one strategically makes use of
knowledge m anagem ent. For example, Churchm an interpreted the viewpoint of
philosophers Leibniz, Locke, and Kant by emphasizing that knowledge resides in the
user and not in the collection of information. “To conceive of knowledge as a collection
of information seems to rob the concept of all of its life. Knowledge resides in the user
and not in the collection. It is how the user reacts to a collection of information that
m atters” (Churchman 1971). He emphasizes the human nature of knowledge creation
in an environment characterized by discontinuous change. Malhotra also discusses the
nature of knowledge creation (M alhotra 2001).
The focus on tacit knowledge being the primary type of knowledge to tap is re-
inforced by Boland, who pointed out that by considering the meaning of knowledge as
“unproblematic, predefined, and prepackaged, they ignore the human dimension of
organizational knowledge creation” (Boland 1987). In 1995, Gill reinforced the view a
different way. He stated that “prepackaged interpretation of knowledge works against
the generation of multiple and contradictory viewpoints that are necessary for meeting
the challenge posed by wicked environments: This may even hamper the firm’s learn-
ing and adaptive capabilities.”
In this chapter, we also discuss knowledge architecture as a framework of the
p e o p le core. We explore the types of people making up the organization, the knowl-
edge they own, and where they use it. Under the k n o w le d g e sh a rin g c o re , the focus is
on deciding where knowledge resides and where it is used, and identify those responsi-
ble for knowledge sharing. Finally, the te c h n o lo g y co re illustrates how to leverage the
Internet and the technical layers for building a technology base for knowledge sharing.
With all these issues, the CKO plays a critical role in the KM process.

■■■■ Knowledge Creation

Most of us live in a world where today's knowledge will not solve tomorrow’s prob-
lems. This means updating our knowledge on a regular basis. You probably have heard
the old joke about the teacher who used the same experience with a course 20 times
rather than having 20 years of teaching experience. The two are not the same. Knowl-
edge update is creating new knowledge based on ongoing experiences in a particular
problem area and then using the new knowledge in combination with the initial knowl-
edge to come up with updated knowledge for knowledge sharing. Unfortunately, most
organizations have a very competitive culture, where no one wants to share. With a
change in culture, values, and improved mutual trust, sharing creates a learning culture
that feeds itself in terms of improving knowledge sharing —giving away, holding in
common, and improving bonds within the firm. In contrast, our view is to look at
explicit and tacit knowledge as the two primary knowledge types and learn how to cap-
ture them for organizational inquiries and growth in an unpredictable environment.
Either type can be created and should be captured. Capturing explicit knowledge is
covered later in the text under data warehousing and data mining. In the next chapter,
we focus on capturing tacit knowledge, and C hapter 6 elaborates on the main tools
 CHAPTER 4 Knowledge Creation and Knowledge Architecture i i i ■ 117

used in knowledge capture. We approach knowledge and “knowing” as deeply per-

sonal. Keep in mind that KM is not a technology. It is an activity enabled by technology
and produced by people. It is also how people share the type of knowledge that will
add special value to the growth and competitive nature of the business. The use of tech-
nology is a critical aid for processing and deploying knowledge whenever it is needed.
An alternative way of creating knowledge is through teamwork. Figure 4.1 shows
how a team translates experience into knowledge. Essentially, a team commits to per-
forming a job over a specific time period, such as weeks or months. A job is more than
a task. It is a series of specific tasks carried out in a specific order, format, or sequence.
In the end, a job is completed, and the results are either successful or disappointing.
The team looks back and compares the experience it had when starting the job to the
outcome. This comparison is what translates experience into knowledge. When it per-
forms the same job in the future, the team takes corrective steps, modifying actions
based on the new knowledge they have acquired.
With either knowledge creation alternative, the outcome is the same. Individuals
or teams begin a commitment to a job with specific knowledge. A fter the job is com-
pleted, a comparison of the “before and after” occurs. In most cases, the outcome is
extended knowledge that should improve the quality and effectiveness of performing
the same job the next time around. In some quarters, this is called maturation or sea -
soned experience. Over time, experience leads to expertise, where one person or one
team becomes known for handling a complex problem very well. This knowledge can
be transferred to others in a reusable format. The receiving individual or team adapts
the knowledge for use in a special context.

FIGURE 4.1: Knowledge Creation and Knowledge Transfer via Teams

118 7777 PART II Knowledge Creation and Capture

Team formation and teamwork begin with experienced individuals working jointly
on a project. In consulting assignments, for example, an experienced individual man-
ages a team of experienced and semiexperienced members (see Box 4.1). The manager
assigns tasks and decides how much knowledge sharing will occur. In a way, tacit
knowledge begins with the individual, not the team. Real knowledge continues to be
held by people (Berry 2000). In a team environment, shared exchange results in collabo-
rative decision making and problem-solving. So, when it comes to knowledge capture,

BOX 4.1

ROLE AND FORMATION OF KM TEAMS

DRAFT YOUR DREAM TEAM based Andersen Consulting suggest staffing teams
with individuals from the business process in
By definition, it takes at least two people to share
question and from the human resources and in-
something. When that som ething is knowledge
formation technology departments. According to
applied to the goals of a large enterprise, the shar-
Christina Schultz, manager of global knowledge
ing multiplies, and the number of people involved
services at A ndersen Consulting, a team needs
must increase. Therefore, most organizations make
members from all three sectors. “You don’t find
establishing a project team an early priority in
all these skills in one person, so you have to
their knowledge management initiatives.
ensure that all are represented,” she says.
Choosing and recruiting the right people for
the team is a vital task that will have enduring
CASTING ROLES
consequences. Accomplishing corporate business
objectives requires a mix of skills, experience, and A team member’s primary duty within the group
stakeholder interests, so the team members usu- is as im portant as his or her qualifications. From
ally will come from multiple disciplines. Team his KM project experience, Merrill has derived six
developers also must consider issues of trust across key roles, which he calls “hats,” that one team
departm ents and business units as well as be- member or another must play. They are as follows:
tween individuals. Business leads: Specialize in understanding
Successful KM initiatives, experts say, focus the measurable result to be obtained
first on people business processes and company Organizational leads: Specialize in communi-
culture. Only later should the emphasis shift to the cation plans, m otivational incentives, and other
technology and tools that facilitate knowledge shar- organizational needs
ing. Members may need expertise in management, Super users: Can influence the user commu-
administration, human resources, and business nity and provide practical feedback
processes as well as in the supporting technologies. Interface design leads: Specialize in computer/
“A knowledge m anagem ent project must human interaction and integrating content into
leverage the strengths of each of these profes- work processes
sional perspectives to be successful,” says Ray Training leads: Specialize in cognitive needs,
Merrill, a consultant at Ariel Performance Cen- learning strategies, and repurposing legacy content
tered Systems, Inc., in Fort Lauderdale, Florida. Technical architects: Specialize in technology
When helping clients to im plem ent knowledge installation and administration
m anagem ent projects, consultants at Chicago-

SOURCE: Excerpted from Robb, Drew. “Draft Your Dream Team,” Knowledge Management, August 2001, pp. 44-50.

fl a fl a a a
 CHAPTER 4 Knowledge Creation and Knowledge Architecture m■ ■ ■ 119

Compensation
Recognition Personality
Ability Utilization
Creativity
G ood Wqrk Environment
Autonomy
Job Security
Moral Values
Vocational
Advancement
Reinforcers
Variety
Achievement
Independence 11
Social Status ________
Attitude

1[
Work Norms

fl 1 I fl fl fl FIGURE 4.2: Impediments to Knowledge Sharing

the approach is to tap the individual’s tacit knowledge, which could be a single expert
or multiple experts, as the situation warrants.
With knowledge transfer and knowledge sharing, there are impediments to con-
sider. W hether it is the individual expert or the “expert” team, knowledge sharing is not
that straightforward. As shown in Figure 4.2, there are factors that encourage or retard
knowledge transfer—from personality and attitude to vocational reinforcers. Whether
knowledge sharing causes change in the organizational culture or the other way
around continues to be a hotly debated area. In either case, a review of the motiva-
tional factors that encourage people to share is im portant if the KM process has a
chance of succeeding.
As shown in the conceptual model in Figure 4.2, personality is one factor in knowl-
edge sharing. For example, people who are extroverts, display self-confidence, and feel
secure tend to share experiences more readily than those who are introspective, self-
centered, or security conscious. A ttitude and work norms are additional considera-
tions. People with a positive attitude, who trust others, and who work in an environ-
ment conducive to knowledge sharing tend to be better at sharing than those working
in a “cutthroat” environment.
Finally, vocational reinforcers are key to knowledge sharing. Based on the authors’
vocational psychology research, people whose vocational needs are met by job reinforcers
are more likely to favor knowledge sharing than those who are deprived of one or more
reinforcers. One may conclude that the totality of attitude, personality, work norms, and
vocational reinforcers determines how likely people are to view knowledge sharing and
knowledge transfer in a positive light.

■■■■ N onaka’s Model of Knowledge Creation

and Transformation
A nother way of looking at knowledge creation and knowledge transformation is to use
N onaka’s model. Nonaka (1995) coined the terms tacit knowledge and explicit knowl-
edge as the two main types of human knowledge. Tacit knowledge is considered the most
120 7 ^ 7 PART II Knowledge Creation and Capture

valuable knowledge. The key to knowledge creation lies in the way it is mobilized and
converted through technology. C onversion of know ledge betw een tacit and explicit
know ledge is shown in Figure 4.3.
Tacit to tacit communication, re ferred to as so c ia liza tio n , takes place betw een
people in m eetings or in team discussions. Such knowledge sharing, transfer, or collabo-
ration often produces no explicit knowledge. E xperience am ong people in face-to-face
business situations is mostly shared, and technology plays a m inim al role. How ever,
th e re is an increasing tre n d for online groupw are tools be used in w ork groups or
teams. In such a synthetic atm osphere, especially in geographically dispersed settings,
team m em bers can conduct meetings, listen to presentations, or simply carry out dis-
cussions. L otus N otes is one exam ple of a softw are product that allows people to share
docum ents and discuss them via video and text-based chat.
Tacit to explicit communication, or ex tern a liza tio n , is essentially articulation
am ong people through dialogue. B rainstorm ing, w here a team carries out discussions
around a specific problem is one example. Com m ents and suggestions are entered into
a PC and displayed on a screen, but no one know s w ho m ade the com m eht.
B rainstorm ing is covered in detail in C hapter 6.
O nline discussions can capture tacit know ledge and apply it to an im m ediate prob-
lem (M arwick 2001). The database becom es a repository of useful knowledge. On the
surface, th e exchange gives th e im pression th at it is purely explicit knowledge.
How ever, the participant first has to decide on the nature of the problem before he or
she can give the best so lu tio n —both prom pting the use of tacit knowledge before such
know ledge becom es explicit. A fter the knowledge has been m ade explicit and has been
sto red in a repository, persons facing a sim ilar problem can consult the database at
their convenience.
Explicit to explicit communication (also re fe rred to as co m m u n ica tio n ) is one
transform ation phase th at is best supported by technology. Explicit know ledge can be
easily c a p tu red and tran sm itted to a w orldw ide audience. For exam ple, sending an
attached m em o or docum ent via e-m ail expedites know ledge sharing in an efficient
and effective way. In this respect, technology helps by m otivating people to capture
and share w hat they have or w hat they know.
Explicit to tacit communication (also re fe rre d to as in tern a liza tio n ) is taking
explicit know ledge such as a re p o rt and deducing new ideas or taking constructive
action. C reating technology to help users derive tacit know ledge from explicit knowl-
edge is an im portant goal of knowledge m anagem ent. By searching and finding associ-
ations, technology m akes inform ation m ore useful by m aking it easier to derive new
tacit know ledge from it. For example, superm arkets profiling custom ers when they use

Tacit to Tacit Tacit to Explicit

(socialization) (externalization )

Team Meetings Dialog Within Team a a ■ l a a

and Discussions Answer Questions
FIGURE 4.3: Conversion of
Knowledge Between Tacit and
Explicit to Tacit Explicit to Explicit Explicit Forms
(internalization) (com m unication )
S O U R C E : Marwick, A. D.
Learn from a Report E-Mail a Report “Knowledge Management
Technology,” IBM Systems
Journal , vol. 40, no. 4,2001, p. 815.
 CH A PTER 4 Knowledge Creation and Knowledge Architecture aaaa 121
a sto re-issu ed card to get discounts is an exam ple of explicit to tacit know ledge.
Through data mining, the store can derive a num ber of new ideas regarding the kind of
product that is popular or the preference of custom ers by age, location, or price. Such
tacit know ledge is critical for com petitive advantage.
In conclusion, one can see that N o n ak a’s m odel divides knowledge creation pro -
cesses into four categories. It focuses on tacit know ledge and the use of technology in
generating or transm itting such knowledge to others. U sing electronic m eetings or chat
room s for collaboration, know ledge sharing, and transfer are well know n in to d ay ’s
know ledge organizations. D esp ite the advan cem en t of technology, one lim itation
rem ains: the ability of technology to support the use of tacit knowledge in face-to-face
m eetings and direct interactions (M arwick 2001). H um an knowledge continues to be a
valuable resource, and technology is expected to trickle slowly into a hum an dom ain
w here know ledge creation and know ledge transfer can be expedited for hum an deci-
sion making regardless of tim e or location.

Knowledge Architecture
Know ledge architecture is a prerequisite to knowledge sharing. We view the infrastruc-
ture as a com bination of people, content, and technology. As shown in Figure 4.4, these
com ponents are interd ep en d en t and inseparable. People with knowledge provide con-
tent, relying on technology to transfer and share knowledge. This com bination provides
the efficiency and perform ance to m anaging the know ledge core of the corporation.

THE PEOPLE CORE

K now ledge m anagem ent is abo u t people and the way they creatively perform in an
environm ent conducive to know ledge sharing. By p eo p le , we m ean knowledge w ork-
ers, managers, custom ers, and suppliers. The first step in knowledge architecture is to
evalu ate th e c u rre n t in fo rm atio n and docum ents people use, the applications they
need, the people they contact for solutions, the associates they collaborate with, the e-
mail they send and receive, and the database they access. All these sources m ake up an
em ployee profile, which is later used as the basis for designing a knowledge m anage-
m ent system th at best serve p eo p le’s long-term needs via technology.
The goal of profiling the people part of an organization is to get a handle on existing
knowledge exchanges that the organization uses and ways to capture them. As a proce-
dure, it involves not only asking the right questions but also providing answers that become
part of the KM system infrastructure. Discovering whom to profile, in what sequence, and
for what duration and using the right knowledge capture tool are discussed in Chapter 5.

■ ■ ■ ■ ■ ■

FIGURE 4.4: Conceptual Mew

of Knowledge Management
122 7777 PART II Knowledge Creation and Capture

Profiling can be approached in several ways. The worst way is to send a survey or
questionnaire by e-mail. W hen it comes to knowledge capture, the most im portant p re-
requisite is selecting the em ployees and planning the questions. Planning whom to
elicit know ledge from m eans deciding w hether to talk with senior em ployees only,
younger or new er employees, or some mix based on criteria that m ake sense. The next
prerequisite is type of question to ask. Each question should be phrased and intended
to elicit inform ation accurately and completely.
Finally, the w hole idea b ehind assessing the people core is to do a good job in
assigning job content to the right person and to ensure that the flow of inform ation
th at once was o b structed by departm ents now flows to the right people at the right
time. For exam ple, custom er service need no longer have an exclusive on troubleshoot-
ing inform ation, any m ore than the legal departm ent should have an exclusive on set-
tling custom er litigation. To expedite knowledge sharing, a knowledge netw ork should
be m apped in such a way as to assign people authority and responsibilities for specific
kinds of know ledge content. It means:
1. Identifying know ledge centers (for example, sales, products, or m arketing).
2. A ctivating know ledge co n ten t satellites —low er level know ledge centers that
hierarchically fall under higher level centers (for example, custom er support or
product docum entation in repositories).
3. Properly staffing experts over each knowledge center. The outcom e is a corpora-
tion-wide know ledge structure that blends the best of existing organization struc-
ture and the best of technology—e-mail, intranets, extranets, and netw orking—to
ensure a sm ooth, 24-hour knowledge sharing environm ent.

IDENTIFYING KNOWLEDGE CENTERS

Once you have determ ined the know ledge that people need, the next step is to iden-
tify w here know ledge resides and how to capture it. By kn o w led g e center ; we m ean
areas in the organization w here know ledge is available for know ledge capture. For
exam ple, in Figure 4.5, a retail organization might have at least four knowledge cen-
ters: sales, custom er service, hum an resources, and m arketing. These centers become
the skeletal fram ew ork for knowledge capture. They also serve to identify individual
experts or expert team s in each center who could be candidates for the knowledge
capture process.

Activating Knowledge Content Satellites

As m entioned earlier, this step simply breaks down each know ledge center into
lower, m ore m anageable levels, satellites, or areas. The focus is on searching for knowl-
edge content, not evaluating departm ents or divisions in the way we evaluate the tradi-
tional organizational structure.

Assigning Experts to Each Knowledge Area

O nce the new fram ew ork is m ade final, each know ledge satellite should be as-
signed a m anager to ensure integrity of inform ation content, access, and update (see Fig-
ure 4.5). In the assignm ent process, it is im portant to identify m anagem ent-type people
for each satellite know ledge center; these people will provide contact sources for
know ledge capture. This enforces discipline and quality of inform ation, because m an-
agers have vested ow nership in their respective departm ents.
O w nership is critical to know ledge capture, know ledge transfer, and knowledge
im plem entation. In a typical organization, departm ents tend to be territorial. So-called
 CHA PTER 4 Knowledge Creation and Knowledge Architecture i i ■ i 123

tu rf b attles occur regularly over b u dget size or control of sensitive processes. This
includes the kind of know ledge a d ep artm e n t owns. T hese reasons justify assigning
departm ent ow nership to know ledge content and knowledge process. In Figure 4.6, a
w ell-defined know ledge content ensures discipline and order to a com pany’s know l-
edge base. You can w ork your way from the m ajor departm ents down to lower levels in
each departm ent. For exam ple, “O p eratio n s” shows the m ain know ledge areas (teller
schedules, depo sits and w ithdraw als, and electro n ic balancing). Likew ise, “L o an s”
owns a variety of loans and loan processing (com m ercial, unsecured, personal, and
auto). For troubleshooting, “C ustom er Service” has an exclusive on issuing traveler’s
checks, settling custom er com plaints, ordering checkbooks, and opening new accounts.
From the departm ent level, we can m ap out each knowledge content in greater detail.
Such a step is carried out in knowledge capture.
A d jacen t or in te rd e p e n d e n t d ep artm e n ts are ex pected to be co o p erativ e and
read y can d id ates for know ledge sharing. For exam ple, the bank tellers th at accept

FIGURE 4.6: Knowledge Content by Department in a Bank

Knowledge What Why Troubleshooting

Department Operations Loans Customer Service

Owner

Knowledge Teller Schedule Commercial Loans Traveler’s Checks

Content Electronic Balancing Unsecured Loans Handling Complaints
Deposits/Withdrawals Personal Loans Order Checkbooks
Auto Loans Open New Accounts

Details Details Details

124 7777 PART II Knowledge Creation and Capture

m ortgage paym ents are accom m odating the loan departm ent that issued the loan in
the first place. We need to first identify the com m onality of knowledge and then sell
the “goodness of fit” betw een departm ents for assuring sm ooth knowledge sharing and
know ledge transfer.

THE TECHNICAL CORE

The goal of the technical core is to enhance com m unication and ensure effective
know ledge sharing. C om m unications is a special area, w here technology provides
unlim ited o p p o rtu n ities for m anaging tacit know ledge. C om m unications netw orks
create links w ithin and across databases. Like the telep h o n e, they open up a rich
m edium for inform al com m unication am ong em ployees and m anagers alike. E-mail,
for exam ple, has already dem onstrated the rich exchange within and across organiza-
tions worldwide.
The phrase technical core is used here to refer to the totality of hardw are and soft-
ware and to the specialized hum an resources required to enable knowledge. Speed,
accuracy, reliability, integrity, and security are all expected attributes of successful tech-
nology under the technical core. R em em ber that technology enables knowledge deliv-
ery, and it is people who enable knowledge. Because the organization is a knowledge
netw ork, the goal of know ledge econom y is to push em ployees tow ard greater effi-
ciency and productivity by making b etter use of the knowledge they have.
O nce in place, a knowledge core becom es a netw ork of technologies designed to
work on top of the com pany's existing netw ork. The approach to designing the tech-
nology core is leveraging and m ore leveraging of the organization’s existing technol-
ogy. In designing a know ledge m anagem ent system, it is prudent to tie in what already
exists and blend it with w hat is recom m ended.
W hen it comes to leveraging, keep the following in mind:

• Focus on real know ledge of people rather than artificial intelligence that was
once unique to expert systems but which has now becom e a flop.
• C oncentrate on a know ledge m anagem ent system that facilitates finding the
sources of know-how, not just the know-how. A user should be able to readily
locate people and their respective expertise and use available knowledge as he or
she sees fit.
• A successful know ledge m anagem ent system should be tagged to collaboration. It
w ould be superfluous to think of know ledge sharing, knowledge transfer, or con-
tinued creativity w ithout collaborative success.
• People are no longer expected to adapt to the requirem ents of the machine, but
the oth er way round. Today’s hum an intelligence should m ake use of the way the
m achine (com puters, networks, and the like) supports intelligent thinking, innova-
tions, and creativity.
• C oncentrate on realism, not perfection. A knowledge m anagem ent system should
begin with w hat you have and try to im prove it. Disposing of current infrastruc-
ture and replacing it with a new one invites a host of problem s that could backfire
on the organization fo ^a long time.
• Allow the user to decide on w hether the proposed knowledge m anagem ent sys-
tem m akes sense and how well inform ation is organized and ready for day-to-
day access.

The best way to visualize the technical architecture as the building block of KM is
in term s of a layer system, starting with the user interface (least technical) and going all
 CH A PTER 4 Knowledge Creation and Knowledge Architecture aaaa 125
the way to the repositories th at form the m ost technical layer of the technical core.
They represent technologies of the com pany’s intranet and its interface with the W orld
W ide W eb, including the protocols, standards, brow sers, and servers. K now ledge
exchange and com m unications within the com pany and across organizations are m ade
possible via satellite com m unication, telecom m unication netw orks, and wireless n e t-
works. Satellite com panies are setting up new bro ad b an d netw orks to reach people
where telephone service is not available. Telecom m unication com panies are develop-
ing new technologies for higher-bandw idth com m unication across existing networks.
W ireless netw orks are also being converted for In ternet use. Providers for each tech-
nological area play a m ajor role in the expansion of the Internet.
The 7-tier technical layer, form ing the overall building blocks of the knowledge
m anagem ent architecture, is shown in Figure 4.7. Each layer is briefly covered in the
following sections.

The User Interface Layer

The first layer is the interface betw een the user and the KM system, usually rep re-
sented by a W eb browser. It is the top layer in the KM system architecture and the only
layer with which the user interacts, To the user, beyond this layer is simply a black box.
The goal is to move inform ation into and out of the KM system. It simplifies technol-
ogy for the user by the way graphics, tables, and text are displayed on the screen. It rep-
resen ts know ledge w hen displayed in a timely, actionable, and nontechnical user
friendly universal view. U niversal m eans uniform , consistent, and custom ized naviga-
tion. It m eans no barriers to inform ation and tacit know ledge transm itted from reposi-
tories or rem ote sources.
P ractically speaking, technology best su p ports explicit know ledge, although it
should also represent tacit knowledge, as we saw in the 1990s. The user interface layer
m ust provide a way for tacit and explicit knowledge flow. Knowledge transfer betw een
people and technology requires capturing tacit know ledge from experts, storing it in a
know ledge base as an in term ed iate step, and m aking it available to people as tacit
knowledge for solving com plex problem s (see Figure 4.8).
For years, it has been know n in the artificial intelligence (A l) com m unity that
hum an know ledge can n o t be fully codified. T hat is all the m ore reason why tacit
know ledge should be m ade available by face-to-face, e-mail, or oth er m edia. D irect
com m unications allow for context, tone of voice, and nonverbal interpretation. This
goes beyond the storage and electronic com m unication that databases and netw orks
are best known for providing. C apturing tacit know ledge is covered in C hapter 5.
In term s of user interface design, several features should be considered:

• Consistency. Menus, icons, and buttons should have the same meaning or representa-
tion throughout the KM system, including those used in future upgrades. All infor-
mation should be consistent in layout and form at across all areas of the system.
• Relevancy. A ny inform ation displayed should be relevant to the user’s expecta-
tions. It m eans users should be able to custom ize inform ation in the way they
want. It also m eans personalization of the Web site.
• Visual clarity. Ideally, all inform ation should be presented in one screen, with
space-saving features such as pull-down m enus available as default. Proper use of
color, margins, and text density can add quality and clarity of the displayed infor-
m ation. In to d ay ’s digital age, people do not like to read text-heavy screens. In
designing Web pages, for example, there is an 8-second rule that m easures user
tolerance to reading a screen full of inform ation or to waiting for one.
126 7 7 7 7 PART II Knowledge Creation and Capture

User User User

Layer

1 User Interface
(Web browser software installed on each user’s PC)

Authorized Access Control

(security, passwords, firewalls, authentication)

Collaborative Intelligence and Filtering

(intelligent agents, network mining, customization, personalization)

Knowledge-Enabling Applications
(customized applications, yellow pages, skills directories, videoconferencing,
decision support systems, group decision support systems tools)

Transport
(e-mail, Internet/Web site, TCP/IP protocol to manage traffic flow)

Middleware
(specialized software for network management, security)

The Physical Layer

(repositories, cables)

L - 1 fj

Legacy Applications
Databases (payroll) Groupware Data
(document Warehousing
exchange, (data cleansing,
collaboration) data mining)

I FIGURE 4.7: Technical Layer of the KM System

SO U R C E : Adapted from Tiwana, Amrit. The Know ledge Management Toolkit. Upper Saddle River, NJ:
Prentice Hall, 2000, p. 309.

Navigation. The term navigation means how easily a user moves through Web pages,
files, or records in a computer-based system. W hen dealing with different packages
or tools within the KM system, the user’s view should be a single mode of operation
or one where the system tells the user which packages he or she has in operation.
Usability. This feature addresses navigation in term s of how easy it is to use icons
and buttons and the speed of access to files or databases.

W ith this in m ind, it is critical th at talen ted user interface designers be hired to
build this layer, w hether the layer is brow ser-based or client/server-based. The com-
 CHA PTER 4 Knowledge Creation and Knowledge Architecture ■■■■ 127
Inform al Tacit
K now ledge Transfer

Tacit Knowledge Intermediate Tacit Knowledge

fl fl I I I I FIGURE 4.8: Knowledge Transfer

pany’s IT departm ent can be instrum ental in the final version of the design, including
choice of colors, banners, icons, graphics, and so forth. U sability testing by com pany
users is the final test of the acceptability of the user interface.

Authorized Access Layer

The authorized access layer m aintains security and ensures authorized access to
the know ledge captured and stored in repositories through the com pany’s intranet, the
Internet, and extranet. In essence, the com pany’s traditional client and com m unication
structure is replaced by repositories and know ledge bases that can be accessed from
anyw here and at any time.
Figure 4.9 shows the technical access layer, with focus on the use of the Internet,
intranet, and extranet as p art of the technical infrastructure of the KM system. A com -
p any’s intranet is the internal netw ork of com m unication systems m odified around the
In te rn e t. It is a way of thinking ab o u t how people in a business w ork and transfer
know ledge. Technically, it is a netw ork connecting a set of clients by using standard
In tern et protocols. In tern et technologies are superior to the conventional m eans for
developing internal com m unication systems. The W eb browser, for example, is readily
available and a fam iliar access tool for im m ediate use. D ocum ents are easily handled

fl fl fl fl fl fl FIGURE 4.9: Technical Access Layer

Internet Intranet Extranet

/ Company Employees \
/ Suppliers \
I • Vendors J
\ • Partners /
Public • Customers
at Large

• News/Events • Human Resource • Product Information

• Marketing Information • Sales Information
• E-Commerce • Production • Collaboration/
• Careers Information Cooperation
• Sales Information
• Strategic Plans
128 7777 PART II Knowledge Creation and Capture

and m ultiple m edia can be supported as well. M anagers can share docum ents on inter-
departm ental and in trad ep artm ental levels. W ith various operating systems, netw ork
protocols, and application suites, trying to ensure hom ogeneity in m anaging docum ents
can be quite a challenge. In tranets can handle all these problem s with ease.
In contrast, an extranet is an intranet with extensions that allow clearly identified
custom ers or top suppliers to reach and access com pany-related technical educational
inform ation. For exam ple, in one electrical corporation the ex tran et is accessible to
950 electricians and small electrical outlet shops via the com pany’s special Web page.
A n electrician enters a specially assigned password to access inform ation about new
products and special deals that are available to high-volum e buyers. Based on 7 m onths
of practice, the com pany’s N et-based practice paid dividends. The com pany cut down
on phone calls and fax orders, and gave sales reps in the field online support when
needed. It even im proved shipm ent schedules and deliveries, regardless of distance.
It should be noted that in today’s In ternet economy, a com pany’s success depends
less on the resources within its walls and m ore on its relationships with suppliers, ven-
dors, and custom ers. Effective m anagem ent of the com pany’s product traffic is m ar-
shalling the resources of its em ployees in line with those of o th er companies. In this
respect, relationships becom e assets, just as the com pany’s buildings, inform ation, and
know ledge (Tapscott 2001).
In tranets and extranets have a lot in common. A s shown in Figure 4.10, intranets
are m ore localized within a firm and move data m ore quickly than the m ore distributed
extranets. The use of In tern et (prim arily W eb) protocols is m ore common to connect
business users and branches. O n the intranet, Web adm inistrators prescribe access and
policy for a defined group of users. To collaborate via extranet, the applications have to
perform consistently on all represented platforms. The same is true for trading partners
and suppliers, who may be totally unknow n to one another.

FIGURE 4.10: General Extranet Layout

Firewall

Corporate
Intranet

Suppliers
Corporate
Firewall
Intranet
 CH A PTER 4 Knowledge Creation and Knowledge Architecture aaaa 129
The access layer has taken on a special role in the technical core due to the vulner-
ability of intranets and heavy traffic on the Internet. The focus of today’s access layer is
on security, use of protocols like passwords, authentication to ensure users are au th o -
rized, and softw are tools like firewalls to block certain inform ation from leaving the
com pany's repositories or incom ing inform ation from contam inating com pany files.
Firewalls are hardw are and software tools that define, control, and limit access to n et-
works and com puters. They detect intruders, block them from entry, keep track of what
they do and where they originate, notify the system adm inistrator of mischievous acts,
and produce a report. Specifically, a firewall protects against the follow ing
• E-m ail services that are known to be problem s
• U nauthorized interactive log-ins from the outside world
• U ndesirable m aterial such as pornographic images, movies, or literature
• U nauthorized sensitive inform ation leaving the com pany
In contrast, a firewall cannot prevent:
• A ttacks that do not go through the firewall (such as taking data out of the com -
pany by tape or diskette)
• W eak security policies
• Viruses on floppy disks
• Traitors or disgruntled em ployees
There are two main issues related to the access layer: access privileges and back-
ups. In organizing access privileges, a netw ork adm inistrator has to authenticate each
com pany user and define the types of files he or she may access. In term s of backups, a
know ledge base requires a duplicate in case of fire, security violations, virus attack, or
hardw are breakdow n. A com pany whose lifeline depends on such repositories must
strategize a backup schem e and im plem ent it accordingly.
B iom etrics is an evolving technology that is slowly becom ing p art of the access
layer. Biom etrics is the autom atic identification of a person based on physiological or
behavioral characteristics. This science along with voice and fingerprint recognition
should add reliable muscle to the access layer.

Collaborative Intelligence and Filtering Layer

Layer three is the collaborative intelligence and filtering layer within a KM system.
It provides custom ized or personalized views based on stored knowledge. It is designed
to reduce search tim e for info rm ation by com bining the know ledge sought with the
user's profile. A uthorized users registered in the KM system ’s directory can expect infor-
m ation tailored to their view in form at, content, and detail.This is all available through a
search m echanism such as the brow ser or the client/server netw ork of the organization.
Intelligent agents are prob ably the best co n trib u tio n th at A l has m ade to the
Web. A n intelligent agent is an active object that can perceive, reason (learn from past
m istakes), and act in a situation to assist in problem -solving. M ost agents are m ore
m obile than static. They search across servers to find the inform ation requested by the
client (user).
Intelligent agents are clusters of precanned bits of intelligence that relate specific
needs or attributes to oth er agents on a netw ork. They are not real agents; they do not
knowingly cooperate or share in problem solutions. They m ust be program m ed to do
those things, and they “learn" only a few things. They are planned for m any kinds of
services, from spread sh eets to personal inform ation m anagers. O ne can realistically
expect intelligent agents of the future to represent users in searching databases. They
130 7777 PA R T II Knowledge Creadon and Capture

could conceivably arrange meetings, pay bills, and even w ander through virtual shop-
ping malls, suggesting gifts, greeting cards, and so on.
Figure 4.11 illustrates the key differences betw een traditional client/server infra-
structure and th at of mobil agent/com puting. In client/server, there is direct but m ore
frequent interaction betw een the client and the netw ork server. This m eans heavy n e t-
w ork load and possible congestion. In m obil agent/com puting, the agent picks up the
slack; the in teractio n is b etw een the agent and th e server, leaving the overall load
vastly reduced.
Think of the m obil ag en t’s role as that of an agent for a professional athlete. A ny
requests or contract negotiations are handled by the agent, which reduces the direct
volum e of correspondence for the athlete. In agent/com puting, m obility is im portant. A
m obile agent roam s around the In tern et across m ultiple servers looking for the right
inform ation or synthesis. T here are other benefits as well:
• Fault-tolerance. W hen a server or a node in the netw ork fails, an agent is w arned
in advance; the agent can take special action to com plete an activity or relocate
on anoth er server to continue operation.
• R educed overall n etw ork load. As m entioned earlier, agents reduce the num ber of
interactions betw een clients and servers. In doing so, they reduce the load and
delay on the netw ork, releasing the savings to carrying bandw idth-intensive files
such as videoconferencing or sound in the same netw ork.
• H eterogeneous operation. W ith netw orks handling different hardw are and soft-
w are within the sam e netw ork, m obil agents operate seamlessly in the same h e t-
erogeneous environm ent; this saves time and effort.

! ■ ■ « ■ « FIGURE4.il: Contrast Between Client/Server and Mobil Agent/Computing

PC Workstation

Client/Server

Activity

Network Environment

PC Workstation

□
Network Environment
 CHA PTER 4 Knowledge Creation and Knowledge Architecture ■■«* 131
The key com ponents of the collaborative intelligence layer are (1) the registration
directory that develops tailored inform ation based on user profile, (2) mem bership in spe-
cialized services such as news service, sales prom otion, schedules, and custom er support,
and (3) the search facility such as a search engine to assist in finding information. Users
can search for words, subjects, or single queries for information, regardless of where it is
stored. The security feature is that users see only the results for which they have access.
In term s of the prerequisites for an effective collaborative layer, several criteria
are considered:
• Security. This is a critical feature, especially as an organization becom es increas-
ingly distributed worldwide.
• Portability. A collaborative platform must operate in a portable environm ent
across platform s. For exam ple, H T T P and the W eb allow different systems to com -
m unicate with each other, regardless of architecture, location, or language. The
Web brow ser is the m ost universal tool that users em ploy to access repositories.
• Integration with existing systems, including legacy applications. The Web is a good
example, w here a collaborative know ledge-sharing platform integrates with exist-
ing systems and m ainfram e databases.
• Scalability, flexibility, and ease o f use. These three criteria go together in em pha-
sizing the im portance of a know ledge-enabling, upgradable, flexible, and easy-to-
use (and to learn) collaborative platform . It boils down to the platform ’s ability to
filter out irrelevant content and display what the user needs to see.
For years, com panies have been trying to use com puters to capture and m anipulate
know ledge for pro d u ctiv ity and perform ance. The collaborative and filtering layer
relies on a com bination of technologies such as expert systems, case-based reasoning
(C B R ), neural netw orks, and intelligent agents.
Expert systems are a branch of artificial intelligence or AL Its main purpose is to
assist a p e rso n ’s thinking process, not m erely to provide inform ation for a person to
consider. W hen a com pany faces a critical decision —such as w h eth er to acquire
anoth er company, m arket a new product, or grant a hefty loan to a new custom er—it
turns to an expert for advice. These experts have years of experience in a particular
area and generally know the probability of success for a particular decision. A difficult
problem requires a seasoned professional, who is invariably an expensive commodity.
Expert systems are decision-making and problem-solving tools. They consist of com-
puter software program s that em ulate or clone the reasoning of a hum an expert in a
problem domain. As shown in Figure 4.12, the knowledge base is the repository of the
rules, facts, and knowledge acquired from the human expert. The knowledge is typically
represented in the form of IF . . . T H E N rules, facts, and assumptions about the problem
the system is designed to solve. The developm ental environm ent simply allows a human
knowledge engineer to acquire the knowledge from the hum an expert. Such knowledge
is entered into a knowledge base that represents the operational environment.
The user environm ent consists of a user friendly interface; a justifier that explains
to the user how and why an answ er is given; an inference engine that is the problem -
solving m echanism for reasoning and inferencing; and a scheduler that has the unique
jo b of c o o rd in atin g and co n trolling rule processing. A ll these co m ponents w ork
together to sim ulate a near-real environm ent of a client-expert relationship.
P roperly used, ex p ert system s can help a person becom e wiser, not just b e tter
informed. A l encom passes many aspects of hum an behavior (such as speech, language,
and m ovem ent); expert systems focus on the task of problem solving. E xpert systems
utilize the research discipline of A l to create a commercial reality that produces benefits
132 7777 PART II Knowledge Creation and Capture

« FIGURE 4.12: Software Components of an Expert System

each day. A n expert system is a branch or a commercial spin-off of AI. It contains practi-
cal knowledge obtained from a hum an expert. Its inform ation is explicit and com prehen-
sible, and the system is able to explain its reasoning on dem and. In this respect, it is
viewed as a problem-solving tool.
E xpert systems have not lived up to expectations and have not revolutionized the
business environm ent. The m ain reason is th at hum an know ledge is too com plex to
understand, capture, or m anipulate. Their applications are also rath er restricted. They
w ork best with two sim ultaneous conditions: express a problem variable in hard term s
(such as num bers or speed) and rules cannot overlap.
C ase-B ased Reasoning. In this particular area of reasoning, old solutions are adapted
to m eet new situations by using old cases to explain those new situations. For example,
attorneys use case precedents to justify or explain their argum ents in new court cases.
D octors often treat patients based on sym ptom s they have treated in previous patients.
E x p erien ce w ith old cases allows an a tto rn ey to decide on how to handle the case
easily and quickly (see Figure 4.13).
A case is a contextual pie^e of know ledge representing an experience. It is know l-
edge at an o p eratio n al level. C ase-based reasoning (C B R ) is useful to experts with
experience in their particular domains, because it helps them reuse the reasoning they
have done in the past. C B R also rem inds the expert to avoid repeating past mistakes.
Neural networks, or neural nets, are inform ation systems m odeled after the hum an
b ra in ’s netw ork of n e u ro n s—the basic processing elem ents of the brain. Each neuron
has a small am ount of local m em ory and is connected to o th er neurons to receive input
and produce output. C om m unications am ong neurons are carried out only through the
in put-output pathways. A n euron evaluates the inputs, determ ines their weights, sums
 CH A PTER 4 Knowledge Creation and Knowledge Architecture aaaa 133

FIGURE 4.13: Case-Based Reasoning

the w eighted inputs, and com pares the total to a threshold. If the sum is greater than
the threshold, the n euron fires. O therw ise, it generates no signal. This m eans it modifies
its activity in response to new input until it produces the right results.

Knowledge-Enabling Application Layer

The fourth layer is the know ledge-enabling application layer. It is also referred to
as the value-added layer; which creates a com petitive edge for the learning organiza-
tion. M ost of such applications provide the users b e tte r ways to do th eir jobs. They
include know ledge bases, discussion databases, sales force au tom ation tools, yellow
pages, decision support, and im aging tools. The ultim ate goal is to show how knowledge
sharing could im prove the lot of em ployees and the organization as a whole.

The Transport Layer

The transport layer is probably the m ost technical layer to im plem ent KM archi-
tecture. It ensures th a t the com pany will becom e a netw ork of relationships, w here
e-m ail and electronic tran sfer of know ledge th ro u g h o u t the organization will be as
ro u tin e as answ ering th e telep h one. The key change is a m ove away from classical
transm ission processing to know ledge transfer and the sharing of expertise.
Included in the tra n sp o rt layer are local area netw orks (L A N s), wide area n e t-
works (W ANs), intranets, extranets, and the Internet. The things to consider are multi-
m edia, U R L s (W eb addresses), graphics, connectivity speeds and bandwidths, search
tools, and m anaging netw ork traffic on a continual basis.

The Middleware Layer

W hen designing a KM system, concern m ust be given to how such a system will
interface with legacy systems and o th er program s or applications th at reside on differ-
ent platform s or operate u nder older operating systems. The term legacy system refers
134 7777 PART II Knowledge Creation and Capture

to applications that once ran on m ainfram es or operated under traditional program -

ming languages, such as C O B O L . The designer needs to address databases and appli-
cations with which the KM system m ust interface. The m iddlew are layer contains a
cluster of program s to do th at job —to provide connections betw een legacy applica-
tions and existing and new systems. Essentially, it m akes it possible to connect betw een
old and new data formats.

The Repositories Layer

The b o tto m layer in the KM a rc h itectu re rep resen ts the physical layer w here
repositories are installed. T hese include intelligent data w arehouses, legacy applica-
tions, operational databases, and special applications for securing traffic m anagem ent
and integrity assurance of the knowledge architecture. Each repository has a structure
appropriate for the type of know ledge stored.
Once the repositories are established, they are linked to form an integrated reposi-
tory. For example, knowledge about sales might be in one repository, while sales best
practices are stored in another repository. A n integrated repository brings together all the
knowledge available from the repositories and creates a profile of the next sales strategy
for the following q u arter or for the rem ainder of the year. A conceptual model of an
enterprise-wide integrated knowledge m anagem ent system is shown in Figure 4.14.
The main advantage of an integrated knowledge repository is that the user does
not care in which repository the know ledge resides. O n the other hand, as users add
content to various repositories with no clear validation procedure, inform ation over-
load could cause all kinds of validity and reliability problems. C ontent in repositories
should be proactively m anaged, not only for adding content, but also for getting rid of
old content and elim inating redundancies.
M anaging an integrated repository m eans knowing w here content comes from and
what inform ation to tap. In a banking installation, for example, the key content centers
include the following:

• O perations d epartm ent (bookkeeping, accounting, payroll, teller work)

• Loan departm en t (com m ercial, autom obile, personal, construction, student, and
collateral loans)
• H um an resources d epartm ent (career planning, perform ance data, skills data,
pension)
• Trust d epartm ent (pension m anagem ent, investm ent, stock trading)
• Safe deposit d epartm ent (box rental, billing, opening/closing new accounts)

As to w hat inform ation to tap, the same financial institution might consider the
following:

• W hat do outside sources say about our com petition?

• Financial and governm ent agencies: financial data, governm ent reports, and
industry data
• C ustom er and business community: newspapers, articles, and cham bers of
com m erce
• Business organizations: vendors, suppliers, users, and industry journals
• W hat inform ation is generated by the com petition?
• N um ber of new branches in the area
• Q uarterly annual reports
• A dvertising volum e and frequency
 CH A PTER 4 Knowledge Creation and Knowledge Architecture 1111 135

“A n yw h ere” Clients

□
Integrative Knowledge
System

Connectivity
Through the Internet ,

□
Application UNIX
Server Server

CT J}
Transaction Web Database
Server Server Server

Web Browser Based

■ I I I I 9 FIGURE 4.14: Enterprise-Wide Integrated Knowledge Management System

BUILD IN-HOUSE, BUY, OR OUTSOURCE?

The q u estio n of w h e th e r to build a know ledge m an ag em en t system in-house from
scratch o r a d o p t an o ff-th e-sh elf system involves several options. A s sum m arized
in Table 4.1, building a know ledge m an ag em en t system in-house from scratch re -
quires ready professionals in system design, w here the developm ent cost is high and the
risk of the system not running on the first try is also high. The m ain benefit is custom iza-
tion, which m any firm s find well w orth the cost. O n the o th e r extrem e, installing an
off-th e-sh elf system m eans quick in stallatio n and low cost, although custom ization
is usually n o t quite right in m ost cases. In betw een are situations w here the end user
delves into system design. In m ost cases, this is not a recom m ended approach to system
developm ent.
The tren d in to d ay ’s softw are is tow ard ready-to-use, generalized softw are pack-
ages. O u tso u rcin g is also a tre n d th a t allow s com panies to c o n c en trate on th eir
strengths while technological design and o th er specialized areas are released to o u t-
siders. B ased on th e a u th o rs’ experience, purchasing a softw are package represents
136 7 7 7 7 PART II Knowledge Creation and Capture

TABLE 4.1 Key Elements In Build Versus Buy Decision

Option Cost Time Factor Customization Comments
In-house Usually high Much longer than off- High, depending on Consultants can be hired to
development the-shelf installation quality of staff improve quality, but the cost
will be higher. Leak of unique
system features by consultants
is a possibility.
Development Usually low Depends on skills set, High to the user This is not recommended for
by end users system priority, and specifications for first-time users.
so forth
Outsourcing Medium to high Shorter than in-house High The competition might
already have the same
knowledge system.
Off-the-shelf Low to medium Nil Usually up to Installation time is short, and
solution 80 percent usable installation is not that difficult.
1 fl 9 fl fl fl

roughly 10 percent to 20 percent of the cost of developing the same in-house. A reliable
KM softw are package offers several advantages:
• Short im plem entation tim e. Installation of a softw are package is com pleted in a
m atter of days rath er than the m onths necessary for “hom egrow n” packages. It also
contains fewer errors and is of higher overall quality than m ost in-house software.
• R educed need fo r resources. In-house software can be costly, and com pletion
dates are hard to predict. Incom plete software projects are not uncom m on
because of turnover am ong program m ing staff.
• L ow er developm en t cost. A package can take years to build, great expense to
com plete, and even m ore time to m aintain. U sers do not have the expertise to
cost-justify developing such packages.
• Greater flexibility. Users are not saddled with a software package, as they would be
with in-house programs. If the package fails the test, there is no obligation to acquire.
Software packages, on the other hand, are not w ithout drawbacks:
• Shorter track record. Many of the software houses have been in the knowledge m an-
agement business less than 5 years. It is difficult to predict how long they will continue.
• A pplication incom patibility. For a sophisticated user, extensive m odification of a
package can be costly. If the user modifies the software, the vendor may no longer
be responsible for any errors that occur.
• L ack o f com petition. Selecting a package is often not straightforw ard. In some
cases, a softw are package is one-of-a-kind. In other cases, it is the only available
softw are for the problem .
W hether the choice is build, buy, or outsource, it is im portant to set criteria for the
selection. In know ledge m anagem ent system developm ent, the following com ponents
are generally agreed to be crucial:
• R eliability has to do with how long software will operate w ithout a failure. It also
relates to how accurate and dependable the results are. Software does not fail. It
does not even w ear out. Invariably, reliability problem s are the result of errors
overlooked during the developm ent process. In contrast, hardw are fails because
of w ear and tear associated with continued use or inadequate m aintenance.
 CHA PTER 4 Knowledge Creation and Knowledge Architecture i i i a 137
• M arket pressure to produce software has had an adverse impact on the reliability of
software in general. To make things worse, first-time users do not know much about
software testing. A dem onstration by a sales representative is often all it takes to pur-
chase the package. For this reason alone, inquiring about vendor support is crucial.
• M odu larity refers to the ease of m odifying the softw are across product lines.
Softw are with high m odularity operates in com puters across product lines. It
should be noted that m ost of to day’s softw are packages are seldom m odified by
the user. In contrast to m odularity, expan dability has to do with how well the soft-
ware can be m odified to m eet a user’s changing requirem ents. Some of the
expandability questions related to know ledge m anagem ent systems are:
• H ow easily can know ledge be changed or relocated?
• Will the know ledge m anagem ent system be unusable if a part of it fails?
• W hat are the recovery features of the know ledge m anagem ent softw are?
R elated to expandability are capacity and flexibility. C apacity refers to the
softw are’s capability for file size, volume of transactions, and additional reports.
Flexibility is the ease of m odifying or extending a softw are package to address
changing requirem ents.
• Usability ensures user friendly software. For exam ple, the popularity of dBase II in
the early 1980s was attrib u ted to the fact that a novice user could create and m ain-
tain a database file in less than 4 training hours. In this respect, it was highly usable.
• P ortability is a m easure of how well the softw are will run on different computers.
A softw are package rated high on portability would also be high on usability.
• Serviceability em phasizes the im portance of vendor services and support. A n
effective package m ust be well docum ented for m aintenance and future enhance-
ments. This includes a description of the system logic, input/output file descrip-
tions and layouts, and a readable user m anual. V endor support m eans technical
support, softw are m aintenance, and updates. O n-site training and conversion sup-
port are part of m ost com m ercial software installations.
• Perform ance is one of the m ost critical criteria of software selection. It is a m ea-
sure of the softw are’s capacity to m eet user requirem ents and perform under
peak loads. The program m ing language of the softw are may be a perform ance
factor. For exam ple, a know ledge m anagem ent package w ritten in C + + or Java
usually outperform s substantially a similar package w ritten in BASIC. However,
it is cheaper and easier to modify a package w ritten in BASIC.
Buying or even o u tso u rcin g a know ledge m an ag em en t system brings up the
im portant question of ownership. M ost license agreem ents lease (not sell) the software
for an indefinite time period. To protect proprietary rights, the vendor does not provide
the source code. H ow ever, the availability of the source code becom es crucial when
dealing with a softw are house that could fold w ithout notice. The questions to consider
on ow nership are as follows:
• W hat is the user paying for?
• W hat restrictions are there to copying the software for com pany branches or
oth er d epartm ents within the firm?
• W ho m odifies the softw are? W hat are the charges? How are m odifications m ade
if the vendor goes out of business?
R elated to the issue of buy versus outsource is evaluating the vendor’s services in
term s of conversion, backup, m aintenance, and upgrade. V endor-assisted conversion
provides prog ram m in g su p p o rt, file conversion, and user training. B acku p ensures
138 .«■. PART II Knowledge Creation and Capture

altern ativ e sites for processing applications in an em ergency. M aintenance ensures

against unexpected system failure. In contrast, upgrade revises the system in order to
m ake it m ore current.

■■■« Implications for Knowledge Management

A successful know ledge m anagem ent system cannot exist w ithout a clearly defined
architecture. The cost of correcting the “wrongs” at a later stage can be prohibitive. In the
early stages of planning such infrastructure, it is im portant to keep the following in mind:
• D ecide whether to build, outsource, or buy the software. There is really no best
answer. In any case, the company has the right to modify the KM system as the situa-
tion warrants. Buying the system rather than leasing also means cash savings from
depreciation and lower cash outlay if the system is kept 5 years or longer. The main
negative, however, is that maintenance, insurance, and upkeep are borne by the user.
• U nderstand what yo u are building in term s o f the com ponents o f the K M architec-
ture. This m eans evaluating the quality, reliability, perform ance, and m anageability
of the user interface, integrative repositories, the collaborative platform , and
know ledge directories. The user interface, for example, provides a unique oppor-
tunity to see how well the users buy into the system. The IT departm ent or an out-
side consultant can do the evaluation and inform m anagem ent of the status quo.
• Focus on optim ization fo r scalability , perform ance, and flexibility o f the K M archi-
tecture. Prior to com m itting the KM system to actual operation, rigorous testing
of the technology is critical. The system should be tested not only to see how well
it works, but also under w hat conditions it begins to fail. This is called force-fail
testin g—a step that requires experience and patience.
• Ensure interoperability. This m eans that the existing technology and its affiliate
com ponents, intranet, extranet, databases, legacy systems, and PCs should all
w ork together in an interactive and integrative environm ent. The alternative
could be fragm ented applications, w here one application is totally oblivious to the
technical and know ledge-based activities of the others.
A n o th er area of im portance is m anaging a KM project. Choosing and recruiting
the right specialists for the KM project team is a vital task. Team m em bers come from
m ultiple disciplines representing various areas of the firm. The focus should be first on
people, followed by business processes and organizational culture. The technology that
ensures know ledge sharing com es next (see Box 4.2).

1 1 1 1 1 1 BOX 4. 2

DEVELOPING A KM TEAM

WHO IS IN CHARGE? m ust und erstan d the business requirem ents and
problem s the com pany m ust address and be able
W hen m any business units are involved in an
to articulate a broad vision that spans the unique
endeavor, project leadership can becom e a point
needs of the lines of business —and the business
of contention. A team leader m ust be able to bal-
processes and technical in frastru ctu re th a t can
ance the various, som etim es com peting, dem ands
m eet those needs.
on the group while keeping it focused on m eeting
the o rg an izatio n ’s know ledge needs. The lead er (con tin ued)
 CH A PTER 4 Knowledge Creation and Knowledge Architecture aaaa 139

(con tin ued) willing to give it th eir all if they see the project
as a finite effort with a definite end, after which
A very large organization should hand the reins they can return to their norm al duties, observers
of team leadership to a chief know ledge officer say. For this reason, it may be a good idea to set
(CKO). You need someone with the influence and a life span for the team at the outset of the proj-
skills to take the enterprise in the right direction. In ect. The d u ra tio n can vary from a few m onths
any case, team leaders must comm and respect and for p ro jects th a t seek to plan t the seeds for a
trust within the company. This is especially true if m ore am bitious e n d ea v o r to several years for
team members come from different business units. com plex projects that involve reengineering busi-
Team m em bers who do not view a KM p ro - ness processes.
ject as a new p ath to career success will be m ore

SO U R C E : Excerpted from Robb, Drew. K now ledge M anagem ent , August 2001, pp. 44-50.

■ fl fl fl 1 fl

Sum m ary • • • •
• O ne way of creating know ledge is based on experience; the other way is through
team w ork. E ith er way, the outcom e is the same, because the focus is knowledge
sharing. Team form ation and team w ork begin with experienced individuals w ork-
ing jointly on a project.
• K now ledge sharing is influenced by personality, attitude, vocational reinforcers,
and w ork norms. The concept also considers organizational culture, com pany
strategies, and com pany policies. It is not so straightforw ard, because people are
m ore used to know ledge hoarding than know ledge sharing.
• K now ledge architecture is a com bination of people, content, and knowledge shar-
ing. They are in terd ep en d en t and inseparable. People with know ledge provide
content, relying on technology to transfer and share knowledge.
• P art of the people core is profiling. The goal of profiling is to get a handle on
existing know ledge exchanges that the organization relies on and ways to capture
them . The idea is to assign job content to the right person and to ensure that the
flow of inform ation that once was obstructed by departm ents now flows to the
right people at the tim e needed.
• O nce profiling is com pleted, the next step is to identify where knowledge re-
sides and how to capture it. These centers becom e the skeletal fram ew ork
for know ledge capture. A t the sam e time, experts are assigned to each knowl-
edge area to assure continuity and effective m anagem ent of the knowledge
sharing process.
• The goal of the technical core is to enhance com m unication and ensure effective
know ledge sharing. By technical core, we m ean the totality of hardw are and soft-
ware, and the specialized hum an resources required to enable knowledge capture.
O nce in place, a know ledge core becom es a netw ork of technologies designed to
work on top of the com pany’s existing netw ork.
• The best way to visualize the technical architecture as the building block of KM is
in term s of a layer system, starting with the user interface and going all the way to
the repositories that form the m ost technical layer of the technical core. In term s
140 7777 PART II Knowledge Creation and Capture

of the user interface layer, consistency, relevancy, visual clarity, navigation, and
usability are im portant design considerations.
The technical access layer involves the use of the Internet, intranet, and extranet
as p art of the technical infrastructure of the KM system. Each area has unique
contributions to make. Because of the vulnerability of the intranet and extranet,
firewalls and other security devices are installed.
The collaborative intelligence and filtering layer reduces search time for inform a-
tion by com bining the know ledge sought with the u ser’s profile. Intelligent agents
are used to assist in problem -solving by retrieving the right inform ation. For an
effective collaborative layer, security, portability, integration, scalability, flexibility,
and ease of use are im portant design criteria.
The collaborative layer relies on expert systems, CBR, and neural nets to do its
job. E xpert systems are a branch of A l, which assist a perso n ’s thinking process
based on years of captured hum an experience. C BR is based on real-life cases
that allow the decision m aker a chance to com pare his or her decisions to the
ones decided by older cases. N eural nets are inform ation systems m odeled after
the hum an b rain ’s netw ork of neurons. They can be trained to improve on the
quality and accuracy of decisions.
The rem aining layers are know ledge-enabling application* transport, and m iddle-
ware layers. E ach layer gives the user b etter ways to do the job. The transport
layer is the m ost technical layer to im plem ent in KM architecture.
The bottom layer is the one representing the repositories, including intelligent
data w arehouses, legacy applications, operational databases, and special applica-
tions for securing traffic m anagem ent and integrity assurance of the knowledge
architecture.
Building a KM system in-house provides the benefits of custom ization and p er-
sonalization, although at a high cost. The alternative is to outsource or buy an off-
the-shelf package. KM softw are packages usually m ean short im plem entation
time, lower developm ent cost, and greater flexibility. The main draw backs include
shorter track record, application incom patibility, and lack of com petition
Finally, the role of the CKO is crucial in developing KM architecture. The key
functions of such an officer are maximizing the returns on investm ent in knowl-
edge, sharing the best practices, prom oting and im proving com pany innovations,
and minimizing brain drain or knowledge loss at all levels of the organization.
Interpersonal and technical skills are also crucial.

Terms t o K n o w 1111
Artificial intelligence: The science of making machines relevant cases to determine their usefulness in solving a
do things that would require intelligence if done current problem; computer systems that solve new
by humans; the capacity to acquire and apply an problems by analogy with old ones.
understanding gained through experience or Collaborative intelligence: A knowledge management
study in order to imitate or emulate “natural technical layer that provides customized views based
intelligence.” on stored knowledge.
Authentication: Making sure that a cardholder is, in Data warehouse: A repository where relational data are
fact, the person authorized to use the card. specially organized to provide “cleaned” data in a for-
Browser: A program designed to search for and bring in mat understandable to the user.
Internet resources. Expert system: A sophisticated computer program that
CBR (case-based reasoning): A m ethodology that records applies human knowledge in a specific area of expertise
and documents previous cases and then searches the to create solutions to difficult problems.
 CH A PTER 4 Knowledge Creation and Knowledge Architecture a a a a 141
Extranet: Network that connects separate companies with processing elements called neurons; a self-program-
a shared database. ming system that creates a model based on its inputs
Fault-tolerance: In hardware, having a specially designed and outputs.
operating system that keeps the Web site or any appli- People core: The human part of a knowledge manage-
cation running, even when the central processing unit ment environment and the way people creatively per-
goes down. form in that environment.
Firewall: A network node consisting of hardware and Performance: A measure of the software’s capacity to
software to protect or filter certain information enter- meet user requirements and perform under peak loads.
ing the company’s databases or keep select information POP: Point of presence; a physical location on the
from leaving the company. premises of a local exchange carrier at which messages
Inference engine: The “brain” of an expert system; a are transferred or linked to other carriers for routing to
cluster of computer programs that coordinates through their intended destinations.
a scheduler the reasoning and inferencing based on Portability: Ability of software to be used on different
the rules of the knowledge base to produce the solu- hardware and operating systems.
tion or advice. Profiling: In knowledge management systems, it is gener-
Inferencing: Deriving a conclusion based on statements ating a graphic or textual representation of people in
that only imply that conclusion. terms of criteria such as skills or personality traits.
Intelligent agent: A cluster of precanned bits of intelli- Reliability: Has to do with how long software will operate
gence that relate specific needs or attributes to other without a failure.
agents on a network. Repository: A subsystem such as a database used for
Internet: An infrastructure that links thousands of net- data, information, or knowledge storage.
works to one another; the information highway that Scalability: Ability of a computer system, a database infra-
makes the information stored on thousands of com- structure, or a network to be upgraded to new standards.
puters worldwide available to millions of people Serviceability: A criterion that emphasizes the impor-
everywhere. tance of vendor services and support.
Interoperability: Ability to move back and forth between SMTP: simple mail transfer protocol; a TCP/IP protocol
different systems. used on the Internet to deliver messages between elec-
Intranet: A network using TCP/IP to share information tronic mail hosts and to specify message structure.
within an organization. TCP/IP: Transmission control protocol/Internet protocol;
Justifier: Explains the action (line of reasoning) of the standards architecture used on the Internet and in
expert system to the user. many other networks.
Knowledge center: An area in the organization where Technical core: The aspect of the knowledge management
knowledge is available. system that represents the technology and technical
Knowledge creation: Generating experience via team- features that make the system operational.
work, repositories, or other interfaces. Transport layer: The layer that standardizes exchanges
Legacy application: An application that once ran on between the operating systems of the computers in the
mainframes or operated on traditional programming system.
languages such as Cobol. Usability: The degree to which a system is deemed useful
Middleware: A software layer in the technology core that by the end user; a criterion that ensures user friendly
makes connections between old and new data formats. software.
Modularity: Refers to the ease of modifying the software User interface: A layer that standardizes exchanges
across product lines. between the ultimate user and the system.
Navigation: In Web design, it refers to how easily a user Web: Also called World Wide Web; a service on the
moves through Web pages, files, or records in a com- Internet in which users working with browsers can
puter-based system. retrieve Web pages from any Web server around the
Neural net: An information system modeled after the world. Each page is in hypertext format and contains
human brain’s network of electrically interconnected links to other Web pages.

Test Y o u r U n d e r s t a n d in g 1111
1. W hat is know ledge creation?
2. A job is m ore than a task. D o you agree? Give an example.
3. H ow is know ledge created and transferred via team s?
4. Explain the m ain im pedim ents to know ledge sharing.
142 7 7 7 7 PART II Knowledge Creation and Capture

5. Distinguish betw een:

a. transport layer and application layer
b. usability and portability
c. profiling and repository
d. collaborative intelligence and intelligent agent
6. Why do we profile people in knowledge m anagem ent design? Be specific.
7. W hat is the m ain difference betw een an in tran et and an extranet? W here
does the In tern et fit in?
8. Explain in your own words the functions of middleware. How does m iddle-
ware differ from the user interface?
9. W hat does it m ean to leverage technology?
10. Why is it a critical requirem ent that a successful knowledge m anagem ent sys-
tem be tagged to collaboration?
11. People are no longer expected to adapt to the requirem ents of machines, but
m achines are expected to adapt to people. Explain.
12. W hen it comes to leveraging, what should one keep in mind?
13. Explain the basis for the user interface layer and the services it offers.
14. W hat design considerations are unique to the user interface?
15. W hat is involved in authorized access to a know ledge system?
16. Explain the m ain steps in knowledge transfer.
17. How does an intelligent agent differ from filtering? Be specific.
18. Explain briefly the key com ponents of the collaborative intelligence layer.
19. W hat distinguishes a Web brow ser from a client/server netw ork?
20. Cite some prerequisite criteria for an effective collaborative layer.
21. D istinguish between:
a. scalability and portability
b. flexibility and ease of use
c. fault-tolerance and integrity
d. expert systems and CBR
22. W hat are the main drawbacks of expert systems? A re there any unique benefits?
23. How do databases differ from legacy systems?
24. Identify a business organization and cite its key content centers.

Knowledge Exercises • • * *
1. A rrange a visit to a m edium - or large-size firm in your area. In an interview
with a com pany representative, address the following:
a. Identify the satellite knowledge centers.
b. Identify m anagers for each know ledge center within departm ents or
divisions.
c. Identify the com pany’s KM strategies and specific projects that support
overall corporate goals.
d. W rite up strategic plans to higher levels of m anagem ent and get their
buy-in.
2. Access to know ledge is based on profiles derived from the knowledge base.
W hat technology would you recom m end to provide access? W ho will m an-
age such access?
T. Visit the In tern et and review the latest applications based on expert systems,
C BR, and neural nets. W rite a short report showing how these applications
relate to know ledge m anagem ent.
4. Visit the In tern et and look up two articles that explain the role of the CKO.
R ep o rt your findings to class.
 CHA PTER 4 Knowledge Creation and Knowledge Architecture aaaa 143

Ref ere nc es , , im
Berry, John. Real Knowledge Is Held by People. New Malhotra, Yogesh. “Knowledge Management in Inquiring
York: CMP Media, Inc., 2000. Organizations,” 2001, www.brint.com/km, Date
Boland, R. J. “The In-Formation of Information Sys- accessed August 2002.
tems,” in R. J. Boland and R. Hirschhe (eds.), Critical Marwick, A. D. “Knowledge Management Technology,”
Issues in Information Systems Research. Chichester: IB M Systems Journal, vol. 40, no. 4,2001, pp. 814-830.
Wiley, 1987. Nonaka, I., and Takeuchi, H. The K nowledge Creating
Churchman, C. W. The Design o f Inquiring Systems. New Company. Oxford, UK: Oxford University Press, 1995.
York: Basic Books, 1971. Robb, Drew. “Draft Your Dream Team,” Knowledge
Davenport, T. H., and Prusak, Laurence. Working M anagement, August 2001, pp. 44-50.
Knowledge. Boston, MA: Harvard Business School Shruti, Date. “Agencies Create CKO Posts to Get in the
Press, 1998. Know,” G overnm ent Com puter N ew s , November 8,
Flash, Cynthia. “Who Is the CKO?” Knowledge 1999, pp. 1-2.
Management, May 2001, pp. 37-41. Tapscott, Don. G rowing Up Digital: The Rise o f the Net
Gill,T. G. “High-Tech Hidebound: Case Studies of Generation. Boston, MA: Harvard Business School
Information Technologies That Inhibit Organizational Press, 2001.
Learning,” Accounting, Management and Information Tiwana, Amrit. The Knowledge Management Toolkit.
Technologies , vol. 5, no. 1,1995, pp. 41-60. Upper Saddle River, NJ: Prentice Hall, 2000, p. 309.
 Capturing Tacit
Knowledge
Contents

In a Nutshell
W hat Is Knowledge C apture?
Evaluating the E xpert
Levels of Expertise
C apturing Single Versus M ultiple E xperts’ Tacit Knowledge
A dvantages and D raw backs of Using a Single E xpert
Pros and Cons of M ultiple Experts
D eveloping a R elationship with Experts
C reating the Right Im pression
U nderstanding the E x p ert’s Style
Preparing for the Session
Deciding W here to Hold the Sessions
A pproaching M ultiple Experts
Fuzzy R easoning and the Q uality of Knowledge C apture
A nalogies and U ncertainties in Inform ation
U nderstanding Experience
The Language Problem
The Interview As a Tool
Types of Interviews
G uide to a Successful Interview
Setting the Stage and Establishing R apport
Phrasing the Q uestions
Q uestion C onstruction
Things to Avoid
Reliability of Inform ation from Experts
E rrors M ade by the K nowledge D eveloper
Problem s E ncountered D uring the Interview
Ending the Interview
Issues to Assess
R apid Prototyping in Interviews
Benefits and D raw backs
 CH A PTER 5 Capturing Tacit Knowledge i «i i 145
Im plications for Knowledge M anagem ent
Sum m ary
Terms to Know
Test Your U nderstanding
K now ledge Exercises
R eferences

I suppose it was something you said

That caused me to tighten and pu ll away
A n d when you asked, “What is it? ”
I, o f course, said, “Nothing. ”
Whenever I say, “Nothing, ”
You may be very certain there is something.
The something is a cold, hard lump o f " N othing”
— Lois W y s e

In a Nutshell
K now ledge m anagem ent system s derive their pow er from the know ledge they use.
W hen know ledge developers begin the building process, they are confronted with the
first ste p —capturing tacit know ledge (see Figure 5.1). As m entioned in earlier chap-
ters, tacit know ledge resides in the minds of people —the experts in the departm ent,
division, or the firm. A know ledge developer converts hum an know-how into machine-
ready “say-how ” by using an iterative process of articulation, a series of refinem ent
cycles, or rapid prototyping, in which the com puter’s perform ance is com pared to that
of the hum an expert.
C apturing tacit know ledge and converting it into rules that the com puter can use is
a costly business. It requires an extensive time com m itm ent from the dom ain expert
and the special skills of the know ledge developer. A t times, the expert m ight lose inter-
est in the project and even feel like quitting. Perhaps the knowledge developer and the
ex p ert ju st n ever seem to hit it off because of their interp erso n al chem istry, or the
knowledge developer may use the wrong tool or approach.
W orking with experts in capturing their tacit know ledge is not a straightforw ard
routine. For exam ple, the m ethods or tools chosen for know ledge capture depend on
the tem peram ent, personality, and attitude of the expert and w hether the knowledge
autom ation system is being built around a single expert or m ultiple experts. A nother
im portant factor is w hether one or m ore knowledge developers will be involved in the
building process.
B efore beginning the know ledge capture process, a knowledge developer needs to
have an understanding of the e x p ert’s level of expertise. The knowledge developer can
look at several indicators of expertise as well as specific qualifications to determ ine
w hether som eone is an expert. O ne of the most im portant indicators is the ex p ert’s
com m unication skills.
Using either a single or m ultiple experts each has advantages and limitations. For
sm aller know ledge m anagem ent systems, a single expert should suffice. In problem sit-
u ations with d ispersed know ledge, which req u ire synthesis of experience, m ultiple
experts are p referred. In either case, working with experts (and accom m odating the
146 7777 PART II Knowledge Creation and Capture

/
Knowledge
Knowledge
Innovation

Sharing

Goal
Knowledge
Transfer

I I I fl I fl FIGURE 5 . 1 Knowledge Codification in the KM System Life Cycle

ex p erts’ individual styles of expression) requires special skills. The location at which
know ledge cap tu re sessions are held adds a n o th e r dim ension th a t the know ledge
developer m ust consider.

• • ■ ■ What Is Knowledge Capture?

A review of the literature reveals as m any definitions of knowledge capture as there
are authors. Some of the definitions are interesting:

• A n “applied brain d rain ”

• A “m anual craft that depends on the skill and effectiveness of the knowledge
d eveloper”
• The “transfer of problem -solving expertise from some know ledge source to a
repository or a pro g ram ”
 CHA PTER 5 Capturing Tacit Knowledge ■■■a 147
• The “process by which know ledge m anagem ent system developers discover the
know ledge th at com pany experts use to perform the task of in terest”
• A n investigative experim ental process involving interview s and protocol analysis
in o rd er to build a KM system
In this book, we define know ledge ca p tu re as a p ro c e ss b y which the e x p e r t’s
thoughts an d experiences are captured . It is a sort of “m ind autom ation.” In a broader
view, know ledge cap tu re m ay also include capturing know ledge from o th er sources
such as books, technical m anuscripts, and drawings. In fact, these processes lead to the
case-based reasoning used in the building of m any KM systems.
Know ledge capture is a dem anding m ental process in which a know ledge devel-
o p er collaborates with the exp ert to convert expertise into a coded program . T hree
im portant steps are involved:
1. Using an ap propriate tool to elicit inform ation from the expert
2. In te rp re tin g the in form ation and inferring the e x p e rt’s underlying know ledge
and reasoning process
3. U sing the in te rp re ta tio n to build the rules th at rep resen t the e x p e rt’s thought
processes or solutions
The cap tu re of know ledge can be accom plished in m any ways; no single way is
best. The process typically entails several refinem ent cycles. The KM system ’s task is
not simply to display the know ledge but also to codify it at different levels of reasoning
or explanation. Box 5.1 illustrates the challenge in know ledge capture.
K now ledge is usually c a p tu re d w hen the know ledge d ev elo p er interview s the
expert, who answ ers the follow ing questions: W hat do you do as a first step? W hat
inform ation do you consider next? W hat constraints do you look for? These types of

■ ■ a a a a box 5.1

THE CHALLENGE IN KNOWLEDGE CAPTURE

Som etim es even the best know ledge d e v e lo p - because the two know ledge developers w orking
m ent tech n iq u es and th e m ost ta le n te d kn o w l- with him on know ledge capture were not picking
edge developers cannot get the job done, which up on w hat he was saying fast enough.
happened to two Texas Instrum ents (TI) know l- Too am bitious a scope and poor com m unica-
edge developers who w ere signed on to capture tion betw een the know ledge d ev elo p er and
the know ledge of T hom as Kelly. Kelly, a civil expert crippled the progress of this m ost interest-
developer with S outhern C alifornia E dison, is a ing system. Today, Project Kelly has finally wound
genuine expert. H e can alm ost feel the presence dow n. A good p ro to ty p e, b u t n o t a deployable
of d an g er at th e V erm ilion h y d ro electric dam . one, was built. Southern California estim ates that
This task was a com plex one th at relied on many an o th e r $200,000 will have to be p u m ped into
rules of thum b. How did Kelly know that a small V erm ilion dam to m ake it a generic dam know l-
stream of w ater on one side of the dam m eant a edge m anagem ent system. Now it just keeps Tom
blocked d rain at its base? K elly was unhappy, Kelly com pany during his lonely vigils.

SO U RC E: Awad, E. M. Building Expert Systems. Minneapolis, MN: West Publishing, 1996, p. 144.
148 7777 PART II Knowledge Creation and Capture

questions lead the expert through scenarios, or case situations. Then, the knowledge
developer returns to specific points and questions the expert further until all angles of
the problem are explored. H ere is a partial session with a loan officer who is trying to
capture her experience on how she decides to approve or deny a loan:
S teve (know ledge developer): W hat is your m ain job as a loan officer?
S andra (loan officer): I review loan applications and recom m end which
ones qualify for a loan.
St eve: H ow long have you been on this job?
Sa n d r a : E ig h te en years w ith this bank and 7 years with
another bank before coming here.
St eve: The idea of lending and how loans are approved or
denied is intriguing. Suppose you receive a loan ap-
plication. Then what? I m ean, what do you do next?
Sa n d r a : Well, let me see. This is going to take time to explain.
The gist of the process I follow is to m entally plug
the inform ation that the applicant provides against
criteria I use that help me decide w hether to approve
or deny the loan.
St eve: W hat criteria do you use?
Sa n d r a : The m ain criteria are the ap p lic a n t’s em ploym ent
status and how long he or she has been on the job,
m arital status, annual salary and o th e r sources of
incom e, credit history, m onthly expenses, and the
like. There is really no set procedure, and nothing is
set in concrete. It all depends.
St eve: W hat do you m ean?
Sa n d r a : Well, I sort of size up the overall application against
criteria, strategies, potential, the interest of our bank
in the loan, the state of the com petition, and so on. I
tell you, I d o n ’t mind elaborating further, but maybe
we should allocate a n o th e r session or two for the
details. W hat do you think?
St eve: (sensing unease by Sandra) I agree. How about next
Tuesday, the same time, the same place?
Sa n d r a : Fine, unless you hear from me today or tom orrow ,
le t’s go ahead and m eet then.
St eve: Thanks. See you Tuesday. M eantim e, if you have any
set rules that I can work with regarding the criteria or
the decision process, I ’d app reciate your sending
them to me by e-mail at your convenience.
Sa n d r a : I’ll see what I can do.
To c a p tu re th e know ledge used in building com plex KM systems, know ledge
developers use flow charts, flow diagram s, decision trees, decision tables, and o th e r
graphic rep resen tatio n s. U sing the loan cap tu re scenario, a p artial decision tree is
shown in Figure 5.2. You can appreciate the logical flow of a decision tree and how
easy it is to codify tacit know ledge based on graphical flow.
K now ing how ex p erts know w hat they know is the b o tto m line in know ledge
cap tu re. E x p e rt know ledge is cognitively com plex and tacitly pragm atic. It is not
always easy to c a p tu re th ro u g h a tra d itio n a l interview process. In m any cases in
 CHA PTER 5 Capturing Tacit Knowledge aaaa 149

fl fl fl I fl fl FIGURE 5.2: Partial Decision Tree for a Loan Advisor

w hich know ledge c a p tu re was unsuccessful, th e know ledge d ev elo p er did not
quite understand the pragm atic natu re of the expertise. Just as the knowledge devel-
o p er m ay n o t fully u n d erstan d the expert, the expert m ay be equally unclear about
th e know ledge d ev elo p er. E x p erts som etim es perceive know ledge developers as
dom ain novices, w ho re q u ire p atience and w ho m ust go through an apprenticeship
process th at takes them from novice to near-expert status during the building of the
know ledge m anagem ent system. This perception is even m ore reason for the know l-
edge d e v e lo p e r to d ev o te serio us efforts to w ard p re p a ra tio n and a fam iliarity in
the dom ain.
The following suggestions can be used to im prove the know ledge capture process:

1. K now ledge developers should focus on how experts approach a problem . They
m ust look beyond the facts or the heuristics.
2. Know ing th at tru e expertise takes the form of chunked knowledge, know ledge
developers should reevaluate how well they understand the problem dom ain and
how accu rately they are m odeling it. C an they see p a tte rn s and relationships
leading to a solution? T hat is, can the know ledge developer grasp the complexity
150 7777 PART II Knowledge Creation and Capture

of the dom ain? C onceptual tools (discussed in C hapter 6) provide the developer
with a concise, easy way to capture heuristics.
3. The quality of hum an em ulation is best captured through episodic knowledge or
know ledge b ased on p revious experience. T h erefo re, a know ledge developer
should elicit the ex p ert’s know ledge through concrete case situations or scenarios.

■«■ ■ Evaluating the Expert

Tacit knowledge capture requires free access to a cooperative and articulate expert. In
m ost cases, th e know ledge d ev elo p er does not have the luxury of deciding on the
expert. H ow ever, he or she m ust be able to identify real expertise and how well a p ar-
ticular ex p ert’s know -how suits the project (see Box 5.2).
W ithout a good m atch of expert and project, the knowledge developer could spend
hours questioning som eone whose knowledge might be unusable in the proposed sys-
tem. T here are several indicators of expertise:
• Peers regard the ex p ert’s decisions as good decisions. The expert com m ands g e n -
u ine respect.
• W henever a specific problem arises, people in the com pany c o n su lt the expert.
• The expert adm its not knowing the answ er to a problem . H onesty indicates self-
c o n fid e n c e and a rea listic vie w o f lim itation s.
• The expert avoids inform ation that is irrelevant to the dom ain and instead sticks
to the facts and works with a focu s.
• The expert dem onstrates a k n a ck f o r ex p la in in g th in gs and can customize the p re-
sentation of inform ation to the level of the individual listener.
• A n ex p ert’s expertise is apparent to an inform ed listener. A n expert exhibits a
certain depth of detail and an e x c e p tio n a l q u a lity in explan ation s.
• The expert is n o t a rro g a n t about personal credentials, years of experience, or
strong ties with people in power.
In addition, experts usually exhibit specific qualifications:
1. K n o w s w h en to f o llo w h u n ch es a n d w h en to m a k e e x cep tio n s. This quality is a
result of experience. Trial and error teaches an expert when to m ake exceptions

BOX 5.2

CHARACTERISTICS OF AN AEROBIC FITNESS KM SYSTEM

The dom ain expert is the fitness program direc- city. She is quite en ergetic and excited ab out
tor. She is a certified aerobic and fitness instruc- w orking with me as the “e x p e rt” in developing
tor. She has her m asters in sports fitness and has the aerobic fitness know ledge m anagem ent sys-
been w orking in this field for m ore than 10 years tem . H e a th e r is co m m itted to her career and
in h ealth clubs and at university health centers. enjoys helping oth ers achieve fitness w henever
She also certifies the aero b ic in stru cto rs in the she can.

S O U R C E : Wakat, Heather. Interview by Mary Hill, knowledge developer of the aerobic fitness knowledge management system.
University of Virginia, Spring 2000.
 CHA PTER 5 Capturing Tacit Knowledge a a a a 151
and build the episodic memory. U nlike sem antic memory, which is the knowledge
of facts, episodic m em ory is the know ledge of cases based on experience. Case-
based reasoning is effective in building com plex KM systems.
2. Sees the big picture. Experts can sift through information and readily determine which
factors are im portant and which are superfluous. In a study of radiologists reading X-
rays, novices searched for particular features in the X-ray, looking for each feature’s
underlying causes. As the radiologists gained more experience, they began to realize
that certain features were dependent on other features in the X-ray and learned that
the context offered them cues. This process illustrates the complexity of true expertise.
3. P ossesses g o o d com m u n ication skills. A know ledge developer cannot create a
reliable expert system if the expert cannot com m unicate his or her knowledge.
V erbal skills facilitate sm ooth tacit know ledge capture.
4. Tolerates stress. Building KM system s can test an ex p e rt’s patience, particularly
w hen the know ledge developer asks the wrong question, the interview simply stag-
nates, or the dom ain is too difficult for the know ledge developer to understand.
5. Thinks creatively. A n expert should d em onstrate a certain inventiveness w hen
solving a p ro b lem or explaining the reasoning b eh in d a solution. C reative
ap p ro ach es can clarify th e e x p e rt’s th o u g h t processes and ability to convey
know ledge to the know ledge developer.
6. E xhibits self-confidence. E xperts believe in their ability to deliver. A self-confident
expert can accept criticism and react w ithout feeling threaten ed w hen an answer is
not known. Self-confidence also boosts an e x p ert’s capacity to handle stress.
7. M aintains credibility. E xperts who dem onstrate com petency in the way they field
questions gain a rep u tatio n for believability based on the experience others have
had with them .
8. O perates within a sch em a-driven orientation. The expert does serious schem a-
driven or stru ctu red , thinking in an effort to com e up w ith realistic, w orkable
solutions. (Two exam ples are provided in Box 5.3.)
9. Uses ch u nked know ledge. C hunks are groups of item s stored and recalled as a
unit. Several studies have shown that an individual’s ability to recall inform ation
increases p roportionately with the am ount of chunking. For example, a study of

BOX 5.3

EXAMPLES OF SCHEMA-DRIVEN ORIENTATION

1. A graduate student was interviewing with a Big national flights, and the like. For example, inter-
Six accounting firm for a position in information national flight takeoffs are heavier in early to late
systems. One of the partners asked, “How many evening than during the day. Landings are heavier
takeoffs and landings occur during a given day at during mid- to late afternoons.”
O ’Hare Airport?” The recruit snapped, “Fifteen 2. Two chess players were asked to recall moves of a
per hour, on the basis of 4 minutes per flight.” game in progress. One player recalled the moves
When later asked how the question should have by attack and defense strategies; the other player
been answered, the partner replied, “Well, there is recalled the moves by local position. The first
no specific answer. The number of flights depends player was the expert and showed evidence of
on the number of usable runways, the time of the higher-order processing than did the second player.
day, weather conditions, ratio of international to
152 7777 PART II Knowledge Creation and Capture

m aster-level and novice chess players involved in copying b o ard positions of

games in progress show ed that experts turned back to look at the board fewer
tim es th an did novice chess players. The experts took in larger chunks of the
game than did the novices. A n o th er study found that baseball experts tended to
recall entire sequences of plays m uch m ore easily and accurately than did indi-
viduals with only surface knowledge of the game. A possible problem with chunk-
ing, however, is the difficulty m any experts have in describing their knowledge —
segm enting chunked know ledge and verbalizing it.
10. G enerates m otivation and enthusiasm. Enthusiastic experts who enjoy what they
do generally perform well. Indifference and boredom contribute to poor m otiva-
tion and endanger the project.
11. Shares expertise willingly. A “tight” expert will produce a less-than-com plete tacit
know ledge. The m ost effective experts have a volunteering nature and do not
fear being eased out of their job.
12. Em ulates a g o o d teacher’s habits. M ost people rem em ber one particular teacher
because of the way that person delivered information. The best teachers are also good
learners, and good learners make sure knowledge is transferred effectively. A n effec-
tive expert “teaches” how to solve a problem realistically and achieve good results.
Obviously, no person can possibly possess all these attributes. Perhaps the most
im portant ones are acknow ledged expertise, the ability to com m it a large am ount of
time, and a gift for com m unicating knowledge. In return, experts have to be pam pered.

LEVELS OF EXPERTISE
The critical role of com m unication skills cannot be overem phasized. Various studies
have shown th at different levels of expertise influence com m unication quality. O ne can
arbitrarily classify experts into three levels: the highly expert person, the m oderately
expert person, and the new expert.
H ighly expert person s generally give concise explanations. They assume the listener
has enough know ledge about the problem ; therefore, they focus on the key steps, often
skipping vital details. Highly expert persons also tend to rate their own expertise highly.
M oderately expert p ro b le m solvers may be m ore tentative in their explanations, but
they tend to provide detailed explanations. They are quicker to give answers than the
highly expert person and m ore often adapt their description to the level of the know l-
edge developer.
N ew experts are m ore likely to offer answers that are brief and fragm ented, which
suggests shallow know ledge of the dom ain. (See Box 5.4 for a description of character-
istics and capabilities of an e x p ert’s knowledge.)
B ased on the generalizations about the levels of expertise, one can conclude that
th e m ore ex p ert th e p erso n is, the longer th a t person will tak e to com m unicate an
answer. A highly expert person stores years of experience in long-term memory, which
may take longer to recall and transform into a verbal response. The challenge for the
know ledge developer then is to help the expert “unpack” or “dechunk” the know ledge
a ccu rately and quickly. R echecking and cross-validating the e x p e rt’s opinions are
im portant steps to follow in the tacit know ledge capture process.

CAPTURING SINGLE VERSUS MULTIPLE EXPERTS’ TACIT KNOWLEDGE

To en su re th e reliab ility and quality of the KM system , a prim e co n sideration is
w hether to tap one expert or a panel of experts. Based on the au th o rs’ experiences, this
decision is based on several factors: the com plexity of the problem , the criticality of the
 CH A PTER 5 Capturing Tacit Knowledge ■■■■ 153

I fl 1 fl I 9 BOX 5.4 fl fl fl I fl fl

CHARACTERISTICS AND CAPABILITIES OF EXPERT KNOWLEDGE

E xperts are distinguished by th e quality and Specifically, ex p ert know ledge helps problem
quantity of know ledge they possess; they know solvers im prove th eir efficiency by m arshalling
more, and w hat they know m akes them m ore effi- relev an t facts, avoiding com m on errors, m aking
cient and effective. critical distinctions betw een problem types, pru n -
Expert knowledge, w hether applied by hum an ing useless paths of investigation, ordering search,
or machine, works in situations that do not adm it elim inating redundancy, reducing am biguities,
optim al or provably co rrect solutions. E x p e rt elim inating noise in data, exploiting know ledge
knowledge is used by problem solvers to find an from co m plem entary disciplines, and analyzing
acceptable solution that m eets or exceeds require- problem s from different perspectives or levels of
ments with a reasonable expenditure of resources. abstraction.

SO U RC E: Hayes-Roth, F., and Jacobstein, N. “The State of Knowledge-Based Systems,” Com m unications o f the A C M , March
1994. p. 28.

project to the organization, the types of experts available, and the funds allocated for
building the KM system. Each alternative has advantages and limitations.

ADVANTAGES AND DRAWBACKS OF USING A SINGLE EXPERT

There are several advantages to w orking with a single expert:

1. A single expert is ideal when building a sim ple K M sy ste m —one with few rules.
A one-on-one interactive relationship betw een a single know ledge developer
and a single expert prom otes close ties quickly and speeds the knowledge cap-
ture process.
2. A p ro b le m in a restricted do m a in calls f o r a single expert. For exam ple, D avid
Smith, G eneral E lectric’s locom otive expert, knew the problem s of diesel engines
and how to best repair them . His expertise m ade him the best choice for solving a
particularly difficult problem in a restricted domain.
3. A single expert facilitates the logistics asp ect o f co o rd in a tin g arran gem en ts fo r
know ledge capture. O nce an expert agrees to do the work, one can m ore easily
arran g e m eetin g tim es and schedules th an if m ultiple ex perts are involved.
P ro to ty p in g also m oves quickly, because only one expert needs to review the
work and vouch for its readiness.
4. With a single expert, p ro b le m -rela ted o r p erso n a l conflicts are easier to resolve.
C om m unications are usually easier, and the approach to know ledge capture is
m ore consistent w hen only one expert is involved.
5. Single experts tend to share m ore confidentiality with project-related inform ation
than do m ultiple experts. P art of hum an nature is sharing inform ation: A s m ore
people are party to a session, m ore opportunities arise for th a t inform ation to
leak. O f course, m uch depends on both the n ature of the project and the stan-
dards and integrity of the experts.
II Knowledge Creation and Capture

As attractive as the single expert approach is, several draw backs are w orth keep-
ing in mind:
1. Som etim es the expert’s kn ow ledge is n ot easy to capture. If the expert has difficulty
explaining or com m unicating procedures, the project can die quickly, especially if
no backup expertise is available.
2. Single experts p ro v id e a single line o f reasoning, which m akes it difficult to evo k e
in-depth discussion o f the dom ain. If som eone sees a doctor and learns that he or
she has a serious disease, that person would certainly appreciate a second opin-
ion. M ultiple experts are m ore likely to provide a built-in system of checks and
balances (see Box 5.5).
3. Single experts are m ore likely to change scheduled m eetings than experts who are
p a rt o f a team. U nfortunately, people who m ake the best experts are often the
least accessible. O th er projects may limit an ex p ert’s com m itm ent to a new one.
As a result, a reasonable deadline for com pleting the project may be uncertain.
4. E xpert kn ow ledge is som etim es dispersed. Relying on a single expert, especially in
a com plex system, can either create blind spots or result in a system that will have
no users. C onsider, for exam ple, the area of m edicine, in which personnel are
trained separately for radiology, gynecology, and internal medicine. M ost medical
KM system s n eed to draw upon com m unity know ledge bases and in teg rate
expertise from different areas to be successful. No single expert can realistically
address m ore than one area.

PROS AND CO N S OF USING MULTIPLE EXPERTS

C apturing the tacit know ledge of a team of experts has its own advantages and draw -
backs. The advantages include the following:
1. Com plex problem dom ains benefit from the expertise of m ore than one expert,
especially if know ledge is dispersed and a num ber of “p artial” experts are avail-
able. A pool of experts is m ost often needed to capture sufficient reliable know l-
edge to build com plex KM systems. For example, building a KM system to predict
the next direction of a given stock on the New York Stock Exchange requires the
talent of a pool of highly seasoned traders. No single trader could supply a reli-
able answ er to all aspects of this problem .
2. W orking with m ultiple experts stim ulates interaction that often produces synthe-
sis of experience. Bringing in a num ber of experts is bound to enrich the quality
of know ledge captured.
3. Listening to a variety of views allows the knowledge developer to consider alter-
native ways of representing the knowledge. Two heads are usually b ette r than
one, provided they agree to work together tow ard the goals of the project.
4. Form al m eetings are frequently a better environm ent for generating thoughtful
contributions.
Possible draw backs of m ultiple experts are as follows:
1. Coordinating meeting schedules fo r three or m ore experts can be a challenge. Because
an expert is usually in demand, each expert on a team will probably have to compro-
mise in order for group meetings to be arranged. This area is the one in which inter-
personal abilities of the knowledge developer can make or break the prdject.
2. D isa g reem en ts freq u en tly o ccu r a m o n g experts. The g re a te r the num ber of
experts on a panel who m ust agree on a procedure or a solution, the higher the
probability of disagreem ent. If the diversity of opinions does not resolve into a
 CHA PTER 5 Capturing Tacit Knowledge aaaa 155

fl fl 1 9 fl i BOX 5.5

THE ADMISSIONS KM SYSTEM

O ne person was selected as the dom ain expert for and enthusiastic. This allow ed the know ledge
an admissions KM system. The expert is a practic- d ev elo p er to conduct efficient and m eaningful
ing adm issions director of a m ajor university. She interview sessions.
was chosen because she is the sole decision m aker The expert was very patient with the continu-
in the adm issions selection process. Plus, the ing questioning. As a by-product of this work, she
process is unique, m aking o th e r ex p e rts’ know l- has becom e interested in KM systems as a m echa-
edge invalid for this specific problem dom ain. The nism for training new personnel in the admissions
expert also show ed enth u siasm for th e p ro ject process. D espite her cooperation, however, some
and willingness to com m it her tim e and expertise. problem s came up particular to the use of a single
T hroughout the know ledge cap tu re process, expert. O ne difficulty was bias. The know ledge
the director was extrem ely cooperative. She had base will only reflect inputs from her. W ithout a
excellent com m unication skills and conveyed secondary source of expertise, the expert is both
in form atio n in a clear and d escriptive m anner. the supplier of know ledge and final authority who
T hroughout the project, she rem ained interested validates the KM system.

a i 9 9 9 9

consensus, the success of the project could be in jeopardy. D isparity am ong the
experts can also cause the least-senior expert to stifle, edit, or com prom ise an
opinion. The know ledge developer must be sensitive to this issue (see Box 5.6).
3. The greater the num ber o f p eo p le in vo lved , the harder it m ay be to retain confiden-
tiality. As m entioned earlier, both the individuals involved and the nature of the
problem dom ain determ ine the level of confidentiality needed.

9 9 9 9 9 9 BOX 5.6

MULTIPLE EXPERT AGGREGATIONS

In tacit knowledge capture, it is often desirable to are then averaged to establish perform ance qual-
aggregate the judgm ent of multiple experts into a ity of the KM system . A n im p o rtan t p art of the
single KM system . In som e cases, this takes the com parison process should be determ ining if the
form of averaging the judgm ent of those experts. experts have a similar view of the world. If they do
In these situations, it is desirable to determ ine if not, their evaluations of perform ance may differ,
the experts have d ifferent views of the w orld resulting in a m eaningless average perform ance
before their individual judgm ents are aggregated. measure. Alternatively, if all the validating experts
In validation, multiple experts often are em ployed do have similar views, then the validation process
to com pare the perfo rm an ce of ex p ert system s may result in a paradigm myopia.
and o th er hum an actors. O ften those judgm ents

SO U RC E: Awad, E. M. Building Expert Systems. Minneapolis, MN: West Publishing, 1996, p. 168.
156 7777 PART II Knowledge Creation and Capture

TABLE 5.1 Advantages and Drawbacks of Using Single and Multiple Experts
_______________Single Expert_________________________________ M ultiple Experts
Advantages Advantages
1. Ideal for building a simple domain KM 1. Works best for complex problem domains
system 2. Stimulates interaction and synthesis of
2. Works well for problems in a restricted experience
domain 3. Presents a variety of views allowing
3. Simplifies logistical issues for alternative representations of tacit
4. Increases the probability of resolving knowledge
conflicts 4. Generates more thoughtful
5. Decreases risks to confidentiality contributions

Disadvantages Disadvantages
1. Not easy to capture an expert’s tacit 1. Creates scheduling difficulties
knowledge 2. Increases the probability of disagreement
2. Provides only a single line of reasoning 3. Raises confidentiality issues
3. Increases the chance of schedule 4. Requires more than one knowledge
changes developer
4. Does not accommodate dispersed 5. Leads to a “process loss” in determining
knowledge a solution

..................................................... .. .......... m m m m mi m

4. W orking with m u ltiple experts often requires m o re than one kn o w led g e d e v e l-

oper. W ith two interview ers, one can m aintain the topic agenda and the record-
ing while the o th er asks questions and gives feedback to prom ote discussion or
clarify points. B ecause com m itting two know ledge developers to a single job
can be costly and tim e-consum ing, this practice is usually reserved for complex,
priority projects.
5. The overlapping m ental processes o f m ultiple experts can lead to a “process loss .”
W hen two or m ore experts get together, the nature of the discussion tends to over-
lap, and the problem -solving process can be colored by other experts’ opinions.
Sometimes, determ ining which expert is the closest to the solution is difficult.
A sum m ary of the pros and cons of single and m ultiple experts is shown in Table 5.1.

■ Developing a Relationship with Experts

CREATING THE RIGHT IMPRESSION
W orking with experts can be a com plicated business. P erception plays an im portant
role in the early phase of the building process. For example, how does the knowledge
developer know w hether an e x p ert—whose knowledge is sought, but is not paid fo r—
will cooperate fully and reliably? From the ex p ert’s view, the knowledge developer is
building a KM system in an area about which she or he knows little. The expert often
has a poor perception of w hat the know ledge developer can do. To reverse this percep-
tion, the know ledge developer m ust learn quickly and use psychology, com m on sense,
m arketing, and technical skills to gain the e x p ert’s attention and respect.
 CH A PTER 5 Capturing Tacit Knowledge aaaa 157
R elated to this point is underestim ating the expert’s experience. A seasoned expert
will explain things simply and clearly. This simplicity can be deceptive, however, and cre-
ate a false im pression that the problem is easy to solve when in reality the expert has
barely scratched the surface. To quote E d Koch, form er m ayor of New York City, “For
every com plex problem , there is a solution which is simple, economical, and wrong.” For
every simple thing the expert says, the details behind it are likely to be m ore complex. A
knowledge developer m ust keep exploring until the details are understood.

UNDERSTANDING THE EXPERT’S STYLE

E xperts express th eir views in a variety of ways th at require special understanding.
K now ledge developers should acclim ate them selves to the e x p e rt’s style during the
first session. E xperts generally use one of four styles of expression:
1. T he p ro c e d u re ty p e of ex p ert is verbal, logical, and p ro ce d u re -o rie n te d . This
expert shows a m ethodical approach to the solution, with em phasis on structure,
som etim es at the sacrifice of content. The know ledge developer does not have to
encourage the expert to talk but should focus the e x p e rt’s explanations on the
pro b lem at han d , staying w ithin th e guidelines of the pro b lem dom ain.
O therw ise, m ost of th e know ledge d e v e lo p e r’s energy will go tow ard m aking
sense out of the e x p e rt’s details.
2. The storyteller type of expert focuses on the content of the dom ain at the expense
of the solution. This expressive expert probably cam e up through the ranks and
received little atten tio n or recognition. Such a person is usually willing to repeat
a story or a d escrip tio n , b u t will o ften go off on a tan g e n t, w ander from one
aspect of the dom ain to the next, or bounce betw een topics w ithout noticing. The
know ledge developer m ust use structured interviews with this type of expert.
3. The godfather type of expert, by natu re of position and personality, has a com pul-
sion to take over. If allowed, this expert will wrest control from the knowledge
developer. N o single prescription works every tim e for this situation, which will
often w orsen if the know ledge developer attem pts to right the balance of pow er
in the w rong way. In fact, th e w orst scenario occurs w hen a godfather expert is
paired with a m anipulative know ledge developer.
4. The salesperson type of expert spends m ost of the know ledge d ev eloper’s time
“dancing” around the topic, explaining why the ex p ert’s solution is the best. The
know ledge developer can becom e distracted by this e x p ert’s habit of answering a
question with an o th er question. The know ledge developer’s role is to redirect the
conversation and encourage straight answers.

PREPARING FOR THE SESSION

It should be obvious th at before m aking the first appointm ent, the know ledge devel-
oper should know som ething about both the expert (such as personality, tem peram ent,
job experience, and fam iliarity with the problem ) and the problem .
The initial sessions can be the m ost challenging and the m ost critical. N ot knowing
w hat to expect, th e e x p ert m ay becom e p ro tectiv e w hen the know ledge developer
begins to “p e e l” inform ation from the e x p e rt’s memory. The expert may resent any
aggressive prodding. The b u rd en of building trust falls on the know ledge developer,
because th e responsibility of building the KM system also lies with the know ledge
developer, not the expert.
The know ledge developer should becom e fam iliar with the project term inology
and review existing m aterial, such as a p ap er w ritten by the ex p ert or a book th at
158 77TT PART II Knowledge Creation and Capture

details the n atu re and m akeup of the dom ain. This type of research is also a way to
learn the ex p ert’s language. Can you imagine building a real estate KM system w ithout
know ing term s like p rim e rate, variable interest rate, or balloon m ortgage? A n early
grasp of the basics inspires confidence on the part of the expert and allows the know l-
edge developer and expert to be m ore productive.
A n im p o rtan t beginning step in the first session is personal introductions. The
expert is eager to know w hat type of person the know ledge developer is and what the
goals are for the project. The sooner rapport is established, the easier things will move.
H onesty is im portant. The expert and the know ledge developer need to understand
each o th e r’s experience related to the problem at hand. Hum ility avoids a situation in
which constant questions may appear arrogant.

DECIDING WHERE TO HOLD THE SESSIONS

Location is a m atter of preference, but protocol calls for the expert to decide on the
m eeting site. B ecause a series of sessions take place over an extended tim e period,
m ost experts opt for the convenience of m eeting on their own turf. The know ledge
developer may also find benefit in recording the ex p ert’s know ledge in the environ-
m ent w here he or she works. An expert may feel m ore com fortable in having any nec-
essary tools or inform ation within arm ’s reach. M oreover, the know ledge developer’s
willingness to travel to m eet with the expert should generate goodwill for the project.
A n im portant guideline is to m ake sure the m eeting place is quiet and free of in terru p -
tions, such as phones ringing or “urgent” messages.

APPROACHING MULTIPLE EXPERTS

O ne of the challenges in tacit knowledge capture am ong m ultiple experts is deciding
the ap propriate approach. Three approaches are most often used.

The Individual Approach

The know ledge developer holds sessions with one expert at a time, making it eas-
ier for the ex p ert to offer inform ation freely while assuring privacy. The individual
approach is also useful in clarifying knowledge or defusing interpersonal problem s that
crop up when several experts are gathered in one room .

The Approach Using Primary and Secondary Experts

The know ledge d ev elo p er may decide to hold a session with the senior expert
early in the know ledge capture program for clarification of the plan. For a detailed
probing of the do m ain ’s m echanics, he or she may tap other ex perts’ knowledge. In this
approach, the prim ary expert verifies solutions in the event of diverse responses or
conflicting opinions of secondary experts.
A n o th er way of w orking with prim ary and secondary experts is to gather inform a-
tion from the secondary experts first and then consult the prim ary expert to check the
reliability of the tapped knowledge. This bottom -up approach is useful when the pri-
m ary e x p ert is in d em an d and cannot devote full-tim e to the project. The prim ary
ex p ert’s energies can be saved for the crucial steps as needed.

The Small Groups Approach

E xperts are gathered in one place to discuss the problem dom ain and provide a
pool of inform ation. Their responses are m onitored at the sam e time, and the function-
ality of each expert is tested against the expertise of others in the group. The knowl-
 CH A PTER 5 Capturing Tacit Knowledge a a a 1 159
edge developer can decide which expert is m ore cooperative or m ore believable. This
approach requires experience in assessing tapped knowledge, as well as perception and
cognition skills.
O ne of the issues th at the know ledge developer m ust deal with is pow er and its
effect on the e x p ert’s opinions. For instance, how does a “ju n io r” expert openly contra-
dict the senior expert? P rotocol may well require the junior expert to let com m ents go
uncontested, which defeats the purpose of having experts share their knowledge.
It should be noted that group perform ance is closely related to group size. As the
size of a group increases arithm etically, one can expect a geom etric increase in the total
num ber of relationships. Using the form ula n (2 n/ 2 + n - 1), for a group size of n = 3,
the total num ber of relationships is equal to 18. W hen group size is increased to n = 4,
the total num ber of relationships increases to 44. For this reason, the num ber of experts
m ust be lim ited to the know ledge developer’s capacity to handle the additional in ter-
action and feedback. A n increase in group size may also bring out group conflict rather
than ready consensus.

««■« Fuzzy Reasoning and the Quality

of Knowledge Capture
P eo p le reaso n w ith w ords, n o t num bers, which can lead to a p roblem com m only
encountered in interviewing: E xperts are not always easy to understand. Because the
problem dom ain is likely to be technical or complex, decision m aking often involves
rules with m any shades of m eaning that may be fuzzy to the nonexpert. For example,
one can easily understand the following simple rule:

In the U nited States, if yo u are under 60 years o f age, yo u are n ot entitled to

social security benefits.

A retirem ent benefit rule m ight be m ore complex, although still understandable.
For example:

I f y o u are at least 60 years o ld an d have been a state em p lo yee f o r at least

25 years o r at least 62 years o f age an d h ave w o rk ed fu ll-tim e f o r m o re than
5 contiguous years , then yo u are entitled to collect social security benefits p r o -
vided yo u are n ot h andicapped o r yo u are n ot receiving a salaried incom e greater
than $10,000 or collecting u n em ploym en t com pensation fro m a state agency.

People use IF . .. T H E N statem ents in conversations, but problem s that require

expertise are not always straightforw ard, and the solutions cannot be stated in simple
IF . . . T H E N rules. The words used to express solutions may also have m any meanings.
For exam ple, consider the w ord warm and the criteria used to characterize w eather as
warm. The m eaning of warm depends on the tim e of year, tem perature, and location.
W hat is w arm to a person from the m id-A tlantic coast can be judged as hot by som e-
one from th e N o rth o r even called chilly by som eone from the D eep South.
F urtherm ore, a day th at is w arm in w inter can be colder than a day labeled as cold in
the sum m er. B ecause no definite rules distinguish w hat is cold and w hat is hot, this
range of tem p eratu re can be described as fuzzy.
Take an o th er example. W hen an expert com m ents, “The procedure I just explained
w orks m ost of the tim e,” w hat does he or she m ean by “m o st”? Is it any num ber of
tim es over 50 percent? Is it about 75 percent or perhaps m ore?
160 7777 PART II Knowledge Creation and Capture

H ere is a third example. Suppose a heart surgeon said,

I d o n ’t expect heart trouble in a 20-year-old patient. The occurrence is so rare
I w ou ld say it is nil. It isn ’t worth subjecting a p erson that yo u n g to a rigorous
h eart stress test. P atients that you n g, I m o st lik ely xbon’t d o the tests; o ld er
patients Vd p ro b a b ly consider it.

Try pressing the surgeon to explain this sta te m e n t and see how far you get.
B ecause the English language does not always distinguish clearly betw een words or
the m eaning of certain words, the interpretation of words often depends on their con-
text. This fuzziness increases the difficulty of translating w hat the expert has said into
rules that m ake sense.

ANALOGIES AND UNCERTAINTIES IN INFORMATION

W hen experts explain events, they use analogies—com paring a problem to a similar
one encountered in a different setting m onths or years a g o —to m ake a judgm ent. For
exam ple, a professor, after listening to a student in a new course, m ight say,
This student rem inds m e o f a student I h ad in the sam e course last year. She
d id great work. M ost p ro b a b ly this one will d o ju st as well.

This p attern m atching is an o th er fuzzy com parison that m ust be clarified before
the inform ation acquired can be usable. A know ledge developer who runs into this line
of reasoning m ust learn m ore about the ex p ert’s exam ples and pay attention to how
the expert forms analogies, or p attern matching.
In addition to analogy problems, a knowledge developer may also encounter uncer-
tainties in information. A n expert’s knowledge or expertise is the ability to take uncertain
information and use a plausible line of reasoning to clarify the fuzzy details. O ne aspect of
uncertainty is belief. Belief describes the level of credibility. People generally use all kinds
of words to express belief; these words are usually paired with qualifiers such as very,
highly, or extremely. For example, words such as possible, likely, and definite show rela-
tionships between words that express belief. To handle uncertainty expressed in words, the
knowledge developer must draw out such relationships during tacit knowledge capture.

UNDERSTANDING EXPERIENCE
Knowledge developers can benefit imm ensely from an educational grounding in cog-
nitive psychology. The hum an expert usually follows the same basic reasoning princi-
ples as a com puting system. To solve a problem , the expert acquires inform ation from
the outside world and stores it in the brain. W hen a question is asked, the expert operates
on certain stored inform ation through inductive, deductive, or o ther problem -solving
m ethods. The resulting answ er is the culm ination of processing stored inform ation.
W hen know ledge developers ask an expert how he or she arrived at an answer,
they are asking the expert to relive an experience. The right question will evoke the
m em ory of ex periences th a t previously p ro d u ced good solutions. H ow quickly the
expert responds to a question depends on the clarity of the content, w hether the con-
tent was recently used, and how well the expert understood the question.
Visual im agery is often re p resen te d by phrases the expert uses in a descriptive
answer. For example, “som ething tells me,” “sounds O.K. to me,” “I heard that,” “hmm,
let me see —one m ore tim e,” and “the m anual illustrated” are phrases that show the
expert is m entally reacting to an image. W hen the expert uses this type of phrase, the
know ledge dev elo p er should ask, “W hen you say som ething tells you, w hat do you
m ean or w hat do you see?” or “W hy did you say it sounds O.K. to you?”
 CHA PTER 5 Capturing Tacit Knowledge a a ■a 161
A t times, the e x p e rt’s em otions are involved in the problem -solving process. For
exam ple, phrases like “I have a feeling,” umy gut feeling is,” and “my sixth sense tells
m e” serve as the conscious representation of sensory m em ories that have gone dorm ant
for lack of use. The know ledge developer can clarify the content behind these phrases
by asking, “H ow would you describe this feeling?” or “W hat m akes you feel this w ay?”
P eople w ho have difficulty with visual im agery usually have tro u b le answ ering
questions. W hen the know ledge developer does not understand an explanation, he or
she should ask the expert to go through it again and pay closer attention to the content
of the answ er as well as any clues to visual imagery. A know ledge developer may need
to step back to b e tte r u n d e rsta n d how th e e x p ert thinks th ro u g h a problem . A s a
know ledge capture tool, interview ing allows rep eated questioning for knowledge cap-
ture and verification. Interview ing principles are explained later in the chapter.

THE LANGUAGE PROBLEM

The m ain concern during know ledge capture is to be able to understand and interpret
the e x p e rt’s verbal description of inform ation, heuristics, and so on. R ep eated ques-
tioning often clarifies and prom otes com plete answers. B ut how well experts represent
in te rn a l processes varies w ith th e ir com m and of the language and the know ledge
developer’s interview ing skills. For exam ple, take a look at the preceding sentence and
review the structure of the sentence: “H ow well” (how well who? W hat?), “experts rep -
resen t” (do they present? Sum m arize?), “internal processes” (w hat internal processes?
W hat does it m ean ?). The sen tence was w ritten to convey an idea in w ords w hose
m eanings the w riter hopes the reader understands.
Language may be unclear in a num ber of ways:
• C om parative w ords such as taller o r b etter som etim es are left hanging. For ex-
ample, the expert m ight say, “This last solution works b etter for w hat you’re look-
ing for.” The question is, b etter than w hat? The com parison requires clarification.
The know ledge developer might ask, “B etter than w h at?” or “Can you elaborate
on that com parison?” or “ I d o n ’t understand what you m ean by better
• Certain w ords o r com ponents m ay be left ou t o f an explanation. E xperts som e-
tim es think that certain things are so obvious that they do not need to spoon-feed
every detail. A know ledge developer who is not as fam iliar with the dom ain as
the expert assum es may be lost w ithout these details. The know ledge developer
m ust ask the expert to clarify and to elaborate. If the expert com m ents, “To me,
th at is the key point,” the know ledge developer might reply “ That refers to . . . ?”
or “W hat m akes it a key p o in t?” or “How do you know it is the key?”
• A b so lu te w ords an d ph rases m ay be used loosely. For exam ple, be wary when you
hear “always,” “never,” “cannot,” or “no way.” The expert might say, “This
approach will never give you the right com bination,” “T here is no way I can
explain this procedure in 1 hour,” “Your suggestion will never w ork,” or “This
procedure cannot be applied.” The know ledge developer might probe by asking,
“H ow does this approach m ake that com bination im possible?” “Can you imagine
a situation w here it m ight not clear things u p ?” “Then, can you briefly explain the
p ro ced u re?” or “W hat do you m ean this procedure cannot be applied?” These
questions will gently push the expert to reconsider.
• Som e w ords seem to have a built-in am biguity. For exam ple, “The titanium shield
of the missile perform s a protective function during reentry.” The question here is
“W hat is a protective fu nction?” or “protective in reference to w hat?” The know l-
edge developer should either rephrase the question or ask for further clarification.
162 7777 PART II Knowledge Creation and Capture

The main goal of pointing out these am biguities is to w arn the knowledge devel-
o p er to be alert to m eanings as well as to verbal and nonverbal cues. The success of
the sub seq u en t steps in building a KM system is d ep en d en t on accurate know ledge
capture.

■ The Interview As a Tool

The interview is a tool th at is used com m only in the early stages of tacit knowledge
capture. It is the oldest and m ost often used (and abused) tool for capturing or verify-
ing tacit know ledge. B ecause the know ledge c a p tu re d th ro u g h an interview is so
im p o rtan t to KM system developm ent, validity is a critical issue. Validity of a given
question m eans that it is logically correct and is understood the same way by different
people. It also provides consistency in the way questions are asked. The know ledge
d ev elo p er needs to m ake special efforts to elim inate in terv iew er bias th at w ould
reduce validity. O ften the validity of inform ation is m ore reliable if it is freely given.
For this reason, the voluntary n ature of the interview is im portant; it should be a rela-
tionship freely and willingly en tered into by the expert. If the interview and participa-
tion in the know ledge capture process are requirem ents, the expert m ight be giving
inform ation with no assurance of accuracy or reliability, which is the m easure of tru th -
fulness or credibility.
In an interview, the know ledge developer has an opportunity to verify inform ation
and observe the e x p ert’s thought process in action. A m ajor benefit of interviewing is
behavioral analysis. The expert is asked to act out a case or a scenario in the domain.
W hile the e x p ert goes th ro u g h the m otions, the know ledge dev elo p er records the
observations and assesses body language, facial m ovem ents, and other com m unication
cues. A know ledge developer who is ill trained can create or fail to elim inate problem s
with the resulting know ledge base.
The interview as a tacit know ledge capture tool has four prim ary advantages:

1. Its flexibility m akes it a superior tool for exploring areas about which not m uch is
know n concerning w hat questions to ask or how to form ulate questions.
2. It offers a b e tte r oppo rtu n ity than any o th er tool for evaluating the validity of
inform ation acquired. The know ledge developer cannot only hear w hat is said
but also how it is said.
3. It is an effective technique for eliciting inform ation about complex subjects and
for probing an individual’s sentim ents underlying expressed opinions.
4. M any p e o p le en joy being in tervie w e d , regardless of the subject. They usually
cooperate when all they have to do is talk.

The m ajor draw back to interview ing, how ever, is the cost involved in extensive
p rep aratio n time. Interview s also take a lot of time to conduct. W henever a m ore eco-
nom ical way of capturing the same knowledge is available, the interview is generally
not used.
E ven th ough interview ing experts is an art learn ed at school, m ost know ledge
developers develop expertise through experience. Knowledge developers need to pay
attention to the prim ary requirem ents for a successful interview: the developm ent of a
sound, structured strategy and the creation of a friendly, nonthreatening atm osphere
th at puts the expert at ease. A n experienced interview er elicits credible answers that
the expert can offer with no fear of criticism.
 CH A PTER 5 Capturing Tacit Knowledge ■■« a 163
TYPES OF INTERVIEWS
Interview s vary widely in form and structure. From a know ledge capture perspective,
interview s range from the highly unstructured (neither the questions nor their responses
are specified in advance) to th e highly stru c tu re d (th e q u estions and responses are
definitive). O f course, a great deal of variation is possible within this range.
The u n stru c tu re d in terview is used w hen the know ledge d ev elo p er w ants to
explore an issue. It allows experts to answ er spontaneous questions openly (see Fig-
ure 5.3). U n structured techniques are not easy to conduct, and this kind of interview is
difficult to plan. A n expert may in terp ret the lack of structure as perm ission to avoid
preparing for an interview.
T he stru ctu red interview is used w hen the know ledge developer w ants specific
inform ation. It is goal-oriented. In m ultiple expert know ledge capture, the questions
should be p resented with the sam e wording and in the sam e order to each expert. For
exam ple, the question to one expert, “W ould you like to see a prototype of what you
gave m e during the last session?” will not elicit the same response as asking, “How do
you feel about showing the m aterial you gave me during the last session in the form at
of a p ro to ty p e ? ” Standardized questions can im prove the validity of the responses by
ensuring th at all experts are responding to the sam e question.
A stru ctu red question is expressed as a set of alternatives. Several variations of
structured questions can be used:

1. Multiple-choice questions offer specific choices, faster tabulation, and less bias
due to the way the answers are ordered. Because m ost people have a tendency
to favor the first choice item , alternating the o rder of the choices may reduce
bias; how ever, additional time is needed to pick the choice item question (see
Figure 5.4).
2. Dichotomous (yes/no) questions are a special type of m ultiple-choice question.
They offer two answers (see Figure 5.5). The sequence of questions and content is
also im portant.
3. Ranking scale questions ask the expert to arrange items in a list in order of their im-
portance or preference. In Figure 5.6, the question asks the expert to rank five state-
m ents on the basis of how descriptive the statem ent is of the expert’s current job.

Finally, with the sem istructured technique, the know ledge developer asks p red e-
fined questions but allows the expert some freedom in expressing the answers. H ere is
a dialogue from the D iabetic Foot KM System:

K n ow ledge developer: W hat would be your first question to the patient?

D iabetic specialist: W here is the pain?

FIGURE 5.3: Examples of Unstructured (Open-Ended) Questions

From a Publisher KM System:
• Suppose you have the manuscript ready for production. What does the author do
between then and the final binding of the book?
From a Diabetic Foot KM System:
• How can you tell if a diabetic patient has charcot?
From a Loan Approval KM System:
• How do you deal with excessive bad loans after a grace period?
 From Auto Loan KM System:
• If the customer made a final offer within $100 from your minimum to sell a $10,000
car, what do you do next?
____ Talk the matter over with the sales manager
____ Split the difference with the customer
____ Walk away from the deal
____ Throw in amenities

From Diabetic Foot Knowledge Automation System:

• If a diabetic patient complains about foot problems, who should he or she see first?
Podiatrist
____ General practitioner
____ Orthopedic surgeon
____ Physical therapist

From Loan Evaluator KM System:

• Which type of loan brings the highest revenue to your bank?
____ Auto loan
____ Personal loan
____ Mortgage loan
____ Collateral loan

I fl i 9 fl fl FIGURE 5.4: Examples of Multiple-Choice Questions

From Publisher KM System:

• D oes your book contract prohibit you from publishing on the same subject with
another publisher?
____ Yes ____ No

From Real Estate KM System:

• D oes the seller or the buyer pay your commission?
____ Seller _____Buyer

From Diabetic Foot KM System:

• D o patients with neuropathy come for regular checkups?
Yes ____ N o

fl fl fl fl fl fl FIGURE 5.5: Examples of Dichotomous Questions

FIGURE 5.6: An Example of a Ranking Scale Question

Please rank the five statements in each group on the basis of how well they describe your job
as a loan officer. Give a ranking of “1” to the statement that best describes your job; give a
ranking of “2” to the statement that provides the next best description. Continue to rank all
five statements, using a “5” for the statement that describes your job least well.
In my job, I
____ D ecide on auto loans over $10,000
____ Review each application before authorizing credit checks
____ Interview loan applicants as they come in
____ Cosign all checks sent to local auto dealers
____ Work with local auto dealers to attract more loans

164
 CH A PTER 5 Capturing Tacit Knowledge aaaa 165
K n ow ledge developer: W hat possible answers do you expect?
D iabetic specialist: A nkle, big toe, from kneecap to ankle, or “I ’m not sure.”
K n ow ledge developer: Suppose the answ er is “big toe.” W hat do you do next?
D iabetic specialist: I’d look for coloration, swelling, and possible curving of the toe.
K n ow ledge developer: W hat would be your next question?
The process continues until the know ledge developer has no m ore questions to
ask. The ad v antage over the u n stru ctu red technique is th a t the inform ation is well
structured and the inform ation solicited is reasonably clear. A disadvantage is that this
inform ation may not be com plete in other possible ways.

Guide to a Successful Interview

Interview ing should be approached as logically as program m ing. In tacit knowledge
capture using the interview, several steps are considered:
1. Setting the stage and establishing rapport
2. Phrasing questions
3. Listening closely and avoiding argum ents
4. E valuating session outcom es

SETTING THE STAGE AND ESTABLISHING RAPPORT

This first step is an ice-breaking, relaxed, inform al exchange. The know ledge developer
opens the session by focusing on the purpose' of the m eeting, the role of the expert, and
the agenda for review. D uring stage setting, the know ledge developer closely observes
how the e x p ert cooperates. B oth the content and tone of the e x p e rt’s responses are
evaluated. H ow well the interview goes depends on the know ledge developer’s ability
to adapt to the e x p ert’s style. A n expert may be th e frien d ly type, who m akes everyone
involved feel th e process will be easy team w ork; th e tim id type, w ho needs to be
coaxed to talk; or the residen t expert type, who bom bards the know ledge developer
with opinions disguised as facts.
In one respect, tacit know ledge capture is tim e-consum ing and an intrusion into
th e e x p e rt’s privacy. E v en w hen the p ro ced u re has m an ag em en t su p p o rt and the
expert is on the com pany’s payroll, the process is still lengthy and taxing. The expert
m ay anticipate som e harm from the KM system being developed. It is im portant to
establish rap p o rt with the expert at the beginning and m aintain that rapport through-
out the know ledge capture process. A lthough setting the stage has no. hard and fast
rules to follow, the know ledge developer needs to keep in m ind certain guiding factors:
1. H on esty : D o n o t deliberately m islead the expert a b o u t the p u rp o se o f the project. A
careful, w ell-thought-out briefing should provide honest details regarding the
purpose of the project. The briefing should be consistent w hen dealing with m ore
than one expert.
2. C onfidentiality: A ssu re the expert that no inform ation offered will be shared with
u n au thorized personnel. The sole purpose of the session is to build a KM system
in line with the e x p ert’s thinking rath e r than to disclose specific inform ation to
o th er experts or to u p p er m anagem ent.
3. M o d e sty : A v o id sh o w in g o f f k n o w le d g e o r revea lin g in fo rm a tio n fr o m o th er
experts. Such action can color the objectivity of the approach and discourage the
expert from freely giving inform ation.
PART II Knowledge Creation and Capture

4. Efficiency: R espect the tim e schedule or com m itm en ts o f the expert. The session
should not be an extended social event. The expert may not com plain, but other
people who need the expert might.
5. P rofessionalism : P ro p er dress an d dem ean or are im portant. For example, a con-
servative gray suit is a p p ro p ria te a ttire w hen you are interview ing a banking
expert. A casual outfit may be passable when you are interview ing on the assem -
bly line. If the session takes place in the e x p e rt’s environm ent, the know ledge
developer should be a courteous guest.
6. Respect: B e courteous and a g o o d listener. M ore than any other advice, respect for
the ex p ert is a m ust. In te rru p tin g the ex pert can alter the th o ught process or
break the ex p ert’s train of thought.

PHRASING THE QUESTIONS

E xcept in unstructured interviews, each question should be asked in the same order as
it appears on the interview schedule. A ltering the sequence could affect the com para-
bility of th e sessions w ith th e experts. In stru ctu rin g q u estions for the expert, the
know ledge developer needs to consider the level and sequencing of the questions. A
question may be prim ary or secondary. A prim ary question is usually a lead question
that paves the way for m ore detailed questions. For example, in the D iabetic Foot KM
System, a prim ary question is “W hat is the first thing you do w hen a diabetic patient
com es to you with a sw ollen a n k le ? ” It leads to a series of questions concerning
w hether to advise the diabetic person to seek self-help, medical treatm ent, or surgery.
A secondary question usually probes for m ore details. For example, “So, you find a
blister just under the ankle. W hat does it m ean in term s of procedures or treatm ent of
the sw elling?” This is a secondary question, because the question asks for m ore infor-
m ation about the prim ary question —a swollen ankle. Using secondary questions helps
the know ledge developer probe deeper for clarity of details.
Q uestion sequencing determ ines w hether the interview begins with general open-
ended questions followed by secondary questions and closed questions, or vice versa.
E ith er approach has benefits and drawbacks. The open-to-closed question sequence
helps establish rap p o rt and enables the know ledge developer to assess the quality of
the response' before asking m ore focused questions. This assessm ent often results in
discarding unnecessary questions due to redundancy or overlap [see Figure 5.7(a)].
W hen using the closed-to-open sequence, the know ledge developer begins with a
set of well-focused questions, lending a m ore form al tone to the session. It also allows
th e know ledge d ev elo p er to use o p e n -en d e d questio n s to end the interview . This
approach requires substantial advance preparation [see Figure 5.7(b)].

QUESTION CONSTRUCTION
R eg ard less of q u estio n type, a qu estio n m ust be valid enough to elicit a reliable
response. The know ledge developer m ust focus on question content, wording, and for-
m at. H ere is a'sum m ary checklist:
1. Q uestion content
a. Is the question necessary? Is it a part of o th er questions?
b. D oes the question adequately cover the intended area?
c. D oes the expert(s) have p roper inform ation to answ er the question?
d. Is the question biased in a given direction?
e. Is the q u estio n likely to g en erate e m o tio n al feelings th a t m ight color
responses?
 CHA PTER 5 Capturing Tacit Knowledge aaaa 167
(a) Open-to-Closed Question Approach
Begin

Open-Ended Questions

Secondary Questions

Closed Questions

Acquired
End
Knowledge

Open-Ended Questior

Secondary Questions

Closed Questions

Begin
(b) Closed-to-Open Question Approach

1 FIGURE 5.7: Sequencing of Open-Ended and Closed Questions

2. Q uestion wording
a. Is the question w orded to suit the e x p ert’s background and expertise?
b. C an the question be m isinterpreted? W hat else could it m ean?
c. Is the fram e of reference uniform for all experts?
d. Is the wording biased tow ard a specific answer?
e. H ow clear and direct is the question?
3. Q uestion form at
a. C an the question be fram ed in the form of m ultiple choices (answ ered by a
w ord or two or by a num ber) or with a follow-up free answer?
b. Is the response form easy to use and adequate for the job?
c. Is the answ er to the question likely to be influenced by the preceding ques-
tion? T hat is, is it subject to a contam ination effect ?

THINGS TO AVOID
O ne of the dangers of interview ing is generalizing conclusions on the basis of a few ses-
sions. Transcribing notes into rules often results in verification problem s th at arise later
d u rin g th e testin g of th e KM system . E x p e rts are often ann o y ed w hen they k eep
repeating the sam e rules or p ro cedure or keep repeating the sam e inform ation. The
alternative to n ote taking is taping the session. The m ain draw back of this technique is
the tim e it takes to play back the tape. V ideotaping is not recom m ended, because few
experts are com fortable in fro n t of a cam era. In either case, a know ledge developer
should practice with the equipm ent prior to the session to becom e fam iliar with it and
to m ake sure everything is w orking properly.
A n im portant guideline for a know ledge developer is not to aggressively debate
or in te rru p t th e e x p e rt’s discussions or convert the interview into an interrogation.
168 7777 PART II Knowledge Creation and Capture

Inconsistency should be pointed out in an unobtrusive way. A knowledge developer

should also avoid asking questions that put the dom ain expert on the defensive. For ex-
ample, a knowledge developer who realizes the expert is being inconsistent should say,
“How would this solution relate to an earlier one about the same problem ?" rather than
“Based on the earlier solution I have here in my notes, this solution doesn’t make sense.”
Using technical term s such as data mining systems or KM infrastructure often results
in an unfavorable response; it can also put the developer in a defensive position in having
to explain the technology or system. Explanations such as “This is more like a system
project,” “ It’s a system that captures knowledge offered by people with experience like
yourself” or “ It’s a special system application that uses the expert’s knowledge to give
the user som ething to work with to arrive at the same solution” are better alternatives.
A know ledge d eveloper’s com m and of the session is critical to a successful inter-
view. A loss of control is a sign of overaccom m odation, an overbearing expert, or poor
p reparation for the session. D istributing an agenda in advance, keeping the discussion
on the topic, sum m arizing the e x p ert’s com m ents when appropriate, and limiting the
interview to a reasonable time period should help control the session.
Being in control also m eans that a knowledge developer should avoid getting frus-
trated over am biguous answers when they are least expected. For example, a cool head
is n eeded when an expert answers, “ Because th at's the way it w orks —I d o n ’t know
why, but it works all the tim e.” Explaining how some things work is often m ore difficult
than one might imagine. The knowledge developer should avoid excessive in terp reta-
tion of the e x p ert’s answers.
The converse, of course, is pretending to understand som ething. The knowledge
developer needs to recapitulate what he or she understands, thereby reinforcing the
expert. However, starting from scratch is frustrating for everyone. The expert can help
only by knowing exactly w here the knowledge developer got lost. O ne helpful practice
for keeping everyone on track is to draw pictures to illustrate a point or a procedure,
especially if the expert is a developer or a database designer.
In addition to honesty about understanding the expert's information, the knowledge
developer should avoid promising something that cannot or should not be delivered. For
example, during the first session, a knowledge developer should not say, “A couple of ses-
sions and we should be beyond the halfway m ark” if the project is obviously going to
take several more sessions. As an alternative response, “ I don't know how many sessions
we need, but chances are it will be several sessions before we can wrap things up,” is
much better. Figure 5.8 summarizes the things to avoid in a knowledge capture interview.

RELIABILITY OF INFORMATION FROM EXPERTS

The knowledge captured from the expert is presum ed to correspond accurately to the
actual way so lutions are p ro d u ced in the dom ain. H ow ever, several u n controlled
sources of erro r may reduce the inform ation's reliability:
1. The expert's perceptual slant. P erceptual abilities vary; the rep o rts of a single
event given by several experts often have little resem blance to one another.
2. The expert's failure to remember just what happened. As time passes, the exact
details of an event generally becom e m ore difficult to describe.
3. An expert's fear o f the unknown. Experts often distort descriptions of events for
fear that the KM system will replace him or her. These distortions defeat the p u r-
pose of knowledge capture that is based on m utual trust.
4. C om m unication problem s. Some ex perts have difficulty com m unicating their
know ledge. In a related vein, the know ledge developer is som etim es unable to
 CH A PTER 5 Capturing Tacit Knowledge a a a a 169
1. Never aggressively debate or interrupt the expert’s interpretations or opinions.
2. Avoid asking questions that put the expert on the defensive.
3. Stay away from using technical or special terms just to show off.
4. Try to be specific, rather than general.
5. Avoid losing control of the session.
6. Avoid getting frustrated over ambiguous answers.
7. Avoid bringing up items or details that are not on the agenda.
8. Avoid pretending to understand something.
9. Avoid promising something that cannot or should not be delivered.

FIGURE 5.8: Things to Avoid in a Knowledge Capture Interview

elicit the inform ation th at the expert is qualified to provide. V erbalization conveys
the inform ation th at is available in working memory, also called spoken p ro to c o l
5. R ole bias. The risk of role bias increases when the expert is aware of his or her
im portance as the critical elem ent in building the KM system. The expert might
w onder, “W hat type of person should I be as I answ er this question?” and then
select a role to fit the expectations of the know ledge developer. R esearch has
shown that experts with limited formal schooling tend to exhibit “put o n ” behavior
and are m ore likely to distort the knowledge elicited than highly educated experts.
V arious validation and cross-validation m ethods should be applied before cap-
tu re d know ledge can be re p re sen ted . For exam ple, one way to cross-validate an
ex p ert’s opinions is to ask an o ther expert and check for sim ilarities betw een the two
opinions. A n o th er way to validate an opinion is to ask the question again at the next
session to see if the expert gives the sam e answer.

ERRORS MADE BY THE KNOWLEDGE DEVELOPER

In addition to possible bias originating with experts, know ledge developers can also
contribute to the reliability problem . Validity problem s often stem from w hat is called
the interview er effect: Som ething about the know ledge developer colors the response
of the dom ain expert. Some of the effects include the following:
• The age effect: In the eyes of a much older dom ain expert, a 20-som ething knowl-
edge developer is not viewed as experienced enough to take on the building of a
know ledge m anagem ent system. The wrong perception of age can color responses
of the dom ain expert.
• The race effect: U nfortunately, people can be biased because of race. In a profes-
sional environm ent, biases should not exist, but they do. The best option for earn -
ing an ex p ert’s respect is for a know ledge developer to dem onstrate competency,
integrity, and determ ination to do quality work. If race continues to be a serious
barrier, perhaps an alternative expert or an alternative know ledge developer is
the best course of action.
• The gender effect: The gender of the know ledge developer can alter the way the
interview is conducted. R esearch in the social sciences has shown that male
interview ers obtain few er responses than fem ale interviewers, regardless of race,
and white fem ales obtain the highest num ber of responses from men. The excep-
tion is young w om en who interview young men; they get few er responses than
older women.
170 7777 PART II Knowledge Creation and Capture

The know ledge developer m ust watch for biased responses from the expert early
in the capture process and find ways to correct the tendency.

PROBLEMS ENCOUNTERED DURING THE INTERVIEW

O th e r difficulties are not uncom m on during the interview. They range from biased
responses and the way the questions are phrased to personality clashes betw een the
expert and the know ledge developer.

Response Bias
A response bias occurs when experts answer questions on the basis of their inter-
pretation of the question and in response to certain constraints: lack of time, lack of m oti-
vation, perceived hostility, an attem pt to please the knowledge developer, and so on. A
com bination of wording and tone of voice can also prom ote response bias. For example:

K n o w le d g e D e v e lo p e r: Isn’t it true that when a diabetic comes in with a blister,

you pretty m uch decide that she will be making m ore
than one visit?
D o m a in E x p e rt: Well, as a general rule, I ’d say yes.
K n o w le d g e D e v e lo p e r: D on’t you think that following your verification procedure
results in disapproval of loans for minority applications?
D o m a in E x p e rt: I ’d probably say yes, but we have good reasons for this.

The questioner coaxed an unintended “yes” answer by prefacing the question with
“Isn ’t it tru e ” or “D o n 't you think.”
A n o th er response bias stem s from the order of the questions asked, which is called
the contam ination effect. For example, a question that is asked following another ques-
tion is answ ered one way; if it were asked later or earlier in the questioning sequence,
the answer could be totally different.
O ne test to reveal bias is the circular triad. For example, when the expert for the
D iabetic Foot KM System says, “ I consider vascular disease m ore serious than neu -
ropathy,” and then says, “I consider neuropathy m ore serious than charcot,” and then
states, “ I consider charcot m ore serious than vascular disease,” it is called a circular
triad. This type of contam ination is usually the result of long interview sessions, loss of
m otivation or interest, fatigue, or faking, which provides even m ore reason for pretest-
ing the questions.

Inconsistency
The problem of inconsistency is most likely to occur when the knowledge devel-
oper interviews two dom ain experts and is inconsistent when asking the questions. For
the sake of validity, the questions and their order should be standardized. Validity is
related to reliability; they are two faces of the same coin. Validity is addressed when
one asks, “D oes the question m ean the sam e thing to all of the experts being in ter-
view ed?” If it m eans different things to different people, then the question is not valid.
No invalid questions obtain accurate responses. Reliability, on the other hand, occurs
w hen the question elicits the sam e response at different times. The reliability of an
ex p ert’s answ er can be revealed by answering the question, “H ow much credence can
we place in this answ er?” If the question is not valid, one can assume that the answer is
not reliable. Knowledge developers often rephrase questions and ask a variety of ques-
tions about the same process to ensure the reliability of the inform ation captured.
 CH A PTER 5 Capturing Tacit Knowledge aaaa 171
Communication Difficulties
N ot everyone has a knack for explaining things. The know ledge developer may
need to resort to analogies or o ther tools to stim ulate the ex p ert’s thought process. For
example, the know ledge developer may say, “If I understand you correctly, you would
require the application to be a hom eow ner and have an average income of $50,000 to
qualify for a $30,000 car lo an ?” or simply ask the expert “I am not sure I follow what
you just explained. Can you go through it one m ore tim e?”
C om m u n icatio n p ro b lem s can also stem from the know ledge developer, who
may not have been listening or did not understand. H ow ever, when a knowledge devel-
o p er adm its to these lapses, m ost experts do not m ind rep eatin g or simplifying the
poin ts ju st explained. N ext tim e, the know ledge d ev elo p er had b e tte r get it right,
because pretending to u nderstand is a destructive habit. A sking the expert to repeat
the sam e thing three tim es could cause tem pers to flare and th reaten the end of the
whole process.

Hostile Attitude
N othing is m ore problem atic to an interview process than hostility. Various factors
cause hostility: bad chem istry betw een the expert and the know ledge developer, an
ex p ert’s “forced” participation, or time wasted on repeated deadends. Unfortunately,
no quick fixes work for reducing hostility.

Standardized Questions
O n the surface, the task of stan d ardizing q u estions does not a p p e a r difficult.
H ow ever, even if the wording is the same, the way the question is asked, the tone of
voice, and the facial expression can elicit a different response.

Length of Questions and the Duration of the Interview

No w ritten rules provide the ideal n um ber of questions or m inutes in an in ter-
view. M uch depends on the preference of the expert, the expectations of the know l-
edge develo p er, and the n a tu re of the co n strain ts su rro u n d in g the w hole project.
Ideally, each interview should last no m ore than 1 hour. A ny longer, and the partici-
p a n ts’ atte n tio n spans begin to break dow n, and the quality of the thought process
degenerates.

ENDING THE INTERVIEW

E nding an interview requ ires sensitivity to the dom ain e x p e rt’s preferences, use of
p ro p er verbal and nonverbal cues, and efforts to m aintain ties with the expert even
a fter the interview is finished. W hen the know ledge d ev elo p er has elicited all the
know ledge needed, com m on tendency is to cut the session abruptly. This action can
easily be m isinterpreted. The b etter practice is to carefully plan the end of the session.
O ne com m on procedure calls for the know ledge developer to halt the questioning
a few m inutes before the scheduled ending and to sum m arize the key points of the ses-
sion. This practice allows the expert to com m ent and schedule a future session. If the
procedure is carried out naturally, the expert will feel good about time well spent, and
the next session is m ore likely to flow just as smoothly.
Several verbal and nonverbal cues can be used to close an interview. Table 5.2 sum -
m arizes some of the verbal closing cues that have been found to be successful in ending
a know ledge capture session. N onverbal closing cues that may be used with verbal cues
to end a session are listed in Figure 5.9.
172 7 7 7 7 PART II Knowledge Creation and Capture

TABLE 5.2 Verbal Closing Cues for Ending an Interview

_________________ Verbal Cue______________________________________ Meaning
1. This is my summary of the session. D o you have I’m ready to end the session unless
any questions? you have questions.
2. I think I’ve asked all the questions I had in I’m ready to call it a day.
mind. I appreciate your time and input.
3. My time allowance is up. I know you have Time’s up.
another meeting soon.
4. I learned a great deal from this session. I don’t need any more information.
I appreciate your support. The session is over.
5. This covers pretty much what I had in mind. I am ready to end the session.
Did I miss anything?
6. This turned out to be an informative meeting. Session is over. I’d like to schedule
How about scheduling another meeting another one.
sometime next week?
------------------------------------------1 i ■ i ■ a ------------------------------------------

ISSUES TO ASSESS
D uring the interview, several issues may arise. To be prepared for the m ost im portant
issues, a know ledge developer should consider the following questions:
• H ow would one elicit knowledge from experts who cannot say what they mean or
mean what they say? Sometimes, the expert is know ledgeable but has difficulty
explaining things. W hether this difficulty stems from lack of know ledge usually
becom es apparen t during the first m eeting. Unless it is a one-tim e blunder on the
part of the expert, an alternative expert should be considered immediately.
• What does one say or do when the expert says, “Look, I work with shades o f gray
reasoning. I sim ply look at the problem and decide. D o n ’t ask me why or how. ”
U nless a know ledge developer is well prepared to handle unexpected situations,
this type of answ er could very well end the interview. The expert probably knows
the solution so well th at the details are an unconscious part of it. The knowledge
developer m ust w ork around this unconsciousness and look for ways to get at the
details indirectly.
• H ow does one set up the problem domain when one has only a general idea o f
what it should be? O ne rule of thum b is to ask the expert to illustrate or explain—
through scenarios—the general procedure of th.e domain. The knowledge devel-

1 FIGURE 5.9: Nonverbal Cues for Ending an Interview

1. Look at watch and uncross legs.
2. Put cap on pen and uncross legs.
3. Put cap on pen and close folder.
4. If taping session, stop taping and rewind tape.
5. Stop taking notes, uncross legs, put cap on pen, and look at watch (strongest nonverbal cues).
6. Place writing materials in briefcase and close briefcase.
7. Close folder and uncross legs.
8. U se any combination of these cues.
 CH A PTER 5 Capturing Tacit Knowledge aaaa 173
oper should then feed back his or her understanding of w hat the expert said and
w ork from there.
• W hat does one do if the relationship with the dom ain expert turns out to be diffi-
cult? Perhaps the expert is not only uncooperative, but he or she also seems to be
trying to sabotage the project. The expert may react defensively to the knowledge
d eveloper’s queries. The expert may have been told by m anagem ent to participate
in the project. Unless the expert can be coaxed into a m ore positive fram e of
m ind, the resulting sessions could be strained at best. It should be noted that p e r-
sonality differences that surface quickly could work against the capture process.
Personality clashes can be triggered by a num ber of things: a series of jum bled
questions with no end in sight, “grilling” the expert as if on the witness stand, or
rem inding the expert too frequently of prior successes in building expert systems.
If the situation reaches a deadlock, the best solution is to replace the knowledge
developer or scrub the project.
• W hat happens if the expert dislikes the kn ow ledge developer? This issue also
addresses personality clashes. For example, on one project a young knowledge
developer with a Ph.D. was assigned a senior loan officer to capture the decision-
m aking process regarding jum bo loans (loans over $275,000). From the beginning,
the expert was reluctant to cooperate beyond cursory inform ation. Well into the
fourth session, the know ledge developer learned that the expert had not been able
to com plete his doctorate and had difficulty working with younger consultants.
Ways of gaining the respect of an expert vary from person to person. Sometimes,
young know ledge developers do not do well with experts 20 years their senior. In other
cases, the expert may feel his or her tim e is being w asted when talking to a novice. For
these situations, an accom m odating personality, finesse, and advance planning are criti-
cal in reaching the expert. A relationship that does not jell in the first session could
m ake the entire capture process an uphill battle.

■ Rapid Prototyping in Interviews

O ne approach to building KM system s is rapid prototyping, or the build-as-you-go
ap proach, in which know ledge is added with each know ledge cap tu re session. This
increm ental or iterative approach allows the expert to verify the rules as they are built
during the session (see Box 5.7). Some people have likened rapid prototyping to the
m iracle of birth: A n 8-pound baby is the prototype of a 180-pound person, who has
essentially all the sam e features, but with a m ore com plete and m ature configuration.
Each stage of KM system developm ent has its own problem s, requires its own m odifi-
cations, provides its own excitem ent, and results in its own contributions.

BENEFITS AND DRAWBACKS

R apid prototyping can open up com m unication through its dem onstration of the KM
system. B ecause of its instant feedback and m odification, it also reduces risk of failure.
A n o th er benefit is allowing the knowledge developer to learn every time a change is
m ade in the prototype. Its iterative nature encourages discovery of better ways to build
the KM system.
R apid prototyping depends on the team w ork spirit of the know ledge developer,
the expert, and the end user. It is highly interactive, and its continuous feedback leads
to im provem ents and refinem ents all the way to com pletion, which creates a KM sys-
tem th a t co rresp o n d s as closely as possible to its sta te d req uirem ents. A n obvious
174 7 77 7 PART II Knowledge Creation and Capture

■ a a a a fl b o x 5.7 fl a a fl a a

ROLE OF RAPID PROTOTYPING

In some beliefs, everything began in Genesis. G od step-w ise im provem ents —evolution —turns the
created the world in 6 days and rested on the sev- system into the custom er’s desired paradise.
enth. Everything else occurred through evolution. Can you imagine what would have happened
Genesis is an excellent model for rapid evolutionary if G od had tried to specify the req uirem ents
development, because we want to create a deliver- for this earthship and then built it from scratch
able, functioning prototype in as little time as pos- will all the bells and whistles? It could have taken
sible and begin the endless additions, changes, and 150 m illion years just to get started . Instead,
deletions required to keep the system in balance G od chose to ra p id -p ro to typ e the initial version in
and evolving to meet the needs of the business. 6 days and grow it from there.
R apid ev o lu tio n ary d ev elo p m en t relies on Unlike the traditional developm ent life cycle,
speed, simplicity, and a shared vision to create a speed is req u ired m ore than direction in rapid
desired p ro d u ct. P ro to ty p e rs create a w orking prototyping. O nce you are rolling, you can change
system that does work and provides the essential, course at will. If you are not moving, you have no
initial elem en ts of the system . T hen a series of feedback to guide your first steps.

S O U R C E : Excerpted from Arthur, L. J. Im proving Software Quality: A n Insider's Guide to TQM . New York: John Wiley & Sons,
Inc., 1992.

problem , however, is lack of clear guidelines for rapid prototyping use. A nother draw -
back is th at the p ro to ty p e itself can create user expectations that, in turn, becom e
obstacles to further developm ent efforts. As a result, users refuse to buy into it, and the
transition to an operational KM system never materializes.

• ••• Implications for Knowledge Management

For success in building KM systems, several things must fall into place. First, complex
applications require tim e and talent of the knowledge developer. By the time knowl-
edge capture begins, all the prelim inaries should be established, including the budget
approval, the ex p ert’s availability, and the com pany’s com m itm ent.
Many KM systems are built by using teams. A team includes a cham pion, system
users, an expert, one or m ore knowledge developers, and a scribe. The cham pion is a
respected executive who believes in the project and sells the idea to those who control
the purse strings. Executive support is usually assured if the project prom ises a fair
retu rn on investm ent and the availability of expertise in several locations. A prototype
of the KM system can be built to sell the change early in the life cycle of the product,
which also m akes it easier to train the user and gain their acceptance.
A key m anagerial issue centers around who owns the know ledge of the expert.
Som e o rganizations form contracts with th eir experts, u n d er which the know ledge
gained on the job belongs to the firm and cannot be sold to or shared with the com pe-
tition within a specified tim e period after the expert leaves the firm or retires. In the
absence of an agreem ent, the organization cannot claim knowledge ownership, except
in the case of investm ent in special training that actually prom oted the expertise.
 C H A PTER 5 Capturing Tacit Knowledge aai i 175
Interview ing itself carries several m anagerial implications. First, m anaging know l-
edge cap tu re m eans m anaging the e x p e rt’s tim e, which is precious. The know ledge
developer should m ake the m ost efficient use of the ex p ert’s time in the fewest num ber
of sessions. M anagem ent m ust also provide the right atm osphere and support if the
interview ing process is to succeed, which is actually the responsibility of the champion.
Finally, to w hat extent can the expert provide time for the capture phase? Experts who
are on the com pany payroll often draw high salaries. A dvance planning by the knowl-
edge developer is necessary to m ake the most efficient use of the ex p ert’s time.
In term s of rap id p ro to ty p ing, m anagem ent should be e d u cated ab o u t the p o -
tential of the product well before it is ready to be installed. D em onstrating the p ro to -
type at som e point can be a good “selling” technique. W ithout such occasional feed-
back, funds and su p p o rt m ay dw indle, often resulting in half-baked or incom plete
systems.

Su m m a r y 1 ■ 1 ■

• The goal of tacit know ledge capture is to extract problem -solving knowledge
from the hum an expert in o rder to build a KM system. The scope of the dom ain
should be well defined, the dom ain should contain com m on-sense knowledge,
and the resulting KM system must be a m anageable size.
• E xperts have definite qualifications. They include perceptual ability to discern
the relevant from the irrelevant, com m unication skills, creativity, self-confidence,
and credibility. M otivation, enthusiasm , and willingness to share the expertise
are also im portant. The right expert does not guarantee a successful expert sys-
tem , however.
• T hree levels of experts are emphasized: the highly expert person, the m oderately
expert problem solver, and the new expert. Each level carries its own contribu-
tions and constraints. The best situation occurs when the knowledge developer is
able to w ork with the m ost qualified expert available.
• Using a single expert or m ultiple experts both have pros and cons. Single experts
are ideal for simple and relatively quick KM systems; they entail fewer conflicts
to resolve and m ake logistical arrangem ents easier. The m ain risk is that the
credibility of the expert system depends entirely on the thought processes of one
expert. In contrast, m ultiple experts are typically used in problem dom ains with
diffused knowledge. They provide a synthesis of experience that is critical for the
reliability of the system ’s solutions. H ow ever, scheduling difficulties, process loss,
and the need to involve m ore than one knowledge developer can m ake the use of
m ultiple experts problem atic.
• A com m on problem encountered in interview ing is an expert who cannot easily
convey the know ledge even though he or she may be quite com petent at the spe-
cialty. To the nonexpert, an ex p ert’s p attern m atching or analogy can resem ble
fuzzy reasoning.
• A m biguities of language during interview ing or when interpreting the ex p ert’s
descriptions or explanations is a com m on problem a knowledge developer must
be aw are of and work to avoid.
• A structured interview has several variations: multiple-choice, dichotom ous, and
ranking-scale questions. Each variation has its own form at and goals.
• Interview ing should be approached logically, using stage setting, proper phrasing
of questions, good listening, and evaluating the session outcome. Q uestion
sequencing as well as probing the expert for details are also im portant.
176 iiaa PART II Knowledge Creation and Capture

• Reliability of an e x p ert’s inform ation can be affected by question construction,

which requires a careful review of question content, wording, and form at.
Reliability is also affected by uncontrolled sources of errors: the ex p ert’s percep-
tual slant, failure to rem em ber just w hat happened, fear of the unknow n, com m u-
nication problem s, and role bias. E rrors can also originate with the knowledge
developer; these can include age, race, and gender effects.
• Several problem s may be encountered in interviews: response bias, inconsistency,
com m unication difficulties, hostile attitude, and standardized questions.
• Several issues that may arise during the interview m ust be assessed. For example,
how would one get know ledge from experts who do not say w hat they m ean or
m ean w hat they say? How can one set up a problem dom ain while knowing little
about it? W hat does one do when the relationship with the expert turns out to
be difficult?

T e r ms t o K n o w 1 , 1 1
Age effect: Bias against a person due to age. Ranking scale question: A question that asks the respon-
Analogies: Comparing a problem to a similar one dent to arrange items in a list according to preference
encountered previously. or importance.
Closed question: A question that asks for specific Reliability: Dependability; truthfulness of the response or
responses. answer to a given question; credibility; how well the
Dichotomous question: A question answerable by one of KM system delivers solutions with consistency, accu-
two answers, usually yes or no. racy, or integrity; detecting or removing anomaly.
Gender effect: Bias against a person based on gender. Response bias: Bias resulting from the subjective
Interview: A face-to-face interpersonal situation in which responses of the expert to any given question.
a person called the interviewer asks another person Role bias: An altered attitude resulting from the expert’s
questions designed to elicit certain responses about a awareness of his or her importance in the building of a
problem domain. KM system.
Multiple-choice question: A question that offers the Scribe: A person who takes notes during interviews and
expert specific answer choices. maintains all the documentation related to the project.
Novice: An individual with skills and solutions that work Secondary question: A question used to probe for further
some of the time but not all of the time. details or for follow-up on an area under discussion.
Open-ended question: A question that asks for general Uncertainty: Lack of adequate information to make a
rather than specific responses. decision.
Primary question: A question that elicits the most impor- Unstructured interview: An approach in which the ques-
tant information in one area during the interview; leads tions and the alternative responses are open-ended.
to further questions to obtain pertinent details. Validity: The logical correctness of a question, which is
Race effect: Bias against a person due to race. worded in order to elicit the information sought.

Test Y o u r U n d e r s t a n d in g 1 , 1 1

1. In your own words, define tacit knowledge capture. W hat m akes it unique?
2. A re there any particular steps involved in knowledge capture? Explain briefly.
3. “The problem solution should be based on symbolic know ledge rath er than
num erical com putation.” D o you agree? Explain.
4. How would one identify expertise?
5. W orking with experts requires certain skills and experience. W hat sugges-
tions or advice w ould you give to an inexperienced know ledge developer
concerning:
a. w orking with or approaching an expert
b. preparing for the first session
6. W orking with m ultiple experts has definite benefits and limitations. Cite an
exam ple in which the use of m ultiple experts is a must. Explain your choice.
 CH A PTER 5 Capturing Tacit Knowledge i ■i ■ 177
7. U se an exam ple of your own to illustrate the conditions u n der which you
would be willing to build a KM system based on a single expert. Justify your
choice.
8. W hy should the know ledge developer understand the differences am ong the
levels of experts? Isn ’t an expert an expert regardless of level?
9. Explain the relationships betw een an e x p ert’s:
a. m otivation and willingness to share know ledge
b. credibility and perceptual ability
c. creativity and w ell-developed perceptual ability
10. H ow does p attern m atching adversely affect the quality of an interview?
11. In w hat ways is visual im agery helpful in recalling chunks of experience?
12. Distinguish betw een:
a. validity and reliability
b. m ultiple-choice and dichotom ous questions
c. question form at and question content
d. question w ording and question content
13. Briefly explain each of the uncontrolled sources of e rro r th at the know l-
edge developer needs to consider. W hich source do you consider most seri-
ous? Why?
14. W hat is the interview er effect? How likely is the occurrence of this problem ?
15. R eview briefly som e of the problem s e n c o u n te re d during an interview.
W hich problem do you consider m ost serious? W hy?
16. The chapter suggests a guide to a successful interview. D oes it allow enough
flexibility? T hat is, w hat d eterm in es how strictly one should follow the
guide?
17. E xperience teaches that how questions are phrased can determ ine the nature
of the answers. In know ledge capture, explain how the following are related:
a. question phrasing and question sequence
b. prim ary and secondary questions
c. open-ended and closed questions
18. In w hat way is rapid prototyping related to interview ing? Be specific.

K n o w l e d g e E x e r c is e s ' 1 11

1. W hat suggestions can you m ake for im proving know ledge capture?
2. W hich of the prerequisites for know ledge capture described in the chapter
do you think is the m ost crucial? Why?
3. D o you think know ledge capture can be fully or easily autom ated? Why or
why not?
4. “A n expert exhibits a certain depth of detail and exceptional quality in expla-
nations.” How im portant is the “exceptional quality” of an ex p ert’s explana-
tions in know ledge capture? Explain.
5. W hen the expert tells you what he or she does not know, does that indicate
expertise?
6. Why are good com m unications skills so im portant in knowledge capture? Is
this quality a prerequisite for experts only? For knowledge developers? Why?
7. If you were asked to select an expert, how would you proceed? W hat charac-
teristics would you look for? W hat other factors would you consider?
8. “Knowledge developers can benefit immensely from an educational ground-
ing in cognitive psychology.” Do you agree? Why or why not?
9. If you w ere building a KM system using a single e x p ert w ho had been
coaxed for weeks to participate, what things would you avoid in interviewing
such an expert?
178 7 7 7 7 PART II Knowledge Creation and Capture

10. Select two other people to form a group of three. O ne person will be A, the
second person will be B, and the third person will be C. Have A ask a question
of B. B will then ask a question of C. C will ask a question of A. Q uestions
may cover areas pertaining to the course m aterial or a KM system project
you would like to work on. Review the following questions:
a. H ow well have you learned about one another?
b. W hat were you doing when the other two people were talking to each
other?
c. W hat experience did you have when you asked one question at a time?
Be specific.
d. In general, how did you spend m ost of your time?
11. Interview ing is som etim es similar to conversation. G et into a conversation
with an o th er person for 15 minutes. Choose a topic in advance. How many
tim es did the topic change? W hat triggered the change in the topic?
12. W atch an interview on TV betw een a news rep o rter and an expert in the field
(such as a police or g o vernm ent official). R eco rd how the interview was
started. W hat types of questions were asked? D o you agree with the question
sequence and construction? Explain.
13. D eterm ine w hether the following questions are leading questions or neutral
questions:
a. You like to operate on patients, d o n ’t you?
b. W hen was the last time you had less than satisfactory surgery?
c. H ow does an auto loan differ from a com m ercial loan?
d. W ould you classify yourself as an orthopedic surgeon or a specialist in
hip replacem ent?
e. H ow do you feel about applicants lying on their application form?
14. W rite the nam e of one person you adm ire and one person you dislike or dis-
trust. The person could be som eone you know, a celebrity, or a national fig-
ure. M ake a list of traits each person possesses. W hich of these traits are criti-
cal aspects of credibility? H ow w ould the d istru sted person p roceed to
change his or her image in your eyes?
15. Publisher KM System:
The goal of the first session with the expert was to develop a m aster list of the
v ariables th a t the ex p ert con sid ered im p o rta n t in deciding on w hether
a m an u scrip t should be published. These variables w ere condensed into
20 possibilities and then ran k ed in o rd e r of im portance. The ranking was
p resen ted to the expert for verification. The variables w ere reduced to the
five m ost im portant:
a. m arketability of the work
b. review ers’ opinions of the m anuscript
c. au th o r’s prior publishing record
d. a u th o r’s requirem ents before signing a contract with the publisher
e. cost of m anufacturing the book
In p re p a ra tio n for the first interview , the know ledge d ev eloper review ed
published articles abo ut book publishing, w hat m anuscripts get published,
the criteria used in m aking favorable decisions, and the book production
process. This b ack g ro u n d p rovided a n u m b er of p refa to ry questions th a t
w ere included in the first session.
Assignm ent:
For this project, set up a list of questions to ask the expert (book editor)
during the first interview. Explain the reason(s) for choosing these ques-
tions. W hat variables or param eters have you been able to conclude from
the session?
 C H A PTER 5 Capturing Tacit Knowledge i i i ■ 179
16. M ultiple experts w ere used to capture know ledge for the R esidential Real
E state KM System. The reasons are:
a. O ne objective of this KM system is to integrate the expertise in real
estate tax and finance, which requires two separate experts.
b. The procedure of m aking a buy-or-lease decision is a complex and intense
process for a single expert to provide all necessary inform ation to come
up with a decision. The nature of the problem also suggests that to m ake a
decision, one has to consider both tax and financial issues. Thus, multiple
experts are necessary to divide the scope into m anageable pieces.
c. M ultiple experts give the ability to com pare and contrast the knowledge
captured. The real estate finance expert has deep know ledge in the real
estate tax field as well. H e was particularly helpful in confirm ing the tax
inform ation captured from the tax expert. Confirm ing the validity of the
proposed recom m endation and solution is critical considering the contin-
uously changing tax law. O ne expert m ight not be aware of the change,
while the o th er is.
Q uestions:
a. In your opinion, is the use of a m ultiple expert approach justified for this
project? Explain.
b. Can the use of two know ledge developers to tackle the problem be justi-
fied? W hat are the conditions for a m ultiple knowledge developer
approach to building this system? Explain.

Re f e r e n c e s 1111
Arthur, L. J. Im proving Software Quality: An Insider’s Hill, Mary with Dr. Heather Wakat, University of
G uide to TQM . New York: John Wiley & Sons, Virginia, Spring 2000 (personal interview).
Inc., 1992. Johannessen, Jon-Arild, Olaisen, Johan, and Olsen, Bjorn.
Awad, Elias M. Building Expert Systems. Minneapolis, “M ismanagement of Tacit Knowledge,” www.program.
MN: West Publishing Co., 1996, pp. 134,168. forskningsradet.no/skikt/johannessen.php3, Date
Davenport,T. H., and Prusak, Laurence. Working accessed April 2002.
Knowledge. Boston, MA: Harvard Business School Pfeffer, Jeffrey, and Sutton, Robert I. The Knowing-D oing
Press, 2000, pp. 137-140. Gap. Boston, MA: Harvard Business School Press,
Eisenhart, Mary. “Gathering Knowledge While It’s Ripe,” 2000, pp. 1-28.
Knowledge Management, April 2001, pp. 49-54. Sherman, Lee. “Managing the Modern Docum ent,”
Gill, Philip J. “The Ancient Art of Storytelling Emerges Knowledge M anagement, September 2001, pp. 48-52.
A s a Tool for Knowledge M anagement,” K nowledge Stewart, Kathy A ., Storey, Veda C., and Robey, Daniel.
M anagement, May 2001, pp. 24-30. “Plugging the Knowledge Drain: Strategies and
Hayes-Roth, E, and Jacobstein, N. “The State of Technologies for Acquiring Knowledge in Lean
Knowledge-Based Systems,” Communications o f the Organizations,” Journal o f Information Technology
A C M , March 1994, p. 28. M anagement, vol. 10, no. 3-4,1999, pp. 59-67.
 Other Knowledge
Capture
Techniques
Contents

In a Nutshell
Overview
On-Site Observation
Brainstorming
Electronic Brainstorming
Protocol Analysis
What Is a Scenario?
Protocol Procedure of the Diabetic Foot KM System
Consensus Decision Making
The Repertory Grid
Nominal Group Technique (NGT)
The Delphi Method
Concept Mapping
Procedure
Concept Mapping and Semantic Nets
Blackboarding
Implications for Knowledge Management
Summary
Terms to Know
Test Your Understanding
Knowledge Exercises
References

“/ d o n ’t think like y o u do. ”

—D a n ie l K o r k ill
 CHAPTER 6 Other Knowledge Capture Techniques i aaa 181

In a Nutshell
Like any other professional, the knowledge developer must be well versed in the use of
specialized knowledge capture tools. Each tool has a unique purpose, depending on
w hether the capture process revolves around a single expert or m ultiple experts.
Among the tools are computer-based tools designed to prom ote accuracy and integrity
of the knowledge capture process.
A knowledge developer might want to observe the problem-solving process in a
situation in which an on-site observation would be a proper tool. When dealing with
two or more experts, brainstorm ing followed by consensus decision making is pre-
ferred. Whatever tool is chosen, the knowledge developer must follow a preestablished
procedure for conducting the session.
A unique knowledge acquisition technique for single experts is protocol analy-
sis or the think-aloud method. In essence, it involves listening to the spoken protocol
of the expert. During the session, the expert speaks aloud whatever thoughts come
to mind while answering a question or solving a problem . In contrast, a repertory
grid uses a grid or a scale to represent the expert’s way of looking at a particular
problem. The number of gradations on a scale can be increased to reflect a more accu-
rate rating.
O ther techniques for multiple experts include the nom inal group and Delphi
methods. Each m ethod has unique features and a procedure to make its use opera-
tional. The nominal group method is an interface between consensus and brainstorm-
ing; the Delphi method is a polling of experts’ opinions via questionnaires.
An attractive computer-based method for multiple experts is the blackboard. Also
called groupware, this global memory structure promotes privacy and equality among
participating experts. It attempts to get the experts to agree on a solution, and it is ideal
for complex problem-solving involving several experts.

Overview
A knowledge developer must recognize that many complex problems cannot be solved
by single experts. For example, the prototype for the Diabetic Foot KM System
described in the preceding chapter will involve two other specialists in the field to
agree on medical and surgical treatm ent before the KM system can be medically certi-
fied. To be sure that knowledge is reliable, multiple experts are often required.
When dealing with multiple experts, the knowledge developer must be famil-
iar with the tools and techniques that are unique to the task. C hapter 5 focused on
ways to ensure a successful interview with a domain expert. The use of this tool re-
quires a customized approach to both the expert and the process. This chapter exam-
ines other tools used in knowledge capture: on-site observation, protocol analysis, and
consensus m ethodologies, including brainstorm ing, consensus decision making, the
repertory grid, nominal group technique, and the Delphi method for multiple experts.
A more recent knowledge capture tool, blackboarding, will be discussed at the end of
the chapter.
Acquiring knowledge from multiple experts requires experience and special-purpose
techniques. W ithout this combination, serious problems often arise, such as experts
walking out, faking solutions to end the session, and so on.
182 77 7 7 PART II Knowledge Creation and Capture

■■■■ On-Site Observation

This specialized knowledge capture technique has been borrowed from the social sci-
ences, where the knowledge developer observes the daily work of a domain expert.
On-site observation, or action protocol, is a process of observing, interpreting, and
recording an expert’s problem-solving behavior while it takes place. It requires con-
centration on the overall steps that the domain expert takes as well as the more subtle
details of the process. In addition to observing and recording the expert’s behavior, the
knowledge developer asks the expert questions about the problem-solving process.
On-site observation carries with it a certain protocol. The knowledge developer
does more listening than talking; avoids giving advice and does not pass judgment on
what is observed, even if it is incorrect; and does not argue with the expert while the
expert is performing the task.
On-site observation was used by one knowledge developer to build a teller KM
system for a commercial bank, a project that took 2 months to complete. She sat in a
booth adjacent to a senior teller with 7 years of experience. The purpose of the proj-
ect was to build a KM system to determ ine when a check exceeded the set limits
($500, in this case) and what to do about the problem without requiring an officer’s
signature. This particular teller’s success rate (the check did not bounce) for the deci-
sion was 95 percent; the average bank officer’s success rate for the same decision was
70 percent.
In this example, on-site observation gave the knowledge engineer a visual or a live
exposure to the kind of information that can also be gathered through the interview. In
the teller KM project, the knowledge developer conducted brief informal interviews
with the teller when no customers were waiting. The answers during those interviews
verified the information picked up through participant observation.
O bservation of behavior enables the knowledge developer to seek knowledge
within the working world of the expert. In comparison to the interview, observation
places the knowledge developer closer to the actual steps and procedures used by the
expert to solve the problem. One problem with this capture technique is that some
experts do not like to be observed. They prefer to talk about their thought process
rather than show them in practice. Sometimes, experts fear that observation will give
away years of experience in one quick look.
The reaction of others in the observation setting can also be a distracting problem.
During the early stages of the teller KM system project, other tellers’ curious and occa-
sionally hostile feelings were conspicuous. One teller with 18 years’ experience was
quite irritated at the bank for allowing such a “ridiculous automation of the human
mind.” She guarded her own actions closely after commenting during a lunch break
that the teller “expert” was being used as a guinea pig and wondering, “How can they
tell us to follow procedure, and here they are trying to autom ate the procedure?”
However, in the case of the teller KM system project, the teller expert thrived on
attention. She was not at all distracted by the presence of the knowledge developer or
the reservations of coworkers. Participant observation went on for 2 weeks without
any problems.
Another problem with on-site observation is the accuracy or completeness of the
captured knowledge. Although the general assumption may be that the recording of an
event and the event itself occur simultaneously, in reality a time gap separates what is
observed and its recording. The event is first transcribed into a mental image before
that image is translated into the knowledge developer’s own words. It becomes a retro-
spective process, which inevitably invites errors.
 CHAPTER 6 Other Knowledge Capture Techniques * a * * 183
To form a mental picture of the event, the knowledge developer attempts to inte-
grate his or her perception of the situation with the expert’s perception. That mental
picture is then recorded. What happens is a continuous shuttle process; the knowledge
developer mentally moves back and forth from the initial impression of the event to
the later evaluation of the event. What is finally recorded is the evaluation made dur-
ing this retrospective period. Because a time lapse can make details of a situation less
clear, the information is not always valid.

Brainstorming
Unlike on-site observation, which focuses on the work of a single expert, brainstorming—
an unstructured approach to generating ideas about a problem —invites two or more
experts into a session in which discussions are carried out and a variety of opinions are
tossed around. The primary goal of this process is to think up creative solutions to
problems. In brainstorming, all possible solutions are considered equally. The emphasis
is on the frequency of responses during the session. Anything related to the topic can
be brought up, and everything is valued. Questions can be raised for clarification, but
no evaluation is made at the moment.
In brainstorming, the first look is on idea generation, followed by idea evaluation.
Similarities begin to emerge across opinions, which are then grouped logically and
evaluated by asking several questions:
• If one followed up on this idea, what benefits would ensue? (All ideas are item-
ized or prioritized.)
• What problem(s) would a selected idea solve?
• What new problem(s) would arise?
In the evaluation phase, the knowledge developer explains each idea and treats
any comments or criticisms accordingly.
The general procedure for conducting a brainstorming session is as follows:
1. Introduce the brainstorming session. Explain what it is and what it is not designed
to accomplish, the role of each participant, the “rules of the game,” and the
expected outcomes. Starting with the wrong objective or the wrong step can
doom the entire process.
2. Give the experts a problem to consider. The problem approved by the organiza-
tion is in the experts’ domain of expertise. The knowledge developer must give
them time to think it through and then be a good listener and show enthusiasm c
but also set reasonable time limits.
3. Prom pt the experts to generate ideas. The experts can do this either by calling out
their ideas or by establishing some order in which each expert will have a turn to
speak. The knowledge developer must keep pace with the expert.
4. Watch fo r signs o f convergence. Ideas often trigger counteropinions or reinforce-
ments that should eventually coax the experts to converge onto the final four, the
final two, and the final solution. When experts begin to pass or the rate and qual-
ity of ideas presented declines, the process moves on to convergence.
If experts cannot agree on the final solution from two or three alternatives, the
knowledge developer may call for a vote or a consensus in order to reach agreement
on the best final solution; this does not guarantee success. The intensity of the discus-
sion just prior to such a vote, the tem peram ent of the experts, and how willing they are
to resolve conflicts all influence whether a consensus can be reached.
184 7 7 7 7 PART II Knowledge Creation and Capture

ELECTRONIC BRAINSTORMING
A relatively new development in brainstorming is a computer-aided approach to deal-
ing with multiple experts. Desks in a U-shaped layout hold PCs networked through a
software tool that serves as a catalyst in the meeting, promotes instant exchange of
ideas between experts, and sorts and condenses those ideas into an organized format.
Such a tool also allows experts to elaborate and vote on ideas (see Figure 6.1).
For example, an electronic brainstorm ing process begins with a presession
plan that identifies objectives and structures the agenda, which is presented to the
experts for approval. During the live session, each expert chooses a PC and engages in
a predefined approach to resolving a focused issue and then generates ideas or plans.
The experts gain leverage from anonymity, focus on content (not personalities), and
engage in parallel and simultaneous communication. This format allows two or more
experts to provide opinions through their PCs without having to wait their turn. The
software displays the comments or suggestions on a huge screen without identifying
the source.
This method protects the shy expert and prevents tagging comments to individuals.
The overall benefits include improved communication, effective discussion of sensitive
issues, shorter meetings, and closure of m eeting with concise recommendations for
action. The sequence of steps is summarized in Figure 6.2. The expert’s ideas are priori-

FIGURE 6.1: Electronic Brainstorming Environment

Whiteboard Projection Screen Whiteboard

Printer
 CHAPTER 6 Other Knowledge Capture Techniques ■■■■ 185

FIGURE 6.2: Brainstorming Process

tized.This eventually leads to convergence and setting final specifications. The result is
joint ownership of the solution.
To illustrate the process used in electronic brainstorming, a group of graduate pro-
fessors met one day to discuss the field research topics for MIS graduate students. The
1-day session using IB M ’s Electronic B rainstorm er began with a focused topic:
G enerate a portfolio of field research projects for graduate MIS students. The facilita-
tor (the role taken by the knowledge developer) asked each professor to list as many
topics as desired through the PC. Each professor was assigned an ID that was not visi-
ble on the screen when the questions were entered or displayed.
Table 6.1 shows the resulting input of the seven participants.
After the questions had been listed on the screen, they were discussed, and redun-
dancies were removed. Afterwards, the key topics remained, as shown in Figure 6.3.
The questions listed were reviewed and arranged in order of priority. Then a vote was
taken. The resulting matrix shown in Table 6.2 provides the voting results.

■■ Protocol Analysis
Suppose you want to understand the diagnostic process of a medical expert; the knowl-
edge he or she uses; and the cognitive actions they take. How would you go about it?
One obvious approach is to ask the expert questions about diagnosis. Chances are the
expert will not find it easy to answer questions. One of the authors found out that his
orthopedic surgeon is more used to doing the job than explaining it. The surgeon tried to
explain diagnosing a diabetic foot in terms of the formal procedure he learned in medi-
cal school, which is not quite the diagnosis he follows with each patient. The alternative
is to observe an examination of a real patient and then listen to the spoken protocol.
186 7 7 7 7 PART II Knowledge Creation and Capture

TABLE 6.1 Questions Generated by Seven Experts Anonymously

Problem Domain: Generate a portfolio o f field research fo r graduate MIS students.___________
Participant 1:
1.1 D evelop an interactive graphical data model query selection
1.2 Analyze group decision support systems; e.g., team focus
1.3 D evelop a manufacturing database to support Comm 326
1.4 D evelop an executive information system that uses real-time data feeds
1.5 A nalyze D ave Smith’s office (Can he possibly be productive in his environment?)

Participant 2:
2.1 Work with physical plant on cost estimating system
2.2 Assist athletic department in establishing an injury database
2.3 Work with local IBM office to sell more PCs
2.4 D evelop database of nine in-class speakers from industry

Participant 3:
3.1 Work with medical imaging specialists on patient database
3.2 Explore use of COBOL to access distributed database
3.3 Establish a database of MIS internships

Participant 4:
4.1 Analyze communication media (e.g., electronic and/or voice mail), perhaps using the
Commerce School as a case study
4.2 Study the allocation o f computer resources in the University of the Comm School
4.3 D eterm ine the effect of various levels of DSS on player performance of a management
simulation game
4.4 R econcile use of university and Comm School e-mail
4.5 Evaluate the impact of G DSS on Comm student group productivity

Participant 5:
5.1 Study graphics applications in GDSS (Can we integrate graphics input?)
5.2 A nalyze issues of migration to graphic user interfaces for end users
5.3 Automate commerce student registration

Participant 6:
6.1 A nalyze new CASE tools
6.2 Examine use of simulation tools for dataflow modeling
6.3 Evaluate use of GDSS tools in an academic environment

Participant 7:
7.1 Explore implementation issues of electronic mail
7.2 Simulate knowledge acquisition session with multiple experts
7.3 Implement a neural network to evaluate mortgage applications
a ■ ■ a ■ a ---------------------------------------
 CHAPTER 6 Other Knowledge Capture Techniques ■■■« 187
A nalyze group decision support systems (1.2)
D evelop executive information systems (1.4)
D o projects that add to technical capabilities o f our school
Simulate rapid prototyping through brainstorming with experts
Implement e-mail for graduate MIS students (4.4)
Simulate knowledge capture session for multiple experts (7.2)
Build interactive knowledge-based system to train med students (3.1)
D evelop electronic rosters and grad submissions (7.1)
Evaluate current case tool (6.1)

■ i ■ l i i FIGURE 6.3: Final Set of Questions Agreed Upon by the Experts

In working with multiple experts, the problem-solving process is bound to vary.

Two experts might give the same answer, but they might use two different problem -
solving strategies. A think-aloud protocol provides a clear insight as to how each expert
arrived at the solution through the individual expert’s verbalizations. The think-aloud
method avoids interpretation by the experts.
Box 6.1 illustrates how two solvers interpreted an arithmetic word problem. The
two solvers arrived at the same answer, but each used a different solving strategy. The
cognitive processes of experts are difficult to obtain by other means. The protocol

TABLE 6.2 Voting and Rank Key Topics

Number o f Votes in Each Position
Key
Topics 1 2 3 4 5 6 7 8 9 Mean
Anal 4 1 — 2 — — — — — 2.00
Exec 1 2 1 1 — — 1 1 — 3.86
Do p 2 — — 2 — 1 1 — 1 4.57
Doin — — 1 — 5 1 — — — 4.86
Impl - 1 2 - - 2 - 1 1 5.29
Know — 2 — — — 3 1 1 — 5.29
Buil — 1 — 1 1 — 2 2 — 5.86
Deve — — 2 — — — 2 2 1 6.43
Shou — — 1 1 1 — — — 4 6.86

Rank
Sum
Analysis o f group decision 56
Executive information system 43
D o projects that add to 38
D oing rapid prototyping 36
Implement e-mail for 33
Knowledge acquisition session 33
Build interactive knowledge 29
D evelop electronic roster 22
188 . . . . PART II Knowledge Creation and Capture

■ ■ 1 1 1 1 BOX 6.1 i 1 I 1 i 1

COGNITIVE APPROACH TO PROBLEM SOLVING

A father, a m other, and their daughter have a 3. If father is 25 years old and mother is 25, then the
combined age of 70 years. The father is three times daughter is 20 years old.
as old as the daughter. The mother is the same age 4. This doesn’t make sense, because 25-year-old par-
as the father. How old is the daughter? ents cannot have a 20-year-old daughter.
5. Also, the father is not 3 times the age of the
SOLVER 1 daughter.
1. This sounds com plicated—I’ll give it a shot. 6. Father and mother must be older.
2. Assume F for father, M for mother, and D for 7. If father and mother are 27 years old, then the
daughter. daughter must be 16.
3. F + M + D = 70 8. The daughter is not 3 times the age of the father.
4. F is 3 times D. 9. A 27-year-old father cannot logically have a 16-
year-old daughter.
5. M = F
6. So far, I have three equations and three unknowns. 10. This is not the right combination. I’ll have to
change ages.
7. D + 3 times D (or F) + 3 times D (or M) = 70
11. What if the father is 30 years and the mother is 30
8. 7 D = 70 years? This will make the daughter 10 years old.
9. D = 10 12. This adds up to 70 all right.
10. A ges of father and mother total 70 -1 0 or 60. 13. A t 30, the father is exactly 3 times the age of the
11. Because both are the same age, father is 30 years daughter, who is 10 years old.
old and mother is 30 years old. 14. The mother is also the same age as the father,
which is 30.
SOLVER 2
15. The whole thing makes sense. A 30-year-old
1. Father, mother, and daughter are 70 years old, total. father can have a 10-year-old daughter.
2. How can this be?

a a i i fl a

analysis or think-aloud method shows clearly how the solvers approached the problem
one step at a time. It is a useful technique for knowledge capture.

WHAT IS A SCENARIO?
In protocol analysis, protocols (also known as cases or sc e n a rio s) are collected by ask-
ing experts to solve a problem and verbalize what goes through their heads, stating
directly what they think. In other words, the expert keeps talking, speaking out loud
whatever thoughts come to mind, while they answer a question or solve a problem.
Unlike their role in other techniques, knowledge developers do not interrupt or ask
questions in the interim. The solving process is carried out in an automatic fashion
while the expert talks. Structuring the inform ation elicited occurs later when the
knowledge developer analyzes the protocol.
The term scen a rio is used to refer to a detailed and sometimes complex sequence
of events. Mission scenarios, for example, are a detailed specification of a flight plan, a
 CHAPTER 6 Other Knowledge Capture Techniques i ■« 1 189
series of aircraft maneuvers and pilot tasks. These can be used to train pilots or to con-
sider the performance requirements of an aircraft.
A scenario refers to a situation or, more precisely (because it has a temporal com-
ponent), an episode. A scenario involves individuals, objects, and courses of events.
O ne im portant characteristic of scenarios is that they describe particular states of
affairs and events. They can be used to describe possible worlds in which certain yet to
be considered actualities could occur. This is the case with mission scenarios, where a
pilot is being trained for certain anticipated events. It is also true of design scenarios, in
which a designer imagines what it would be like to interact with a system using a par-
ticular kind of device.
The purpose of a scenario is to provide an explicit concrete vision of how some
human activity could be supported by technology. Scenarios are one good reference
point for making design decisions. By showing the actual circumstances under which
people work, scenarios provide guidelines on how a technology should perform. The
very act of making things explicit and clear also helps in making design choices.
For a scenario to be good, it must realistically depict some actual human activity. It
must “bring to life” for its users the activity that is to be supported. Scenario develop-
ers need interviewing and writing skills as well as technical imagination and under-
standing. These skills could be distributed betw een two people; for example, one
researcher might be an especially good interviewer, while the other has strong writing
skills and technical depth.

PROTOCOL PROCEDURE OF THE DIABETIC FOOT KM SYSTEM

This KM system stores knowledge that directs a diabetic foot patient to seek either
self-help, medical help, or surgical help.
Case Description
1. Sex female
2. Age 45-48
3. Complaint blister does not heal; pain in left
foot when standing or walking
4. How much does patient weigh? 230 pounds
5. How long has patient had the blister? 2 weeks
6. Has patient had blisters before? no
7. Does the blister get worse when walking? yes
8. Where is pain located? left foot below ankle
9. How strong is the pain? quite strong
10. How long has patient had pain? over a week
11. Did patient use any medication for pain? no
12. Does pain change with movement? no
13. What does patient think is wrong? poor shoe fit
Protocol
1. This is a woman in her mid- to late 40s.
2. Being quite overweight and a diabetic, blisters are common occurrences.
3. Pain is symptomatic of the blister.
4. Patient is experiencing this blister for the first time. She’s probably more
worried than in pain.
190 7 7 7 7 PART II Knowledge Creation and Capture

5. Being diabetic, blisters take a long time to heal. It is not likely to get worse.

40. I don’t see skin broken or pus accumulating, which is a good sign.
41. I’m going to recommend NSD and soaking the foot in warm water before going
to bed and after getting up in the morning.
42. Her husband is going to have to help.
43. I’m going to recommend that patient wear wide-toed shoes.

64. So, for the moment, I am going to tell the patient to see me in 2 weeks.
65. Right now, I wouldn’t recommend any medical treatment. Surgery is the last
thing on my mind.
66. I’ll relay this diagnosis and decision to patient.
In summary, think-aloud or protocol analysis is a knowledge capture method that
consists of asking experts to think aloud while going through a problem solution and
analyzing the resulting verbal protocols. It is an effective source of information on cog-
nitive processes and for building knowledge management systems. In fact, think-aloud
makes the expert (in this example, the surgeon) cognizant of the process being
described, which is a contribution to the validity of the practice or procedure used in the
diagnosis. It also provides a wealth of information toward knowledge representation.

*■■■ Consensus Decision Making

The term consensus refers to a clear agreement regarding the best solution to a prob-
lem. As a tool, consensus decision making follows brainstorming. It is effective only
when each expert in the team has had adequate opportunity to air their views. To arrive
at a consensus, the knowledge engineer conducting the exercise tries to rally the
experts toward one or two alternatives while fostering the experts’ feelings of the owner-
ship of the alternative(s) so that they want to support it.
In consensus decision making, a knowledge developer follows a procedure
designed to ensure fairness and standardization in the way experts arrive at a consensus:
1. The knowledge developer explains the solutions generated earlier, possibly
through brainstorming. After a problem is presented to the panel of experts, they
are asked to vote on the alternative solutions they generated. Voting takes place in
rounds, with each round devoted to reducing the number of alternative solutions.
2. In round 1, the knowledge developer explains that each expert has three votes to
cast: one vote per option for a maximum of three options among the total number of
options. The options with fewer than the agreed-upon number of votes are deleted.
3. In round 2, the knowledge developer questions the experts to see if any of the
remaining options can be deleted or merged. Then, the experts are given two
votes each to vote on the remaining options. That is, each expert votes on two of
the total number of options that remain after round 1.
4. As with round 1, those options that receive fewer than the required number of
votes are deleted. This process of voting by rounds continues until only two options
 CHAPTER 6 Other Knowledge Capture Techniques a t « a 191
are left. During the last round, each expert is allowed to cast only one vote for the
final option. The option that carries the majority vote is selected as the solution.
The consensus method is quite effective and works well most of the time, but it is not
without problems. First, it is a bit tedious and can take hours; at any time during the round,
an expert may call for a discussion of the remaining options before further voting. Second,
some experts complain that each option unfairly carries the same weight; some options
are more serious than others and should be adopted or rejected by every expert on the
panel. Third, experts have been known to walk out on the whole capture process because
of the rigidity of the consensus method. Although this tool promotes a democratic process
of decision making, some experts allow personalities and personal agendas to interfere.

The Repertory Grid

The repertory grid is another tool used in knowledge capture. The domain expert is
viewed as a scientist who classifies and categorizes a problem domain using his or her
own model. The grid is used to capture and evaluate the expert’s model or the way the
expert works through the solution. Two experts in the same problem domain will pro-
duce different sets of results that are personal and subjective. Experts see problems
based on reasoning that has stood the test of time and are able to use this deep knowl-
edge in the problem-solving process. For example:
N o v ic e l o a n c u s t o m e r : The bank tried to charge more interest than they
promised. Now I am in a bind.
Ex pe r t l o a n c u st o m e r : The issue is whether their last interest quote was an esti-
mate or a binding rate for the loan
A novice may have a general sense of the effects of certain actions, but the expert fore-
sees cause and effect in a specific way. The goal of knowledge capture is to identify the
underlying principles the experts use so that they can be captured and later used by the user.
The repertory grid is a representation of the experts’ way of looking at a particular
problem. A grid is a scale or a bipolar construct on which elements are placed within gra-
dations. The knowledge developer elicits the constructs and then asks the domain expert
to provide a set of examples, called elements. Each element is rated according to the con-
structs that have been provided. The following are examples of bipolar constructs:

Novice Expert

Unsatisfactory Satisfactory

G radations break down the scale for more accurate ratings. In the following
unsatisfactory-satisfactory construct, a 1-5 point scale is used.

Unsatisfactory Satisfactory
1 2 3 4 5

Once a particular scale has been adopted, it should stay the same throughout the
grid, although the terms that describe the rating vary from construct to construct. The
rating is useful in comparing individual rankings. For example, if John is rated “1,” Ann
is rated “3,” and Bob is rated “4,” it does not mean that Bob is four times as satisfactory
192 7 7 7 7 PART II Knowledge Creation and Capture

as John. It only means that Bob is more satisfactory than either John or Ann. It is a
comparative subjective rating of elements along a scale.
One of the benefits of the repertory grid is that it may prompt the expert to think more
seriously about the problem and how to solve it. The main drawback is that it tends to be a
difficult tool to manage effectively when large grids are accompanied by complex details.
Large grids defeat the tool’s goals of clarity, simplicity, and manageability. For this reason,
knowledge developers normally use the grid in the early stages of knowledge capture.
The practical use of the repertory grid can be demonstrated by the following inter-
view, in which the human resources director compares bank tellers in order to decide
on potential promotions or transfers. The expert is the human resources director. The
comparative rating of the tellers is shown in Figure 6.4.
KD: As I understand it, we’re here to discuss the selection process and selec-
tion criteria for the bank’s tellers.
Ex pe r t : Yes, I’m interested in reviewing all of the tellers to get a better under-
standing of our selection procedure.
KD : Which tellers do you consider prime candidates for the purpose of the
selection procedure?
Ex pe r t : I have in mind Dixie, John, Barry, Curt, Lester, and Joanne.
KD: If you take the first three tellers (Dixie, John, and Barry), which two are
about the same but differ from the third?
Ex pe r t : I’d say Dixie and John are similar, but Barry is quite different.
KD: If experience is the similarity factor, what word would describe their
dissimilarity?
Ex pe r t : Inexperience
KD: Suppose we place “experience” and “inexperience” on a scale of 1 to 3,
where 1 represents “inexperienced” and 3 represents “experienced.”
How would you rate the tellers?
Ex pe r t : Dixie 3
John 3
Barry 1
Curt 1
Lester 1
Joanne 1

FIGURE 6.4: A Repertory Grid Rating Job Performance of Bank Tellers

Construct T1 T2 T3 T4 T5 T6 Construct

1 Inexperience 3 3 1 1 1 1 Experienced
2 Academically Ill- Academically Qualified
Qualified 2 1 2 1 1 3 Good Appearance
3 Poor Appearance 3 2 1 2 1 3 Punctual
4 Late 2 3 2 3 1 1 Extroverted
5 Introverted 2 3 3 2 1 1
T1 Dixie
T2 John
T3 Barry
T4 Curt
T5 Lester
T6 Joanne
 CHAPTER 6 Other Knowledge Capture Techniques iiai 193
KD: Let us now take the last three tellers, Curt, Lester, and Joanne. How
would you compare them?
Ex pe r t : These tellers, I’d want to compare in terms of academic qualifications. I’d
say Joanne is better qualified than Curt or Lester. Using the 1 to 3 scale
you suggested earlier, I’d rate the six tellers as follows:
Dixie 2
John 1
Barry 2
Curt 1
Lester 1
Joanne 3
KD: Next, let’s take Joanne, Dixie, and Barry. How would you rate them?
Ex pe r t : I guess I’d want to rate them based on appearance —how well they look
in the eyes of the customers. Based on custom er feedback, I’d rate
Joanne and Dixie the same. Barry is not in the same league. I’d rate them
as follows:
Dixie 3
John 2
Barry 1
Curt 2
Lester 1
Joanne 3
KD: Let's take one more group—John, Curt, and Lester.
E x pe r t : These tellers remind me of attendance. John and Curt are hardly ever
late to work. Lester is often late. So, using the 1 to 3 scale (1 is tardy; 3 is
prompt). I’d rate them as follows:
Dixie 2
John 3
Barry 2
Curt 3
Lester 1
Joanne 1
KD: Do you have other ratings you want to include?
Ex pe r t : Just one dealing with John, Barry, and Lester. John and Barry are extro-
verts; Lester isn’t. I’d use the 1 to 3 scale (1 = introvert; 3 = extrovert)
and rate them as follows:
Dixie 2
John 3
Barry 3
Curt 2
Lester 1
Joanne 1

Nominal Group Technique (NGT)

In some problem domains, more than one expert might be available as a source of
knowledge for building the KM system. However, for situations in which several
experts have overlapping expertise, each expert's opinion must be interpreted in line
with the problem domain. In fact, a single expert is used precisely to avoid potential
194 7 7 7 7 PART II Knowledge Creation and Capture

contradictions between experts and possible m isinterpretations on the part of the

knowledge developer. The nominal group technique and Delphi m ethod have been
shown to mitigate some of the process losses associated with multiple experts.
An alternative to the consensus technique, the nominal group technique (NGT)
provides an interface between consensus and brainstorming. In this technique, the
panel of experts becomes a “nominal” group whose meetings are structured in order to
effectively pool individual judgment.
NGT is an ideawriting or idea generation technique. Ideawriting is a structured
group approach used to develop ideas and explore their meaning for clarity and speci-
ficity; the result is a written report.
The NGT procedure begins as follows:

1. The knowledge developer explains the technique and provides to each expert
information about the problem or alternative solutions.
2. Instead of discussing the problem, the knowledge developer asks each expert to
list on paper the pros and cons of the problem or alternative solutions. This is
silent generation of ideas in writing.
3. The knowledge developer compiles a list of all pros and cons without making
comments regarding their merit. Any overlapping ideas are either reworded or
deleted.
4. Each expert is given a copy of the compiled pros and cons and asked to rank
them on the basis of their priorities.
5. The knowledge developer then leads a discussion of the pros and cons and their
respective ranks. The discussion focuses on the priorities placed on each item and
the reasoning behind them, which leads to a listing of possible solutions.
6. The knowledge engineer compiles the alternative solutions. This is followed by a
group discussion that should lead to agreement of the “best” solution.

Unfortunately, NGT can be time-consuming and tedious. It has been known to

promote impatience among the experts who must listen to all kinds of discussions with
other experts, some of whom will lobby aggressively for their own ideas. One way to
expedite this procedure is to have the experts vote as a group on the set of possible
problems to work on, and later on which solution(s) to use
On the other hand, as experts share expertise, things can really jell in adopting the
best solution. NGT is ideal in situations of uncertainty regarding the nature of the
problem domain. It is also effective in multiple expert knowledge capture, especially
when minimizing the differences in status among experts is important.
NGT is similar to brainstorming, except that it is based on the understanding that
certain group goals can be best achieved by writing rather than by discussion. Each
expert has an equal chance to express ideas in parallel with other experts in the group.
Because discussion is accommodated in sequential order, NGT can be a more efficient
and productive approach than brainstorming.

• • •• The Delphi Method

A nother tool used in m ultiple expert knowledge capture is the Delphi method.
Essentially, it is a survey of experts. Experts are polled concerning a given problem
domain. A series of questionnaires are used to pool the experts’ responses in order to
solve a difficult problem. Each expert’s contributions are shared with the rest of the
experts by using the results from each questionnaire to construct the next questionnaire.
 CHAPTER 6 Other Knowledge Capture Techniques aaaa 195
The Delphi method gets its name from the ancient Greek oracle at Delphi, who
was said to be able to look into the future. It works this way:
1. A panel of experts is asked to prepare an anonymous opinion about a focused
problem domain. Each person is usually given a short written explanation of the
problem. For example, one problem could be to estimate when air cargo revenues
of XYZ airline would equal its passenger revenues.
2. For the second round, each expert is given a summary of the results of the first
round and is asked to make a second (still anonymous) estim ate on the same
issue based on the additional information and reasons provided in the summary
of the preceding anonymous opinions. Knowing how the other experts responded,
an expert may stick to or change his or her initial answer.
3. Step 2 is repeated two or more times, and then a final summary is prepared. By
this point, all extreme estimates will have been deleted, and those that remain
will converge on a narrow range of answers.
The Delphi method has three important features:
1. A nonym ous response. Because opinions are obtained anonymously, the danger
that the response of one expert will bias another is removed. For example, a top
executive will have no influence on a subordinate who may be part of the group.
2. C ontrolled feedback. Through a controlled set of rounds, the Delphi m ethod
allows each expert to rethink any previous answers in light of the anonymous
feedback received from other experts.
3. Statistical group response. The final opinion of the experts is an aggregate of each
expert’s response in the final round.
The Delphi method has two main limitations. First, experts often lack the neces-
sary knowledge on which to base their final judgment. The problem could be one of
communication. Second, poorly designed questionnaires have been known to be less
than effective in developing an understanding of the complexity of the problem
domain. The knowledge developer must exercise care in assessing in advance the likely
consequences of using the Delphi method.

Concept Mapping
A unique tool to represent knowledge in graphs is concept mapping. It is a network of con-
cepts that consists of nodes and links. A node represents a concept, and a link represents
the relationship between concepts (see Figure 6.5). Similar to a semantic network, concept
mapping can be done for several reasons: to design a complex structure (such as large Web
sites); to generate ideas ( for example, brainstorming); to communicate complex ideas
(such as finding petroleum); or to diagnose misunderstanding. Initially developed by J. D.
Novak at Cornell University in the 1960s, concept mapping is designed to transform new
concepts and propositions into existing cognitive structures related to knowledge capture.
As can be seen, concept mapping is a structured conceptualization, where a topic
of interest involving input from one or more experts produces a clear pictorial view of
their ideas, their concepts, and how they are interrelated. It is an effective way for a
group to function without losing their individuality. In contrast to concept mapping is
mind mapping, which consists of a key word or concept. The knowledge developer
draws around the central word several main ideas that relate to that word. From each
“child” word, several main related ideas are further drawn. A mind map has only one
main concept; a concept map may have several.
196 7 7 7 7 PART II Knowledge Creation and Capture

fl I 1 fl 9 I FIGURE 6.5: Conceptual Map—An Example

SOURCE: Adapted from Lanzing, Jan. “Concept Mapping.” Article posted on users.edte.utwente.nl/lanzing/
cm_home.htm, Date accessed October 18,2002.

PROCEDURE
Using a concept map as a tool involves a 6-step procedure. As shown in Figure 6.6,
these steps are as follows:
• Preparation, which centers around identifying the participants and developing a
focus for the knowledge capture project. For example, in the Diabetic Foot KM
System, the orthopedic surgeons decided to focus on defining gangrene before
discussing it. A schedule for the mapping is also agreed upon.
• Idea generation. In this step, the participants generate a series of statements that
address the focus and expected outcomes as a result of the participation. This is
where brainstorming and nominal group tools may be useful.

fl fl fl fl fl fl FIGURE 6.6: Steps in Concept Mapping

1
Preparation of Project
(participants, focus, schedule)

5
Interpretation 4
(cluster analysis) Statement
Representation
 CHAPTER 6 Other Knowledge Capture Techniques aaa a 197
• Statement structuring. Once the statements are assembled, they are sorted into piles
of similar ones, with one statement on each standard-size card. Then, each pile is
given a unique descriptive label, and each participant rates each statement on an
agreed-upon scale, such as 1 to 5 (1 = unimportant and 5 = extremely important).
• Representation. This step represents idea analysis. It is where the sort and ratings
are represented in map form. One form of analysis, called cluster analysis, takes
the output of the point scaling and partitions the map into clusters. The state-
ments that are specific outcomes are viewed as outcome clusters.
• Interpretation. In this step, the knowledge developer or facilitator works with the
participants to arrive at their own labels and how each map is interpreted.
• Utilization. This is where the maps are actually made operational, and the results
are displayed.

CONCEPT MAPPING AND SEMANTIC NETS

One graphic tool used to encode relationships to capture knowledge is the semantic
net. Similar to concept mapping, a semantic net is a collection of nodes linked together
to form a net. Through a net, a knowledge developer can graphically represent descrip-
tive or declarative knowledge. Each idea of interest is represented by a node linked by
lines called arcs that show relationships between nodes. It is merely a network of con-
cepts and relationships.
Figure 6.7 presents a partial semantic net representing an airline’s personnel KM
system; it shows a specialist in pilot testing. The knowledge of the expert (Jim Harding)
is expressed in basic elements rather than complex statements. The personnel knowl-
edge base includes elements such as O lesek’s (flight instructor) license, a list of the

9 I « fl 1 « FIGURE 6.7: Semantic Net of a Personnel Knowledge Base

198 7 7 7 7 PART II Knowledge Creation and Capture

characteristics of each element, and a way to link things together. For example, the
knowledge base would indicate that the insignia on O lesek’s shirtsleeve signifies a
flight instructor.

■ Blackboarding
Imagine bringing a group of experts together in a room with a large blackboard. The
experts work together to solve a problem, using the blackboard as their work space.
Initial data are written on the blackboard for all to see. Each expert has an equal
chance to contribute to the solution via the blackboard. The process of blackboarding
(also called groupware) continues until the problem has been solved.
One im portant assumption of a blackboard system is that all participants are
experts, but they have acquired their own expertise in situations different from those of
the other experts in the group. Because each expert’s experience is unique, no one
need feel either inferior or superior in offering a possible solution. The essence of this
technique is the independence of expertise in an atmosphere that discourages compli-
ance or intimidation.
There are several characteristics of blackboarding:

1. Diverse approaches to problem-solving. The fact that the experts have different
ways of thinking does not prevent them from solving the problem. Each expert
(called a knowledge source , or KS) in this technique is similar to a black box
whose internal thinking process is hidden from other KSs in the meeting. It does
not really matter how each KS thinks through a solution. Each KS can make con-
tributions within the framework of the blackboard.
2. Common language fo r interaction. The KSs participating in this technique should
share a common language (including use of diagrams and charts) so that they can
interact, interpret, and contribute to the final solution. Private jargon or abbrevi-
ated phrases should be discouraged, because they limit the involvement of other
KSs in the solution process.
3. Flexible representation o f information. An expert’s suggestion of an alternative or
a partial solution in graphics versus text is allowable by the software as long as it
can be understood by all experts concerned. No prior restrictions on the informa-
tion limit its content or how it should be represented on the blackboard.
4. Efficient storage and location o f information. Virtually all of the information con-
tributed by the KSs is stored somewhere in the blackboard. To ensure quick
access to an updated version of the blackboard, the blackboard is organized into
regions, each representing a specific type of information. Once the type of infor-
mation required is determined, the KS can go directly to the suitable blackboard
region and scan the information stored in it. Remember that experts do not inter-
act with one another directly. Each KS watches the blackboard, looking for the
right opportunity to contribute to the solution. KSs who see a new change on the
blackboard can decide whether they agree with it, and if not, they can go ahead
and contribute their opinion.
KSs are also motivated by events other than those displayed on the black-
board. For example, a KS might inform the blackboard about an event of per-
sonal interest, which is entered on the blackboard. The blackboard then directly
considers the KS owner whenever that type of event occurs.
 CHAPTER 6 Other Knowledge Capture Techniques i aaa 199
5. O rganized participation. One benefit to blackboard technology is that two or
more experts cannot respond to an event simultaneously. Only one expert at a
time is allowed, which promotes control and gives each KS a chance to mull over
the changes before they decide to contribute. The knowledge developer can be
used to restore order by considering each KS’s request to approach the black-
board with the proposed change.
6. Iterative approach to problem -solving. The solution to the problem domain is
approached step by step, like building blocks from bottom up or from basic to
advanced. In this situation, no single KS has a full answer to the problem. Each
KS refines and adds something of value to someone else’s contribution, and the
whole group moves toward the solution incrementally. A summary of the charac-
teristics of blackboarding is provided in Figure 6.8.

A typical blackboard system consists of three parts: knowledge sources (KSs), the
blackboard, and a control mechanism. The inference engine and the knowledge base
are part of the blackboard system. A knowledge source is the expert—a unique module
that has the knowledge to solve the problem.
The blackboard is a global m emory structure, a database, or a repository that
stores all partial problem solutions and other data that are in various stages of comple-
tion. It serves as a communication medium and triggers a KS into action or controls the
necessary information for the blackboarding process.
B lackboard processing is based on the concept of independent cooperating
experts. Each KS is an independent expert who observes the status of the blackboard
and tries to contribute a higher level partial solution based on the knowledge it has and
how well such knowledge applies to the current blackboard state. In this respect, the
blackboard is useful for structuring complex problem-solving jobs that require multi-
ple experts.
The control mechanism coordinates the flow and pattern of the problem solution.
It monitors the changes on the blackboard and decides on the next action(s) to take.
The problem-solving behavior of the whole process is encoded in the control module
(see Figure 6.9).
The blackboard approach is useful in situations involving m ultiple expertise,
diverse knowledge representations, or uncertain knowledge. It can be particularly
valuable when working with complex applications or prototyping an application.
Unfortunately, blackboard systems have too short a history to provide a good rationale
for advocating their use. This is compounded by a lack of commercial software specifi-
cally designed for building blackboard applications and a shortage of developers expe-
rienced in this area.

FIGURE 6.8: Key Characteristics of Blackboard Systems

Diverse approaches to problem-solving

Common language for interaction
Flexible representation of information
Efficient storage and location of information
Organized participation
Iterative nature of problem-solving
200 7 7 7 7 PART II Knowledge Creation and Capture

«■■■ Implications for Knowledge Management

The knowledge developer has a number of challenges in choosing knowledge capture
tools. First, the technique that best taps the knowledge of the domain expert must be
determined. Certain techniques are ideal for single experts, and others are more suit-
able for multiple experts. Second, advance planning and preparation for knowledge
capture can best be carried out through using the most efficient tools, thereby making
the best use of the expert’s time and resources.
The managerial aspects of tool selection have to do with the organization’s com-
mitment to providing proper training and support for the knowledge developer. The
knowledge developer should be an effective m anager of time. Selecting the right
knowledge capture techniques means saving time and ensuring reliable representation
of the knowledge. The knowledge developer must also be able to communicate with
experts, regardless of their background or expertise. Communication breakdown could
spell disaster for any project, large or small.
Finally, the knowledge developer’s success depends on cultivating good relations
with management in general. In other words, a knowledge developer must be aware of
the organization’s politics, its grapevine, and what steps are necessary to get systems
accepted. The support of a champion could make the entire knowledge process easier.
M anagement support also includes continued commitment to fund the project.
Lack of proper financial support can lead to a low-quality product, an incomplete sys-
tem, or dissatisfied users. In the final analysis, management must see the KM system
project as an investment from beginning to end and gauge its support according to the
gain it hopes to realize.

Summary 1 1 1 1
• On-site observation is a process of observing, interpreting, and recording based
on observation of the expert’s problem-solving behavior while it takes place. The
technique carries with it a certain protocol. It requires more listening than talking
and enables the knowledge developer to seek knowledge within the working
world of the expert. The main problem is that some experts do not like to be
observed. Some question as to the accuracy of the captured knowledge may also
arise. The shuttle process allows room for possible errors.
 CHAPTER 6 Other Knowledge Capture Techniques a«a 1 201
Brainstorming invites two or more experts into an idea generation session in
which discussions are carried out and a variety of opinions are tossed around. The
procedure includes introducing the brainstorming session, presenting a problem
for the experts to consider, prompting the experts to generate ideas, and watching
for signs of convergence. This technique dictates that the expert cannot be cut
short, yet some experts talk too long for comfort. Electronic brainstorming tends
to improve the efficiency with which brainstorming is conducted and evaluated.
Protocol analysis or think-aloud is a knowledge capture method that allows cases
and scenarios to be collected by asking experts to solve a problem while verbaliz-
ing what goes through their heads—stating directly what they think. The expert
speaks out loud whatever thoughts come to mind while answering a question or
solving a problem.
Consensus decision making is used to arrive at a clear agreement regarding the
best solution to a problem. As a tool, it follows brainstorming. Experts converge
toward one or two alternatives for a final consensus. This technique is quite effec-
tive and works well most of the time, but it is a bit tedious and can take hours to
conclude. One of the drawbacks is that options unfairly carry the same weight.
Some participants are uncomfortable with its rigidity.
The repertory grid is used to capture and evaluate the way the expert works
through a solution to a problem domain. The grid is a scale with elements
described by the gradations. Gradations break down the scale to allow for more
accurate rating. Once adopted, the scale should stay the same throughout the
grid, although the terms that describe the rating vary from construct to construct.
One of the benefits of the grid is that it may prompt the expert who sees it on
paper to think more seriously about the problem and how to solve it. On the
other hand, the grid can be difficult to manage.
An alternative to the consensus technique is the nominal group technique or
NGT. It is an ideawriting or idea generation technique used as an interface
between consensus and brainstorming. A panel of experts becomes a nominal
group. The idea is to structure small group meetings so that individual judgments
can be effectively pooled. NGT is time-consuming and tedious. One way to expe-
dite this procedure is to have the experts vote as a group on the set of possible
problems to work on, and later on which solution(s) to use. The tool is ideal in
situations of uncertainty regarding the nature of the problem domain.
The Delphi m ethod is a series of surveys that poll experts concerning a given
problem domain. Questionnaires pool the experts’ responses in order to solve a
difficult problem. The procedure begins by asking a panel of experts to prepare
individual anonymous opinions about a focused problem domain. For the second
round, each expert receives a summary of the results of the first round and is
asked to make a second estimate on the same issue based on the additional infor-
mation provided ir^ the summary. This second step is repeated two or more times,
and then a final surtimary is prepared. The main feature of this tool is anonymous
response, controlled feedback, and statistical group response. A poorly designed
questionnaire can be a serious constraint.
Blackboarding offers a diversity of approaches to problem-solving and a common
language for interaction. It also has the characteristics of flexible representation
of information, efficient storage and location of information, organized participa-
tion, and an iterative approach to problem-solving. The blackboard approach is
useful in situations involving multiple expertise, diverse knowledge representa-
tions, or uncertain knowledge.
202 7 7 7 7 PART II Knowledge Creation and Capture

Terms to K n o w 1 1 1 1
Active participant observation: Switching back and forth Ideawriting: A structured group approach used to
between asking questions and observing the domain develop ideas, explore their*meaning for clarity and
expert. specificity, and produce a written report.
Blackboard: A shared database in which various knowl- Knowledge source: A unique module that has the knowl-
edge sources work together to solve a problem. edge to solve the problem.
Blackboarding: Experts work together in a common work Nominal group technique (NGT): Small group meetings
area to come up with a solution. in which individual judgment can often be effectively
Brainstorming: A n unstructured approach to generating pooled; interface between consensus and brainstorming.
ideas by which two or more experts generate ideas On -site observation: Observing, interpreting, and record-
about a problem domain. ing an expert’s problem-solving behavior as it occurs in
Consensus: Finding a clear agreement on the best solu- the expert’s domain.
tion to a problem. Passive participant observation: Act of only observing the
Consensus decision making: A knowledge developer con- domain expert, in which the knowledge engineer
ducts the exercise after brainstorming to rally the refrains from asking questions as the situation arises
experts toward one or two alternatives and to convey during the observation.
the impression that all of them are part owners of the Protocol analysis: Systematic collection and analysis of
alternative(s). protocols; synonymous with think-aloud protocol.
Control mechanism: Coordinates the flow and pattern of Repertory grid: Knowledge capture tool by which the
the problem solution in a blackboard model. problem domain is classified and categorized around
Delphi method: A series of surveys that poll experts con- the domain expert's own model: a representation of the
cerning a given problem domain. expert’s way of looking at a particular problem.
Electronic brainstorming: A computer-aided approach to Shuttle process: The back and forth aspect of the knowl-
dealing with multiple experts through a network of edge developer’s recollection of an event as it was men-
PCs, which promotes live exchange of ideas between tally registered at the time of the retrospection.
experts; the ideas can be sorted and condensed into an Think-aloud: Knowledge capture method in which the
organized format. expert speaks out loud whatever thoughts come to
Grid: A scale or a bipolar construct on which elem ents mind while answering a question or solving a problem;
are placed within gradations. synonymous with protocol analysis.

Test Y o u r U nderstanding 1 , 1 1
1. Explain in your own words the distinctive features of on-site observation.
What kinds of problems does it pose?
2. Distinguish between:
a. brainstorming and consensus decision making
b. protocol analysis and Delphi method
c. repertory grid and nominal group technique
d. blackboarding and electronic brainstorming
3. How is brainstorming conducted? Provide an example.
4. If no interruptions are allowed or no questions asked while the expert is
answering a question or solving a problem, how does the facilitator control a
protocol analysis session?
5. In what way does consensus decision making follow brainstorming? Be specific.
6. Give an example of your own to illustrate the procedure followed in consen-
sus decision making.
7. “The grid is used to acquire and evaluate the expert’s model —the way the
expert works through the solution/’ Do you agree? What is unique about the
grid? What are some of its drawbacks?
8. In what way is the nominal group technique an alternative to the consensus
technique?
9. Illustrate by an example of your own the procedure followed in NGT.
 CHAPTER 6 Other Knowledge Capture Technique: 203
10. If you were asked to apply the Delphi method involving four experts, what
procedure would you follow? W hat limitations would you expect in this
knowledge capture technique?
11. B lackboarding offers certain characteristics. Explain each briefly. What
makes up the basic model?

Knowledge Exercises 11 * »
1. Divide into small groups of five to seven people. Select a group discussion
leader and a person to record responses. Use the brainstorming guidelines to
conduct a 5-minute brainstorming session on the following topic. Your goal is
to identify creative solutions to the problem:
Employees in large companies often complain that personal worth per-
ception is low. They feel that the company does not overtly reward them for
their contributions or set procedures that allow them to be most productive
and creative.
2. Based on the situation presented in problem 1, complete the following tasks:
a. Brainstorm how the company can reward efforts and increase the percep-
tion of personal worth other than issuing pay increases.
b. Use nominal group technique to find the best solution to the employee
personal worth perception problem. Consider the solutions from the
brainstorming activity and select the “best” solution from that set.
c. Use consensus decision making with the goal of selecting a solution to the
employee personal worth perception problem to which all members of
the group can commit.
3. Publisher KM System:
Knowledge capture uses a variety of tools. In the Publisher KM System
(discussed in Chapter 5), several knowledge capture tools were used.
Knowledge was captured from a series of four interviews with the expert.
The first two interviews encompassed fairly broad questions about the expert’s
decision-making process. The last two interviews elicited more detailed infor-
mation about the expert’s decision process by discussing the outcome of spe-
cific sample case scenario decisions, to which confidence levels were assigned.
In one interview, an attem pt was made at rapid prototyping, but the for-
mat of the table used was difficult for the expert to understand. It was not
used as productively as expected.
An induction table (similar to a decision table) was developed, which
included possible combinations of the five variables and their outcom es—
accept, defer, or reject a book m anuscript (see Table 6.3). Evaluate the
approach taken in the knowledge capture process to determine the rationale
and reliability of the tools used.
4. A total of six knowledge capture sessions were conducted with the expert at
the expert’s location, except for the few occasions when the session required
use of the system.
In the first session, the knowledge developer became familiar with the
domain expert, elaborating on the admissions materials previously provided.
This overview made it possible to establish a basis of knowledge, which
evolved into the concept dictionary. In the second session, the interview was
used to extract basic param eters and then had the expert organize these
param eters according to her view of the domain. This organization was trans-
formed into a visual concept diagram of the domain. The concept diagram
was referred to and refined throughout the entire acquisition process. This
diagram provided a clear direction for the acquisition process. It also became
the foundation for the knowledge representation structure (see Figure 6.10).
204 7 7 7 7 PART II Knowledge Creation and Capture

TABLE 6.3 Induction Table

Author’s Reviewers’ Author’s Publisher’s Confidence
R u le # Market Reputation Comments Requirements Decision Level

0 High Excellent Offer contract 0.90

1 High Average Very good Average or below average Offer contract 0.80
2 High Average Very good A bove average Offer contract 0.85
3 High Average Mixed Average or below average Offer contract
4 High Average Mixed A bove average G et sample chapters
5 High Average B elow average Average or below average Get sample chapters
6 High Average Below average A bove average Reject manuscript
7 High Poor Very good A bove average Offer contract 0.75
8 High Poor Very good Average Offer contract 0.85
9 High Poor Very good B elow average Offer contract
10 Moderate Poor Mixed A bove average Reject manuscript
11 Moderate Poor Mixed Average or below average G et sample chapters
12 Moderate Poor B elow average A bove average Reject manuscript
13 Moderate Poor B elow average Average Reject manuscript
14 Moderate Poor B elow average B elow average Reject manuscript
15 Moderate Average Very good A bove average Offer contract 0.65
16 Moderate Average Very good Average Offer contract 0.75
17 Moderate Average Very good B elow average Offer contract 0.90
18 Moderate Average Mixed A bove average Get sample chapters
19 Moderate Average Mixed Average Offer contract 0.70
20 Moderate Average Mixed B elow average Offer contract 0.75
21 Moderate Average B elow average A bove average Reject manuscript
22 Moderate Average B elow average Average Reject manuscript
23 Moderate Average B elow average B elow average Reject manuscript
24 Moderate Poor Very good A bove average Offer contract 0.70
25 Moderate Poor Very good Average Offer contract 0.80
26 M oderate Poor Very good B elow average Offer contract
27 Moderate Excellent Very good Offer contract 0.90
28 Moderate Excellent Mixed Average or below average Offer contract 0.90
29 Moderate Excellent Mixed A bove average Get sample chapters
30 Moderate Excellent Below average Get sample chapters
31 Low Excellent Very good A bove average Offer contract 0.60
32 Low Excellent Very good Average Offer contract 0.70
33 Low Excellent Very good Below average Offer contract 0.70
34 Low Excellent Mixed Average or below average Get sample chapters
35 Low Excellent Mixed A bove average Reject manuscript
36 Low Excellent Below average Reject manuscript
37 Low Average Very good Below average Offer contract 0.70
38 Low Average Very good Average Offer contract 0.60
39 Low Average Very good A bove average Reject manuscript
40 Low Average Mixed Average or below average Get sample chapters
m ■ a ■ a a
 CHAPTER 6 Other Knowledge Capture Techniques aaaa 205

fl fl 9 1 9 9 FIGURE 6.10: Concept Diagram for Undergraduate Admission

O ther tools employed were process tracing and case scenarios. I used
process tracing by asking the domain expert to provide an example of a typical
admissions decision. From following her decision-making process, I began to
understand how the parameters interacted with one another. This helped me
assign parameters to groupings based on the domain concept diagram. This
also alerted me to the minimum and maximum values that could be assigned to
each parameter. In addition, I was able to derive the minimum requirements
for consideration in which the applicant would be immediately rejected. A list
of parameters is shown in Table 6.4.
In order to capture the expert’s heuristics, I devised a matrix of parameters
and their potential values. This allowed me to vary parameters in relation to
each other. This proved to be a formidable task. Ten multivalued parameters,
each with three or more permissible ranges, produced more scenarios than
could easily be generated via manual means. The difficulty of this task also had
implications for completeness of the rule base; therefore, scenarios could easily
be overlooked. To assist me in generating scenarios, I discovered Logic Gem, a
logic processor software that made life easier for the rest of the capture process.
Questions
a. Based on the nature of the problem domain, evaluate the tools used in
this project.
b. How im portant is listing param eters early in the knowledge capture
phase? Explain.
c. Convert the concept diagram to a concept diagram for Publisher Advisor.
Hint: Focus should be on the author’s past record, standing in the univer-
sity, motivation as reflected in the num ber of books he or she has pub-
lished, and so forth.

Re f e r e n c e s , i "
A iken, Milam, Krosp, Jay, Shirani, Ashraf, and Martin, Aouad, G. F., et al. “Knowledge Elicitations Using
Jeanette. “Electronic Brainstorming in Small and Large Protocol Analysis in a Workshop Environment,”
Groups,” In form ation an d M an agem en t , September C onstru ction s M an agem en t a n d E c o n o m ic s , May 1994,
1994, pp. 141-149. pp. 271-278.
206 7 7 7 7 PART II Knowledge Creation and Capture

TABLE 6.4 ADMIT Parameters

Name Type Values
ENGLISH Number 0-3
SCIENCE String Biology, Chemistry, Physics
LA N G U A G E Boolean Yes/No
US HIST Boolean Yes/No
W ORLD Boolean Yes/No
MATH String Algebra I, Algebra II, Geometry
GRADES String A -A , A-B, B-C, C-C, C-D, D -F
PSAT Number 0-160
SAT Number 0-1600
AP_ESSAY String Excellent, Good, Fine, Poor
TM_ESSAY String Excellent, Good, Fine, Poor
M OTIVATION String Over Achiever, On Target, Under
CH A LLENG E Boolean Achiever
D ESIR E String Yes/No
P .SU P P O R T String Enthusiastic, Positive, Unsure, Negative
EXTR A_C Boolean Excellent, Average, Weak
SCHOOL String Yes/No
DISCIPLINE String Above, Equal, Below
DEM EANO R String Acceptable, Unacceptable
M INORITY Boolean Excellent, Average, Difficult
LEG ACY Boolean Yes/No
FACKID Boolean Yes/No
STAYED Boolean Yes/No
BOARD Boolean Yes/No
SIBENROLL Boolean Yes/No
NEW Boolean Yes/No
Yes/No

C H A N G ES (from Session #4)

1. Rem ove RELIG ION
2. R em ove FORM ER
3. Add NEW
4. Rem ove ERB, SSAT
5. Rem ove ACCSIB (same as
SIBENROLL)
6. Rem ove individual language,
add L A N G U A G E
 CHAPTER 6 Other Knowledge Capture Techniques i ■a a 207
Awad, E. M. Building Expert Systems. Minneapolis, MN: Kambouri, Maria, Koppen, Mathieu, and Vilano, Michael.
West Publishing Company, 1996, pp. 214-238. “Knowledge Assessment: Tapping Human Expertise by
Dixon, Nancy M. C om m on Knowledge: How C om - the Q U ERY Routine," International Journal o f
panies Thrive by Sharing What They Know. Hum an-Computer Studies, January 1994, pp. 119-151.
Boston, MA: Harvard Business School Press, 2000, Lanzing, Jan. “Concept Mapping,” users.edte.utwente.nl/
pp. 127-141. lanzing/cm_home.htm, Date accessed October 18,2002.
El-Sherif, Helmy H., and Tang, Victor W. “TeamFocus— Ono, Ryota, and Wedemeyer, Dan L. “Assessing the
Electronic Brainstorming,” Training and Management Validity of the Delphi Technique,” Futures, April 1994,
D evelopm ent M ethods, vol. 8, no. 3,1994. pp. 289-304.
Garvin, David A. Learning in Action: A Guide to Putting Tiwana, Amrit. The Knowledge Management Toolkit. Upper
the Learning Organization to Work. Boston, MA: Saddle River, NJ: Prentice Hall, 2000, pp. 147-193.
Harvard Business School Press, 1999, pp. 155-165. Ward, Lewis. “Collaborative KM Tools: Putting Customer
Horibe, Frances. Managing Knowledge Workers. New Care Online,” K nowledge Management, April 2001,
York: John Wiley & Sons, Inc., 1999, pp. 57-90. pp. CS1-CS6.
 Knowledge
Codification
Contents

In a Nutshell
What Is Knowledge Codification?
Why Codify?
Diagnosis
Instruction/Training
Interpretation
Planning/Scheduling
Prediction
Things to Remember
Role of Planning
Modes of Knowledge Conversion
How to Codify Knowledge
Codifying Tacit Knowledge
Codification Tools and Procedures
Knowledge Maps
Decision Tables
Decision Trees
Frames
Production Rules
Case-Based Reasoning
Knowledge-Based Agents
The Knowledge Developer’s Skill Set
Knowledge Requirements
Skills Requirements
Implications for Knowledge Management
Summary
Terms to Know
210 7 7 7 7 PART III Knowledge Codification and System Implementation

Test Your Understanding

Knowledge Exercises
References

A ll you have to do is write one true sentence.

Write the truest sentence that you know.
— E r n e s t H e m i n g w a y , ,4 M o v e a b l e Fe a s t

•••• In a Nutshell
In the previous chapter, we talked about knowledge capture and the tools that make
this initial KM system building process possible. The next step is to find a way to codify
and organize knowledge into a form for others to use when needed. Getting the right
knowledge to the right people at the right time is the whole idea behind knowledge
codification (see Figure 7.1).
Codification comes in many forms, depending on the type of knowledge and its
specificity. A common way of displaying knowledge processes is via knowledge maps.
Others refer to the process as storyboarding knowledge. Essentially, the goal is to link
KM projects to identifiable corporate objectives that add value to the company’s bot-
tom line. The overall emphasis is on intellectual capital, company users, and the con-
sumer. The knowledge map diagrams the relationships among people, processes, and
knowledge on a project basis, departm ent-specific or company-wide. This chapter
details the reasoning and mechanics of knowledge mapping, tacit and explicit knowl-
edge codification, and the tools of codification as a step for knowledge transfer.
Knowledge and skills requirements of the knowledge developer are also covered.
In addition to codification justification and process, the chapter highlights the knowl-
edge developer's skill set —the characteristics and criteria for codifying tacit knowledge —
because it originates with people and is interpreted by the people who explain it. The key
point is that it is one thing to know what to codify and how to codify, but it is another thing
to know how to deal with people’s attitudes and motivation during the knowledge capture
phase to ensure reliable knowledge for codification. Without the interpersonal skills, the
question of reliability of the knowledge capture would be severely questioned.

What Is Knowledge Codification?

As an emerging trend, knowledge management systems are slowly succeeding in stan-
dardizing terminology and meanings for the common user. As a matter of procedure,
after knowledge is captured, it is organized and codified in a m anner amenable for
transfer and effective use. Knowledge codification is organizing and representing
knowledge before it is accessed by authorized personnel. The organizing part is usually
in the form of a decision tree, a decision table, or a frame, as we shall see later in the
chapter. Codification must be in a form and a structure that will build the knowledge
base. It must make it accessible, explicit, and easy to access.
From a knowledge management view, codification is converting tacit knowledge to
explicit knowledge in a usable form for organizational members. From an information
system view, it is converting undocumented to documented information. Regardless of
the view, codification is making corporate-specific knowledge (tacit and explicit) visi-
 CHAPTER 7 Knowledge Codification ■■■a 211

Shells, Tables,
Tools, Frames,
Maps, Rules

Knowledge
Capture Knowledge
(creation) / Logical Testing, \
Codification
( User Acceptance j

I fl i 9 ■ C FIGURE 7.1: Knowledge Codification in the KM System Life Cycle

ble, accessible, and usable for value-added decision making, no m atter what form it
may take. This means that:
• Tacit knowledge such as human expertise is identified and leveraged through a
form that delivers the highest return to the business. It may be through knowl-
edge-sharing events, organized directories, yellow pages, or other means that will
connect the ones who need the expertise to the source of expertise.
• Explicit knowledge should be organized, categorized, indexed, and accessed via
the company’s intranet or some other means to make it visible, accessible, and
usable. As conceptualized in Figure 7.2, we cultivate new organizational commu-
nities and keep them aligned around the company’s intranet. This includes the
social, interpersonal, financial, managerial, legal, and marketing systems. Such a
“constellation” encompasses all knowledge roles perform ed within the firm. The
connectors are facilitators that make the linkages among the “islands” viable and
bind the constellation together.
212 7 7 7 7 PART III Knowledge Codification and System Implementation

Facilitators

Boundary Spanners

FIGURE 7.2: Select Communities in the Intranet Knowledge of the Enterprise

Think of a legal knowledge base, where precedents are codified and made explicit
to authorized users —lawyers, judges, and lay people. Such knowledge can be made
available on diskette and CD-ROMs or it can be downloaded from the Internet for a
charge. Precedents may be available as a model, a knowledge map, a simulation, or a
set of rules. Each approach has unique features and limitations. Technology expedites
action, but no form at encompasses the tacit knowledge of attorneys and judges.
Codified knowledge simply makes available articulated legal knowledge.

i i a « Why Codify?
Like a chameleon, organizations constantly adapt to changing situations—changes in
market, changes in customer tastes and preferences, turnover among company special-
ists and experts, and new developm ents in technology. In codifying knowledge, the
resulting knowledge base serves in several im portant training and decision-making
areas. Among the im portant ones are the following.

DIAGNOSIS
A diagnostic KM system is given identifiable information through the user’s observation or
experience. Built into the system’s knowledge base is a list of all identifiable symptoms of
specific causal factors. For a medical diagnostic system, for example, the goal is to identify
the patient’s illness. It starts by assuming that the patient has a particular illness. It proceeds
by reviewing the rules and their actions and asks the patient for additional information
about unique symptoms in order to create a description of the illness. If it cannot prove an
illness exists on the basis of the symptoms, the system takes up another possible illness as
the assumed illness and proceeds in the same fashion. All previous responses by the patient
are retained in working memory so that repetitive questioning of the patient is avoided.

INSTRUCTION/TRAINING
The basic concept of this type of codified knowledge base is to prom ote training of
junior or entry-level personnel based on captured knowledge of senior employees.
Most training KM systems have explanatory capabilities; they display the reasoning
 CHAPTER 7 Knowledge Codification a a i a 213
behind their solutions. Some systems allow students to pose hypothetical “what if” sce-
narios, enabling them to learn by exploration.

INTERPRETATION
Interpretive codified knowledge systems compare aspects of an operation to preset
standards. Typically, they use sensor data to infer the status of an ongoing process or to
describe a given situation. Much of the reasoning is coupled with confidence factors.
For example, one knowledge rule in an auto knowledge advisor reads, “If water tem -
perature is over 50 percent of max and oil pressure is less than 9, then there is .78 cer-
tainty that engine is low on oil.” In such a case, the system flashes a warning message,
“CHECK EN GINE OIL.”

PLANNING/SCHEDULING
A planning KM system maps out an entire course of action before any steps are taken.
The system creates detailed lists of sequential tasks necessary to achieve two or more
specific goals. An example is a system used by the U.S Air Force in the Afghanistan
war. It makes 5-day projections of the missions to be launched and materiel to be used
in reducing the enemy’s military capability over a designated geographical area. Once
a plan is completed, the system refines each step to arrive at the optimal schedule. The
system rejects any out-of-line approach that is not within the prescribed constraints.

PREDICTION
Predictive KM systems infer the likely outcome of a given situation and flash a proper
warning or suggestions for corrective action. The system evaluates a set of facts or con-
ditions and compares it by analogy to precedent states stored in memory. The knowl-
edge base contains several patterns on which predictions are based. These systems are
used in hurricane prediction, for example. They are capable of estimating damage to
real estate, and they can predict the nature of the hurricane by evaluating wind force
and ocean and atmospheric conditions.
In summary, justification for codifying knowledge is. to address the areas where
codified knowledge results in efficient, productive, and value-added knowledge bases
and knowledge applications. It means improving an organization’s competitive advan-
tage by reducing its dependence on human experts, who are expensive, hard to come
by, inconsistent, and mortal. Codified knowledge and the resulting knowledge base
mean allow organizational users to perform at a level closer to that of a human spe-
cialist in the domain. In the least, knowledge bases can be used as training grounds for
junior employees to go through the thought process of the seasoned specialists and
learn from it.

THINGS TO REMEMBER
As a process, knowledge codification is neither easy nor straightforward. There are
things for the knowledge developer to remember before beginning to codify. As shown
in Figure 7.3, they include the following:
• Recorded knowledge is difficult to access, either because it is fragmented or
poorly organized.
• Diffusion of new knowledge is too slow.
• Knowledge is not shared, but hoarded. This has political implications, especially
when knowledge cuts across departments.
• People are oblivious to who has what knowledge.
214 7 7 7 7 PART III Knowledge Codification and System Implementation

New Knowledge
Diffusion Too Slow

Location

Knowledge
/ Not
/ Shared People Do \
f Form / Time '\ Not Know
/ Knowledge \ Where
V Dimension J 1 Knowledge
Resides /
Difficult to
Know Who Knowledge
Has Knowledge Not Shared
Recorded (content)
Knowledge
Difficult to
Access

FIGURE 7.3: Knowledge Dimensions and Bottlenecks

• Knowledge is often
• not present in the proper form (form)
• not available when needed (time)
• not present where the knowledge process is carried out (location)
• not complete (content)

With these constraints, a knowledge developer should plan carefully, as we shall

see later in the chapter.

■■ ■■ Modes of Knowledge Conversion

Before we delve further into knowledge codification, it is im portant to distinguish
between knowledge capture and knowledge creation and how they are treated in the
knowledge codification phase. In knowledge creation, the basic assumption is that
individuals create knowledge and that organizational knowledge creation is the ampli-
fication of the knowledge created by organizational members. In contrast, knowledge
capture is viewed as part of knowledge creation. In knowledge capture, usefulness is
often more important than originality and must be followed by codification and adap-
tation to the organization’s needs.
The literature also looks at tacit and explicit knowledge as the two major types of
knowledge and combines them into four modes of knowledge conversion. As shown in
Table 7.1, they are as follows:
 CHAPTER 7 Knowledge Codification i 1 i ■ 215

TABLE 7.1 Modes of Knowledge Conversion

From To Tacit Explicit
Tacit Socialization Externalization
Explicit Internalization Combination
Source: Adapted from Marwick, A. D. “Knowledge Management Technology,”
IBM Systems Journal, vol. 40, no. 4, 2001, p. 815.

1 fl 1 i 1 fl

• From tacit to tacit knowledge produces socialization, where experience is what a

knowledge developer looks for in knowledge capture. Observation and practice
are two im portant knowledge capture tools.
• From tacit to explicit knowledge is externalizing, explaining, or clarifying tacit
knowledge via analogies, models, or metaphors. Explicit knowledge can then be
stored in knowledge bases or repositories for day-to-day use.
• From explicit to tacit knowledge is internalizing or fitting explicit knowledge into
tacit knowledge. An example is learning to ride a bicycle by practicing again
and again.
• From explicit to explicit knowledge is more or less combining, reorganizing, cate-
gorizing, or sorting different bodies of explicit knowledge to lead to new knowl-
edge. An example of this would be sorting an employee m aster file to see how
many employees are older than 65 years of age, fall between the ages of 50 and
64, and so forth.

Knowledge capture is part of knowledge creation, which requires group work, recog-
nition of the value of tacit knowledge, and support for socialization and internalization.
Knowledge creation also requires experience, which over time leads to expertise. Human
experts become candidates for tapping their tacit knowledge via knowledge capture.

How to Codify Knowledge

The key question in this chapter is how to retain knowledge and its distinctive proper-
ties. Before any codification, an organization must focus on the following:
• What organizational goals will codified knowledge serve? For example, if the
intent is to improve quality sales, then the sales knowledge base should be codi-
fied to improve sales quality. Because codifying puts knowledge in a usable form,
a corporation should have a specific strategic reason for its use. Firsthand knowl-
edge alone is not enough. Different goals for a knowledge management project
require different perspectives. At this point, we need to assess an organization’s
knowledge and needs in order to develop a strategy for fulfilling its unmet needs.
• What knowledge exists in the organization that addresses these goals? It means
starting with a survey of various aspects of knowledge—presence, availability,
transfer, use, and so forth. At this stage, we need to understand the knowledge
and work practices of an organization and then use these insights to help trans-
form the organization into what it intends to be. Such knowledge must be
216 ■ ■ • ■ PART III Knowledge Codification and System Implementation

identified and put into some usable form before sharing it outside the organi-
zation. The goal is to look for relevance of knowledge, not completeness.
Knowledge codification requires specific aims rather than merely making
knowledge available.
• How useful is existing knowledge for codification? As we shall see later, one way
to make existing knowledge useful is via knowledge maps —diagramming the
logic trail to a major decision.
• How would one codify knowledge? That is, what medium or tool would be the
most appropriate for codification? Mapping corporate knowledge is an essential
part of the codification process. To do the job, the knowledge developer must
identify the type of knowledge —rich versus schematic, tacit versus explicit, rule-
based or frame-based, and so forth. What can be done with the knowledge
depends on the level of importance, its type, and how labor-intensive it will be
for proper evaluation.

CODIFYING TACIT KNOWLEDGE

We should know by now that tacit knowledge is complex. It is the type of knowledge
that was developed and internalized in the mind of the human expert over a long time
period. Because of that, it is not easy to codify it in its entirety in a knowledge base or
a repository. The distinctive style of the expert can barely be described or externalized
in a set of rules as part of the codified knowledge. In brief, it requires special knowl-
edge capture skills to codify tacit knowledge effectively.
One conclusion is that certain knowledge is more of an art than a science, and art
is difficult to codify into rules and formulations. To process and fully codify an expert’s
years of intuitive experience would be too laborious and futile. That is why when the
knowledge base project is completed, the best that the knowledge developer can say
about it is “ It’s good enough"; it is neither complete nor final. Maturation with human
knowledge over time continues to be the missing area in codified tacit knowledge. If
you do not believe this, try to explain to a child how to ride a bicycle.
For these reasons, we codify tacit knowledge by relying heavily on a willing expert
with the knowledge and a knowledge developer to capture such knowledge. Dealing
with experts is not easy, but once chosen, a knowledge developer needs to assess the
expert’s style of expression. The expert may be the procedure type who thrives on
explaining by procedure, the storyteller type, the parental type, or the salesperson type
who sells his or her way of solving a problem. The knowledge developer must adopt
knowledge capture and codification methods accordingly.
It should be noted that many organizations lack the transparency of company-
wide knowledge. Critical knowledge is often available, but no one seems to know
where to find it. Because of that, knowledge identification and knowledge capture
improve knowledge transparency, which is a prerequisite for knowledge sharing. To
identify knowledge, one needs to create a codified form of it. In this respect, codifica-
tion means representing knowledge in a useful and meaningful way so that it can be
transferred and shared by any authorized user of an organization.

■■■■ Codification Tools and Procedures

Several different ways of encoding facts and relationships to codify knowledge exist.
Some schemes are graphical, others are tabular, and others are descriptive statements
of facts. Knowledge can be categorized, described, modeled, mapped, or embedded in
 CHAPTER 7 Knowledge Codification i 1 a« 217
rules and recipes. The codification schemes included in this section are knowledge
maps, decision tables, decision trees, rules, case-based reasoning, and intelligent soft-
ware agents. The knowledge developer’s skill set is also covered.
Before getting into codification tools and procedures, it is important to note that
knowledge managers must decide on the business goals that the codified knowledge
will serve and identify existing knowledge appropriate to the goals. They must also
evaluate codified knowledge for usefulness and appropriateness for codification and
the medium for codification and distribution.

KNOWLEDGE MAPS
In tacit knowledge capture, it has been said, “You can’t verbalize unless you visualize.”
One way to codify knowledge is to visualize it via knowledge maps. Knowledge maps
arose from a belief that people act on things that they understand and accept. They
simply do not like to be told what to think. This means self-determined change is sus-
tainable. Instead of telling employees what to do to succeed, knowledge maps are tools
to transform employees into knowledge partners for effective decision making.
A knowledge map is a visual representation, not a repository of knowledge.
Typically, it is a straightforward directory that points to people, documents, and reposi-
tories. It does not contain them, however. The main purpose of knowledge maps is to
direct people where to go when they need certain expertise. Such maps usually recog-
nize explicit and tacit knowledge captured in documents, repositories, and in experts’
heads. Knowledge maps can represent explicit or tacit, formal or informal, documented
or undocumented, and internal or external knowledge. In terms of tacit knowledge,
codification effort is generally limited to locating the person with the knowledge,
pointing the seeker to it, and making the expert available for interaction.
W ithout a “knowledge highway” map to access knowledge locations, finding
knowledge would be difficult. In addition to the guiding function, knowledge maps
may also identify strengths to exploit and missing knowledge gaps to fill. For example,
codifying a corporate strategic plan for a new product may make it clear that the sales
force has been using a successful procedure that was never part of the plan. Not cap-
turing and storing but using knowledge is the aim of knowledge maps. They should be
integrated into knowledge management tools such as intranets and groupware to sup-
port operational processes.

How Knowledge Maps Work

As a process, there are five key elements that work together to ensure a knowledge
map experience:
• The map depicts visually the business issue or problem at hand. It is designed
to be fun, easy to follow, engaging, and clearly reflecting the material to be
covered.
• The pace of the group’s collaborative discussions should be guided by a set of
questions that hopefully create shared knowledge. The process allows each
employee to find out how the process or the problem applies to him or to her and
what the employee should be doing differently.
• Facts are presented to the group to focus on the realities of the problem. They
often come up with the same conclusions that senior leaders do.
218 ■ ■ ■ ■ PART III Knowledge Codification and System Implementation

• The nature of the collaborative discussion among peers should be an open and
trusting environment, facilitated by a coach. Otherwise, knowledge maps will
not work.
• Finally, postsession follow-up activities are reviewed, and conclusions are drawn.

The Building Cycle

Building a knowledge map is easy. Once you know where knowledge resides for a
problem, you simply point to it and add instructions on how to get there. Keep in mind
that quality knowledge must first be available. Then it must be accurate and easy to use
and navigate. A company’s intranet is a common medium for publishing knowledge
maps. Often, the human resources package interface with the company’s intranet pro-
vides knowledge capability. In one partial knowledge map, this author was using
Microsoft Word, but the tab somehow would skip 10 spaces instead of the normal five.
A fter clicking on a special help key, a screen asked him to enter the problem. When he
entered “tab problem,” the response was “call Brad at 4-3775.” Brad happened to be
the Word specialist in the department. He was also the faculty-IT liaison.
A more complex student health compensation map is illustrated in Figure 7.4. The
benefits of this map are as follows:

• Sensible reduction o f time to prepare paperw ork and approval to be reimbursed

fo r medical expenses.
• Reduction o f division among departments. The knowledge map cuts across several
departm ents—departm ent where the employee works, the human resources
department, payroll, company-approved physicians, the hospital, and the health
insurance carrier.
• Helps the employee acknowledge various knowledge sources.

■ fl 9 1 fl fl FIGURE 7.4: Question Knowledge Map for Organizing the Content

Have you Call health insurance at 800 555-1111;

Bill goes directly

to health insurance.

x Health insurance___ If worried, call hospital at 555-6200;

will pay hospital. ask for Art Tremaine in billing.

Have you
checked out? A fW ™ Have rail RfX) 555-1111
ment.

Did you pay

hospital bill? Did reimbursement_____ Check with
reflect in next pay? Joan in HR.
Drop off hospital"
bill to Joan in HR.
Payroll will process
reimbursement.
 CHAPTER 7 Knowledge Codification ■a a a 219
Employees that experience a knowledge map learning process are asked to chal-
lenge their attitudes, seek understanding, and collaborate with others to acquire new
knowledge about products, customers, and internal performance. They can then come
up with new conclusions and internalize specific action steps in dealing with business
issues. They will also discover what it will take to create success for the employer and
for their career paths.
There are different approaches to codifying knowledge into systems. The more
bounded and rule-based the knowledge, the easier it is to embed it into a computer-
based system. This includes knowledge-based systems, expert systems, case-based sys-
tems, and software agents. Each system is explained later in the chapter. It should be
noted that a knowledge system should not be viewed as automating expert tasks but
rather informing about such tasks. A knowledge system is not necessarily an expert sys-
tem either, although we can codify tacit knowledge in an expert system format, using
rules or heuristics.

Select Building Guidelines

A knowledge map is often put together in bits and pieces—some facts are docu-
mented and others are informal. In a large organization, asking the question more than
once can cause waste. FAQs (frequently asked questions) are appropriate when tagged
to a computer-based environment, but FAQs also tend to get longer and more tedious.
Being primarily text-oriented, certain questions are not easy or quick to look up. In
contrast, knowledge maps are generally graphic, more specific, and easier to follow
than FAQs. The focus of a knowledge map is on procedure rather than on fragmented
questions and answers.
To illustrate, revisit Figure 7.4, which shows a procedure employees must follow to
get reimbursed for hospital care. A knowledge map m aker initially follows a trail of
fragmented recommendations, beginning with the human resources departm ent as the
primary knowledge source. Then, he or she talks to another person in charge of health
insurance claims and then follows up with the people in compensation and payroll,
respectively. The widening cascade of referrals eventually links the fragmented knowl-
edge in some sequence that makes sense to employees who expect to be reimbursed
after a hospital stay.
Another popular knowledge map in the human resources area is a skills planner,
where the goal is to match employees to jobs. Once the knowledge map maker has an
idea what knowledge is required of employees, they can be selected for educational
training and other offerings. Building the map requires four steps:
1. D evelop a structure o f the knowledge requirements. For example, the requirement
might be 1 year of experience in teller operations for basic knowledge and
5 years of experience in teller operations plus 2 years of Excel for a junior offi-
cer’s position.
2. Define the knowledge required o f specific jobs. For example, knowledge required
may be 3 years of cost accounting in retailing and a minimum of 1 year of super-
visory experience for entry-level management.
3. Rate employee performance by knowledge competency.
4. Link the knowledge map to a training program fo r career developm ent and job
advancement.
With this in mind, a knowledge map may be set up as a 4-tier knowledge structure
to evaluate employee competency and identify within the structure tacit and explicit
knowledge for each competency. As shown in Figure 7.5, basic knowledge is expected
220 7 7 7 7 PART III Knowledge Codification and System Implementation

Knowledge Structure Knowledge Category

Global Company-wide knowledge Applies to all company employees. For example,

,k competency knowledge of company policies and business practices.

Department or division Applies to all employees within a function or a department.

level competency For example, every employee in the IT department should
be well versed in computer technology.

Working knowledge High-level skills unique to a specific job type.

competency For example, a loan officer to decide whether
to approve a loan.

'' Basic level knowledge Foundation knowledge. For example, ability to send and
Local competency receive e-mail as part of a job.

fl fl fl 1 I I FIGURE 7.5: Knowledge Structure and Categories for Knowledge Mapping

as entry-level knowledge competency. For example, each employee is expected to

know how to send and receive e-mail. The next higher level is working-level knowl-
edge competency, when high-level skills are unique for a specific job type. For example,
a loan officer of a bank should be able to decide whether to approve a loan.
After basic and job-specific knowledge competency comes the division or depart-
ment level knowledge type, where certain knowledge applies to all employees in that
departm ent or division. For example, all IT employees should be familiar with com-
puter technology. Finally, the highest knowledge type competency is company-wide
knowledge, where all company employees should be familiar with company policies
and practices.
The resulting knowledge map is loaded in a computer-based environment, where a
manager building a new team may query the computer database. The manager may ask
the database to “SELECT all candidates with advanced skills above 90 percent of the
knowledge competency for job = senior analyst and location = Richmond, Virginia.”
The system goes to a knowledge repository via the company’s intranet and lists on a
screen (in a m atter of seconds) those who meet the criteria.
A nother popular area where a knowledge map can be used is plotting company
strategy against the com pany’s current knowledge to determ ine the knowledge
required to successfully implement strategy. In Figure 7.6, a company begins by identi-
fying the strategy that it must implement for a new product or a new approach to com-
pete. Next, it specifies the knowledge required to execute the strategy. These two items
establish the strategy-knowledge link before searching within the company to scan for
the strategic and knowledge gaps to plug. Comparing what the company knows to the
required knowledge measures the knowledge gap. In contrast, comparing the “must
d o” strategy to what the company “can d o ” determ ines the strategy gap. In other
words, a knowledge map describes a company’s strategy against the current company
knowledge base and displays the missing knowledge that will ensure successful execu-
tion of the strategy.
Overall, one can conclude that the criteria of a good knowledge map are clarity of
purpose, ease of use, and accuracy of content. Although technology is involved in its
application and availability, technology alone cannot guarantee that the knowledge
map will be accepted or used effectively in an organization. Most maps have a political
catch. They cannot be interpreted as favoring one view over others. They cannot be
perceived as giving away power or influence, either. In the final result, a knowledge
map should represent equity and fairness in the way it is designed. It is not the terri-
 CHAPTER 7 Knowledge Codification ■■■■ 221

FIGURE 7.6: Use of a Knowledge Map in Executing Strategies

SOURCE: Adapted from Tiwana, Amrit. The Knowledge Management Toolkit. Upper Saddle River, NJ:
Prentice Hall, 2000, p. 153.

tory, but it can influence the territory in the way it defines it and represents it. FAQs on
knowledge maps are presented in www.appliedleaminglabs.com/kmaps/faqs.html.

DECISION TABLES
A nother scheme used for knowledge codification is the decision table. A decision table
is more like a spreadsheet; it is divided into two parts: (1) a list of conditions and their
respective values and (2) a list of conclusions. The various conditions are matched
against the conclusions. A simple example that determ ines the discount policy of a
book publisher is shown in Table 7.2.
Note that the answers are represented by a “Y ” to signify “yes,” an “N ” to signify
“no,” or a blank to show that the condition involved has not been tested. In the action
entry quadrant, an “X ” or check mark indicates the response to the answer(s) entered
in the condition entry quadrant. Each column also represents a decision or a rule. For
example, rule 1 states:
IF customer is bookstore and order_size > 6 copies
THEN allow 25% discount
The decision table example provides six decisions and, therefore, six knowledge rules.

DECISION TREES
Another knowledge codification scheme is the decision tree. A decision tree is a hierar-
chically arranged semantic network and is closely related to a decision table. It is com-
posed of nodes representing goals and links that represent decisions or outcomes. A deci-
sion tree is read from left to right, with the root being on the left. All nodes except the
root node are instances of the primary goal. To illustrate, the decision table in Table 7.2 is
shown as a decision tree in Figure 7.7. The primary goal is to determine a discount policy.
The equivalent rule is IF customer is a bookstore and order size is greater than 6 copies,
THEN discount is 25 percent. Several other rules can be extracted from the same tree.
222 7 7 7 7 PART III Knowledge Codification and System Implementation

TABLE 7.2 Decision Table of Publishing Company Discount Policy

Condition Entry
Condition Stub 1 2 3 4 5 6
Customer is bookstore Y Y N N N N
Order size > 6 copies Y N N N N N
Customer is librarian/individual Y Y Y Y
IF Order size 50 copies or more Y N N N
(condition) Order size 20-49 copies Y N N
Order size 6-19 copies Y N
A llow 25% discount X
A llow 15% discount X
A llow 10% discount X
THEN A llow 5% discount X
(action) A llow no discount X X
Action Stub Action Entry
m m m m m m

FIGURE 7.7: Decision Tree Representing Publishing Company Discounts Policy

 CHAPTER 7 Knowledge Codification m■ ■ ■ 223
Decision tables and decision trees are commonly used as a first step before actual
codification. Decision tables are especially useful in verifying rules, a task for which
special software is available. A major advantage of a decision tree is its ability to verify
logic graphically in problems involving complex situations that result in a limited num-
ber of actions. A major drawback of decision trees is the lack of information in the tree
format that might indicate what other combinations of conditions could be included.
This is where the decision table is useful.

FRAMES
A frame is a structure or a codification scheme for organizing knowledge through pre-
vious experience. It handles a combination of declarative and operational knowledge,
which makes it easier to understand the problem domain. A frame represents knowl-
edge about an entity in the real world, such as an employee, a person, or person type.
Frames were developed by Marvin Minsky of MIT, and they are based on the idea
that people use analogical reasoning to build new systems by taking a previous collec-
tion of structures from memory. A fram e can contain inform ation about various
aspects of the situations that people describe.
A frame is like a cookbook recipe: Its “slots” contain both the ingredients for the
recipe and the procedural details (“cook over medium heat,” “toast until brown,”) to
make the data operational or to fill the slots, within or between frames. The idea is to
catalog the requirements for membership of certain elements of a knowledge scheme. In
other words, it is a data structure with a name, a type, and a set of attributes, called slots.
Frames have two key elements:
• A slot is a specific object being described or an attribute of an entity. For ex-
ample, in a personnel knowledge base, some of the slots are “instructor,”
“unique feature of employee verification,” and “training certification.”
• A facet is the value of an object or a slot. As shown in Figure 7.8, the facet
(value) of slot type “instructor” is “Olesek”; “verification” is “training license”;
and so forth.
Figure 7.9 shows a frame describing automobiles. The frame consists of six slots
(attributes): manufacturer, city of origin, year, color, miles per gallon, and owner. One
or more facets are associated with each slot. They include range (set of possible val-
ues), default (value to assume if none were specified), i f needed (how to determine
actual value), and i f changed (what to do if the slot’s value changes). If-needed and if-
changed are called demons or procedures that allow procedural knowledge to be com-
bined with the declarative knowledge part of the frame. They specify the actions that
should take place if certain conditions occur during processing. If-needed means “the
demon lays waiting until needed,” and at that point it is “triggered.”

FIGURE 7.8: A Frame Representing Olesek’s Training License

Object: License
Slot Facet
Instructor Olesek
Verification Training license
U nique feature of verification State seal
Teaching certification In air rescue
224 7 7 7 7 PART III Knowledge Codification and System Implementation

Object: Autom obile

Slot
Manufacturer Range: Default
City of origin Range: Default
Year Range: (1960-2001)
(value cannot be modified)
Color Range: (white, black, burgundy, blue)
If-needed: (examine title or view -autom obile)
Miles per gallon Range: (10-35)
Owner Range: (person’s name)

FIGURE 7.9: A Frame Representing the Object of Automobile

When all the slots are filled with values, the frame is considered instantiated, which
means an instance of the frame is created. The value of the slot could be a string, an
integer, or a pointer to another frame instance.
The problem-solving technique used by a frame-based system is called matching,
in which the values associated with a given entity are matched with the slot values of
frames. With a sufficient match, the frame-based system assumes that an instance has
occurred.
Frames represent knowledge about a particular idea in one place. They provide a
framework or structure, in which new information can be interpreted. They are espe-
cially helpful in dealing with hierarchical knowledge, because their organization of
knowledge makes inferences easy to describe.
Frames can also be linked —one frame may inherit properties of a higher-level
frame. In the example shown in Figure 7.10, a “performance review” is a type of “ses-
sion” with an “informal” nature. The director, attendance, and frequency are attributes
special to the performance review. Relying on the concept of “inheritance,” the child
frames of “session” and “m anager” are a specialized type derived from the more gen-
eral parent frame. Because of this relationship, one can assume that the child frames
inherit attributes of the parent frame. The “child” may also have properties of its own
not shared by the parent. The inheritance mechanism means only information specific
to a given frame will be specified. Knowledge-based systems using frames determine
the particular process for operating inheritance relationships.

PRODUCTION RULES
A popular form of tacit knowledge codification is based on production rules, com-
monly known as rules. Rules represent the major elements of a modular knowledge
codification scheme, especially when using specialized computer-based software pack-
ages. Rules are conditional statements that are easy to understand and write; they spec-
ify an action to be taken if a certain condition is true. They also express relationships
between param eters or variables. In expert systems vocabulary, production rules are
also called premise-action, cause-effect, hypothesis-action, condition-action, test-result,
I F . .. THEN, or I F . .. T H E N . . . ELSE. In the latter form, IF is the rule’s antecedent,
THEN is some consequent, and ELSE is some other consequent.
 CHAPTER 7 Knowledge Codification a a a a 225
Child Frame
Session
Type
Nature: Informal
Time: 10:10 a .m .
Place: Office

Parent Frame
Performance Review
Type Is a Session
Nature: Informal
Director Is,a Manager
Attendance: ?
Frequency: ?

Child Frame
Manager
Type Is a Conference
Seniority: Member of Evaluation
Review Committee

FIGURE 7.10: Set of Interlinked Frames

The basic idea of using rules is to codify tacit knowledge in the form of premise-
action pairs as follows:
Syntax: IF (premise) THEN (action)
Example: IF income is “average” and pay_history is “good”
THEN recommendation is “approve loan”
Knowledge-based rules differ from the traditional if-then programming statements.
In knowledge-based systems, rules are relatively independent of one another and are
based on heuristics or experiential reasoning rather than algorithms. Rules can carry
some level of uncertainty. A certainty factor is represented as CNF or CF and is synony-
mous with a confidence level, which is a subjective quantification of an expert’s judg-
ment. In the preceding example, the rule means: “IF the applicant’s income is average
and pay history, which implies a previous loan, is good, THEN the recommendation is
approval of the loan at the 80 percent level of confidence.” Eighty percent is the default.
Figure 7.11 specifies the structure of a rule, which consists of a premise and an action. The
premise has three expressions or clauses, each of which has an attribute, an object, and a
value. The action has one expression or a clause with an attribute, an object, or a value.

A Premise
The IF . . . TH EN structure of the rule language is represented by prem ise and
action clauses or statements. The premise is a Boolean (yes/no, true/false) expression
that must be evaluated as true for the rule to be applied. It is composed of one or more
statem ents separated by AND or O R connectors. A specialized software package
226 ■ ■ i ■ PART III Knowledge Codification and System Implementation

Premise: IF Cloud site is black and wind force > 35 mph and
wave height > 18 feet
Action: THEN There is .6 evidence that classification of storm is
hurricane.

Attribute Object Value

Premise IF Site Cloud Black
Force Wind > 35 mph
Height Wave > 1 8 feet
Action: TH EN Classification Storm Hurricane

FIGURE 7.11: Structure and Elements of a Rule

processes a rule by testing its premise for truth or falsity. The test must be satisfied
before any action of a rule can be executed. Examples of a rule premise are as follows:
If G PA is “high” THEN . . .
IF car_motor is “bad” or car_body is “fair” THEN . . .
IF income > $50,000 and assets > $300,000 or payment_history is “good”
TH EN ...
In multiple premises, processing proceeds from left to right. AND has a higher
precedence than OR. Parentheses may be used to change or specify precedence. The
default order of precedence is X and /(divided by), + and —, left association. For exam-
ple 2 + 3 X 5 - 6/8 X 7 is evaluated as ([2 + ( 3 x 5)] - (6/8 X 7)}.

An Action
The action part of a rule is the second component, separated from the premise by
the keyword THEN. The action clause consists of a statement or a series of statements
separated by ANDs or commas (,) and is executed if the premise is true. It could be
either a list of commands to be carried out if the premise evaluates to true or it could
be a Boolean expression that evaluates to true whenever the premise is true. If any part
of the premise were false, then the action would not be true or would be true at a spe-
cific level of uncertainty. Examples of rule action include the following:
IF owns auto = “false” THEN potential_insurance = “high”
IF income > $100,000 THEN federal_tax_rate > .35
IF loan holder - “late payer” THEN add “penalty = .5” to premium
For multiple action clauses, statements must be linked with AND or commas. For
example:
IF GPA - “H IG H ” and SAT > 1400
THEN admission = true and financial_aid = $6,000
In this example, two values are assigned to admission with financial aid. Two
actions are also fired. In most business-oriented knowledge-based systems, rules are
used almost exclusively. They are a natural expression of “what to do” knowledge or
“how to do” knowledge. They are easy to understand and to represent the expert’s
know-how. Their main drawback is the way they are expressed at a fine level of detail.
Such fine-grained representation makes the know-how involved difficult to under-
stand. Also, the amount of knowledge that can be expressed in a single rule is limited.
This limitation depends largely on the shell. For example, EXSYS Corvid (a knowledge-
 CHAPTER 7 Knowledge Codification aaaa 227
based software package) allows as many as 126 conditions in the IF and the THEN
parts of a rule.

Loan Advisor—An Illustration

To follow up on the partial loan advisor represented in a decision tree in Figure 5-2,
the partial tree can easily be converted into a set of rules using EXSYS Corvid, as follows:
lemployment = yes
2 em p jen g th = 4_to_9,10_to_18, over_19
3 [salary] > = 15000
4 credit_history = Good, Fair
employment = yes
AND em p jen g th = 4_to_9,10_to_18, over_19
AND [salary] > = 15000
AND credit_history = Good, Fair
5 [expenses] <(.75*([salary]/12))
6 —> [approve] = 1
7[approve] ! = 1
8 —> [deny] = 1
9 > RESULTS
Rule-based systems made their debut when expert systems were popular in the
1990s. Expert systems represent human knowledge by a formalized rule-based pro-
gramming setup. They play a limited role in knowledge codification. From a practical
point, the more unambiguous and rule-based tacit knowledge is, the more easily it can
be embedded in an expert system environment.

Role of Planning
For some time, writing rules was viewed as being synonymous with expert systems
development. Later, complex projects required greater coordination and organization.
A lack of planning was reflected in poor quality solutions, incomplete answers, or in-
efficient systems. Planning is an important part of the knowledge management develop-
ment process. It ensures proper transition from knowledge capture to knowledge codi-
fication, verification, validation, and im plem entation. For example, in knowledge
codification, knowledge developers sort out the rules, decide on their sequence and con-
tent, determine which rules belong in which modules, and so on. A knowledge devel-
oper is also responsible for planning the development and execution phases and how
the user will interface with the resulting KM system —activities that are difficult to
carry out without a plan.
In knowledge-based systems, planning involves the following aspects:
• Breaking the entire KM system into modules or manageable components
• Looking at partial solutions
• Linking partial solutions through rules and procedures to arrive at final solutions
that make sense
• Deciding on the programming language
• Selecting the right software package
• Arranging for the verification and validation of the system
• Developing user interface and consultation facilities before installing the KM system
• Promoting clarity and flexibility
228 7 7 7 7 PART III Knowledge Codification and System Implementation

• Reducing unnecessary risks

• Making rules easier to review and understand
Unfortunately, plans can go awry, which means that the knowledge developer must
monitor the results, compare plans with system results, assess the discrepancies, and
make the necessary changes to ensure system integrity. Without this type of monitor-
ing, the value of planning is lost.
Role of Inferencing in Knowledge-Based Codification
In an expert system environment, once the knowledge base is built, the system
begins to make inferences. Inferencing means deriving a conclusion based on state-
m ents that only imply that conclusion. The program that manages the inferencing
strategies for controlling the application of knowledge is the inference engine. It is the
reasoning component that decides when certain inference rules should be accessed or
fired (executed). Reasoning is the process of applying knowledge to arrive at solutions.
To reason is to think clearly and logically, to draw reasonable inferences or conclusions
from known or assumed facts. It works through the interaction of rules and data.
To illustrate inferencing, read the following sentence:
Brothers and sisters have I none ,
But that m an ’s father is my father’s son.
Drawing the conclusion, “That man is my son,” may be viewed as an inference
from the riddle, based on our knowledge of relationships between brothers and sisters,
parents and children, and so on.
Take an example related to reasoning:
The victims were stabbed to death in a dance hall in the mall. The suspect was on
a nonstop flight to Chicago when the murder occurred.
You would probably conclude that the suspect was innocent. This typical day-to-
day reasoning illustrates two important phenomena:
1. Reasoning depends not just on the premise but also on general knowledge. In the
example, the suspect could not be in two places at the same time. There are no
dance halls on board flights to Chicago (or to anywhere else).
2. People generally draw informative conclusions. Although several valid conclu-
sions may follow from any set of premises, most of them are probably trivial.
No one can say precisely what mental processes lead to the conclusion that the sus-
pect is innocent, because humans can only observe the results of another’s conscious
thoughts. In any case, people take for granted that deduction depends on a mental logic
containing rules of inference based on experience.
Although rule-based systems look straightforward, their application is somewhat
restricted. They work when one knows the variables in the problem and can express
them in hard terms (numbers, dollars). The rules to apply must cover all the variables,
and they cannot overlap. Before the rule-based system is operational, the rules must be
validated to ensure reliability. Table 7.3 summarizes the relative fit of rule-based sys-
tems in a KM environment.

CASE-BASED REASONING
One type of reasoning that is a promising candidate for knowledge m anagement is
case-based reasoning or reasoning by analogy. A case is knowledge at an operational
level. It is an episodic description of a problem and its associated solution. Case-based
 CHAPTER 7 Knowledge Codification • ■■■ 229

TABLE 7.3 Relevant Fit of Rule-Based System in a KM Environment

Attributes Limitations in a KM Environment
Flexibility Rule-based systems are OK for small systems, but inflexible
in bases with complicated problems requiring new variables.
Scalability With maturity over time, rule-based systems require change
in rules, which means working with the expert again.
Speed o f developm ent Rule-based systems can be developed quickly, provided you
have a willing expert. Unfortunately, knowledge capture
is very time-consuming.
D ependence on domain expert Knowledge capture comes in stages, often requiring several
drafts before the system is ready for final testing.
Response time A s rule-based systems increase in complexity, response
time usually takes a hit. This means upgrading the hard-
ware and software, which can be costly.

reasoning or CBR is reasoning from relevant past cases in a manner similar to humans’
use of past experiences to arrive at conclusions. An expert, for example, relies on previ-
ous cases for solving a current problem through a process of recalling the closest cases
in memory, making inferences based on comparisons of the cases, and asking questions
when inferences cannot be made. From an algorithmic view, CBR is a technique that
records and documents cases and then searches the appropriate cases to determ ine
their usefulness in solving new cases presented to the expert (see Figure 7.12).
Day-to-day examples of CBR are numerous:
• A doctor sees a patient with unusual symptoms. The doctor remembers previous
patients with similar symptoms and prescribes the old diagnosis as a solution.
• An auto mechanic facing an unusual electrical problem is likely to recall other
similar cases involving the same year or make and consider whether the old solu-
tion corrects the current problem.
• When updating a corporate pension fund, an estate-planning attorney relies on
previous pension fund experience to expedite the completion of the current one.
Referring to old cases is especially advantageous when dealing with recurring situ-
ations. Because no old case is exactly the same as the new one, old solutions must often

FIGURE 7.12: General CBR Process

230 7 7 7 7 PART III Knowledge Codification and System Implementation

be adapted to fit a new case. In every case, some reasoning and some learning, are
involved. The success rate of problem solving improves with recurring CBR.
One obvious advantage of CBR over rule-based systems is its time savings.
Knowledge developers can spin their wheels in manually capturing knowledge from
human experts for building the KM system. In CBR, less time is needed to retrieve rep-
resentative cases for performing knowledge development. CBR allows the reasoner to
come up with solutions to problems quickly, without having to start from scratch. CBR
is most advantageous when rules are difficult to formulate and when cases are readily
available, particularly when the domain expert is uncertain, not available, inconsistent,
or uncooperative.
A nother advantage is its maintenance requirements. In rule-based systems, the
cost of maintaining and enhancing the system stem from the difficulty in structuring
rules into a workable system. In CBR, maintenance entails simply adding more cases
that offer features similar to the case in question.
CBR has a downside, as well. Using CBR means rigorous initial planning that
takes into account all possible variables. Also, because cases age, it is almost mandatory
that recent cases be incorporated into the case base on a regular basis. Although a case
provides accurate and consistent documentation, it lacks the dynamics or the best deci-
sion based on maturation that a human expert can provide.
The goal of CBR is to bring up the most similar historical case that matches the
current case —not an easy task. For example, in the analogy “a flashlight battery is like
a reservoir,” the size, color, shape, or make of the battery is not the relevant aspect.
The relevant aspect is the fact that both the battery and a reservoir store and pro-
vide energy as power. In this situation, only relationships dealing with these two attrib-
utes would be useful in this analogy.
Another problem is the difficulty in mechanizing a person’s insight and ability to
see an analogy between a previously encountered case and a current case. No general
representation procedure prescribes how to reason by analogy. As a consequence, rea-
soning by analogy tends to introduce errors, biases, and inconsistencies into the formu-
lation of rule-based systems. To improve the usefulness, and reliability of rule-based
systems, a better understanding of human reasoning is needed.
A third problem is m aintenance of appropriate indexes. Adding new cases and
reclassifying the case library most often expand knowledge. A case library broad
enough for most CBR applications requires considerable database storage with an
efficient retrieval system. CBR can be fully automated, depending on how much cre-
ativity is required to derive solutions and how much complexity is involved.
Regardless of the level of automation, the biggest factor in making CBR work is its
case library. Table 7.4 uses the attributes in Table 7.3 to summarize the limitations of
CBR in a KM environment.

KNOWLEDGE-BASED AGENTS
In software language, an intelligent agent is a program located in a computer-based
system that is capable of performing autonomous action in a timely fashion. It is not
intended to replace hum an intelligence, but to support it. In addition to acting in
response to their environm ent, intelligent agents are capable of exhibiting goal-
directed behavior by taking initiative. They can also be programmed to interact with
other agents (or humans) by some agent communication language.
In terms of knowledge-based systems, an agent can be programmed to learn from
the user behavior and deduce future behavior in order to assist the user. Because they
can take some control, agents need to have access to decision knowledge. Such knowl-
 CHAPTER 7 Knowledge Codification m■ a ■ 231

TABLE 7.4 Attributes of CBR in a KM Environment

Attribute Lim itations in a K M Environment
Flexibility A s the number o f attributes increases, CBR begins to show
degradation and inaccuracy, especially when there is
interaction among multiple attributes.
Scalability Attributes are not easily scalable. A ll future attributes must
be carefully defined for accuracy.
D ependence on domain expert A n expert must be employed to ensure attribute matching
and retrieval.
Response time Adding more cases could result in degradation in the system’s
response time.
Solution accuracy A s more cases are added to the case base, accuracy improves.
----------------------------------------- a ■ ■ ■ ■ a ------------------------------------------

edge can be about the user as a personal assistant or the domain, as is the case in
expert systems.
Information agents are unique in the way they search on behalf of their users or
other agents. The search involves evaluating, retrieving, or integrating inform ation
from multiple information sources (for example, knowledge bases). They learn user
preferences and provide appropriate information based on these preferences. They can
autom atically execute special tasks such as filtering unw anted e-mail messages. In
essence, they learn by “looking over the shoulder” of the user while he or she is work-
ing. They can provide suggestions when the user reaches a useful level of success in
their predictions.

■■■ ■ The Knowledge Developer’s Skill Set

To perform knowledge codification, a knowledge developer is expected to have vari-
ous technical, procedural, and interpersonal skills. Some skills are taught; others are
innate. Table 7.5 summarizes the knowledge and skills requirements of the knowledge
developer. The literature is still sketchy on this subject. The list is culled from a review
of the research and a brief survey of practicing knowledge developers in two select
organizations. O ther surveys being conducted will help standardize the job require-
ments of this important job.

KNOWLEDGE REQUIREMENTS
Computer Technology
The knowledge developer should be well-versed in hardware and software, how
operating systems work, and issues regarding legal matters, security, testing, documen-
tation, and maintenance. A general background in science and computer architecture
is helpful.

Domain-Specific Knowledge
The knowledge developer must be familiar with the nature of the problem, the
business of the user organization, and the factors surrounding the problem domain,
such as the organization’s political climate, level of m anagem ent support, and the
users’ computer literacy. For example, if the problem domain is mortgage loan analysis,
the knowledge developer should have a basic understanding of lending practices and
232 77 7 7 PART III Knowledge Codification and System Implementation

TABLE 7.5 Knowledge and Skill Requirements of the Knowledge Developer

Knowledge Required Skills Required
• Computer technology and operating systems • Interpersonal communication
• Domain-specific knowledge • Ability to articulate the project’s rationale
• Knowledge repositories and data mining • Rapid prototyping
• Cognitive psychology
a i fl ■ a i

procedures before approaching the loan specialists. A knowledge developer who

builds knowledge bases for oil exploration would probably be the wrong choice for this
assignment.

Knowledge Repositories and Data Mining

Among the technical prerequisites for a knowledge developer are familiarity with
knowledge repositories and what the market offers in terms of off-the-shelf packages
for knowledge management, data warehousing, and data mining.

Cognitive Psychology
The knowledge developer should have a basic familiarity with the principles of
cognition, the understanding of knowledge in general and of how human experts think
or reason. Cognitive skills are im portant in capturing and codifying complex, ill-
defined, and judgmental knowledge and understanding the domain expert’s thought
processes such as decision making under conditions of uncertainty. Background in this
field improves communications, interviewing, and listening skills.

SKILLS REQUIREMENTS
Interpersonal Communication
The most pervasive factor in a knowledge developer’s success is the ability to com-
municate. If he or she cannot understand and relate the expert’s intent or meaning
accurately, the reliability of the resulting knowledge base will be questionable.
Persons with poor interpersonal skills cannot be effective knowledge developers;
their deficiency cannot help but detract from their relationships with experts, users,
management, and others. This also applies to domain experts. They must enunciate
concepts and procedures in simple, clear terms so that the knowledge can be codified
accurately. Interpersonal skills enable one to work with people to achieve common
goals or to help others work together as a team. Interpersonal behavior is closely
related to personality and dictates a person’s management style.

Ability to Articulate the Project’s Rationale

Part of the knowledge developer’s job is to discuss project feasibility, justify invest-
ment in the project, and sell the proposed change to management. Persuasive articula-
tion of the rationale is developed with experience. Actually, this ability is important in
working with management at all levels.

Rapid Prototyping Skills

The knowledge developer will occasionally sit down with the user or the specialist
with the tacit knowledge to get feedback on how the knowledge base is shaping up.
Rapid prototyping is an iterative process in which the knowledge developer shows
 CHAPTER 7 Knowledge Codification ■a i a 233
what the knowledge base looks like on the basis of the knowledge codified to date. The
knowledge developer can make changes to the system on the spot in the presence of
the audience. Rapid prototyping requires experience and quick thinking, and it pro-
motes respect and trust between the knowledge developer and the expert or the user,
respectively.

Personality Attributes
When it comes to personality attributes, knowledge developers should exhibit the
attributes described here.

• Intelligence. An intelligent knowledge developer develops KM systems efficiently.

He or she is constantly learning from experience and from exposure to the vari-
ous specialists and users. Intelligence enables him or her to keep abreast of
advances in technology and to learn and think quickly and well, especially when
various procedures must be integrated.
• Creativity. A creative mind enables a person to generate ideas and react quickly
to new methods for solving problems. Creative knowledge developers are
more adept at selecting appropriate reasoning strategies and ways of captur-
ing knowledge.
• Tolerance fo r ambiguity. The willingness to probe further and further into an
unclear issue or procedure until it is clear and usable demands a degree of
patience that some people do not possess. This process takes time, and the expert
or knower may require much encouragement.
• Realism. Experienced knowledge developers understand that not every expert
they deal with will be cooperative or motivated and that some projects will
be more onerous than others. These knowledge developers tend to be more
tolerant.
• Cognitive ability. The ability to reason by analysis, association, and inference char-
acterizes knowledge developers. They are able to “see through things,” to perceive
the reality of a situation.
• Persistence. Knowledge developers have the determination to try new things and
maintain their commitment to a project. Self-confidence is a component of this
persistence. When dealing with ill-structured problems, the knowledge developer
must remain positive, saying, “I can do it if I keep at it.” The poem in Box 7.1
addresses the importance of a positive state of mind.
• Logical thinking. The ability to learn new technology and continue to improve
the codification process is essential to a knowledge developer. When codify-
ing knowledge, he or she must be willing to argue reasonably in order to
resolve inconsistencies. Otherwise, the resulting system will be illogical and
unreasonable.
• G ood sense o f humor and optimism. Light-hearted, cheerful people are easier
to work with and make attractive collaborators.

Job Roles
The knowledge developer plays a variety of roles, depending on the nature of the
knowledge project, the type of knowledge, and the organization’s climate. Some of the
key roles are those of change agent, investigator, diplomat, architect, builder, tester,
accountant, and salesperson. He or she may also be a project planner, contractor, moti-
vator, quality inspector, and adept at conflict resolution. These roles relate to personal-
ity attributes as well as knowledge and skills requirements.
234 7 7 7 7 PART III Knowledge Codification and System Implementation

I 1 1 1 fl 1 BOX 7.1

ALL IN THE STATE OF THE MIND

If you think you are beaten, you are; Think big and your deeds will grow;
If you think you dare not, you don’t; Think small and you’ll fall behind;
If you would like to win and don’t think you can, Think that you can and you will.
It’s almost a cinch you won’t. It’s all in the state of the mind.
If you think you’ll lose, you’re lost; If you think you are outclassed, you are;
For out in the world you'll find You’ve got to think high to rise.
Success begins with a fellow’s will. You’ve got to be sure of yourself before
It’s all in the state of the mind. You can ever win a prize.
Full many a race is lost Life’s battles don’t always go
Ere even a step is run, To the stronger or faster man;
And many a coward fails But sooner or later, the man who wins
Ere even his work is begun. Is the fellow who thinks he can.

SOURCE: Awad, E. M. Expert Systems. Minneapolis, MN: West Publishing Co., 19%, p. 129.

a i I i a a-

■ Implications for Knowledge Management

In traditional terms, management means planning, organizing, directing, and coordi-
nating the efforts of people, procedures, and technology in an organizational setting.
The work of knowledge codification involves more than codification schemes or tools.
It means planning the codification phase, planning the use of specific tools within the
framework of a programming environment, and making the most efficient use of the
knowledge developer’s and the expert’s time and talent.
In addition to the front-end phase of m anagem ent, the knowledge developer
should sense the user's expectations and determ ine how the user can be part of the
codification phase of the knowledge management system development life cycle, espe-
cially during knowledge codification. Partnership or cooperation can be accomplished
through rapid prototyping or inviting the user to join in the day-to-day routines of
codification. In addition, the knowledge developer needs to foster the relationship with
the champion, whose role continues to be vital to the future acceptance and success of
the knowledge-based system.
Finally, in preparation for the detailed writing of the coding, rules, and procedures,
the knowledge developer would be well advised to begin promoting the goodness of fit
and exciting features of the knowledge management system to the key person in the
area where the system will be installed. In fact, the selling effort is best begun early in
the process, especially when the knowledge-based system is likely to take months
rather than weeks to build. The ultimate goal is user acceptance and willing commit-
ment to long-term use of the system.
From the human resources and career path view, encouraging signs suggest a
career for the knowledge developer. Building knowledge management systems can
provide higher pay and the opportunities for doing more creative work. Professionals
in the IT field who want to move to a career in knowledge development should deter-
 CHAPTER 7 Knowledge Codification aai a 235
mine the level of support for building knowledge m anagem ent systems in their
employing organization. They may be able to stay with their current employer as a
knowledge developer.
Given the increasing demand for knowledge developers, an organization can now
consider long-range training programs, by which the talent for knowledge develop-
ment may be tapped from its existing pool of IT and other professional areas in the
business. Another alternative is for the organization to support academic IT programs
that offer courses in knowledge management or knowledge engineering.
Knowledge development allows a firm to preserve its intellectual capital for problem-
solving and competitive advantage. A learning company understands and capitalizes
on its human resources. To incorporate knowledge m anagement without knowledge
development as the core competency would be like building automobiles without quali-
fied assemblers.
The knowledge developer is a person who is committed to this evolving technol-
ogy and who pursues a goal of building quality systems. Without quality assurance, reli-
ability, and integrity, the whole process remains questionable. For knowledge develop-
ment to become an established area of system building, organizations must provide
rewards sufficient to minimize voluntary resignations. Knowledge developers are
expensive to train and hard to replace. Motivating technical people requires regular
performance, monitoring, and feedback for reinforcement. Providing opportunities for
advancement, creativity, utilization of abilities, recognition, higher compensation, and
the like promotes employee loyalty to the firm and its objectives. This level of concern
for people counts in the final results.

S u mma r y 1 ■■1
• Knowledge codification is organizing and coordinating knowledge before the
user can access it. It must be in a form and a structure meaningful for access at
any time, from anywhere, by any authorized person.
• Codification is a prerequisite to knowledge transfer. The resulting knowledge
base serves several useful areas such as diagnosis, scheduling, planning, instruc-
tion, and training. However, there are bottlenecks such as difficulty in accessing
knowledge and slow diffusion of new knowledge. Bottlenecks can also be created
when knowledge is not in the proper form, not available in a timely manner,
incomplete, or not present at the right location.
• The four modes of knowledge conversion tag the four possible relationships
between tacit and explicit knowledge and the resulting outcomes of socialization,
externalization, internalization, and combination.
• Before codification begins, an organization must address several questions: What
organizational goals will codified knowledge serve? What knowledge exists in the
organization that addresses these goals? How useful is existing knowledge for
codification? How would one codify knowledge?
• Codifying tacit knowledge is complex and is more of an art than a science. Several
different ways of encoding facts and relationships to codify knowledge exist.
They include knowledge maps, decision tables, decision trees, frames, produc-
tion rules, and software agents. Each has a unique approach to knowledge
codification.
• A knowledge map is like a knowledge highway to access knowledge location. It
identifies strengths to exploit and missing knowledge gaps to fill. Building the
map requires three steps: Developing a structure of the knowledge requirements,
236 7 7 7 7 PART III Knowledge Codification and System Implementation

defining the knowledge required of specific jobs, and rating employee perfor-
mance against knowledge requirements.
A decision table scheme is divided into two parts: (1) a list of conditions and
their respective values and (2) a list of conclusions. In contrast, a decision tree is
a hierarchically arranged semantic network that is closely related to a decision
table. It is composed of nodes representing goals and links that represent deci-
sions or outcomes.
A frame is a structure for organizing knowledge. It handles a combination of
declarative and operational knowledge and makes the current problem domain
easier to understand. It represents knowledge about an entity in the real world
and catalogs the membership requirements of certain elements of a knowledge
scheme. The two key elements of a frame are a slot, which is a specific object
being described or an attribute about an entity that holds a value or facets, and a
facet, the value of an object or slot. When all slots are filled with facets, an
instance of a frame is created.
Rules represent the major elements of a modular knowledge codification scheme.
They are conditional statements that specify an action to be taken, if a certain
condition is true. They express relationships between parameters and represent
the human expert’s tacit knowledge in the form of premise-action pairs.
Knowledge developers are more than talented programmers or knowledge cap-
ture people. They are professionals with skills in tapping human expert knowl-
edge and using tools to codify that knowledge. They perform several key func-
tions. They identify knowledge-based applications, develop project proposals,
capture knowledge, oversee testing and implementation, perform user training,
and maintain knowledge systems.
Knowledge development requires certain knowledge and skills. Knowledge
requirements include computer technology knowledge of the domain, program-
ming knowledge using special tools, and cognitive psychology. Skills requirements
include communication and interpersonal skills.
The personality attributes of the knowledge developer typically include intelli-
gence, creativity, patience, realism, persistence, and a good sense of humor.

Te r ms t o Kn o w * 1 * »
Action: Consists of a list of commands to be carried out if Certainty factor: A measurement of belief or a subjective
the premise evaluates to true, or a Boolean expression quantification of an expert's judgment and intentions.
that evaluates to true whenever the premise does; con- Decision table: A list of conditions with their respective
clusions to be drawn with some appropriate degree of values matched against a list of conclusions.
certainty or instructions to be carried out. Decision tree: A hierarchically arranged semantic net-
Agent: A lso called intelligent agent; a bit of intelligence work that is closely related to a decision table; com-
that represents specific attributes to other agents in a posed of nodes representing goals and links that repre-
knowledge base or on a network. sents decisions or outcomes.
Case: Knowledge at an operational level; episodic Facet: The value of an object or a slot.
description of a problem and its associated Frame: A structure or a codification scheme for organiz-
solution. ing knowledge through previous experience.
Case-based reasoning: A m ethodology that records and Inference engine: The “brain” of a knowledge-based
documents previous cases and then searches the rele- system; a cluster of computer programs that
vant case(s) to determine their usefulness in solving a coordinate reasoning and inferencing based on
current problem; computer systems that solve new the rules of the knowledge base to produce the
problems by analogy with old ones. solution or advice.
 CHAPTER 7 Knowledge Codification i 1 i i 237
Inferencing: Deriving a conclusion based on statements Production rule: Knowledge codification method in
that only imply that conclusion. which knowledge is formalized into rules.
Inheritance: An instance of a particular class is assumed Rapid prototyping: Spontaneous, on-the-spot, iterative
to have all the properties of more general classes of approach to building knowledge-based systems;
which it is a member. an iterative process by which the knowledge devel-
Instantiate: Create a frame by filling its slots. oper shows the domain expert or the user what
Knowledge codification: Organizing and coordinating the system looks like based on the knowledge cap-
knowledge in a form and a structure meaningful to tured to date.
the user. Reasoning: The process of applying knowledge to arrive
Knowledge developer: A specialist responsible for plan- at solutions based on the interactions between rules
ning the developm ent and execution phases of a knowl- and data.
edge management system. Rule: A conditional statement that specifies an action to
Knowledge map: A directory that points to people, docu- be taken if a certain condition is true; a formal way of
ments, and repositories. specifying a recommendation directive, or strategy,
Logic: The study of reasoning; the scientific study of the expressed as IF (premise).
process of reasoning and the set of rules and proce- Rule of inference: A deductive structure that determines
dures used in the reasoning process. what can be inferred if certain relations are taken to be
Matching: The values associated with a given entity true; facts known to be true and used to derive other
paired with the slot values of the frames to signify that facts that must also be true.
an instance has occurred. Slot: A specific object being described or an attribute of
Premise: A Boolean expression that must be evaluated as an entity.
true for the rule to be applied.

Test Y o u r U n d e r s t a n d in g 1 a a 1
1. What assumptions can be made about knowledge codification?
2. Distinguish between:
a. rule premise and action
b. slot and facet
c. decision table and decision tree
3. Two key elements make up a frame. Use an example of your own to illustrate
these elements.
4. How can frames be linked?
5. Summarize the pros and cons of decision tables versus decision trees. Under
what conditions would you use one tool over the other?
6. In your own words, write a short essay explaining the importance of knowl-
edge codification.
7. In your own words, explain knowledge codification. How does it differ from
knowledge creation?
8. Present justification for knowledge codification. How important a step does
it play in building knowledge management systems?
9. How im portant is planning in knowledge codification? Be specific.
10. In what way(s) do the four modes of knowledge conversion result in social-
ization and externalization?
11. What is a knowledge map? How does it differ from a decision tree?
12. Discuss the role of inferencing in knowledge-based codification.
13. Why do we call software agents “agents”? Where do they fit when it comes to
knowledge codification?

Kn o w l e d g e Ex e r c is e s 1111
1. Using the Internet, research case-based reasoning. W rite a 2-page essay
describing the CBR process.
2. Review the employment ads in two consecutive Sunday editions of a major
newspaper. What pattern in knowledge management jobs may be significant?
Knowledge Codification and System Implementation

What types of jobs relate to knowledge management? What types of organi-

zations are looking for knowledge developers or specialists in the field?
3. The knowledge developer’s work is to model the expertise, not the expert.
Do you agree with this sentence? Discuss its meaning.
4. Write a basic set of three rules showing how rules are constructed. Use at
least one AND and one OR operator in your rules.
5. If you were trying to explain to a novice the meaning and role of rapid pro-
totyping, what would you say? Give an example to illustrate.
6. Examine the following profiles of knowledge developers and discuss their
qualifications:
a. Erika and Lori have been contracted to build a knowledge-base system for
a country club. The system is designed to improve the process of evaluating
the applications that the club receives for snack bar worker, lifeguard, and
coach. Erika and Lori hold bachelor of science degrees in IT and are well
versed in two knowledge software packages. Erika is quite congenial and
loves to talk to people. Lori is more introspective and enjoys building
knowledge systems without involvement with the user or the company’s
human experts. In addition, they are both well versed in Pascal and C+ +
Neither has ever built a knowledge-based system for a club.
b. Jeanine holds a master of science in management information systems
from a university in Florida. She is well versed in expert systems and
database design. Her favorite database package is Oracle. She enjoys
working with people and seems to thrive under stress. She has developed
several small knowledge bases. She tends to be pushy when she needs
information. People who worked with her in the past said she is good at
what she does but can be intimidating. She gives the impression that she
has to be in command at all times, using others to do the “leg work.”
c. Bob is 26 years old and has a master of science degree in computer science
from a premier university in New York. He served in the military for 3 years
prior to going to graduate school. Someone who worked with Bob on a
recent knowledge-based project for the military commented, “Bob has pa-
tience for complexity. He is a good listener and is highly intelligent, but he
has a tendency to spin off on his own when the project falls behind schedule,
leaving the rest of his team behind.” He has trouble delegating authority. He
knows his technology inside out and feels he can do the work on his own.
d. Ray earned a doctorate in information technology from a midwestern
graduate school in 1974. He has a m aster’s degree in social psychology
from the same school. Until the mid-1980s, Ray designed information sys-
tems in the banking sector. He started working with software agents and
knowledge-based systems in 1997 and found it both challenging and
lucrative. Recently, he worked on a medical application with a renowned
specialist. The two did not get along well. Ray accused the expert of being
uncooperative, and the expert told Ray he was arrogant. The knowledge-
based medical project ended in disaster.
7. Someone suggested two types of potential knowledge developers:
“Send me a well-developed computer programmer or a programmer compe-
tent in several languages, and we’ll make him or her into a successful knowl-
edge developer.”
and
“Send me a talented generalist with well-developed interpersonal skills—or
somewhat more delicately,‘user friendly person’—and a rigorously analytical
mind, and we’ll team him or her with a competent knowledge developer.”
In your opinion, which approach would be more successful in knowledge
development? Why?
 CHAPTER 7 Knowledge Codification aaaa 239
8. Develop a set of general frames to codify the following:
a. A horse
b. A student
c. An airline pilot
Use these frames and describe the following:
a. Flashdance, an 18-hand thoroughbred
b. Brenda, a medium-height, fourth-year liberal arts student
c. Fred, a 30-year veteran airline captain

Re f e r e n c e s 1 1 1 1
Awad, E. M. Expert Systems. Minneapolis, MN: West Financial Systems, Inc., www.dmreview.com/whitepaper/
Publishing Co., 1996, p. 129. wid274.pdf, Date accessed August 2002.
Brown, E. Donald, et al. “Expert Systems,” February 1999, Lamont, Judith. “Expert Systems and KM Are a Natural
agents.gsfc.nasa.gov/ papers/pdf/es.pdf, D ate accessed Team,” K M W orld, October 2000, www.kmworld.com/
April 2002. publications/magazine/index.cfm?action = readarticle&
Davenport, Thomas H., and Prusak, Laurence. Working A rticle_ID =944& Publication_ID =42, Date accessed
Knowledge. Boston, MA: Harvard Business School August 2002.
Press, 2000, pp. 68-87. Marwick, A. D. “Knowledge Management Technology,”
Dixon, Nancy M. Comm on Knowledge. Boston, MA: IBM Systems Journal, vol. 40, no. 4,2001, p. 815.
Harvard Business School Press, 2000, pp. 143-160. Tiwana, Amrit. The Knowledge Management Toolkit. Upper
Gallupe, R. Brent. “Knowledge Management Systems: Sur- Saddle River, NJ: Prentice Hall, 2000, pp. 146-185.
veying the Landscape,” April 2000, business.queensu.ca/ Applied Learning Labs, www.appliedlearninglabs.com/
docs/fp_00-04.pdf, Date accessed August 2002. kmaps/faqs.html. Date accessed August 2002.
Haley, Paul, and LeFebvre, Peter. “Knowledge M anage-
ment and Knowledge Automation Systems,” Gallagher
 I I I I I I I C H A P T E R I I I fl I I I

System Testing
and Deployment
Contents

In a Nutshell
Quality and Quality Assurance
Knowledge Testing
Key Definitions
Issues to Consider
Approaches to Logical Testing
Circular Errors
Redundancy Errors
Unusable Knowledge
Subsumption Errors
Inconsistent Knowledge
Approaches to User Acceptance Testing
Test Team and Plan
User Acceptance Test Criteria
User Acceptance Test Techniques
Managing the Testing Phase
KM System Deployment
Issues Related to Deployment
Selection of the Knowledge Base Problem
Ease of Understanding the KM System
Knowledge Transfer
Integration Alternatives
The Issue of Maintenance
Organizational Factors
O ther Factors
Role of the Champion
User Training and Deployment
Preparing for KM System Training
Combating Resistance to Change
Postimplementation Review
Security Considerations
Implications for Knowledge Management

240
 CH A PTER 8 System Testing and D eploym ent i 1 i ■ 241

Summary
Terms to Know
Test Your Understanding
Knowledge Exercises
References

The quest fo r quality is a race without a finishing line.

—N o r m a n R ic k a r d

X e r o x V ic e Pr e s id e n t o f Q u a l it y

■■■■ In a Nutshell
The time comes to put a KM system to the test. What good is a system that has not
proven its worth or reliability? Imagine using one in a critical real-time application,
such as m anned space missions, without proving that it will not make catastrophic
errors. At each level of the KM system life cycle, testing and evaluation of the finished
product are critical steps before deployment (see Figure 8.1).
System reliability is one of the most important testing issues in knowledge-based
systems. Reliability refers to how well the system delivers the information or solution
with consistency, accuracy, and integrity. It also means detecting and removing anomalies
(such as redundancy, ambivalence, and deficiency). As KM systems become larger and
more complex, reliability is emerging as a significant issue in developing KM systems.
Two prime considerations behind reliability are quality assurance and m aintain-
ability of the system after deployment. The primary goal of the knowledge developer is
to ensure that the system says what it means and means what it says. In other words,
quality representation of the knowledge requires exhaustive testing. It provides a way
of checking that the user’s requirements are not misunderstood at the conceptual, cap-
ture, or codification phases of the KM system development.
The building process of most KM systems often generates anxiety and concern. In
the rush to build and deliver the system, many knowledge developers cut corners when
it comes to testing. The reasons range from not knowing how to test to running out of
time or funds. Regardless of the reasons, ethical and legal obligations prompt knowl-
edge developers to subject the system to rigorous testing prior to deployment.
Testing KM systems calls for informal, subjective, time-consuming, and often arbi-
trary creation and execution of test cases. In this chapter, the various aspects of testing
are discussed: who should do the testing; what testing criteria should be used; issues in
testing of the KM system; and how to promote objectivity and order.
In this chapter, we shall find that the logical part of testing is a process that ana-
lyzes the program(s) and their relationship to one another. In contrast, user acceptance
testing is a process that assumes that the KM system works fine, but it tests for how
well the system meets user requirements. The problems with the two kinds of tests are
lack of standardization and reliable specification—deciding what constitutes an error
and the sources of test cases. For logical testing, cases must address circular and redun-
dant errors and verify system functionality. User acceptance testing begins with a per-
son or a testing team armed with test criteria using realistic test cases.
No single technique detects all logical or clerical errors in a KM system. Some
tools work better than others, depending on the type of KM system. Known user accep-
tance tools include face validation and test cases provided by the knowledge developer
or the user. The key issue in user acceptance testing is planning and managing a test
242 . . . . PART III Knowledge Codification and System Implementation

/
Knowledge
Knowledge
Innovation

Sharing

Knowledge
Transfer
Database

Knowledge
Base

FIGURE 8.1 : Knowledge Testing and Deployment in the KM System Building Life Cycle

plan that includes decisions on how, when, and where to evaluate the knowledge base
and who should do the testing; these are consistent evaluation criteria for various kinds
of errors. Once planned, testing becomes a m atter of time.
The latter part of the chapter covers im plementation issues and procedures. An
important aspect is user training. The questions often pondered are “How should the
user be trained?” “W here should training be conducted and for how long?” “Who
should do the training?” Obviously, how user friendly and how well documented the
system is has much to do with training ease and success. User motivation, proper fund-
ing, and the role of the champion are additional success factors.

■ Quality and Quality Assurance

Developing quality KM systems is a goal shared by knowledge developers as well as
users. The quality of performance must be demonstrated in a systematic way in order
to promote user acceptance of the technology. The KM system must meet or exceed
 C H A PTER 8 System Testing and Deploym ent a a a « 243

user expectations. Performance capabilities are equated with the quality of the knowl-
edge (tacit and explicit) stored in the knowledge base.
The definition of quality depends on whether the viewpoint is that of the expert,
the user, or the knowledge developer. For the expert, quality is a reasoning process that
gives reliable and accurate solutions within the framework of the KM system. For the
user, quality is the system’s ability to be quick, easy to use, easy to understand, and for-
giving when the user makes mistakes. For the knowledge developer, quality is how well
the knowledge source and the user’s expectations are codified into the knowledge base.

Knowledge Testing
The most challenging part of building KM systems is testing. The basic motivation is to
control performance, efficiency, and quality of the knowledge base. The goal is compli-
ance with user expectations, human expertise, and system functioning. When a KM sys-
tem is built via prototyping, each phase of the building process can be properly tested
during building instead of waiting until the end, as is usually done in testing infor-
mation systems. The comments made at each stage could save time, money, and the
unnecessary embarrassment of reworking the final version of the system. Prototypes
are perhaps the least used form of “rejection insurance” that a knowledge developer or
a development team can ever acquire.

KEY DEFINITIONS
The two main types of testing for knowledge bases and knowledge base systems are
logical testing and user acceptance testing. Logical testing answers the questions, “Are
we building the system right?” “Was the system developed using well-established
knowledge developm ent principles?” It means checking the system to make sure it
gives solutions or results correctly. It looks for anomalies, such as redundant or defi-
cient knowledge, and ensures a bug-free knowledge base.
Logical testing addresses the intrinsic properties of the KM system and its compo-
nents. Intrinsic refers to the syntactic or mechanical aspects of the knowledge base pro-
grams. A logical process analyzes codified tacit and explicit knowledge for sequence,
structure, and specifications. Because no external factors are considered, the quality of
this type of testing can be objectively measured by answering questions such as “What
is . . . ?” “How many . . . ?” This is usually done in a static (paper and pencil) mode,
even when the system is not yet operational. Its goal is error-free codification.
User acceptance testing follows logical testing and checks the system’s behavior in
a realistic environm ent. As a user acceptance phase, these tests ignore the internal
mechanics of the system. Meaning and content of the codified tacit knowledge must
meet codified criteria of adequacy. The quality measurement of user acceptance testing
tends to be subjective. For example, questions based on tacit knowledge such as “What
do you think o f .. . ?” “How well do you like . . . ? are often asked when conducting
user acceptance testing (see Box 8.1).
It is important to note that user acceptance testing is carried out only for technical
verification. When the system is m odular, each module goes through both steps in
building the KM system. See Box 8.2 for an illustration of user acceptance testing.
In today’s practice, no standardization, specific tools, or methodologies guide the
testing of a knowledge-base system. The process is complex and unique at the same
time. Intelligence and knowledge are hard to define, let alone evaluate. Consequently,
the whole testing process is subjective and error-prone. Table 8.1 provides a summary
of select features of logical acceptance and user acceptance testing criteria.
244 7777 PART III Knowledge Codification and System Implementation

1 I fl fl fl fl BOX 8.1 » 1 i » i 1

TESTING THE TRAVEL PROFILER

KM SYSTEM TESTING During logical testing, we found sets of test
cases to run through the system, ran them against
For user acceptance testing, we assumed the KM
a baseline of the software's configuration, and
system to be a black box—we could see what went
saw how the software reacted. Company experts
in and what came out, but not internal processes.
developed the test cases for us, and we used them
Knowledge developers, experts, and potential
not only to determine if the system was running
users provided the system with inputs and deter-
properly, but also to ensure that any modifica-
mined if the outputs were correct. For logical
tions we had made had not altered correct opera-
(technical) checking, we perform ed two proce-
tion of the system from prior versions. In other
dures with hard copy of the program and its
words, we had test cases to ensure that our exten-
knowledge base. First, we read through the lines of
sions were correct and that the work of the past
code one by one to determine if each was correct.
developm ents had not been adversely affected
Second, we took examples of what would be a set
(otherwise known as regression testing). Each
of correct inputs and manually walked through
time new pieces of logic were added (for ex-
the knowledge base to see which rules would be
ample, the addition of handling trips to Cuba),
applicable; if they would be used correctly, that the
new test cases were developed by company
function of these rules was correct; and that their
experts and run by the knowledge developers and
left-hand sides were necessary and sufficient to
potential users.
determine the appropriate trueness of the right-
We ran another set of tests to ensure that
hand side of each rule. We also checked to make
travel agents would understand and be able to use
sure that all of the interfaces between the knowl-
the system and that it would fit their intuitive
edge base and the user's view were correct.
understanding of the travel problem. From this
We look at user acceptance testing as a pro-
standpoint, we were testing the system not for its
cess of substantiating that a system performs with
correctness, but for its usability by the company’s
an acceptable level of accuracy. Logical testing
27 travel agents in seven locations. In all cases, we
involves w hether system perform ance satisfies
were able to get the application to the point where
the requirem ents and w hether the right system
the logic was correct enough for the company
has been built. We validated our system each time
experts, efficient enough for the knowledge devel-
we tested it with the knowers. For every inputted
opers, and usable enough for the travel agents.
problem, we compared the system's output with
the knowers' decisions.

SOURCE: Awad. E. M. "Building Travel Advisor System, Using EXSYS Resolver." Unpublished manuscript, June 2001.

fl 1 I fl I 1

ISSUES TO CONSIDER
The best KM system test is not so much ensuring how well the system functions, but
determining under what conditions it will begin to fail. This is called force-fail testing.
During planning, the knowledge developer puts together a strategy that uses a minimal
set of test cases to discover a maximal set of failures. It is a time-consuming and diffi-
cult process to follow. In fact, it is impractical when marginal cost is measured against
marginal benefits.
 CH A PTER 8 System Testing and D eploym ent u i ■ i 2 4 5

« fl 9 I fl fl BOX 8.2

ADEQUACY AS A USER ACCEPTANCE TESTING CRITERION

Adequacy is a m easurem ent of the fraction of In this knowledge base, adequacy was ex-
actual conditions included in the knowledge base pressed as a simple fraction. The breath gas moni-
system. For example, the breath gas m onitoring toring system was able to diagnose 12 of 27 im-
system diagnosed 12 waveform patterns of C 0 2. p ortant physiological conditions, so the system
Each waveform pattern corresponded to a partic- was 100 X (12/27) = 44% adequate. As an alter-
ular physiological condition. The adequacy of the native, subjective weights may be added to par-
system was judged according to the total desired ticular conditions that are more im portant to rec-
number of conditions to diagnose. ognize (such as cessation of breathing).

SOURCE: Awad, E. M. “Validation and Verification o f a Breath Gas Monitoring S y s t e m Unpublished manuscript, August 2001.

There are several issues to consider in knowledge-based testing:

• Subjective nature o f tacit knowledge. Intelligence is difficult to measure. With
human experts or a team of specialists, intelligence changes over time due to
ongoing upgrades and m aturation of abilities and judgment. Solutions for many
problems often call for subjective decisions rather than straightforward catego-
rizations of right or wrong. Because tacit knowledge always gives an answer based
on the heuristics, it does not “blow up,” as do traditional information systems.
• Lack o f reliable specifications. Ill-defined specifications make knowledge-base
testing arbitrary and unreliable. Without testable specifications, a knowledge
developer might give in to the tendency to make up specifications for what the
system can do instead of figuring out the conditions under which the system will
not function well. This is an important distinction to keep in mind.
• Problem o f establishing consistency and correctness. To make sure the performance
of the knowledge-base system is correct, the knowledge developer needs to verify
both the knowledge base and the reasoning processed by the programs. The

TABLE 8.1 Select Features of User A cceptance and Logical Testing

Logical (Knowledge User Acceptance
Developer Acceptance) (User Acceptance)
Meaning Building the system right Building the right system
Time performed During knowledge codification During prototyping or when the
or prototyping system is technically operational
Sequence First Second
Nature of checking Syntactic (intrinsic properties Semantic (extrinsic properties of
of the system) the system)
Quality measurement Primarily objective (technical) Primarily subjective (judgmental)
fl fl fl fl fl fl
246 7777 PART III Knowledge Codification and System Implementation

complex tasks of verifying correctness and consistency are challenging; a knowl-

edge-base system is heavy on tacit knowledge and continues to give answers even
though part of the knowledge is incorrect.
• Negligence in testing. A vicious cycle begins when no one requires system testing.
Part of the problem stems from the difficulty in providing reliable test cases.
Another reason is that knowledge-based systems with so many possible states or
conditions make exhaustive testing impossible. As a result, user acceptance and
logical testing take on “surface” commitment.
• Lack o f time for knowledge developers to test the system. Many parameters with many
potential values cause a combinatorial explosion of rules and test cases, making ex-
haustive testing difficult. With the simplest tool, tabular checking, rules are grouped
into tables similar to I F . . . THEN decision tables. Inside each table, a comprehensive
comparison of the rules is made, which checks for inconsistency, redundancy, sub-
sumption, and the like. These types of errors are explained later in the chapter.
• Complexity o f user interfaces. Some KM systems have complex user interfaces,
where inputs are not easy to reproduce, making the potential solution highly inac-
curate. The user interface is an important part of the KM system development
effort. Through it, the user determines how easy, user friendly, or useful the sys-
tem is and, therefore, how successful it will be.
One conclusion that knowledge developers should take from these challenges is
the need for a seriously designed logical testing strategy. It is a formal statement on
how to proceed with the logical testing phase. The test design should m eet all the
requirements of the client. A force-fail test strategy requires careful choice of test cases
that look for missing, ambiguous, or redundant rules with the potential for incorrect
answers. These types of rules are discussed later in the chapter.

■■■» Approaches to Logical Testing

Two approaches can be used to verify knowledge-base systems:
• Verify knowledge base formation.
• Verify knowledge base functionality.
Under knowledge base formation, the structure of the knowledge as it relates to
circular or redundant errors is verified. Consistency, correctness, and completeness of
knowledge base rules are also verified. Logical testing of functionality has to do with
confidence and reliability of the knowledge base (see Figure 8.2). Rules and attributes
are defined in Table 8.2.

CIRCULAR ERRORS
Carelessness often results in rules that employ circular reasoning. A circular rule tends to be
contradictory in meaning or logic. For example, suppose rule 1 states that A is greater than B,
and rule 2 states that B is greater than C. If rule 3 states that C is greater than A, then rule 3 is
circular and contains an error in logic. The best solution in this case is to remove rule 3.
H ere is another example of a circular error:
R5 IF John and Rob are co-owners of company A
THEN John and Rob own the same company
R6 IF John and Rob own the same company
THEN John and Rob are co-owners.
 CH A PTER 8 System Testing and D eploym ent aaaa 247

- Verify Knowledge Base Formation

-Structural Verification Anomalies

—— • Circular Rules

r- Unusable Rules
— Duplicate
— Redundant Rules —
— Subsumed

Verification - - Verification of Content

• Completeness
• Consistency
• Correctness

— Verify Knowledge Base Functionality

--------- Confidence
--------- Reliability

FIGURE 8.2: Approaches to Logical Testing

TABLE 8.2 Attributes in Logical Testing

Circular A situation where the action(s) of one piece of knowledge may lead
back somehow to the condition statement of the same piece of
knowledge
Completeness Answers the question, “Does the system deal with all possible situa-
tions with its knowledge base environment reasonably well?”
Confidence The level of trust or dependence one can put on the system’s answers
to problems; related to system integrity, with emphasis on system
reliability and consistently accurate answers
Correctness Described as accuracy
Consistency A check to ensure that the system produces similar answers to all
input data at all times with no contradiction
Inconsistency An anomaly or an error that occurs when an item of knowledge has
different value(s) within the knowledge base
Redundancy Two or more knowledge facts or rules that are written differently but
provide the same solution; a duplication of knowledge; absence of
redundancy provides greater specificity to the notion of completeness
Reliability How well the system delivers solutions with consistency, accuracy, and
integrity; the probability of failure-free operation of the knowledge
base system in a specified environment for a specified time period;
synonymous with dependability or trustworthiness
Subsumption error The conclusion(s) of two pieces of knowledge are the same, except
that one piece of knowledge has fewer condition statements
■ ■ ■ ■ a fl
248 7777 PART III Knowledge Codification and System Implementation

If R6 finds a match, its premise part will cause R5 to match and fire. The problem
here is that the premise part of R5 matches the action of R6, causing R6 to fire and so
on, resulting in useless firing of the two rules.
As a simple example of logical testing, consider a KM system that has several cases
involving a house with a living room. We can evaluate cases to determine that there is
something incorrect with a case that has a bathtub in the living room. There is probably
redundant knowledge in the case if the living room contains two sofas. Likewise, the
case of a living room without a sofa probably would be considered incomplete. Finally,
we might find it inconsistent if two identical items were labeled the same name. For
example, a sofa for one item is labeled a couch for another item.

REDUNDANCY ERRORS
Redundancy errors are ones that offer different approaches to the same problem.
There is actually duplication of knowledge. One gets the same results either way—the
“dinosaur” effect. For example, study the following rules:
IF (Salary > 60,000 OR Home > 100,000)
AND (Loan < 10,000 AND Mortgage < 40,000)
THEN Max = 100,000
IF (Salary > 60,000 OR Home > 100,000)
THEN Assets = ‘yes’
IF (Salary >50,000 AND Dividend > 10,000)
THEN Assets = ‘yes’
IF (Loan < 10,000 AND Mortgage < 40,000)
AND Assets = ‘yes’
THEN Max = 100,000
In this example, we find two approaches to the same problem of determining Max.
We also do not need to use dividends in the second approach to establish assets.
Redundancy can cause difficulties in maintaining the KM system. One version of the
rule could be revised; another version could be left as is.

UNUSABLE KNOWLEDGE
An unusable knowledge in the form of a rule is one that only “fires” (executes) if the
conditions succeed or fail, one that never fires, or one that has one or more contradic-
tions. They are all in error and, therefore, unusable. An example of an unusable rule
that is always false is:
IF humidity is 5% AND heavy_rain = yes
THEN conclusion
This rule cannot succeed. If a rule premise (IF) will be false for all valid values, the
rule will never succeed. In this example, heavy rain cannot occur in 5 percent humidity.
Unusable rules are also called conflicting rules.

SUBSUMPTION ERRORS
In subsumption, if one rule is true, one knows the second rule is always true. Consider
this example:
R l: IF A AND B,THEN D
R2: IF A AND B AND C, THEN D
 CH A PTER 8 System Testing and D eploym ent i i ■ ■ 249

If R1 is true, then R2 is useless, because R2 is subsumed by R l. Both have identical

conclusions, but R2 contains an additional (and thus unnecessary) premise clause.

INCONSISTENT KNOWLEDGE
In inconsistency errors, the same inputs yield different results. For example,
IF (Salary > 50,000 OR Home > 100,000)
AND (Loan < 10,000 AND Mortgage < 40,000)
THEN Max - 100,000
IF (Salary > 50,000 OR Home > 60,000)
AND (Loan < 3,000 AND Mortgage < 20,000)
THEN Max = 50,000
A user will get different answers, depending on which rule is used. Tests conducted
for consistency can also be used to test for completeness. The whole idea is to make
sure all the parameters included in the rules are codified correctly and consistently.

■■«« Approaches to User A ccep tan ce Testing

As m entioned earlier, user acceptance testing follows logical testing. Its focus is on
testing the behavior of the integrated KM system to see if it is satisfactory in the eyes
of the user. Even when individual pieces of knowledge are correct, something in their
interactions can cause the system to act up.
User acceptance testing includes several steps:
1. Select a person or a team for testing. A testing plan should indicate who is to do the
testing. Should it be the knowledge developer? The user? Top management? The
plan should determine the user acceptance technique(s) appropriate for the system.
2. Decide on user acceptance test criteria. The key user acceptance criteria are ease of
use, accuracy, adequacy, adaptability, face validity, and performance. See Table 8.3
for definitions of the criteria.
3. Develop a set o f test cases unique to the system. The test cases should include subtle
and obvious cases, boundary conditions, and obviously meaningless combinations of
valid and invalid information. The test results are then evaluated against the criteria.
4. Maintain a log on various versions o f the tests and test results. Documentation can
be very important; errors are more easily detected at one test level by simply
going back to the previous version. Without this backup, the system would have
to be tested from the beginning every time an error occurs —a frustrating and
time-consuming job.
5. Field-test the system. Once the system has passed the test process, it should be
field-tested with user involvement. Field-testing should go smoothly.

TEST TEAM AND PLAN

A successful testing effort requires a com mitm ent to hum an resources beyond the
knowledge developer or the human expert. Commitment begins with management sup-
port and a test team with a test plan. The team oversees test activities. The team should:
• Be independent of the design or codification of the system
• Understand computer technology and knowledge base infrastructure
• Be well versed in the organization’s business and how the system being tested will
fit into the designated area of operation
250 7771 PART III Knowledge Codification and System Implementation

TABLE 8.3 Select Criteria for User Acceptance Testing

Accuracy D oes the outcom e of the system match reality or expectations? Emphasis
here is on correctness of the outcome, which is measured by comparing
the number of correct answers against known answers.
Adaptability How well does the system adapt to changing situations? This criterion
relates to flexibility and capability to modify facts or rules when new
information becomes available. Can the system be customized for par-
ticular user needs?
Adequacy Are the solutions produced by the KM system satisfactory or “good
enough” to be acceptable to the user? It is a measurement of the fraction
of actual conditions included in the knowledge base.
A ppeal How well does the KM system match the user’s intuition and promote use-
ful ideas. Usability is also related to practicability, marketability, and user
friendliness.
Availability D o other KM systems that solve the same problem exist? Availability is
related to appeal.
Ease of use D oes the system incorporate human engineering and high quality in its
human-machine interaction/interface?
Face validity How well does the system pass the real-life test on the surface? It is
another term for credibility.
Performance D oes the system function as expected? It is related to reliability and quality
assurance and the overall quality of the solutions provided by the KM
system.
Reliability See Table 8.2.
Robustness Is the system able to function near the limits of its coverage?
Technical/ How well do the system ’s technical and operational states meet the user’s
operational requirements? It indicates the capability of the knowledge base to produce
test empirically correct answers.
------------------------------------------------------- 1 I fl fl fl f l ------------------------------------------------------

Team members normally represent a variety of the users’ areas of operation, espe-
cially in the final testing phase. The entire procedure should be guided by a plan that
defines the roles and responsibilities for everyone involved in the system’s life cycle,
and it should provide documentation with respect to testing. The plan should:

• Identify the aspects of the KM system that require testing

• Determine when user acceptance testing is appropriate
• Define the test method that will lead to user and company approval
• Provide the guidelines for documenting test results

Having the knowledge developer do the user acceptance testing poses limita-
tions. In user acceptance testing, the problem for most knowledge developers is that
they have only limited knowledge in developing test cases that meet user acceptance
criteria. The human expert can be consulted to approve the procedure and the ac-
ceptability of the solution. In the end, whoever ends up being the tester depends on
the availability of expertise, motivation, time, and funds allocated for the testing
phase. A test team representing the users’ interests and based on a plan would be the
best choice.
 C H A PTER 8 System Testing and D eploym ent i i i i 251

USER ACCEPTANCE TEST CRITERIA

User acceptance test criteria guide the focus of what to test. As a general rule, all sys-
tems are valid against certain criteria and invalid against others. The question is how
does one know what criteria to use for evaluating the KM system? This is especially
im portant when certain systems (such as a medical diagnosis knowledge base) require
more stringent user acceptance test criteria than other systems (such as one that assists
in the selection of roses for Valentine’s day). The alternative criteria compare system
results against historical cases, against the expert’s prediction of the final results, or
against some theoretical standard. In any case, user acceptance test techniques contain
imperfections and cannot guarantee 100 percent performance.
To illustrate, suppose test cases are chosen from a KM system test plan. Several
questions are raised:
• How many test cases should be used to assure that the system is satisfactory to
the desired confidence?
• How should the test cases be developed objectively and reliably?
• What objective measures should be determined for the selected criteria?
• Should a performance measure be included—other than that of the human
expert —against which test case results should be evaluated?
• What mix of easy versus difficult test cases should be considered in the final test
cases? That is, what type of test cases should be used?
• How should the test cases be evaluated? By a panel of experts? By knowledge
developers?
• If multiple experts are involved, how should the KM system be validated against
differing human opinions?
Addressing criteria measures can assist in test planning. Sound and reliable mea-
sures also mean easier deploym ent and im proved user acceptance. Actually, user
involvement in testing the user interface generally increases user acceptance of the fin-
ished product.

USER ACCEPTANCE TEST TECHNIQUES

No single user acceptance test technique is best for detecting all errors in a KM system.
Some techniques work better for certain types of problems and for certain types of
knowledge codification (for example, rule-based versus frames or decision trees).
Several test techniques are worth reviewing:
• Face validation
• Test cases
• Subsystem validation

Face Validation
In face validation, a knowledge developer, along with users and company experts,
evaluates the perform ance of the KM system at its face value. System solutions are
compared with those of the human experts, and a value judgment is made about the
reliability of the results.
Face validation can be used to test chunks of knowledge at any phase of the devel-
opment life cycle. It is mainly a group effort of the knowledge developer, the expert,
and the end user. This ad hoc tool is quite useful for testing user-system interfaces, user
friendliness, and assist facilities. However, it is not a rigorous user acceptance tech-
nique. As such, it offers no measurable assurance of performance or reliability.
252 7777 PART III Knowledge Codification and System Implementation

Test Cases
A popular test technique is executing a set of prepared test cases on the KM sys-
tem. The results are examined for agreement with those of an independent expert or a
panel of experts that solves the same problem. Only system inputs and outputs are sig-
nificant (see Box 8.3).
In addition to those provided by the human expert, test cases may be provided by
the knowledge developer and the user. Those provided by the developer have been
known to be more effective than those provided by others. The domain expert, the

« a « fl i fl box 8.3 a fl a i fl a

TEST CASES FOR THE KNOWLEDGE RECRUITING PROFILER

The test design was created to ensure that all the utilization of the knowledge base improved the
user’s requirements were met. We created a form manager’s hiring process by ensuring that Ms. Bur-
from which to collect decision data from individ- gess considered all of the required variables before
ual past cases. The form included all seven varia- making her decision. This system removed ambi-
bles with the corresponding choices, an area to guity and inefficiencies in the decision-making pro-
write in Ms. Burgess’s (personnel director) past cess. In essence, it succeeded in eliminating both
decisions, and an area to write in what the knowl- Type I and Type II errors recognized in the cases
edge hiring profiler recommended. Thirty copies being utilized. We tested this by running each of
of the form were given to the manager to be filled the 30 cases through the hiring advisor using the
out. A set of cases was selected from her past following criteria:
decisions to represent the good decisions and bad
decisions. In order to make this process effective, • The past cases with applicants identified as good
em ployees who were actually hired were deter-
Ms. Burgess looked through some of her previous
mined to be qualified by the recruiting profiler.
applications to see what she had decided.
Good decisions included those applicants • The past cases with em ployees who were not
hired but were qualified were determined to be
with whom Ms. Burgess was satisfied with her
qualified by the recruiting profiler.
decision to either hire or not to hire. Likewise,
bad decisions included applicants with whom the • The past cases with applicants who were hired
and ended up being poor performers were not
manager was dissatisfied with her final decision.
determined to be qualified by the recruiting
Bad decisions resulted from either a Type I error
profiler.
or a Type II error. Results of case testing are as
• The past cases with applicants who were not hired
follows:
because it was believed they would be poor per-
SAM PLE N um ber N um ber formers were not determined to be qualified by
SIZ E - 30 H ired N ot H ired the recruiting profiler.

Type I error 9 0
The chief limitations of the previous decision-
Type II error 0 1
making process included the lack of emphasis on
W ithout the use of the hiring profiler, there ability and the lack of standardization in the wage
was a 33.3 percent (10 out of 30) error rate. The placements.

SOURCE: Awad, E. M. “Test Cases For the Knowledge Recruiting P r o f i le r Unpublished manuscript, January 2002.

I I 1 1 I fl
 CH A PTER 8 System Testing and D eploym ent a ■ a ■ 253

knowledge developer, and the user are somewhat like three blind men describing an
elephant. Each looks at a different aspect of the KM system and tries to make sense
out of it. Test cases from all three sources are considered to be the best combination,
because they provide a m ore com prehensive test of the system. They also tend to
increase the objectivity of the testing process.
Like other techniques, testing a knowledge base that is based on test cases has its
own share of limitations. When the expert who writes the test cases is also the tester,
the credibility of the test process can be undermined. Bias—based perhaps on person-
ality conflicts or on ignoring nontechnical aspects such as user interface or ease of
use—can filter into them, especially if test case generation is done in an ad hoc manner.

■««« M anaging the Testing Phase

The critical issue in KM system testing is planning for completeness of testing that will
vouch for system adequacy and lead to user acceptance. A nother im portant issue is
managing the test plan. This includes the following tasks:

• Decide when, what, how, and where to evaluate the knowledge base. The reliability
of answers, explanation, training capability, and the like must be evaluated.
Timing and duration of testing in the hands of specialists are crucial to a success-
ful system test and eventual implementation.
• Decide who should do the logical and user acceptance testing. All indicators point
to having the expert handle the whole process. The next best approach is to have
an outside expert or a team do independent tests of the knowledge base.
• Draft a set o f evaluation criteria in advance. This task involves knowing what to
test for. One can get the system to run, but what are you proving or looking for?
What are the hard criteria?
• Decide what should be recorded during the test. When it comes to testing rule-
based tacit knowledge, among the important statistics to record are the following:
• Rules that always fire and succeed
• Rules that always fire and fail
• Rules that never fire
• Test cases that have failed
• Review training cases, whether they are provided by the expert, the knowledge
developer, or the user. Appropriateness of the test cases should be emphasized in
this task.
• Test all rules. This task ensures that every param eter is correct. Testing looks for
two types of errors in rules:
• Type I error—A rule that fails to fire when it is supposed to fire. This is tanta-
mount to failing to detect a problem and similar to a false negative in statistics.
• Type II error—A rule that fires when it is not supposed to fire. This is similar
to a false positive in statistics.

Finally, generating test cases is so time-consuming and costly that this phase is
often viewed as a chore rather than a challenge. It takes a certain know-how and objec-
tivity to generate test cases. Even then, users have been known to accept the system,
regardless of how lax the test process has been. With this attitude, testing can go for-
ward with no difficulty.
254 7777 PART III Knowledge Codification and System Implementation

«■ ■■ KM System Deployment
The goal of every KM system development is successful deployment. Even the best
system will fail if deploym ent concerns are not addressed. At issue is how well the
knowledge developer coordinates the deployment with the end user. Depending on
how knowledge codification was carried out, the end user might have been a regular
participant all along, which means deployment stands a good chance of user accep-
tance and successful deployment.
Deployment is affected by organizational, technical, procedural, behavioral, eco-
nomic, and political factors, especially where the organization is large and the existing
technology is too established to allow change to occur easily. KM system deployment is
especially complex. In this next section, we examine deployment issues as they relate
to managing KM system projects as well as user training. Knowledge transfer is cov-
ered in the next chapter.
The two aspects of deployment are the transfer of the KM system as a technology
from the knowledge developer to the organization’s operating unit and the transfer of
the system’s skills from the knowledge developer to the operating person in the orga-
nization. This step may also include training an organization’s operating unit for system
maintenance and upgrade. The issues, prerequisites, and techniques for both of these
aspects will also be discussed in this section.

»■«■ Issues Related to Deployment

Once the system is designed and fully tested, the final step is the physical transfer or
deployment of the technology to the organization’s operating unit. Whether or not the
developer will be the one to deploy, certain issues need to be addressed at the beginning
of each KM system development project. These relate to the selection of the knowledge
base problem, ease of understanding the KM system, and organizational factors.

SELECTION OF THE KNOWLEDGE BASE PROBLEM

The success of KM system deployment has a great deal to do with the way the problem
domain was selected. For example, if the domain was selected with end user involve-
ment and management support, one can expect smooth deployment and cooperation
in user training. KM system users can be assured of system success if:
• The user has prior experience with computer applications, which tends to reduce
the inherent fear that automated systems will replace people or add stress to
their job life.
• The user has been actively involved in defining and identifying the specific func-
tions of the KM system.
• The user is involved in user acceptance testing and the final evaluation of the KM
system to assure its accuracy, completeness, and reliability.
• The payoff from the development of the KM system is high and measurable and
leaves little room for arguments that it is draining the budget or is unnecessary.
• The KM system can be implemented in the working environment without much
interruption of ongoing activities.
• The champion has done a great job selling the user’s staff on the potential contri-
butions of the KM system.
By now, the importance of involving the user in the early phases of the develop-
ment process should be evident. Using rapid prototyping can also offer a way of selling
 CH A PTER 8 System Testing and Deploym ent a «aa 255
the system and its potential benefits when it is deployed. Front-end commitment
requires dedicated users, managers, company experts, and a champion to firm up the
planning and provide the support necessary to carry out knowledge codification. Once
commitment is secured, deployment is generally easier to accomplish (see Box 8.4).

EASE OF UNDERSTANDING THE KM SYSTEM

The acid test of knowledge codification is when the system is deployed and the user
realizes how easy it is to understand. D ocum entation also plays a key role. When
changes must be made, a troubleshooting guide, a step-by-step procedure, or a flow-
chart in the form of an easy-to-follow technical or operations manual can guide the
user into unlocking the system and correcting glitches.
Reliable docum entation, especially during user training, is especially im portant
during deploym ent. Docum entation heavy on illustrations, examples, and graphics
reduces training time. Of course, a human trainer who provides face-to-face, inter-
active, on-site training is always an asset to the training process.
With this in mind, successful KM system deploym ent depends on several user-
related factors:
• Level o f motivation o f the user. Good documentation cannot compensate for poor
attitude and low motivation. Promoting motivation and commitment takes time
and must be worked on in advance.
• Computer literacy and technical background o f the user. A computer-literate user
can be easier to work with than someone with no stich background. First-time
users often require education and training before they are able to support devel-
opment and use of knowledge-based systems.
• Communication skills o f the trainer. Selling people on change is sometimes con-
sidered more of an art than a science. Communication skills can make the differ-
ence between a user’s acceptance of or resistance to the new system.
• Time availability and funding for training. A training program that is run on a shoe-
string is usually a loser. Squeezing training time to the bare minimum often results
in trainee impatience, resistance to learning, or nonuse of the system. Training
should be part of the deployment phase offered around the convenience of the user.
• Place o f training. The location of training can make a difference. On-site versus off-
site training continues to be an issue, with pluses and minuses for each alternative.

■ i a a « i BOX 8.4

INSTALLATION AND TRAINING ON TRAVEL PROFILER

Participation in the deployment and maintenance to say “no” to using the system. The way it turned
of Travel Profiler included two senior managers out, entry-level agents found the system a great
who had been involved in the knowledge capture training tool that reinforced what they do on the
phase. This was quite a break for implementation, job. The actual installation and basic training took
because they made it difficult for the travel agents less than 2 days on a weekend.

SOURCE: Awad, E. M. “Building Travel Advisor System, Using EXSYS R e s o lv e r Unpublished manuscript, June 2001.

fl fl fl i fl a
256 T7T7 PART III Knowledge Codification and System Implementation

Off-site training is generally dedicated, uninterrupted learning. Its positive bene-

fits include privacy and focus on the projects. The feasibility of off-site training
depends on distance, location, and funding. In contrast, on-site training means no
out-of-town transportation or room and board expenses. However, it can be inter-
rupted by telephone calls, secretaries, and uninvited gawkers.
• Ease and duration o f training. This aspect depends on the caliber of the trainer
and the attitude and motivation of the trainees. “Chemistry” often affects how
well all parties work with one another. Also, the training period should be reason-
able and able to meet measurable goals. A long, drawn-out 3-week training period
does not promote the same excitement as a 4-day session.
• Ease o f access and explanatory facilities o f the K M system. KM systems should be
easy to access and work with. A feature that provides adequate online explana-
tions satisfies most users and provides convincing evidence of the integrity of the
answers provided by the system.
• Ease o f maintenance and system update. At this stage, good documentation and easy-
to-follow procedures in a module-oriented KM system can make the difference
between easy maintenance and a nightmare. Maintenance ensures that the system
stays on course. It means making corrections to meet the initial system requirements.
Enhancement means upgrading the system to meet a new set of requirements.
• Payoff to the organization. A KM system’s benefit to the organization is usually
measured in cost reduction, improvement in sales or overall performance, and so
on. Measured payoff early in the development life cycle promotes successful
deployment.
• Role o f the champion. Solid top management support and a champion pushing for
system adoption can make the difference between a successful and a lukewarm
installation.
These success factors of KM system deployment are interrelated. Organizational
and technical factors also come into play (see Figure 8.3).

KNOWLEDGE TRANSFER
Two approaches can be used to transfer KM system technology in implementation. In
the first approach, the KM system is actually transferred from the knowledge devel-
oper directly to the working unit (end user) in the organization. The other approach
simply means installing the system on the resident hardware. It includes dropping off
the user’s manual and help facility, a technical manual containing a line-by-line print-
out of the knowledge codification, and copies of the knowledge base for backup. This
step may or may not include user training. Actually, training the user on the KM system
is relatively easier than training on a conventional information system because of the
unique technical and procedural facilities embedded in the KM system.
In the second approach, which transfers KM system technology skills, the knowl-
edge developer transfers to a group within the organization the knowledge and neces-
sary know-how to maintain and upgrade the system in the future. Such a transfer may
include special technical training in the diagnosis and test procedures necessary to
make changes in the system. Once com pleted, the developer is out of the picture,
except perhaps in a consultative role.
Transfer of KM systems can be approached in one of two ways: (1) an abrupt, one-
time transfer that results in a permanent installation on a specific day or (2) a gradual
transfer over a given time period. In the latter case, most often through rapid prototyp-
ing, a receiving group becomes part of the developer’s team. Once one phase of the
 CH A PTER 8 System Testing and D eploym ent aaaa 257

9 fl fl 1 I 1 FIGURE 8.3: Success Factors in KM System Deployment

system is verified and ready to use, responsibility is transferred to the receiving group,
which then examines the makeup of the partial system. The transfer involves the shar-
ing of training, methodologies, experiences, and techniques. Eventually, the rest of the
system is transferred, allowing the receiving group to take full responsibility for the
system’s operation and maintenance.
Implementation can also be approached as a stand-alone KM system installation
on a PC or as a fully integrated application that interfaces with other applications or
databases. KM systems should be designed on platforms that are compatible with other
KM systems in the organization. The compatibility feature makes explanation or justi-
fication of the new system easier. It also makes user training easier, because most users
presumably look at the new system as “a system just like the others.”

INTEGRATION ALTERNATIVES
A KM system can be integrated into the organization’s existing operations through the
following methods:
• Technical integration occurs through the company’s local area network environ-
ment, the resident mainframe, or existing information system infrastructure. This
job is not exactly straightforward. M anagement must decide on a system’s level of
priority for access in the network. For example, in one organization’s IT department,
a LAN-based KM system was given a low priority access, because the course that
258 7777 PART III Knowledge Codification and System Implementation

used the system was taught only during weekends. During regular workdays, em-
ployees frequently experienced considerable delays (more than 20 seconds) in
response time. They complained that work assignments were always late coming
off the printer.
• Knowledge-sharing integration is required, for example, in a firm that needs to
make a KM system available to each of its 17 branches in an equitable way. This
level of availability often requires upgrading the local area network, the main-
frame, or lines to ensure equitable service. For example, in the case of the Travel
Profiler, after the travel agency installed a LAN, it made plans to have the KM
system available to all travel agents at the same time with the same response time.
• Decision-making flow integration suggests that the way the KM system assesses a
problem situation should match the user’s style of thinking. Providing such a
match can be crucial to user acceptance and satisfaction.
• Workflow reengineering considerations come into play when implementation of a
new KM system triggers changes in the workplace or within jobs in the user’s
domain (such as merging jobs or deleting positions). This type of situation can
occur because the actual building of the product leads the KM system and the
user, during their interactive sessions, to develop new insight into the problem
domain and to a more creative way of dealing with it. One can expect a certain
amount of backlash as a result of job reengineering.
The attractiveness of change depends greatly on m anagement’s perception of its
potential contribution to the productivity of the area in which the new system will
reside. The concept of reengineering is evolutionary and usually takes weeks or even
months to firm up. In other words, for a KM system to survive, it must be amenable to
change. It must also provide built-in flexibility to accommodate the kind of change
expected by the user.

THE ISSUE OF MAINTENANCE

An important aspect of KM system development that gets little attention is mainte­
nance. The term maintenance means “making the necessary corrections that continue
to m eet the user’s expectations.” In rule-based KM systems, for example, it means
adding, removing, and rewriting certain rules so that the reasoning is more consistent
with today’s expert’s solutions. Although this type of change is carried out during the
initial testing phase, one can expect corrections resulting from the user's real-time
experience with the system. Whatever mode the knowledge base is codified in, m atura-
tion over time tells us that knowledge needs close attention to reflect the best ways to
update, transfer, and share knowledge.
System maintenance can be improved in several ways. For example, if the knowl-
edge base is organized into several well-defined modules, one can easily go to a specific
module and make the necessary changes. In this respect, a single developer in charge of
knowledge codification can promote consistency and uniformity in programming the
knowledge base.
For knowledge-based system deployment to succeed, it must facilitate easy and
effective maintenance in several ways:
• The system must have a feature to allow changes to be added or deleted as
needed, which brings up the importance of the modular approach to KM systems.
• The system must be capable of allowing inconsistent, conflicting, or redundant
errors to be easily identified. Removing redundancies ensures improved system
performance.
 C H A PTER 8 System Testing and D eploym ent i i 1 i 259

• The system’s help facilities must satisfy the user’s requirements as closely as pos-
sible. User acceptance of the solution is easier with the availability of clear pre-
sentation of the solutions.
• The availability of a qualified person or a team can ensure that maintenance is
carried out effectively and on schedule. With large organization-wide KM sys-
tems, maintenance could be a full-time job.
To assure successful maintenance, responsibility for it needs to be assumed by a
qualified person within the organization. Otherwise, the maintenance function should
be contracted to a freelance specialist. Part of the maintenance function establishes a
change-order procedure with proper docum entation that includes the authorized
requester, the person who made the change, and the outcome of the change.

ORGANIZATIONAL FACTORS
Of all the deployment issues, a prime consideration is a firm commitment to the project
by top management and the user. It means strong, senior-level leadership, because there
will be challenges that arise. Management can support system deployment by providing
adequate funding, ensuring the availability of technology and personnel, and allowing
the champion to function throughout the development process. The second organiza-
tional factor is user participation in the building process. Participation improves com-
mitment, enhances system quality, and fosters a sense of ownership of the KM system.
Commitment is not the same as contribution, however. Anyone can make a contri-
bution through system participation without having to make a commitment to system
use. Commitment to a project means taking responsibility for its outcome, which is
expected to contribute to successful deployment. The anecdote in Box 8.5 illustrates
the distinction.
O ther organizational factors include organizational politics and organizational cli-
mate. This could mean turf war. Different departm ents (such as purchasing and pro-
duction) may support the overall goal of the new KM system but have conflicting pri-
orities. Politics is jockeying for leverage to influence o n e’s domain and control
procedures, technology, or the direction of an area of operation. The prevalence of poli-
tics, especially in a negative sense, is often ignored or deemphasized. The knowledge
developer must remain neutral within the political arena of the firm.

* « 1 ■ 1 1 BOX 8.5

THE CHICKEN AN D THE PIG VENTURE

A chicken and a pig grew up together on a farm ers will be to get the freshest ham and eggs in
and came to know each other very well. One day, the area.”
the two met and discussed a joint venture. The slower-thinking pig pondered this inge-
The quick-thinking chicken told the pig, nious proposal and, after a long pause, said, “My
“I have a great idea for a breakfast restaurant dear friend, your proposal sounds great, although
near a local school. Between the two of us, we lopsided. All you’re offering to do is to make a
have all the ingredients we need —eggs, ham, contribution to the venture. You’re asking me to
and sausage. Just thjnk how excited the custom- make a commitment .”
260 7777 PART III Knowledge Codification and System Implementation

The organization’s climate dictates whether a KM system’s time has come. The key
questions that the knowledge developer should ask are “Is today, this week, or this
month the right time to implement the product?” “Is the organization ready for this
particular environm ent?” Timing and readiness factors must be assessed in advance.
For example, in banking, most computer-based systems are installed during weekends
so as not to interfere with the weekday business activities. A KM system, including
training, should be completed in an environment devoid of interruptions.
User readiness can also influence deployment success. Foot dragging and delaying
the inevitable are not solutions. Users who know that work will accumulate during their
training make poor trainees. This suggests that deployment must be planned in advance.

OTHER FACTORS
There are several other implementation factors to consider. The return on investment
(R O I) factor is not easy to measure. If it is simply “greater custom er loyalty and
improvement in employee satisfaction,” it is a hard sell with the board of directors. The
costs are not hard to quantify, but the benefits are intangible and soft. A company may
decide to spend a $1 million on a KM system, but if it does not make its quarterly fig-
ures, the quick fix is to shelve expensive projects. When that happens, momentum is
lost, consultants leave, and the KM system dies a natural death.
Another factor worth considering is quality information. The number-one require-
ment for a successful KM system is an infrastructure capable of sharing real-time and
accurate information. Content is a critical part of KM system implementation, espe-
cially when the knowledge developer is trying to meet deployment deadlines, cut down
on costs, or do a quick sell to the user.

ROLE OF THE CH A M PIO N

When all is said and done, the underlying factor for successful implementation of a KM
system is the role of the champion. A champion is someone in an organization who,
because of position, power, influence, or control, secures the organization’s support of
the new system from inception to deployment. Without a champion, a KM system is
vulnerable throughout the developm ent process. The cham pion needs to be at the
executive level or a senior steering committee to act as the project's board of directors.
Politics, budgetary problems, deficient support, or conflict of interest can stand in the
way of deployment.
To illustrate, a KM system designated to help VISA/MasterCard personnel decide
on approval of applicants met stiff resistance from the departm ent’s manager. The
problem caught the attention of the bank's operations vice president, who had a good
rapport with the officers of the bank as well as the chairman of the board. The vice presi-
dent had a long talk with the manager and assured her that the new KM system would
“speed things up” rather than replace her final judgment. The vice president concluded
the meeting by saying, “Look, let's give this project a try and see how it turns out."
When time came for deployment, the knowledge developer met with the manager
and asked to proceed with training her staff on the new system. The manager replied,
“That’s not necessary. Because I’ve gone through the process and am familiar with the
way it works, just load the package on the server and I’ll take it from there.''
Concerned about possible delays or improper training, the knowledge developer
took up the m atter with the bank's vice president, who called a meeting with the senior
staff of the VISA/M asterCard division. At the meeting, the vice president explained
the importance of the KM system and how it would improve the reliability of issuing
 CH A PTER 8 System Testing and Deploym ent aaaa 261

credit cards to prospective applicants. He appointed a staff member to undergo the ini-
tial training. She, in turn, trained a volunteer as backup.
Training went well. After the system was successfully installed on the local area
network’s server, the vice president called for an area meeting to announce the official
operation of the applicant assistant and to recognize the contributions of the division
manager in the development process. A similar announcement was also placed in the
bank’s employee newsletter on the intranet later that month.
In a situation where resistance could kill a worthwhile project, a champion makes
a difference.The KM system cost the bank $80,000. Based on the bank’s past records, it
paid for itself in 7 months by eliminating high-risk applicants.

User Training and Deployment

A major component of system deployment is user training. The level and duration of
training depend on the user’s knowledge level and the system’s level of complexity.
User training is also influenced by the requirements of the KM system. Knowledge-
based system requirements range from simple and user friendly to advanced, which re-
quires special vendor training. A user friendly KM system is almost self-instructional,
with menu-driven features and easy-to-follow online and hard copy manuals. The dura-
tion of training is usually measured in hours to days. Advanced company-wide systems
require knowledge-based infrastructure, use of queries, and so forth, which take days
to weeks to learn, even for the experienced user. The level of user training depends on
the following:
• The user’s knowledge of knowledge-based systems
• The complexity of the KM system and how well it accommodates the user
• The trainer’s technical experience and communication skills
• The environment where training is carried out
One way or another, training must be customized to the specific user after consid-
ering capabilities, experience, and system complexity. In situations in which both the
user and the system are new, training procedures are introduced in stages. For example,
training first-time users on a stand-alone KM system begins with showing the user how
to turn on the system, enter a query, load files, display solutions, update rules, and so on.
The most effective training is supported by a well written help file, easy-to-use
explanatory facility, and job aids. A help feature can be an invaluable training docu-
ment, especially if the user is geographically isolated. It is usually highly illustrated and
contains an index for reference. Graphics, photographs, and templates provide quick
learning and are invaluable teaching aids.

PREPARING FOR KM SYSTEM TRAINING

Introducing a KM system into the organization affects personnel responsibilities and,
in some cases, can alter the structure of an organization. The very concepts underlying
a KM system often change the way employees think about tasks. When a system is
introduced, the initial goal is to educate the user about the new system. A strategic edu-
cation plan prepares an organization to adopt it as it becomes ready to deploy. Such a
plan takes place before development and cannot be a substitute for training. Hands-on
training provides specific skills to operate a system. The combination of a strategic edu-
cation plan and well-planned training means quicker organizational acceptance, which
assures successful deployment.
262 7777 PART III Knowledge Codification and System Implementation

One way to promote successful KM system deployment is to follow a few simple steps:
• Define how the KM system agrees with the organizational mission.
• Demonstrate how the KM system can help meet organizational goals.
• Allocate adequate resources to a feasible project.
• Advocate positive effects of KM systems, but do not create unrealistic expectations.
• Perform cost/benefit analyses of KM system technology and sell the company on
the system’s benefits and potential.
To raise awareness, a company might distribute Post-it notepads reading “KM
System, (name of the company)” to all employees everywhere.

CO M BATING RESISTANCE TO CHANGE

A big mistake many companies make is that they assume that the intrinsic value of a
KM system will be welcomed with open arms. This assumption is often shot to pieces.
Regardless of what is being converted or how well deployment is carried out, final
implementation means change, and people tend to resist change. KM system deploy-
ment is the initiation of a new order of things. People understandably become anxious
when they do not know what the system will do and how it will impact their job and
career path. The result is often stress and increased resistance to change. The resistors
are at several organizational levels:
• Knowledge hoarders are usually ambivalent about the value and potential of
knowledge sharing. Those on the way to retirement worry less, but they may still
lack motivation unless they are properly compensated for their efforts.
• Company employees in general resent lack of recognition (sometimes compensa-
tion), especially when they have been involved in building the KM system from
its inception.
• Troublemakers are those left out of the system building cycle or chronic com-
plainers who tend to obstruct the installation, cause delays, or force cancellation
of the system.
• Narrow-minded “superstars ” are usually technical people in the organization’s IT
division. They often resist any change that they did not initiate or approve in
advance. Others veto a project not in their area of interest. This is where lack of
management support can spell doom for a KM system.
Resistance is displayed in three psychological reactions:
1. Projection: Hostility toward peers
2. Avoidance: Withdrawal from the scene, such as calling in sick
3. Aggression: Killing the system because of uncertainty about its direct benefits
One category that explains resistance to change is the extrinsic or intrinsic value
that users place on knowledge and decision making. In most organizations, knowledge
means power. Although knowledge developers build systems that promote knowledge
sharing through knowledge bases, company experts who stand to lose monopoly on
knowledge may resist a new KM system installation. Most resistance relates to the per-
ceived impact of the new system on one’s job or status in the organization.
Resistance also has much to do with the individual personality, the organizational
structure, and group relations within the area where the KM system will be in use. User
education, professional training, and participation in the building process can help
reduce resistance to change.
 CH A PTER 8 System Testing and D eploym ent ■ a a a 263

Another resistance category comes from the system itself rather than people. If a
KM system offers poor user interface and requires extensive training to master, it could
spell disaster for effective deployment. There is no doubt that user friendly, knowledge-
based systems are a sure way to overcome resistance to KM system deployment. Ideally,
what is needed is a user who has an open mind about change and is “friendly” to tech-
nology and the concept of knowledge sharing via technology. Overconcentration on
technology and overlooking behavioral issues have resulted in many KM system failures.
Because a major user concern in system deployment is how to work the system,
users frequently ask, “W hat functions are available?” “How do I access each func-
tion?” “How do I know if the system has answered my questions correctly?” Another
user concern is how the KM system selects the right knowledge—how it reaches con-
clusions or lines up with the problem at hand. The knowledge developer must demon-
strate the system and provide detailed training in a timely manner.
With these ideas in mind, there are various m ethods to prom ote KM system
deployment:
1. User-attitude survey
2. Communication training
3. Training sessions
4. Role negotiation
In a user-attitude survey, opinions are collected from actual users to learn how well
they liked the system and how closely it met their requirements. Poor communication
skills could be a problem, but the survey is still worth the effort. Communication train­
ing can prove to be invaluable for enhancing user-knowledge developer relationships
and successful system deployment.
For small to medium-sized KM systems, training sessions are normally run by the
knowledge developer. For larger systems, they are conducted by a knowledge specialist
who is expected to be a skilled communicator. In either case, the trainer should address
the user’s training needs and gear the pace of training accordingly. Some users may
learn the system in one day; others take much longer.
Resistance to change becomes obvious when users perceive adverse changes in
their jobs. An interesting technique, called role negotiation, attem pts to clarify what
the user expects the altered job to offer. Once understood, users have been known to
accept their roles in the change more readily.
In summary, for a new KM system to ensure user support, knowledge developers
and users must improve communication channels and jointly discuss the new system’s
features and how the change can improve their jobs. Users should also participate in all
phases of deployment. Sensitivity to user expectations is a step toward “deployment
without tears.”

Postimplementation Review
A fter the KM system has been deployed and the operation is up and running, the
effect of the new system on the organization should be carefully evaluated. System
impact must be assessed in terms of its effect on people, procedures, and performance
of the business. More specifically, the main areas of concern are quality of decision
making, attitude offend users, and cost of knowledge processing. For example, a post-
implementation study of the Travel Profiler showed an increase in the number of over-
all travelers accommodated by a factor of 2.5. It also showed that travel agent time had
improved by 20 percent.
264 7777 PART III Knowledge Codification and System Implementation

In the postimplementation stage, several questions should be considered:

1. How has the KM system changed the accuracy and timeliness of decision making?
2. Has the new system caused organizational changes? How constructive have the
changes been?
3. How has the new system affected the attitude of the end users?
4. Has the new system changed the cost of operating the business? In what way?
5. Has the new system affected the organization’s decision-making process? What
measurable results can be demonstrated?
6. In what way has the new system affected relationships among end users in the
organization?
The objective is to evaluate the KM system against standards and determine how
well it meets company requirements. This process is actually related to user acceptance
evaluation. The user initiates the review, which prompts a procedure for maintenance
or enhancement.

SECURITY CONSIDERATIONS
Safeguarding the KM system against unauthorized access is a necessary system deploy-
ment issue. Similar to a conventional information system, at a minimum, the new sys-
tem should provide password or protocol protection so that only users with the correct
code can log onto the system. One aspect of user training is to ensure that passwords
and security procedures are consistently observed. Beyond this point, restricted access
regarding the update of the knowledge base itself can be secured through the existing
LAN or the server in operation.

■ Implications for Knowledge M anagem ent

The main goal of KM system testing is to assure continuous quality and high perfor-
mance. The term performance implies quality. The model shown in Figure 8.4 suggests
that the quality of a KM system is affected by the nature of the problem, technical con-
siderations, the people involved, and organizational factors. For example, if the prob-
lem is complex and the organizational climate or politics is amenable to the prospec-
tive system, then a knowledge developer can expect cooperation from the expert and
the user during knowledge codification. This cooperation should contribute to quality
testing. The end result should be a high performance, quality-oriented KM system.
One of the main issues raised during deployment is maintenance. Maintenance of
KM systems continues to be a nebulous area. Some of the questions to be addressed by
management are as follows:
• Who will be in charge of maintenance ?
• What skills should the maintenance person have? What is the best way to train
him or her?
• What incentives should be provided to ensure quality maintenance?
• What type of support and funding is needed?
• What relationship should be established between the maintenance of a KM sys-
tem and the IT staff of the organization?
A nother managerial issue to consider is how one would know whether deploy-
ment will be a success. In addition to technical considerations, one must be aware of the
people factor. People rallying behind technology can increase the success of deploy-
ment. Additionally, a KM system cannot succeed if the issue of cost of technology and
 CHAPTER 8 System Testing and Deployment ■ a a i 265
People

Domain
Expert

People KM System Technology

Quality

User Knowledge
Developer

Organizational
Climate

fl fl fl fl fl fl FIGURE 8.4: Internal and External Factors Affecting Knowledge-Based System Quality

how it is going to be absorbed has not been resolved. Any of these issues can inhibit
successful deployment.
Other managerial issues affect the direction and outcome of KM system projects.
For example, the depth of the knowledge required to solve the problem is often an
unknown commodity due to the nature of the expertise and the limitations of the
knowledge developer. Forcing clarification too early in the building phase can result in
a less-than-satisfactory system. Therefore, managing expectations and scoping the sys-
tem are crucial early tasks.
Finally, the role of the hum an expert in system deploym ent is also im portant.
Management should properly compensate the expert for efforts in developing a KM
system. A manager may also play a proactive role in the deployment phase by simply
sharing interest or excitement in the system. A lack of this type of enthusiasm and
overt support puts a damper on the whole project.

S u mma r y 1111
• KM system reliability is one of the most im portant issues in knowledge-based sys-
tems. Reliability means how well the system delivers information with consis-
tency, accuracy, and integrity. The prime emphasis beyond reliability is quality
assurance and maintainability of the system after installation.
• The most important part of building KM systems is testing. Logical testing
answers the question, “Are we building the system right?” and its goal is an error-
free program or software. User acceptance testing answers the question, “Are we
building the right system?” and tests the system’s behavior in a realistic environ-
ment. Such testing is done only after the system is operational.
266 ■ ■ ■ ■ PART III Knowledge Codification and System Implementation

• Some important challenges or issues that are considered during testing include
the following:
• The subjective nature of knowledge-based testing
• Lack of reliable specifications
• The problem of establishing consistency and correctness
• The determination of what constitutes an error
• Sources of test data
• The danger of negligence in testing
• Adequacy of automatic logical testing tools
• The complexity of user interfaces
• A circular error tends to be contradictory in meaning or logic. Redundancy errors
offer different approaches to the same problem. An unusable rule always fires
and fails, never fires, or has one or more contradictions. In a subsumption error, if
a rule is true, one knows the second rule is always true. In inconsistency errors,
the same inputs yield different results.
• The steps for user acceptance testing are as follows:
• Select a person or a team for testing.
• Decide on user acceptance testing criteria.
• Determine objective measures for the selected criteria.
• Develop a set of test cases and scenarios unique to the problem.
• Maintain a log or various versions of the test and test results.
• Some user acceptance testing tools are face validation, test cases, and subsystem
validation.
• Three im portant deployment issues are worth noting: selection of the problem,
knowledge codification and deployment, and organizational factors.
• Technology transfer includes actual transfer of the KM system from the devel-
oper to the working unit or the transfer of KM system technology skills.
Deployment can be approached as a stand-alone installation on a PC or as a fully
integrated application that interfaces with other applications.
• A KM system can be integrated into an organization in a number of ways:
• Technical integration
• Knowledge-sharing integration
• Decision-making flow integration
• Reengineering considerations
• The level of user training depends on the user’s knowledge of the KM system, com-
plexity of the system, the trainer’s experience and communication skills, and the
environment in which training is carried out. In any case, training should be geared
to the specific user and, in some cases, introduced in stages for a lasting impact.
• Regardless of what organizational aspects are being changed, deployment means
change, and people in general resist change. The resistors included experts, non-
experts, troublemakers, and narrow-minded technical superstars. Resistance is
displayed in the form of projection, avoidance, and aggression and has much to do
with the individual personality, the organizational structure in which the user works,
and the group relations in the area where the system will be installed. User edu-
cation, training, and participation can help reduce or control resistance to change.
• Deployment can be promoted in a number of ways: user-attitude survey, commu-
nication training, training sessions, and role negotiation. In any case, sensitivity to
user expectations is a step toward implementation without tears.
 C H A PTER 8 System Testing and D eploym ent aaaa 267

Terms t o K n o w 1 1 1 1
Aggression: Resistance to KM systems through em ployee Pupil user: Unskilled worker trying to learn or gain some
sabotage of the system. understanding of the captured knowledge.
Avoidance: Resistance to KM systems through em ployee Redundancy rule: A rule that offers a different approach
withdrawal from the job or scene. to the same problem; duplication or meaning the same.
Circular rule: A rule that has embedded contradiction in Reliability: Dependability; truthfulness of the response or
meaning or logic. answer to a given question; credibility; how well the
Deployment: Physical transfer of the technology to the system delivers solutions with consistency, accuracy, or
organization’s operating unit. integrity; detecting or removing anomaly.
Enhancement: Upgrading the system to m eet a new set of Subsumption: In rules, if one knows one rule is true, one
requirements. also knows the second rule is always true.
Exhaustive testing: A procedure in which all possible Technical integration: Integrating an expert system
combinations of input values are tested. into an existing operation through the firm’s local
Face validity: Testing a system at its face value; comparing area network environment, the resident mainframe,
a human domain expert’s value judgment to test results workstations, and other information system
for reliability. applications.
Inconsistency: A rule that has the same input but differ- Testing: Time-intensive logical and user acceptance test-
ent results. ing of KM systems.
Knowledge-sharing integration: Integrating the expert Unusable rule: A rule that only fires if the conditions suc-
system in such a way that it can be accessed by differ- ceed, one that never fires, or one that has one or more
ent branches of the company, allowing them to share contradictions.
information. User acceptance testing: A system test to ensure the right
Logical testing: A system test to ensure the proper func- system from the user’s perspective; a system that meets
tioning of the system; addresses the intrinsic properties the user’s expectations; beta test; user acceptance test.
of the KM system; alpha test. User-attitude survey: A survey in which opinions are col-
Projection: Resistance to KM systems through em ployee lected from users to learn how well they like the system
display of hostility toward peers. and how closely the system meets user requirements.

Test Y o u r Understanding 1 1 , 1
1. What is the goal of logical testing? What is the goal of user acceptance
testing?
2. In your own words, explain the various phases of KM system testing.
3. Distinguish between:
a. logical and user acceptance testing
b. circular and redundant rules
c. force-fail test and exhaustive test
4. The chapter discusses two approaches to verifying knowledge-based systems.
Discuss each of them briefly.
5. How does verification of knowledge base formation differ from knowledge
base functionality?
6. Give examples of the following:
a. circular error
b. redundant error
c. subsumed error
d. unusable rule
7. In terms of securing test cases for user acceptance testing, where should the
test cases come from? How would one assure their reliability or relevance?
8. Who should do the user acceptance testing? Should it be the end user or the
domain expert? The knowledge developer? Explain.
9. In what way(s) is each of the following validation tools useful:
a. face validation
b. test cases
268 7777 PART III Knowledge Codification and System Implementation

10. If you were coordinating user acceptance and logical testing, what guideline
would be useful before the actual test?
11. How can system users be assured of system success?
12. In what way(s) is knowledge codification related to user acceptance testing?
13. “Successful KM system implementation depends on several factors.” Briefly
explain each factor.
14. Elaborate on the role of maintenance in the life of an expert system.
15. Distinguish between:
a. maintenance and enhancement
b. pupil and tutor user
c. projection and avoidance
16. How would one approach transfer of technology in system implementation?
Which approach do you suggest for implementing a small system on a stand-
alone PC? Why?
17. There are several alternatives for integrating KM systems into the organiza-
tion. Briefly explain each alternative. U nder what condition(s) would one
choose one alternative over the others?
18. Why is place of training important? What about duration of training? Explain.
19. What determines whether user training will be successful? Be specific.
20. How would one promote successful KM system implementation?
21. What is postimplementation review? Why is it important to carry out this step?

Kn o w l edge Ex e r c is e s 1,11
1. Discuss the key issues during logical and user acceptance testing. Is one issue
necessarily more critical than all others? Be specific.
2. Discuss the role of security in KM system use. How important is KM system
security compared to security measures in conventional information systems?
3. If you were asked to coordinate the implementation phase of a KM system in
a department where users are not that interested in the change, what approach
or procedure would you follow to combat resistance to change? Cite a specific
situation or the conditions under which you would handle the assignment.
4. If a KM system is user friendly and has the necessary help facility, how nec-
essary is a user’s manual?
5. How important are organizational factors in system implementation?
6. Discuss the main steps to user acceptance testing. How and why are they
important?
7. “A strategic education plan prepares an organization to adopt a KM system.”
Do you agree? Why or why not?
8. If testing is never that perfect or complete, why test anyway?
9. When a KM system is finally installed and ready to go, is the knowledge
developer released from further obligations to the system? Why?
10. Consider the following rules:
R1 IF 0 = 10
AND b = 15
THEN c = 20
R2 IF a = 10
THEN c —20
Of the two rules, which rule is subsumed by the other?
11. Consider the following rules:
IF tem perature > warm AND
humidity is high AND
atmospheric pressure is low
THEN there will be thunderstorms
 CH A PTER 8 System Testing and D eploym ent 1 «« i 269

If tem perature > warm AND

humidity is high
Then there will be thunderstorms
Which rule is subsumed by the other? Why?
12. Consider the following rules:
R1 IF a = 10
AND b = 15
THEN c = 20
R2 IF b = 15
AND a = 10
THEN c —20
Are these two rules redundant? Why? Which rule is made redundant by the
other?
13. Consider the following rules:
R1 If£i = 10
AND b = 15
THEN c — 20
R2 IF 0 = 10
AND b = 15
THEN c — 35
These rules are said to be in conflict. Do you agree? Why or why not?
14. Consider the following rules:
R1 IF tem perature > warm AND
humidity is high
THEN there will be no thunderstorms
R2 IF humidity is high AND
tem perature > warm
THEN there will be no thunderstorms
In what way(s) are these rules redundant? Will they always succeed?
15. Consider the following rules:
R1 IF humidity is high AND
tem perature > warm
THEN there will be thunderstorms
R2 IF tem perature > warm AND
humidity is high
THEN there will be sandstorms
Are these rules redundant? In what way(s)?
16. Identify whether the following rules are redundant, conflicting, circular, or
subsumed rules:
a. R1 IF a = x R2 I F b = y
AND b = y AND a = x
THEN c = z THEN c = z
b. R1 IF a = x R2 IF b = y
AND b = y AND a = x
THEN c = z THEN c = w
c. R1 IF a = x R2 IF b = y
AND b = y THEN c = z
THEN c = z
d. R1 IF a = x R2 IF b = y R3 IF decision = yes
THEN b = y AND c = z THEN a = x
THEN decision = yes
17. A KM system is being designed to assist a computer retailer’s customer ser-
vice clerks with managing calls related to replacement of Pentium 4 chips in
PCs. How should the performance of the KM system be evaluated? Explain.
270 7777 PART III Knowledge Codification and System Implementation

18. Implementing the Real-Estate Advisor:

In November 2000, a real-estate firm expressed interest in a KM system
to train new agents and advise senior ones on the best match between a
prospective home owner and the available homes in the area. The firm had
14 offices in eight major cities in the Commonwealth of Virginia. A contract
was drawn to do the work with deployment by mid-October.
The resulting knowledge base consisted of 165 rules involving 17 parame-
ters. Testing the package took 2 months before it was ready for deployment. The
first step was to meet with two volunteer agents from the main office to demon-
strate what the system could do and how easy it was to use. They asked how
long it would take. The reply was “3 to 4 hours, total.” They suggested a meeting
with the knowledge developer the following Sunday to give the system a try.
On Sunday morning, the two real-estate agents came on time and were
eager to see what the KM system could do to improve their sales potential.
After 2 hours of practice, they began to make suggestions about expanding
the system to include variables that had not been raised earlier. For example,
they wanted the system to ask if the prospective buyer would object to living
in an integrated neighborhood, whether the buyer had transportation to and
from work, and so forth.
The result of the 3-hour session was frustration. The agents were courte-
ous, but they did not indicate that they would use the system. Because they
worked on commission, there were under no immediate obligation to try it.
A strategy meeting with the president of the firm led to the decision to hold a
session with the 11 real-estate agents the following week and discuss the deploy-
ment plan. During the 1-hour session, several questions were raised regarding the
system’s capabilities and how it was built. Luckily, the single most experienced
agent used in building the system was a highly respected agent with 18 years of
successful experience. The expert was there and offered his opinion of the system.
Suddenly, the president cut into the discussion and said, “Look, we can
beat the system over the head and still not agree on a direction. I have an
idea for you to consider. Why don’t you give this system a try, and after you
feel you have a handle on the way it works, see if it can help you match a
buyer with the right house. I am willing to increase anyone’s commission
from 50 percent to 60 percent for every documented match that results in a
sale. I am willing to do this for 90 days.”
By the end of the session, 6 out of the 11 agents underwent a 1-day train-
ing session at company expense. That was in February 2001. By June, 46 suc-
cessful matches had been made that resulted in 38 documented sales.
The word got around to the branches regarding the usefulness of the
R eal-Estate Profiler. During the succeeding 5 months, several suggestions
were incorporated into the system under a maintenance agreement with the
firm. In December 2001, the system became available to the branches under
the 60 percent commission plan. Training became easier, as more agents
expressed interest in the system.
Questions
1. Evaluate the approach used in deploying the Real-Estate Advisor.
2. Should one expect changes in the system to take place during deploy-
ment? What went wrong, if anything? Explain.
3. In what other ways might this system have been deployed?

Re f e r e n c e s 1111
Awad, E. M. “Building Travel A dvisor; with EXSYS Awad, E. M. “ Test Cases For the Knowledge Recruiting
R e s o l v e r Unpublished manuscript, June 2001. A d v i s o r Unpublished manuscript, January 2002.
 CH A PTER 8 System Testing and D eploym ent a a a a 271

Awad, E. M. “ Validating and Verification of a Breath Gas Gupta, Uma G. “Successful Development Strategies:
Monitoring System,” Unpublished manuscript, Au- Moving to an Operational Environment,” Information
gust 2001. Systems Management, vol. 9, no. 1,1992, pp. 21-27.
Ayel, M., and Laurent, J. P “SACCO-SYCOJET: Two Horibe, Frances. Managing Knowledge Workers. New
Different Ways of Verifying Knowledge-Based York: John Wiley & Sons, Inc., 1999, pp. 109-126.
Systems,” in M. Ayel and J. P. Laurent (eds.), Verification Jafar, Musa, and Bahill,Terry A. “Interactive Verification
and Validation of Knowledge-Based Systems. New York: of Knowledge-Based Systems,” IEEE Expert, Febru-
John Wiley & Sons, Inc., 1991, pp. 63-76. ary 1993, pp. 25-32.
Freedman, Roy S. “Knowledge-Based Software Testing,” Tiwana, Amrit. The Knowledge Management Toolkit.
The Journal of Knowledge Engineering, Spring 1991, Upper Saddle River, NJ: Prentice Hall, 2000, pp. 99-109.
pp. 47-65.
 Knowledge Transfer
and Knowledge
Sharing
Contents

In a Nutshell
Knowledge Transfer As a Step in a Process
The Knowing-Doing Gap
Prerequisites for Transfer
Instill an Atmosphere of Trust in the Organization
Transfer Methods
Nature of the Problem
Transfer Strategies
Inhibitors of Knowledge Transfer
How Knowledge Is Transferred
Role of the Internet in Knowledge Transfer
Internet Service Providers
Stability and Reliability of the Web
Unique Benefits of the Internet
Limitations
Implications for Knowledge Management
Summary
Terms to Know
Test Your Understanding
Knowledge Exercises
References

A m an has no ears
fo r that to which experience
has given him no access.
—F r ie d r ic h N ie tz s c h e
 C H A PTER 9 Knowledge Transfer and Knowledge Sharing a a a a 273

In a Nutshell
We have seen that for knowledge to be shared, it must first be captured, codified, and
deployed in a format acceptable to the user. The goal is to turn knowledge into action
or to transform individualized learning into organizational learning. This is why this
material was placed after the KM system life cycle. In most instances, the life cycle is a
prerequisite to knowledge transfer and knowledge sharing
Knowledge sharing is more than simply knowing the right thing to do. Knowledge and
information are obviously critical to performance. However, we live in a new millennium,
where there are a growing number of organizations in the business of capturing and transfer-
ring best practices. The goal is to narrow the gap between what they know and what they do.
Part of a changing environment is first to gather intelligence (such as information,
added value, and tacit knowledge) that triggers insight (responsiveness) to generate innova-
tion in products, services, and processes. As part of the KM system life cycle, once knowl-
edge is captured, codified, tested, and deployed, the next step is to demonstrate how it is
transferred to the right party in the right format and at the time needed. Knowledge trans-
fer is a prerequisite for knowledge sharing for competitive advantage, performance, and
profitability. Knowledge sharing is making available what is now known. Knowledge trans-
fer’s role in the building life cycle is shown in Figure 9.1. It may come from knowledge bases,
databases, or via the Internet. A knowledge-based system that helps a user answer technical
questions on the job is an example of knowledge that is being shared with that user.
Of course, there are other ways to transfer and share knowledge. Knowledge may be
transferred between persons and computers, computers and computers, teams and individ-
uals, or between individuals. In this chapter, each option is covered in detail to reflect the
multifaceted nature of knowledge transfer. We also discuss transfer strategies and methods,
inhibitors to knowledge transfer, and how explicit and tacit knowledge are transferred.

Knowledge Transfer As a Step in a Process

Knowledge transfer is part of organizational life. It is carried out whether the process is
managed or not. It is transmission of knowledge (experience, lessons learned, know-how)
and use of transmitted knowledge. It is conveying the knowledge of one source to another
source. Merely making knowledge available is not considered to be knowledge transfer. It
is transmission plus absorption. The goal is to promote and facilitate knowledge sharing,
collaboration, and networking. It is access to scarce resources, new insights, new expertise,
cross fertilization of knowledge, and creating an organizational environment of excel-
lence. Collaboration means the ability to connect diverse assets into unique capabilities in
pursuit of new opportunities for personal, group, and organizational growth.
Knowledge transfer is done directly by working together, communicating, learning
by doing, apprenticing, through face-to-face discussions, or em bedding knowledge
through procedures, mentoring, or docum ents exchange. It is not done solely from
knowledge bases or repositories. Knowledge can be transferred from repositories to
people, from teams to individuals, and between individuals. When an IT employee asks
another how to read an HTM L page, she is requesting a transfer of knowledge. If a
loan officer asks another in the office down the hall if he has ever dealt with a graduate
student loan, the second loan officer, if interested, is transferring his knowledge. In any
case, even transmission and use of knowledge have no useful value if the acquired
knowledge does not result in some new behavior or value-added learning.
It is obvious that knowledge transfer is part of everyday organizational life.
Discussing a problem with someone next door is generally by convenience. It is rare
274 7777 PART III Knowledge Codification and System Implementation

I I fl ■ ■ I FIGURE 9.1: Knowledge Transfer in the KM System Building Life Cycle

that one looks organization-wide for the best expert or the one with the deepest
knowledge. This brings up the concept of bounded rationality. H erbert Simon, a
renowned behavioral scientist at Carnegie Mellon University, coined the term to rec-
ognize hum an limits to a person’s knowledge of all alternatives and their conse-
quences. The hope is to get “good enough” inform ation from someone nearby. The
“good enough” view is referred to as satisficing rather than optimizing.
To transfer knowledge, we consider three factors:

1. W here knowledge is transferred from —knowledge bases, innovations, lessons

learned, programs, books, experts, articles, or data warehouses
2. The media used in knowledge transfer—local area networks, wireless transmis-
sion, secure versus unsecured lines, or encrypted or plain text
3. Where knowledge is transferred to —apprentice, user, another computer, a team,
any authorized individual, a manager, a customer, or an autom ated product maker
 C H A PTER 9 Knowledge Transfer and Knowledge Sharing a a a a 275

It should be noted that only a limited amount of human expertise can be captured
as explicit knowledge to be part of a knowledge base application. This means that most
of human expertise is tacit knowledge, with practical limits as to how much or what
types of knowledge can be transferred between people or within a team configuration.
In any case, knowledge transfer facilitates knowledge sharing. The collective
knowledge of an organization is gleaned from the learning of its members as well as its
stakeholders and customers. Knowledge transfer enhances knowledge flows through
collaborative technologies, especially the intranet and the Internet. Hard, high-tech
communications technologies and networks serve as the enablers of knowledge trans-
fer and assure constant availability of knowledge on demand. This is best enforced by
support from top management (see Box 9.1).
Figure 9.2 shows a partial view of select knowledge transfers and their respective
knowledge m anagem ent applications. Knowledge transfers distribute to points of
action—support education and training, act as customer service representatives, and
so forth. In contrast, knowledge applications use knowledge to deliver' products and
services—assist knowledge workers to perform unique functions in a real-time envi-
ronment, make knowledge available to solve difficult problems, and the like.

a 9 ■ fl fl 9 BOX 9.1 9 9 9 9 9 9

A FRAMEWORK FOR KNOWLEDGE TRANSFER

One of the challenges of managing an E R P valuable knowledge from software or the opera-
(enterprise resource planning) implementation is tional system.
understanding that not all entities within the To gain the benefits inherent in this auxiliary
organization will instinctively see the value of knowledge, organizations must create and deploy a
knowledge capture and management. Farsighted knowledge system. The starting point for this system
executives, however, are beginning to appreciate is the development of a knowledge transfer frame-
this shift to a knowledge-based economy and are work. Such a framework identifies which informa-
investing wisely in building both knowledge con- tion you should retain, who will eventually use that
tent and the technologies that facilitate sharing, information, and how you will capture that informa-
management, and access to knowledge. tion at each stage of the ER P implementation. In
An ER P system contains a great deal of codi- addition, the knowledge transfer framework pro-
fied knowledge, both in the software structures vides a model of the knowledge environment over
provided by the vendor and in process knowl- the life cycle of a system implementation and illus-
edge and business rules developed by the enter- trates the relationships and dependencies of people,
prise. However, a corresponding set of auxili- business processes and events, and knowledge
ary knowledge is captured only in the form of domains. With this framework in place, you can
project documents (formal and informal), train- anticipate and plan for knowledge requirements of
ing documents, data models, additional business subsequent phases of the ERP project. You can also
rules and formulas, and in the brains of the proj- develop mechanisms for transferring or transform-
ect team. You cannot easily access or derive this ing the knowledge base from one phase to another.

SOURCE: Excerpted from Clark, Linda E. “A Stitch in Time,” Intelligent Enterprise Magazine, October 20,2000, p. 2.
276 aaaa PART III Knowledge Codification and System Implementation

Knowledge Transfer Knowledge Application

a a a a ■ ■ FIGURE 9.2: A Partial View of a KM System for Knowledge Transfer and Knowledge Application

Related to this view is the role of knowledge-based system applications in support

of knowledge management. Knowledge-based applications impact both the quality of
results obtained and the effort required to complete knowledge management tasks. Of
course, knowledge-based applications are no match for the human mind, where tacit
knowledge resides. Only a limited amount of a person’s expertise can be elicited for
incorporation into knowledge-based applications. However, there is a growing trend in
artificial intelligence that is beginning to add more “sm art” to such applications, which
are becoming closer and closer to the human way of decision making.

THE K N O W IN G -D O IN G GAP
Many organizations know what to do, but for various reasons ignore the information
that is available and perform differently—creating the “knowing-doing” gap. Recog-
nizing this problem should help firms make corrections and set up a knowledge transfer
environment for all employees to benefit. Even knowledge is not always a solution. To
illustrate, in the spring of 2001, one of us gave a 1-day seminar to senior officers of a
M iami-based banking group. The topic was “Employees Come First” —a topic that is
param ount in a service industry like banking. It has been known for years that
employee attitudes and morale affect turnover, customer satisfaction, and profitability.
 CH A PTER 9 Knowledge Transfer and Knowledge Sharing 1 1 1 1 277

At the end of the seminar, one senior executive vice president of loans made com-
plimentary remarks about the good ideas she got from the session, but added that none
of those ideas would work at the bank. The bank’s president, she said, had no appreci-
ation for the rank-and-file staff (tellers, bookkeepers, and customer service clerks) of
the bank. Every time he passed by her employees, in the hallways or lunchroom, all he
wanted was to make sure employees did not take too much time off. At the end of the
year, it was always a big fight to get him to approve salary increases based on perfor-
mance that was carefully assessed by the human resources department. All he worried
about was what his bonus would look like based on the bottom line. Short of replacing
the president, there was not much that could be done to implement change.
An hour later, there were two senior officers from another bank still in the seminar
room. Although their bank recently implemented performance measures due to a recent
acquisition of a smaller bank, they had condensed their notes on paper napkins to take
back to the office for follow-up. One officer said, “Although we have activated much of
what you covered, the seminar gave us a different way of finalizing the performance
measures. To tell you the truth, we did a lousy job in sharing the information with those
whose jobs will be affected by the change.” The other officer was already on her cell
phone, reporting to her boss the changes she wanted to make first thing next Monday.
The first episode illustrates a knowledge-hoarding, m anipulative organization
under tight top management control. In contrast, the latter episode is an example of a
learning organization on the rise. The spirit of cooperation and collaboration was obvi-
ous in the way the officers worked together, clarified their notes, contacted the head
office, and showed interest in doing something positive for the good of the organiza-
tion. It is an environment where knowledge transfer and knowledge sharing should be
no problem.

PREREQUISITES FOR TRANSFER

The terms transfer and share are interrelated. Knowing is considered deeply personal.
In one respect, it asks som eone to give something that is his or her own property.
Knowledge transfer is a mechanistic term, which provides knowledge for someone else.
The term share is an exchange of knowledge between individuals, between or within
teams, or betw een individuals and knowledge bases, repositories, and so forth.
Knowledge sharing recognizes the personal nature of people’s knowledge gained from
experience. What triggered the writing of this book is the growing trend of more and
more organizations that began doing more than mere talking about sharing knowl-
edge. They started to install knowledge bases and processes to transfer knowledge to
others to get around the problem of reinventing the wheel. E-mail notices from col-
leagues such as “Does anyone know where to . . . ?” and “Would anyone know how
to . . . ?” are examples of using the technology to promote knowledge sharing.
It should be noted that technology alone is not a sure prerequisite for knowledge
transfer or knowledge sharing. The myth that “once you build it, they will use it” does
not work that well in most organizations. The image of a huge high-cost repository that
contains all the knowledge ever needed offers a feeling of com fort and control.
However, neither cooperation nor retrieval occurs with much excitement. When orga-
nizations find that not much is happening, they begin to use the incentive system to
encourage participation. The consulting firm KPMG coined the “give to get” scheme,
where employees cannot take any knowledge w ithout putting som ething into the
knowledge base or into the relationship that provided them the knowledge. In any
case, for knowledge transfer to work, it takes a change in culture, in politics, and atti-
tude to make things happen.
278 7777 PART III Knowledge Codification and System Implementation

As can be seen, transferring knowledge into action is not straightforward. There

are interpersonal, political, leadership, and organizational issues to consider. There
are also assumptions to be made about the organization and its processes. However,
there are recurring episodes and guidelines that help us evaluate the gap between
knowledge transfer and knowledge sharing and how to overcome it.

Instill an Atmosphere of Trust in the Organization

Trust is the foundation of knowledge transfer. It is a psychological state, where
people feel confident about sharing ideas, experiences, and relationships with others.
For decades, people in organizations have learned to hoard knowledge rather than to
share it. There is still inherent fear that once knowledge is shared, the knower (person
who has the knowledge) becomes vulnerable. In a local retail store, the manager asked
the shoe departm ent supervisor of 9 years if she would train a new employee before
she would be a candidate to manage another departm ent in a store being built in a new
mall on the other end of town. Toward the end of the 5-week on-the-job training, the
store manager heard from headquarters that the proposed store would not open due to
recent financial difficulties of the chain. The decision was made to retain the newly
trained supervisor for less salary. The word got around that neither the manager nor the
retail chain could be trusted. It was not the first time this type of episode took place.

Fix the Culture to Accommodate the Change

The trust factor has cultural and political implications. An organizational culture
that prom otes and rewairds cooperation and collaboration is a better candidate for
knowledge sharing than a competitive culture that drives employees to performance
through the fear of punishment (such as layoffs or dismissal). There is the case of a
2000 graduate from a premier MBA school who accepted a 6-digit salary with a fast
growing consulting firm. When the recession hit in 2001, the company laid off 40 per-
cent of its staff without notifying anyone. Early in the year, this graduate sensed things
were not on the level. Managers kept information to themselves, and partners were
working feverishly to attract new clients. He soon decided to keep to himself his
technology-based knowledge required for installing two major projects for Fortune 500
clients. That strategy saved him from being axed by the sudden layoffs.
Culture is usually embedded in a company’s core values, policies, mission, consistent
behavior, treatment of employees, and tradition. A trusting organization tends to attract
trustworthy employees who perform well within the cultural values of the business. Those
who do not fit voluntarily leave or are soon asked to resign. With that in mind, the “learn-
ing organization” notion should be transmitted through the organization’s culture. Becker
(1998) cites several dimensions of values and beliefs making up an organization’s culture:
• Authority • Motivation
• Collaboration • Mistake tolerance
• Commitment • Participation
• Compensation • Partnering
• Competence • Teams
• Conflict resolution • Truth, openness
• Consistency • Self-management
• Cooperation • Risk tolerance
• Creativity • Innovation
• Empowerment • Change
• Fairness • Focus
 CH A PTER 9 Knowledge Transfer and Knowledge Sharing aaaa 279

Positive cultural values include the following:

• Leadership. For example, a leader is a champion who promotes change and one
who is highly regarded by the employees.
• Understanding company mission. For example, a company can have a clear mis-
sion statem ent regarding knowledge management and how it plans to embed i t ' i
the organization.
• Culturally internalized management practices. For example, a company can have
enthusiastic management that is optimistic about the changing environment.
• Culturally internalized operational practices. For example, a company can make
decisions and solve problems by using collaboration and teamwork.
• Culturally driven forces. For example, employees must trust one another in a
shared learning atmosphere. Trust in others generates honesty and fairness for all
participants.
In contrast, the key factors that retard cultural values include the following:
• Culturally driven forces. For example, specialists may hoard knowledge or deny
that they have sought after knowledge. It is also the attitude of “I don’t care” or
“It is not my job.”
• Understanding company priorities. For example, employees and management
have a fuzzy understanding of the company’s mission, policies, and strategies. Any
out-of-the-ordinary ideas or innovations are perceived as unnecessary work.
• Questionable values. For example, the company may value money or financial
goals as the ultimate goal of the business.
• Questionable beliefs . For example, a business may view employees as being
replaceable (Wiig 1993).
There is the chicken-or-egg issue of whether to fix the culture and then get the
employees to share or vice versa. There is no consensus on one approach or the other.
The myth that knowledge sharing happens only in organizations that have a collabora-
tive culture does not pass m uster in every case. The authors are of the belief that a
learning culture makes it both convenient and efficient to implement knowledge trans-
fer in the interest of knowledge sharing. Knowledge exchange, no doubt, impacts the
culture, but going that route could mean a long haul toward fixing the culture on a per-
m anent basis. Either alternative presupposes an ongoing environment of trust in the
approach or process.

Push Reasoning Before Process

It never seems to fail that employees undergoing training show greater interest in
process (“how to do”) than the reasoning (“why”) behind process. For example, in
database training using Microsoft Access, greater attention is given when trainees
learn how to create tables, access data, and enter records than is paid to knowing why a
database is designed and why you need a database for decision making. We experience
the same situation every time we teach the database m anagement system in school.
Students think it is an answer rather than a process.
When people are hired into an organization, supervisors and managers introduce
themselves, tell the newcomer who they are, what the company philosophy is, and what
they hope the new employee will be expected to achieve. That should be the first thing
to provide before employees are shown how to do the job. In our many studies in orga-
nization design, we have heard employees say, “I have been with the company 15 years,
and I have yet to meet the president or know who the board members are.” We often
280 7 7 7 7 PART III Knowledge Codification and System Implementation

respond,“You should tell management how you l'eel so that they understand that it is
not enough to know what to do, but for whom and why you’re doing it.” A learning
organization begins with employee queries, feedback, and transfer of experience
across the board.
A learning organization may also begin with a set of guidelines about how it will
operate. In one large Virginia bank, top management have set three core values for
tellers in 38 branches that guide their behavior: fairness, cordial treatment, and willing-
ness to share experience with other tellers. The same bank has a set of core assump-
tions about its employees in general:
• Respected as humans and treated fairly
• Intrinsically creative and capable of thinking and learning, if given a chance
• Can be responsible and be held accountable
• Will make mistakes and can gain experience from trial and error
• Willing to make a contribution, if given a chance
In summary, it is important to remember that when it comes to knowledge trans-
fer, people should be equipped with an attitude and a philosophy and that precedence
should not be allowed to substitute for thinking. What creates longevity and combats
turnover in a business is making use of the untapped potential within the employees
and transferring the training to help each employee work at peak performance. As a
result, a firm can learn and communicate with newcomers, regardless of job or distance,
and do so in a way consistent with its philosophy and core values. This is what creates
success and boosts performance.

Remember That Doing Is Far Better Than Talking

Once a philosophy is understood, the next guideline is to teach and train employ-
ees as a way of knowing—not merely talking, conceptualizing, and thinking. Learning
from doing things together brings up the concept of knowledge transfer via teams. In
most of today’s consulting firms, for example, junior staff just out of college are
assigned to teams with field knowledge to be mentored by senior team members as a
step toward more responsibilities in the future. Likewise, new graduates from medical
schools serve residency under certified professionals before they can take on the chal-
lenge of independent decision making. This is where they receive experience, knowl-
edge, know-how, and the many ways to handle complex decisions.
Teaching, explaining, and mentoring are ways of transforming and sharing knowl-
edge. So is doing the work and experimenting with different ways to do a job. Once you
do it, you will know. Being involved in the actual process is the best way to learn it.
Consider the case of a new car salesman who served once as an auto mechanic. He
could share his knowledge with a prospective customer about the way the car works,
how reliable it is, and where you get value for the car under consideration. In a way, this
is creativity and a unique way of using knowledge to serve a goal. If the job becomes
routine, motivation, morale, and attitude begin to degenerate toward the company, the
product, and self. In the case of the auto salesman, he has a way of sizing up the cus-
tomer, customizes his pitch, and delivers accordingly. His creativity and focus on cus-
tomized selling earned him the salesman of the year award year after year.
We live in an era of distance learning, where employees can acquire knowledge by
learning on their own. There are also jobs that are best learned by first-hand, tangible,
physical experience. The traditional way of teaching and learning is for the trainer to
stand up, show concepts and procedures through transparencies or slides, and expect
people to go back to the job and do something with it. A better way of learning is to
 C H A PTER 9 Knowledge Transfer and Knowledge Sharing a a a a 281

form a team with experienced members who show novices why a job is needed and
how it is done before they are released to do the job. Unfortunately, knowing by doing
is a costly and time-consuming way of transferring knowledge. Going that route, how-
ever, will certainly shrink the knowing-doing gap in most jobs.
Actions speak louder than words and are more effective than concepts or theory
not tested by experience. Talk without action is no longer workable in a learning orga-
nization. To think that because a decision is made through discussion that action will
follow is also losing viability.

Know How the Firm Handles Mistakes

Mistakes are bound to be made, and they are a cost for the firm to incur. Tolerating
and forgiving mistakes in the interest of learning are important. A story is told about a
young IBM executive who caused a $10 million mistake in the way he managed a high-
risk project. When brought before Dr. Watson, the founder of IBM, the young man
snapped, “I guess I am here because you will want my resignation.” Dr. Watson replied,
“You have to be kidding. We just blew $10 million to train you.”
There is rarely any “doing” without mistakes. In fact, some of the best learning in
business is by trial and error, which implies some failure or taking steps backward to
move forward. The question is in how the firm reacts. Being human, certain mistakes
are forgivable as long as the employees owns up to them. There is the story of a dairy
company that set a guideline for its truck drivers to continue on payroll. It is called “the
3-star life cycle” or three red stars and you are fired. If a truck driver is involved in an
avoidable accident, a red star is placed next to his name on the employee bulletin
board for all employees to see. The accident report is placed in his employee file with a
warning to be more careful the next time. In the second avoidable truck accident,
another red star is placed on the bulletin board with a warning that the next accident
will be the last straw —he will lose his job. When the third accident takes place, he is
asked to leave, with the final paycheck to follow in the mail.
Tolerance for mistakes allows room for learning to take place. The truck driver
example shows two chances for the driver to learn from the previous mistake.
However, there are risks involved in empowering employees to be thinking drivers to
stay alert under stressful conditions. The upshot is that companies that succeed in turn-
ing knowledge into action usually inspire respect and admiration, rather than fear or
intimidation. Managers and executives who bridge authority and power differences
induce less fear. General Electric authorized its executives and employees in the mid-
1990s to come to work in casual attire in order to reduce the visibility of hierarchical
signs. Even the special executive dining room got merged with the cafeteria to elimi-
nate the status gap between managers and employees.

Ensure That Cooperation and Collaboration Are Not Competition

or Internal Rivalry
For years, we have known competition to be the great motivator of our economic
system. We have even incorporated this theme into the way we manage employees and
reward the few based on the most efficient, the most productive, the best contributor,
or the most loyal. As a result of the internal rivalry, to compete within the firm, employ-
ees became knowledge hoarders rather than knowledge sharers.
Today’s growing trend is that the best projects are accomplished via teamwork,
cooperation, and collaboration of company specialists committed to produce the best
product or service against outside competition. The product is based on common effort
and shared goals; the success of each employee depends on joint cooperation and
282 7777 PART III Knowledge Codification and System Implementation

knowledge sharing by everyone in the group. Em ployees begin to learn from one
another and build an attitude or a belief that the best way for them to succeed is if their
peers succeed. They also begin to build a culture of cooperation that makes the final
product competitive in the marketplace.
Turning away from internal rivalry to cooperation and collaboration is good for
any business that can see beyond the horizon. This is especially true when teamwork is
based on commonality of goals and the aspiration of its members. When teams are
properly formed, team members freely exchange ideas, knowledge, and resources and
manage any conflict that stands in the way of doing an excellent job as a team. With the
increasingly complex nature of global business in a business organization, it is team -
work and groupware that count. It is fine to hire and retain human talent, but talented
people cannot do multifaceted projects alone. They need talented help.
In summary, one can easily foresee the benefits and potential of turning knowl-
edge into action in businesses that have replaced fear and internal rivalry with trust, a
sharing attitude, cooperation, and m utual respect for peers and the organization.
M arket competition continues to be a factor in organizational performance, but with
knowledge sharing, the focus is now on fighting the com petition rather than one
another inside the organization.

Identify What Counts and What Makes Sense

An organization relies on reports and computer-based information to assess what
happens over time. All reports represent historical events, transactions, or develop-
ments. Unfortunately, nothing is presented that analyzes the reasoning behind the
information or why the resulting business turned out the way it did. In addition, there
seems to be no prioritization of what should be acted on or what changes should be
made that make sense to create a better future for the firm.
In many studies conducted in this area, it was found that organizations spend more
time on outcome than on process. Managers look at outcomes, but they do not take
time to discuss why it is so good or so bad. For knowledge management to succeed,
focus should be on the stock of knowledge—processes and process improvement that
might lead to a better outcome.
Take the example of a bank president who asked the board of directors for
approval to replace the drive-in machines because tellers had been reporting increas-
ing complaints from customers that the lines were long and the service was slow. The
board was about to approve the $160,000 replacement cost when a board member who
was a specialist in the field asked if she could take a look at the facility before approval
was made. After a 20-minute on-site observation and a brief discussion with the tellers,
it turned out that the problem was not in the machines but with the landscaping that
narrowed the six lanes to two lanes upon entering the drive-in lot. This forced a lot of
the drive-in traffic to wait on the road rather than in the bank’s drive-in area.
Removing the bushes, widening the entrance to four lanes, and giving out dog biscuits
to customer’s dogs in the cars brought customer complaints to a halt.

Take a Close Look at the Managers and How They View

Knowledge Transfer
Managers and leaders set the tone for how a department, a division, or the entire
firm should perform. Within the framework of knowledge management practices, man-
agers are expected to create a culture or an organizational climate that sets work
norms and values the transfer of knowledge to produce value-added products and ser-
vices. It is an effective way to bridge the knowledge-action gap.
 CH A PTER 9 Knowledge Transfer and Knowledge Sharing a i a * 283

In order to make the change possible, managers are put through specialized train-
ing to understand the transformation process. Knowing about the knowing-doing gap
is superficial without the action that must follow. The mind-set has to change before
change can be made in the business. Much depends on m anagem ent practices that
come from the company’s philosophy that values innovation, creativity, and an attitude
of action. Those who have been with the organization a long time may find it difficult
to adopt the change. As a result, they have to be replaced, transferred to stand-alone
jobs, or offered an early retirem ent package.

Assess Employee Job Satisfaction and the Stability

of the Workplace
The success of knowledge transfer and knowledge sharing depends to a great extent
on how satisfied employees are on the job and the turnover rate in the workplace. For
knowledge transfer to be effective, employees have to work in a culture that accommo-
dates transfer, and they must apply and share transferred knowledge in an environment
of trust and trustworthiness. Regardless of these factors, employees must be satisfied
with the status quo if they are to cooperate and collaborate with others as a team.
Job satisfaction is derived from the degree of match between an employee’s voca-
tional needs and the requirem ents of the job. The m ajor kpew n vocational needs
include the following:
• Ability utilization • Recognition
• Achievement • Responsibility
• Activity • Security
• Advancement • Status
• Authority • Supervision—human relations
• Compensation • Supervision—technical
• Creativity • Variety
• Independence • Working conditions
• Moral values
As proposed in Figure 9.3, job satisfaction is a function of the degree to which
employee vocational needs are met by what the job offers to satisfy the needs. For
example, if an employee values recognition highly and the job does not recognize
284 7777 PART III Knowledge Codification and System Implementation

employee performance, the mismatch makes an employee wonder if it is worth contin-

uing on the job. The same view applies to all other needs listed earlier. The sum of the
squared difference of these needs becomes an index of satisfaction. The lower the sat-
isfaction index, the higher the employee’s job satisfaction.

■■■■ Transfer Methods

Once knowledge is captured and codified or otherwise becomes available, the final
step is to transfer it for use by others in the organization. Recipients could be a team, a
group, or individuals. Knowledge transfer makes it possible to convert experience into
knowledge. As shown in Figure 9.4, a team sets out to perform a specific task. Team
action results in an outcome (a finished product, a completed repair, and so forth).This
experience is captured and fed back to the team in a form and a language the members
understand. The new knowledge serves to reinforce or improve the quality and perfor-
mance of the team the next time it faces a similar task. It is also transferred to a reposi-
tory for new recipients whenever it is needed.
It should be obvious that there is no single best transfer method for all tasks. Much
depends on factors such as the nature of the problem, the type of knowledge to trans-
fer, and the barriers to transfer. Each factor is briefly explained.

NATURE OF THE PROBLEM

One factor to consider in transfer method selection is the nature of the problem. Some
problems are routine problems; others are nonroutine. There are problems that recur
once a day; others occur once a month or rarely. There are problems that are complex
and critical; others are basic and can wait. Finally, there are problems that have a com-
bination of constraints. A typical example is a customer who brings her year-old auto-
mobile to the dealer because of squeaky brakes. On the surface, the problem seems
nonroutine, especially when the brakes never squeaked during the first year of opera-
tion. However, this kind of problem could occur in cars that are driven in a hot climate

I fl fl 1 I 9 FIGURE 9.4: Converting Experience into Knowledge

 C H A PTER 9 Knowledge Transfer and Knowledge Sharing a i a ■ 285

or could occur when braking down a long steep hill in heavy rain. Sand, dust, grease
spots, or other m atter accumulating on the drums could be routine factors to consider
in solving the problem.
The procedure for repairs is considered routine, because all the mechanic has to do is
to replace the brake pads, the drums, or both. To do so, he pulls the wheels and the brake
assembly; he looks for foreign deposits, inspects the hubs, and checks the brake lines for
leaks. However, this often turns out to be the wrong solution. An experienced mechanic
may quickly find nothing wrong with the brakes or the deposits anywhere. He knows from
similar cases with the same model, though, that after 10,000 to 15,000 miles of driving in
the local climate, the hubs get hot and have a tendency to warp or get out of round. Once
that happens, it cause unevenness in the braking process; therefore, the squeak. The best
action is to replace the front hubs, hopefully with a better quality product.
The nonroutine, infrequent nature of the problem found a solution with the expe-
rienced mechanic. If he is working with or collaborating with other less experienced
mechanics, the tacit knowledge of the experienced m em ber can be verbally and
directly transferred to the rest of the team for future use. If the squeak problem hap-
pens to be a chronic complaint and more customers bring in their cars for repairs, the
diagnosis and procedure can be autom ated or stored in a knowledge base to be
accessed when needed. Even when the mechanics take notes about the new procedure,
they are transferring new knowledge on paper for later reference or use. In any case,
we experience knowledge transfer as a step for knowledge sharing.

TRANSFER STRATEGIES
Knowledge management implies formalized knowledge transfer and requires specific
strategies to ensure successful transfer. Knowledge is transferred via documents, an
intranet, groupware, databases, knowledge bases, and a lot more. By far, the most
effective channel for knowledge transfer is face-to-face meetings in the business place
(see Box 9.2).
One formalized strategy is codifying and storing tacit knowledge that is neither
easy to transfer nor guaranteed to be up to date, compared to the human expert from
whom such knowledge is captured. Perhaps the best way to absorb tacit knowledge is
to be in the domain where tacit knowledge is practiced. This is done through job rota-
tions, job training and retraining, and devoting specialized focus on on-site learning. It
involves on-site decision making, absorbing the mechanics and the heuristics as they
occur, and coming up with a new knowledge base that emulates the domain in a unique
way. The main limitation of such strategy is time, especially when certain problems can-
not wait a year or two before knowledge becomes available for a solution.
The traditional management attitude in most of today’s organizations is “stick your
nose in the job, minimize the chatter, and get to work.” For example, one multinational
firm placed a sign by each watercooler that read, “For safety reasons, the company dis-
courages any prolonged meetings or discussions around the watercoolers or anywhere in
the hallways, cafeteria, or around the elevators.” When asked by one manager the reason
for the notice, a senior executive replied, “It means don’t kill time when you should be in
your office at work.” For years, employees have used the watercooler or cafeteria (and
rest rooms) as a place to exchange views and resolve problems. The company in question
had no conference rooms or a place for group discussions. Management had the attitude
that such meetings promote gossip and feed the rumor mill.
In contrast to the traditional view, the knowledge management view is not “stop
talking and get to work,” but “start talking and get to work.” The watercooler get-
together could be the enabler of serendipity. W atercooler-type chats are a hit with
286 7T 77 PART III Knowledge Codification and System Implementation

i a fl i a a bo x 9.2 a a a a a a

W IRING A CORPORATION’S KNOWLEDGE INFRASTRUCTURE

W IRING THE CORPORATE BRAIN (think of in-house Fulbrights) for projects that will
make different business units work together; and
Born of the merger in 1996 of two venerable Basel-
hosting periodic knowledge fairs to spark ideas.
based companies, Sandoz and Ciba-Geigy, Novartis
Jack Staeheli, who heads this effort, prefers to call
is the newest and richest beneficiary of B asel’s
it knowledge networking, because that sounds less
confluence of scientific learning, industry, and cul-
hierarchical than knowledge management (the
ture. From two tiny m akers of specialized dye-
usual term for a strategy to make the whole of a
stuffs 150 years ago, Novartis has become a $24 bil-
workforce’s contribution greater than the sum of
lion corporation that calls itself a “life sciences
its individuals’ knowledge). Likewise, Staeheli’s
company,” a conceptual catchall for an entity whose
title is “officer of the science boards.” Senior offi-
products include such diverse items as pharm a-
cer of corporate knowledge management, the role
ceuticals, genetically engineered seeds, baby food,
he acknowledges to outsiders and which he in fact
contact lenses, and animal health products.
performs, would sound too much like he con-
However, Novartis’s breadth of products and
trolled the process, he says. He prefers to see him-
its success in the global economy make for its
self as a shepherd rather than a director. The
greatest challenge: How can a company that
wording is more than symbolic. Novartis’s presi-
depends on innovation to survive find ways not
dent, Dan Vasella, asserts that knowledge transfer
just to share but to increase its knowledge? This
and creation will determine the company’s future
problem faces any knowledge-intensive, global
competitiveness. “Knowledge networking” shows
industry. Novartis addresses it with a three-
that Novartis understands that making the most of
pronged strategy: using Web-based technology to
an organization’s intellectual assets is more of a
foster collaboration within its worldwide work-
cultural than a technological challenge.
force; creating a system of developm ent grants

SOURCE: Excerpted from Abramson, Gary. “Wiring the Corporate Brain,” Enterprise Magazine, March 15,1999, pp. 2-8.

a a a i a a

teams transferring or sharing focused knowledge during a break away from the office.
Spontaneous, brief sessions are door openers for generating ideas and exchanging
views; they have the potential for generating new ways of problem-solving. This is one
approach to creating exchanges that could produce unique results.
Knowledge transfer should adopt or adjust the organization’s culture to facilitate
knowledge transfer and knowledge sharing. Management in the Japanese culture sets
up talk or conference rooms to encourage dynamic knowledge exchange. Japanese
m anagers spend time together after work. From group dinners to social clubs and
sports, they practice knowledge transfer. They function as a mechanism for establishing
trust and take time for constructive criticism. They prefer meeting face to face over
e-mail and on-the-job experience over learning via a computer-based knowledge base.
The whole process, which includes company standards of perform ance and highly
respected protocols, is followed on a regular basis with value-added results.
With this in mind, it would be difficult to fit the Japanese culture with American
business. In so many ways, e-mail is an American invention for efficiency and produc-
tivity. The notion of “Why call for something you can attach to an e-mail message?” is
 CH A PTER 9 Knowledge Transfer and Knowledge Sharing a i « a 287

the American way of promoting performance at the cost of informal or personal knowl-
edge transfer. An employee who spends hours leafing through and responding to e-mail
is viewed as being hard at work. An employee found reading a professional magazine as
a way of capturing knowledge is frowned upon as being “waste at work.” A company
that supports knowledge exchange, on the one hand, and discourages reading or talking
on company time confuses many employees. Irrespective of culture, “slack time” should
be used for knowledge sharing to the advantage of employees and company alike.

INHIBITORS OF KNOWLEDGE TRANSFER

There are several organizational and cultural factors that inhibit reliable knowledge
transfer. Some inhibitors are easier to correct than others. The following are the key
areas where friction occurs and ways to correct it.
• L a c k o f trust. A major element in the success of any knowledge transfer project is
the commonality of language people have with one another. A shared language
can only strengthen trust among the participants, whether they are professors,
surgeons, attorneys, or night watchmen of a high-rise building. Trust develops
after people share proprietary knowledge and feel the reciprocal to be at least
intrinsically rewarding. The more people assess one another and have a feel for
where they stand, the easier it is to decide how much they can trust one another.
In the final analysis, we shall find that cultural, language, status (who gives the
knowledge), and location mismatches could be “make or break” of trust for suc-
cessful knowledge transfer.
• L a c k o f tim e a n d co n feren ce p laces. People tend to find knowledge nearby or get
the closest source, not the best one. To expedite knowledge transfer, people must
show interest, take time, and find meeting places where they can exchange ideas.
The “I don’t have time because I am snowed under” excuse is rarely a valid -
excuse. A volunteering spirit of cooperation, collaboration, and coordination is a
powerful enabler of knowledge transfer.
• S tatu s o f the kn ow er. People decide to act on knowledge or information based on
who gives it to them. A senior vice president sent by the president to attend a
union meeting gives an impression that the company cares. Status refers to the
individual’s position within a group or an organization, compared with those of
fellow employees. One attribute that tends to give a person status, for example, is
age. If a company does not want age to be given status, it will have promotions
and reward policies that favor youth and work against age.
Status and status symbols (title, size of office, office views, type of furniture in
the office, the floor the office is located on, the kind of company car assigned, and
so forth) are important management factors that affect the behavior of individuals
and groups. We should look at what status can and cannot do for an organization.
People who consider status and still are reluctant to put transferred knowl-
edge to good use could have several reasons for feeling that way: not respecting
or trusting the knower, not considering the knowledge to be important, pride, lack
of interest, lack of time, sheer stubbornness, and reluctance to take unnecessary
risks. When this happens, it does little good to push knowledge transfer for the
mere show that knowledge is freely shared.
• Q u a lity a n d s p e e d o f transfer. When knowledge is transferred from a reliable,
high-profile source, it is expected to be of the quality that promotes its use.
Unfortunately, that is not always the case. The quality of transfer can be influ-
enced by factors such as mentorship, on-the-job training, job rotation, and career
Ill Knowledge Codification and System Implementation

paths that define the way, timing, and duration of job experiences. In contrast,
knowledge downloaded from a database or a knowledge base is not a match for
the on-site, person-to-person knowledge transfer.
In addition to quality of transfer is the speed of acquiring and making use of
knowledge. This is where technology can be a tremendous help, especially if the
knowledge transferred affects employees and procedures company-wide. Both
quality and the efficiency with which it is transmitted and received would be of
concern to management in deciding how well the firm uses knowledge capital.

HOW KNOWLEDGE IS TRANSFERRED

Once the foregoing elements are put into place, knowledge transfer should turn into
new action. The question is what types of knowledge can be transferred. The three
main types are as follows:
• Collective sequential transfer
• Explicit interteam knowledge transfer
• Tacit knowledge transfer
Each category is explained next. When it comes to design considerations, each type
requires a unique design, depending on the recipient, the type of knowledge trans-
ferred, and the complexity of the project. The goal is to ensure that the delivery mech-
anism will meet the requirements of the users of the knowledge.

Collective Sequential Transfer

This is one of the easiest and most straightforward types of knowledge transfer. An
ongoing team specialized in one specific task moves to other locations and performs the
same tasks. For example, in building new homes in the suburbs, septic systems are installed
by a team that does nothing else. Based on the local code, a team of three to five members
digs up a designated area with a backhoe; installs pipes, feeders, a trap, and so forth; buries
them; and reseeds the area to blend with the rest of the landscape. This know-how
requires decisions on the size of the pipes, the direction and slope of the flow, and provi-
sions for venting and future repairs. Once the system is completed, the same team reuses
the same knowledge; goes to another new home site; and builds the same septic infra-
structure, with variations to meet the local codes. The procedure remains the same.
Change occurs only if the county or city alters the regulations or the septic codes.
What is unique about collective sequential transfer is that knowledge is trans-
ferred from one site to another by the same team. There is no knowledge transfer from
one team to another team, but there is transfer of knowledge from a team member to
another member within the team. This is where collaboration is carried out. The focus
is on collective knowledge rather than individualized knowledge. With the job carried
out in sequence, each m em ber’s knowledge and performance affects the next member
who has to do the work based on the work completed by the previous member. This
means quicker transfer of knowledge with closer adherence to quality of the resulting
installation. As shown in Figure 9.5, after each job, a team talks about any new knowl-
edge gained, discusses what should go into the next installation, and adjusts the com-
mitment of each member accordingly.
Each time a team performs a job, each member transfers his or her unique experi-
ence into the team so that the resulting transformation can be incorporated into the
entire process. It means evaluating and redesigning collectively the new team knowl-
edge every time a job is completed. Once completed, the team is expected to function
more effectively on the next job.
 C H A PTER 9 Knowledge Transfer and Knowledge Sharing aaaa 289

FIGURE 9.5: Collective Sequential Transfer—Transferring Experience to Knowledge

When performing a job, team members are on constant alert regarding the following:
• How did a m em ber’s task affect others in the team?
• How did a member’s way of performing a task contribute to the performance of others?
• What factors affect a m em ber’s performance on the job?
• How does a m em ber’s job impact the team ’s overall performance?
These observations are usually stored in each m em ber’s mind until the next time
the team meets and exchanges experiences. During the meeting, the sharing of experi-
ence and the transformation of experience into new or updated knowledge is what a
team should take with it onto the next job. The knowledge transfer process also facili-
tates the conversion from personal knowing to group knowing. When that occurs, it
contributes to the achievement of team goals.
Transfer Protocol. In every team setting, there is a protocol or a way to specify how a
team should function. With a focus on collaborative sequential knowledge transfer,
where jobs are interdependent and performed in some sequence, each team member is
expected to carry a “suitcase” of knowledge. There is also an assumption that each
member has vested interest in sharing rather than hogging knowledge. Any activity
carried out by members is done in fulfilling team goals—completing a project collec-
tively, efficiently, and effectively.
The unique features of collaborative sequential knowledge transfer include the
following:
• Team meetings are usually brief, but are held regularly as time and the occasion
permit. Meeting formats are understood, based on the way the team has con-
ducted previous meetings. There is usually a facilitator or an informal leader who
initiates the meeting, monitors the discussions or exchanges, and makes sure
things do not drag unnecessarily. Facilitators are routinely rotated among the
members. Some of the questions raised center around what happened during the
job just completed, whether there were any differences that should be discussed,
and why things happened the way they did.
• Meetings are held with all participants being equal. There are no discrimination
or hierarchical priorities set as to who speaks first, how influential one’s comments
290 7777 PART III Knowledge Codification and System Implementation

are compared to the rest, and the like. With this in mind, team members are
expected to be proactive rather than reactive and assertive rather than laid back,
with no attitude such as “Look, this was not my job, so I can’t comment” or “My
job is to lay pipes, no more no less.”
• Whatever takes place in meetings is kept within the team. There is no reporting
relationship to the organization’s higher-ups. This eliminates the fear that a nega-
tive report might prompt layoffs or dismissals. It means that team members
begin to understand and trust the circumstances under which they can or cannot
share knowledge.
• A unique feature of collaborative sequential knowledge transfer is focus on the
project, not the person or the personality. The approach, the setting, and the seat-
ing arrangements are not army style (one row behind the other) or classroom
style. Team members face one another, and anyone can make a comment anytime
within reason. The goal is to fine-tune performance, improve quality, and assure
integrity of the work performed by the team (see Box 9.3).
Even with all these advantages, this type of transfer also comes with its own set of
problems. A team could be overcommitted or overworked; fatigue or boredom could
set in; or the team might be operating under a tight budget. All of these circumstances
will make it difficult for the team to meet. A nother problem is the level of fluency of

-----------------------------------a a a a a a b o x 9►.3

GUIDELINES FOR FACE-TO-FACE MEETINGSi IN COLLECTIVE SEQUENTIAL TRANSFER

M A K IN G IT WORK out the meeting. Ideally, the facilitator should
have no personal stake in the meeting apart from
Setting up effective face-to-face meetings requires a helping the process to flow.
clear game plan. Here are some pointers to consider:
• Take breaks. Because attention can easily wan-
• Set the agenda. Clearly communicate the purpose der, do not force participants to go much past
and intention of your meeting. Before the m eet- an hour without time to stretch and relax. Focus
ing, tell the participants about the meeting time on discussing specific, manageable tasks in each
and desired goals. session.
• Keep it small. Sometimes a large meeting is neces- • Socialize. To help in building trust, try to sched-
sary, especially if you need to involve participants ule at least one nonbusiness event. When part-
from many disparate locations. In most cases, icipants become familiar with each other in
though, smaller is better. KM experts agree that an informal setting, they are more likely to
six to eight participants are ideal. share their knowledge productively during
work time.
• Invite the right people. Filter your attendee list to
include only participants whose thinking • Show accomplishments. Arrange for participants
responses and direct participation are key to your to leave the session with something tangible in
success. Consider having less crucial personnel hand, even if it is something as simple as a
join remotely in “listening” mode. checked-off copy of the completed agenda. It
serves to mark progress and emphasize the worth
• Facilitate the process. Find a neutral facilitator
of what they have done.
who can keep the participants on track through-

SOURCE: Excerpted from Dell, Pamela. “Getting the Most out of Getting Together,” Knowledge Management, August 2001, p. 1.
 CH A PTER 9 Knowledge Transfer and Knowledge Sharing ■ ■ ■ ■ 2 9 1

the members. It takes a certain amount of communication skills to exchange experi-

ences and to express opinions about new ones.
Finally, it is not so easy to hold a meeting after a full day’s work. Most members are
eager to head home, especially when meeting times are not compensatory. The so-
called “suitcase” of knowledge carried by each team member will have to wait another
day or longer, which tends to age the experience.

Explicit Interteam Transfer

This type of knowledge transfer simply allows a team that has done a job on one site
to share the experience with another team working on a similar job on another site. For
example, take the case of the septic system. A contractor might have two teams of work-
ers working on a septic system for two houses in two separate subdivisions. As shown in
Figure 9.6, team A in subdivision has installed the system more quickly than before, even
with a new twist to the installation. While digging a trench for the major pipe that goes to
the trap, they had to reroute the pipe around the roots of a 120-year-old maple tree.

FIGURE 9.6: Explicit Interteam Knowledge Transfer

House A

o
Team B

Trap

Ro<
Prob

Subdivision Y

Team Members in Each Subdivision

• Pipe Fitter
House B • Backhoe Operator
• Laser Engineer
• Contractor Representative
Team A

Subdivision X
292 7777 PART III Knowledge Codification and System Implementation

They designed a special apron to protect the tree from any damage should the pipe burst
or corrode. This new discovery was shared with another team in subdivision Y, which has
a similar problem but was about to cut the tree and dig up the roots to allow the inflexi-
ble pipes to lay straight to the trap. The new experience from team A was shared in time
to save the tree.
Interteam knowledge transfer is explicit knowledge, because most of the knowl-
edge transferred is routine work and the procedure is precise. Tacit knowledge is trans-
ferred in situations where the team might have found a new or a creative way to get
around problems. In our example, in Figure 9.6, the tree got in the way in both subdivi-
sions. Team A designed an apron to save the tree without compromising the quality of
the work. The apron is something team B never thought of; team B was about to cut the
tree. When the two teams met or a representative of team A met with team B, the
apron design was shared in time to be of use.
Personal interaction is ideal in projects that are timely and can benefit from quick
feedback. It is also useful in situations that reinforce relationships and solidify cooper-
ation and coordination between teams. E-mail or other electronic transmissions of
“news and views” cannot be a satisfactory substitute. It can be useful to supplement
personal interaction in interteam knowledge transfer.
Despite the benefits, there are human relations and other factors that make it dif-
ficult to implement explicit interteam knowledge transfer. One factor is the organiza-
tional subculture of the receiving team. In the case of the apron, team B might feel too
proud to accept any suggestions or ideas from an outside team, especially if the inno-
vation is simple or if it is something that team members should have thought of on
their own. A team may simply be reluctant to share explicit or tacit knowledge, espe-
cially in situations where their innovation and productivity will bring bonuses or spe-
cial awards. Finally, lack of time can be the worst enemy of knowledge transfer or
knowledge sharing. For reasons of their own, the team with the new experience might
not have the motivation to take extra time after work to spoon-feed another team with
the change that it took time and collaboration to invent.

Tacit Knowledge Transfer

This type of knowledge transfer is unique in complex, nonalgorithmic projects,
where knowledge is mentally stored. The team receiving tacit knowledge is different by
location, by experience, by technology, and by cultural norms. This means that what-
ever knowledge is being transferred has to be modified in language, tone, and content
to be usable by the receiving team.
Take the case of Kuwait’s oil wells, which were set on fire by Saddam Hussein’s
Iraqi troops retreating from Kuwait in 1991. To douse the oil fire, specialized teams
from the United States, Russia, the United Kingdom, and Holland attem pted to find a
solution. Fire control depends on the pressure of the oil, the depth of the well, wind
conditions, oil tem perature, and the like. The most successful team was that of Red
Adair of Texas. Red and his team extinguished 117 of the burning oil well fires in the
Ahmadi, Magwa, and Burgan fields—the highest producing fields in Kuwait.
There were logistical problems in the entire fire-fighting operation, which was
unlike the hundreds of other oil fires in the United States, in underwater wild wells, or
in the 1977 Bravo offshore blowout in the North Sea. The years of experience and the
stored tacit knowledge of Adair and his seasoned team allowed them to come up with
the combination that was most successful in capping the oil wells; every other foreign
team failed to do so. After the work was done, the foreign teams were astounded at
A dair’s approach and the combination of standard and creative methods he used to
 CH A PTER 9 Knowledge Transfer and Knowledge Sharing i i ■i 293

Location: United States Location: Kuwait

FIGURE 9.7: Tacit Knowledge Transfer

control the fire and the smoke that had virtually covered the skies and sand dunes of
Kuwait. American ingenuity, skill, and self-discipline combined to master the seem-
ingly impossible. A dair’s tacit knowledge was translated to the Kuwaiti fire-fighting
team in a form different from what they had used before the war. Adair formed a new
company in 1993 to provide consulting services to fire-fighting teams on a global basis.
In every consulting job, he transfers his tacit knowledge to similar teams unrelated to
his and in areas and locations foreign to Texas.
A dair’s team demonstrated the ability to understand a specific situation and then
customize their response to that situation. His team members evaluated the new situa-
tion, which triggered memories of past episodes and their respective solutions for the
final decision on the approach used to control the unique Kuwaiti fires. What distin-
guishes this approach from explicit knowledge transfer is the fact that transfer of tacit
knowledge goes to another team doing a similar kind of work (extinguishing oil fire),
but the knowledge is customized to the location, culture, ability, and constraints of the
receiving team (see Figure 9.7).
One of the problems in tacit knowledge transfer is the understandable difficulty in
tapping tacit knowledge. A dair’s case was no exception. His tem peram ent, tolerance
for ambiguity, and patience for explaining details were barriers to'his knowledge trans-
fer. The process unique to the Kuwaiti fires was very difficult to capture on paper or to
train overnight. Unlike explicit knowledge transfer, tacit knowledge transfer is not
practically transferred electronically.

■■■■ Role of the Internet in Knowledge Transfer

Since 1960, when the first business com puter appeared, inform ation technology has
changed the way commerce is conducted around the globe. From mainfram es to
accounting systems, the personal computer revolution, local area networks, electronic
data interchange, client/server design, and enterprise resource planning have all had a
hand in shaping today’s business organization. The past decade has been the Internet
decade, however, when companies worldwide have embraced a change without equal.
It is a change that promises to have more impact and be more lasting than anything
else we have seen to date. Technology is setting the pace for how a company does busi-
ness, how it exchanges knowledge with suppliers, and how it communicates with cus-
tomers and others in the new marketplace. Any way you look at it, business will never
be the same again.
The prim ary technology for this transform ation is the In te rn e t—a global data,
information, and knowledge network that moves closer and closer to the ubiquity of
the telephone with each passing day. What makes the Internet so powerful is that it is
more about knowledge exchange and communication than it is about technology. It is
294 7777 PART III Knowledge Codification and System Implementation

a medium and a market. It is virtually unstoppable, forcing all kinds of businesses to

reexamine their practices and their futures.
There is one feature of the Internet that you will not find on the telephone. The
Net allows you to send messages to multiple persons at the same time, much like tele-
vision, or radio broadcasting. The Internet began with message communication, but
now it is possible to transmit and receive information containing graphics, voice, pho-
tos, and even full-motion videos. The part of the Internet that can accomplish these
tasks is called the World Wide Web, also known as WWW or the Web.
The Internet today offers a variety of services including e-mail, access to remote
computers, news, and the quick and easy transmission of information between comput-
ers worldwide. Many/federal agencies now allow anyone to access timely information.
These agencies include the Social Security Administration, Veterans Administration,
and the U.S. Postal Service.

INTERNET SERVICE PROVIDERS

To link the rapidly growing commercial Internet landscape, the Internet service
provider (ISP) industry was born in the mid-1990s. It offers a variety of services, includ-
ing the following:
1. Linking consumers and businesses to the Internet (for example, America Online,
Microsoft Network, and CompuServe)
2. Monitoring and maintaining customers’ Web sites
3. Network management and system integration
4. Backbone access services for other ISPs, such as PSI and UUNET
5. Payment systems for online purchases
As public demand for access to the Internet surged, ISPs began to add more lines
and better access to accommodate the traffic. Initially, the cost for Internet access often
exceeded $1,000 per month, but with new ISP arrivals and competition, prices plum-
meted. Many of today’s ISPs offer unlimited access for as low as $5 per month. Many
local governments are funding the use of the Internet because of its political, educa-
tional, and commercial benefits. Once on the Internet, there are no additional charges.
The place of the ISP in the Internet architecture is shown in Figure 9.8.
You can contact anyone, anywhere, anytime for that monthly fee. The exceptions
are Web sites that charge a membership fee or a fee for access to privileged information.
Almost everything one needs on the Internet is free. Among the free services are
the following:
• Hotlists, which tell the user what is popular and what is not
• Comics that focus on entertainment events
• Software archives that list the latest free software available
• Weather services that provide free weather forecasts anywhere in the world
• Magazines and broadcasting stations that constantly update the news
• Searchers that help locate items or subjects on the Internet
• Dictionaries that include thesaurus and “fact” books on almost all subjects
• Government services that publicize what is available from them
The problem for some ISPs is sudden growth without advance planning to accom-
m odate the growth. As a result, response time slows down, which triggers customer
complaints. The challenge is to maintain profitability and meet or beat the competition
while maintaining customer satisfaction. To do all this well requires professional man-
 C H A PTER 9 Knowledge Transfer and Knowledge Sharing 295

Backbone

Network
Access
Points (NAPs)

Regional
Networks

Regional
ISPs

User
Level

FIGURE 9.8: General Internet Network Architecture

agement, a highly skilled technical staff, and a healthy budget to bring the technology
in line with the voracious appetite of today’s consumer. The trick is to ensure a balance
between creativity and control and between managing growth and a stable technical
infrastructure.

STABILITY AN D RELIABILITY OF THE WEB

No single agency or company owns the Internet. Each company on the Internet owns
its own network. Telephone companies and ISPs own the links between these compa-
nies and the Internet. The organization that coordinates Internet functions is the
Internet Society. It does not operate any of the thousands of networks that make up
the Internet, but works with ISPs by providing information to prospective users. This
association’s Internet A rchitecture B oard consists of work groups that focus on
TCP/IP and other protocols. Various committees also handle technical issues and day-
to-day operational aspects of the Internet.
The Web itself, because it resides everywhere and nowhere at the same time, sim-
ply cannot cease functioning by itself. Also, because it is based on the Internet, its sta-
bility is as good as that of the Internet, which is fairly good so far. The Internet is
designed to be indefinitely extendable. Reliability depends primarily on the quality of
the service providers’ equipm ent. Inadequate phone lines, bandwidth, or mediocre
computers can affect the reliability of the overall service.
296 7777 PART III Knowledge Codification and System Implementation

UNIQUE BENEFITS OF THE INTERNET

The Internet is the enabler of knowledge transfer and knowledge sharing worldwide.
Managers use it to glean intelligence about rivals, monitor sales, and so on. At Cisco
Systems, for example, the company’s chief executive required executives from various
departments to demonstrate how they would use the Web. Cisco and other companies like
Dell Computers and Microsoft’s Expedia Travel Service use Internet technology in their
businesses. Dell, for example, sells more than $6 million worth of computers a day from its
Web site, and Expedia’s service generates more than $16 million a month from its Web site.
For all the directions the Internet has taken and all the companies it has spawned,
this technology was first developed as a tool for scientists to exchange knowledge and
keep in touch with one another. This is the way many of us use the Internet today. On
the business side, the rise of the Internet as the enabler of e-commerce is changing how
companies manage their knowledge and business. Closed enterprise systems are giving
way to open system environments, where customers connect to the company’s Web site
and trading partners connect by the extranet and the Internet.
Among the uses and advantages of the Internet as they relate to knowledge man-
agement are the following:

• Doing business fast. Knowledge transfer is conducted in minutes rather than

hours or days, compared to waiting on the phone or meeting as a team. This speed
compresses transfer time and promotes the growth of knowledge sharing.
• Gathering opinions and trying out new ideas. The Internet is an ideal place for try-
ing out new ideas at low cost. Interactive surveys provide quick feedback.
Opinions can be gathered from just about anywhere. Many online opinion polls
provide real-time statistics to the user after a computer package analyzes the
user’s response in real time.
• Leveling the playing field. When a company advertises its products or services on
the Internet, it is on equal footing with larger companies. The nice feature about
the Internet is that it allows your business to appear alongside big-name compa-
nies. Here is a case in point: Caribbean Tour wanted to put its business on the
Internet, but it did not have much money to spend. It designed a basic Web site,
just to have a presence, and spent only $800 per month to gain experience with
online commerce. The Web site broke even after 2 years.
Even mere presence is a benign way of creating new selling opportunities to a
unique type of customer who otherwise would have gone to the competition. The
Internet is a great equalizer. A business might be small, but it encourages that
business to be very good at what it does.
• Providing a superior customer service and support resource. Most Web sites generate
customer feedback in the way of comments, suggestions, and complaints. The chal-
lenge for the online merchant is to have adequate staff to address this feedback in a
timely fashion. A common support resource is an FAQ (frequently asked questions)
list. A FAQ eliminates having staff answer the same questions over and over again. If
a new question comes up, the answer is added to the list. Using e-mail to handle cus-
tomer support also frees company personnel from being tied to a telephone.
• Supporting managerial functions; spreading ideas. The traditional managerial func-
tions of planning, organizing, directing, and controlling require managers to collect,
evaluate, and distribute management information, especially in organizations with
branches worldwide. The Internet sends business information through a company’s
networks and across networks around the globe. E-mail is a convenient tool for man-
agers to reach employees, bosses, customers, and suppliers quickly and at no charge.
 CH A PTER 9 Knowledge Transfer and Knowledge Sharing m■ ■ a 2 9 7

The Internet has spawned discussion groups, chat rooms, and online interactive
sessions in which technical and managerial staff evaluate products and processes and
arrive at value-added decisions that result in lower costs and increased performance.
B etter technical support is one of the key benefits of linking to the Internet.
IBM, for example, offers custom er and technical support and handles software
upgrades on the Internet. There are thousands of free software programs available
for anyone to download.
Market research firms are a natural for the Web. Credit bureaus, lawyers, private detec-
tives, accountants, baby-sitters, and teachers are all examples of people or agencies that use
the Internet for scheduling or advertising their services. The Internet contains thousands of
databases containing information about everything from medicine, vehicles, and food
preparation to hundreds of research and development (R&D) discussion groups. For thou-
sands of research journals, automated searches through current and back issues are avail-
able in minutes rather than the days and weeks it once took to find the information in a
brick -and-mortar library. White papers that research centers place on their Web sites pro-
vide current information about the latest developments in various fields.
Company research is no different. Companies use the Internet to seek information
about customer tastes and preferences, to profile a customer base for a new product, or
to test a new concept to see if it is worth developing. All of this can be done in a mat-
ter of days.

LIMITATIONS
Like any system with unique benefits, the Internet and the World Wide Web also have
unique limitations. The following paragraphs highlight the importance of continuing to
work on these limitations in the interest of advancing use of the Internet in general and
the Web in particular as a knowledge transfer and exchange medium.
The main limitations of Internet use for knowledge exchange are security and pri-
vacy. Key questions that are continually asked by online consumers are “How do I
know I am sharing knowledge on a secure line?” “How do I know the Web site assures
me privacy for the information I am seeking?” In terms of privacy, according to a study
of major Web exchanges by the Federal Trade Commission (FTC), only 20 percent
met FTC standards for protecting consumer privacy. However, the study also found a
90 percent compliance rate by Internet companies (for posting their privacy policies).
In addition to fakes and forgeries, there are other threats. Hackers, worms, Trojan
horses, viruses, and zombies threaten the security and integrity of any information or
knowledge exchange. Viruses are the best-known malicious software. These programs
secretly attach themselves to other programs and then infect and reproduce in a man-
ner similar to biological viruses.
Because hackers (those who access others’ computer systems illegally) continue to
threaten the integrity of files and transactions, hacking schools that teach students how
to hack or break into software to protect their own computer systems have begun to
appear in various cities. In one typical 4-day seminar, students (usually network admin-
istrators) pay $3,500 to learn the tricks of the trade. They are provided with hacking
tools that are available over the Internet for free (for example, an information gather-
ing tool, Sam Spade, and a port-scanning tool, SuperScan, that sends queries to
Internet servers to check their security status). At the completion of the course, stu-
dents sign an affidavit that they will use their experience appropriately.
Despite sophisticated FAQs, e-mail, and other technologies, companies still have a
problem with simple issues like securing information. In customer service, the heavy
demand also puts added pressure on customer service personnel.
298 7777 PART III Knowledge Codification and System Implementation

■■■■ Implications for Knowledge M anagem ent

Knowledge transfer, whether it is face to face, through an electronic knowledge base,
or is team-based, involves management support. Any transfer should be carried out
with the larger goals of the organization in mind. The link between knowledge and
goals should be clearly defined and effectively implemented within the framework of
the organization’s knowledge management strategic plans.
In addition, a knowledge m anagem ent system that accommodates knowledge
transfer on an ongoing basis should have the facility to monitor its progress, report and
promote its benefits to participants, and ensure continuity by the knowledge developer
or the chief knowledge officer (CKO). In the final analysis, the process and outcome
become a compliance based on choice.
If top management is serious about knowledge management and knowledge trans-
fer, several steps should be taken to ensure a reliable infrastructure that will serve the
organization’s long-term needs. First, a candidate departm ent or company division
with interest in knowledge sharing should be selected to set the tone for the enterprise.
As we do in major projects, a committee representing the entity should be formed to
begin strategic thinking about how to build a knowledge m anagement environment
and launch knowledge sharing. This might involve changes in policy, procedures, train-
ing and retraining, and addressing the cultural and work norms to effect a change from
knowledge hoarding to knowledge sharing.
Once a steering committee is in place, the next task is to undertake knowledge
assessment—evaluating existing knowledge, where it resides, how it is being used, and
what part of it needs to be updated, deleted, and so forth. Following this critical step, a
procedure should be designed for knowledge transfer, and the types of knowledge and
how they should be transferred should be determined. Doing that means looking more
closely at existing technology, the company intranet, e-mail technology, and security sta-
tus. To complete the transfer, receiving individuals or receiving teams must begin to
share their knowledge with the right people so that everyone in the division will enrich
their experience, improve their productivity, and add value to the strategic goals of the
organization. In doing so, the organization as an ongoing venture should begin to realize
benefits and grow exponentially in the competitive environment in which it operates.

Su m m a r y 1,11
• Knowledge transfer is part of everyday organizational life and a step in the KM
building life cycle. Knowledge can be transferred from repositories to people,
from teams to individuals, and between individuals.
• To transfer knowledge, we consider where knowledge is transferred from, the
media used in knowledge transfer, and where the knowledge is transferred. In any
case, this transfer facilitates knowledge sharing and enhances knowledge flows
through collaborative technologies.
• Knowledge transfer distributes to points of action. It supports education and
training, substitutes as customer service representatives, and so forth.
• The knowing-doing gap is a situation where an organization knows what to do,
but for various reasons ignores the information available and performs differ-
ently. Knowing about this problem should help firms make corrections and set up
a knowledge transfer environment for all employees to benefit.
• The terms transfer and share are interrelated. Knowing is considered deeply per-
sonal; it is asking someone to give something that is his or her own property. It is
 CH A PTER 9 Knowledge Transfer and Knowledge Sharing aaaa 299

a mechanistic term, which provides knowledge for someone else. The term share
is an exchange of knowledge between individuals, between or within teams, or
between individuals and knowledge bases, repositories, and the like.
Translating knowledge into action is not straightforward. To overcome the gap
between knowledge transfer and knowledge sharing, the following guidelines
are useful:
• Instill an atmosphere of trust in the organization.
• Fix the culture to accommodate the change.
• Push reasoning before process.
• Rem em ber that doing is far better than talking.
• Know how the firm handles mistakes.
• Ensure that cooperation and collaboration are not competition or inter-
nal rivalry.
• Identify what counts and what makes sense.
• Take a close look at the managers and how they view knowledge transfer.
• Assess employee job satisfaction and the stability of the workplace.
There is no single best transfer method for all tasks. Much depends on factors
such as the nature of the problem, the type of knowledge to transfer, and the
barriers to transfer. The upshot of knowledge, transfer is to adopt or adjust the
organization’s culture to facilitate knowledge transfer and knowledge sharing.
Irrespective of culture, “slack tim e” should be used for knowledge sharing to the
advantage of employees and company alike.
The key inhibitors of knowledge transfer are lack of trust, lack of time and con-
ference places, status of the knower, and quality and speed of transfer. Each
inhibitor should be seriously considered before knowledge transfer can be
expected to take hold reliably among participants.
The three main types of knowledge transfer are collective sequential transfer,
explicit interteam knowledge transfer, and tacit knowledge transfer. In collec-
tive sequential transfer, an ongoing team specialized in one specific task
moves to other locations and performs the same task. Once the task has
been completed, the same team reuses the same knowledge, goes to another
new site, builds the same system, and so forth. Team meetings are generally
brief; meetings are held with all participants being equal; whatever takes
place in meetings is kept within the team; and the focus is not on the person,
but the process.
Explicit interteam knowledge transfer allows a team that has done.a job on one
site to share the experience with another team working on a similar job on
another site. Interteam knowledge transfer is explicit knowledge, because most of
the knowledge transferred is routine work and the procedure is precise; it must
follow preset guidelines. Personal interaction is ideal in projects that are timely
and can benefit from quick feedback.
Tacit knowledge transfer is unique in complex, nonalgorithmic projects, where
knowledge is mentally stored. The team receiving tacit knowledge is different by
location, by experience, by technology, and by cultural norms. This means that
whatever knowledge is being transferred has to be modified in language, tone,
and content to be usable by the receiving team. One of the problems with this
type of knowledge transfer is the difficulty in tapping tacit knowledge. Unlike
explicit knowledge transfer, tacit knowledge transfer is not practically transferred
electronically.
300 7777 PART III Knowledge Codification and System Implementation

Term s t o K n o w 1111
Authority: The right to decide or to act. ing the experience with another team working on a
Belief: A way of thinking; feeling or impression. similar job on another site.
Bounded rationality: Tendency to select less than the Knowing-doing gap: Ignoring available information and
best objective or alternative; engage in a limited search performing a given job differently.
for alternative solutions; have inadequate information Knowledge transfer: Transmission of knowledge and use
and control of the factors influencing the outcom es of of transmitted knowledge.
a decision. Learning organization: A n organization that uses and
Collaboration: Sharing in the decision-making process for shares its knowledge to promote creativity and innova-
solutions considered to be mutually beneficial. tion; learning from the resulting experiences to advance
Collective sequential transfer: Transferring knowledge and grow in a competitive environment.
from one team specializing in one task to other loca- Satisficing: The practice of selecting an acceptable objec-
tions where the same task is needed. tive or alternative, which might be easier to identify
Culture: Those acquired and learned behaviors that have and obtain, less controversial, or otherwise safer than
persisted in human groups through traditions; the the best object or alternative available.
shared characteristics and values that distinguish the Tacit knowledge transfer: Transferring knowledge on a
members of one group of people from those of another. complex project stored in a human’s mind.
Empowerment: The leader’s sharing of power with subor- Value: A basic belief about a condition that has consider-
dinates. able import and meaning to persons within a society
Explicit interteam transfer: The transfer of explicit knowl- and is relatively stable over time.
edge from a team that has done a job on one site; shar-

Test Y o u r U n d e r s t a n d in g 1 , 1 1
1. Where does knowledge transfer fit in the KM system building life cycle?
2. In what way(s) is knowledge transfer a step in a process?
3. “Knowledge transfer is not only from knowledge bases or repositories.” Do
you agree? Why or why not?
4. What is “bounded rationality”? How does it relate to satisficing?
5. Cite and explain briefly the factors that affect knowledge transfer.
6. Distinguish between:
a. explicit interteam knowledge transfer and tacit knowledge transfer
b. values and beliefs
c. authority and empowerment
d. knowing-doing gap and bounded rationality
7. Explain briefly the key prerequisites for knowledge transfer.
8. List and briefly explain some of the positive cultural values in an organization.
9. How does employee satisfaction affect knowledge transfer?
10. Explain the factors that determine the effectiveness of knowledge transfer.
11. Elaborate on the key inhibitors of knowledge transfer.
12. What are the distinctive features of collective sequential transfer?
13. Give an example of your own tacit knowledge transfer.
14. Summarize the uses and limitations of the Internet as they relate to knowl-
edge management.

K n o w le d g e E x e r c is e s 11,1
1. How would one instill trust in an organization?
2. In your opinion, does one fix the culture and then get the employees to share
or vice versa?
3. Discuss the difference(s) between collaboration and commitment. How are
they related?
4. In what way(s) is it difficult to fit the Japanese culture onto American business?
5. How important do you think trust is in the process of knowledge transfer?
 CHAPTER 9 Knowledge Transfer and Knowledge Sharing a ■■a 301

R e fe r e n c e s 1 1 1 1
Abramson, Gary. “Wiring the Corporate Brain,” Horibe, Frances. Managing Knowledge Workers. New
Enterprise M agazine , March 15,1999, pp. 2-8. York: John Wiley & Sons, Inc., 1999, pp. 21-44.
Becker, Rick. “Taking the Misery Out of Experiential Koulopoulos,Thomas M., and Frappaolo, Carl. Smart
Training,” Training, February 1998, p. 80. Things to K now A bou t Knowledge Management. Oxford,
Clark, Linda E. “A Stitch in Time,” Intelligent Enterprise UK: Capstone Publishing, Ltd., 1999, pp. 127-165.
M agazine , October 20,2000, p. 2. Pfeffer, Jeffrey, and Sutton, Robert I. The K nowing-D oing
Dell, Pamela. “Getting the Most out of Getting Together,” Gap. Boston, MA: Harvard Business School Press,
Knowledge Management, August 2001, p. 1. 2000, pp. 89-106.
Dixon, Nancy M. Com m on Knowledge: H ow C om - Wiig, Karl. Knowledge M anagement Foundations—
panies Thrive by Sharing What They Know. Thinking A b o u t Thinking—H ow People and
Boston, MA: Harvard Business School Press, Organizations Create, Represent, and Use Knowledge.
2000, pp. 33-142. Arlington, TX: Schema Press, 1993.
 Knowledge Transfer
in the E-World
Contents

In a Nutshell
The E-World
Intranets
Extranets and Knowledge Exchange
Groupware
Groupware Applications
E-Business
The Value Chain
Supply Chain Management (SCM) and Knowledge Exchange
Implications for Knowledge Management
Summary
Terms to Know
Test Your Understanding
Knowledge Exercises
References

The real questions are how

D o we stay connected?
H ow do we share knowledge?
How do we function any time ,
any where, no matter what?
— R o b e r t B uckm an

In a Nutshell
Knowledge management does not operate in a vacuum any more than it is only about
technology. A learning organization is driven by intelligence, and knowledge is volun-
tarily shared via teams, between individuals, face to face, and electronically. Technology
is an inescapable part of most knowledge management success stories. The company’s
 CHAPTER 10 Knowledge Transfer in the E-W orld 1 a ■■ 303
technology—including an intranet, groupware, and extranet—plays a major role in
knowledge transfer and knowledge sharing when distance is a problem within the firm
and between firms worldwide. External stakeholders such as vendors, customers, and gov-
ernment are considered partners in a company’s global integrated knowledge manage-
ment system, where knowledge sharing is part of a geographically dispersed environment.
One area where knowledge transfer has gained popularity is in e-business and
e-commerce. It is an environment where external stakeholders form a networked rela-
tionship with a company’s knowledge bases or organizational members. In e-business,
there are e-merchants, e-vendors, and the knowledge automation tools that make each
relationship a value-added reality. Supply chain management is the architecture that
makes such relationships p ossible.In this chapter, we discuss knowledge transfer
through technology that facilitates knowledge sharing. These electronic relationships
involve the company’s intranet and the Internet. Intranets and extranets as part of an
integrated knowledge management environment are also discussed.
In term s of knowledge transfer, e-business helps create knowledge m arkets
through intelligent exchange of information, strategies, and processes. Small groups
inside big firms can be funded with seed money to develop new ideas. DaimlerChrysler
created small teams to look for new trends and products. A Silicon Valley team is doing
consumer research on electric cars and advising car designers. Sooner or later, these
groups create electronic linkups with outside affiliates.
Today’s electronic marketplaces improve knowledge sharing between merchants
and customers and prom ote quick, just-in-time deliveries. Customers and merchants
save money; are online 24 hours a day, 7 days a week; experience no traffic jams and no
crowds; and do not have to carry heavy shopping bags. However, security continues to
be a problem for online knowledge transfer. In a 2000 Economist article, 95 percent of
Americans expressed reluctance to give out their credit card number via the Internet.

■■ ■ ■ The E-World
By e-world, we refer to the electronic facilities available to a company that accommo-
date knowledge transfer and knowledge exchange, regardless of place, location, or dia-
logue. The facilities include the intranet, extranet, groupware, e-commerce, and e-business.
These are the commonly used technologies that address knowledge management solu-
tions. Most of them handle the explicit side of knowledge transfer.

INTRANETS
Intranet is a term used when we apply Internet technology to serve the internal needs
of an organization. It can be viewed as Internet-like capability at the internal organiza-
tional level. The user can simply point and click to access the available information. The
drill is easy. Click on an icon, a button, or a link on the screen and go to different pages.
Internet technology is superior to conventional internal communication sys-
tems. The Web browser, for example, is a readily available and familiar access tool.
Documents are handled easily, and multiple media can be supported as well. Mid- to
large-sized organizations are spending thousands of dollars just to keep their documents
under control. Managers constantly share documents and knowledge on an inter- and
intradepartmental level. With various operating systems, network protocols, and applica-
tion suites, trying to ensure homogeneity in managing knowledge can be quite a chal-
lenge. Intranets handle all transfers and exchanges with ease.
The following scenario is typical of an intranet. Flameless Electric, the second
largest m anufacturer of electrical parts in Virginia, needed to roll out and sell two
304 77T7 PART III Knowledge Codification and System Implementation

m ajor products in 2002. Its strategy used the Internet, intranet, and extranet. The
Internet part was easy. The com pany’s Web site displayed the products, which were
explained by a company expert for customers to see and order. The company’s intranet
was accessible only by its 743 knowledge workers. Its priority was to support its 112 sales
reps statewide. The intranet supplied them with marketing and technical information
about the products and an autom ated sales application that minimized paperwork,
regardless of location or size of order. They were also in touch with product designers
and other experts who could explain each product as needed.
The third part was Flameless’ extranet. An extranet is an intranet with extensions
that allow clearly identified customers or top suppliers to access company-related tech-
nical and educational knowledge bases. The com pany’s extranet was accessible to
950 electricians and small electrical parts dealers via a special company Web page. An
electrician, for example, couid enter an assigned password to access information about
new products and special deals that were available to high-volume buyers. There was
also a knowledge base, called CRM (custom er relationship m anagem ent), that
addressed customer relations. The system stores customer complaints and weaves their
solutions into more intelligent answers for future complaints. Answers to customer
complaints are based on refined past knowledge of the same.
A fter 7 m onths of use, Flam eless’ N et-based system was paying dividends. It
cut down on phone calls and fax orders and gave sales reps in the field immedi-
ate online support for the knowledge they needed to transact business. It even im-
proved shipment schedules and deliveries. A summary view of that intranet is a way of
thinking about how knowledge workers in a business transfer and share knowledge
(see Figure 10.1).
Intranet operation is a communication project designed by a technical staff. It is a
network of people, not of wired machines. The focus is the message, not the media.
Concentrating on the technology of the intranet is like an author worrying about the
presses and typesetting rather than the manuscript.

fl 1 1 I fl fl FIGURE 10.1: Corporate Intranet—A Conceptual Model

Production Team —
New Product
 CHAPTER 10 Knowledge Transfer in the E-World l a i i 305
Benefits
An intranet provides many benefits and has distinctive features. It links knowledge
workers and smart managers around the clock and automates intraorganizational traf-
fic. Today’s face-to-face com m unication and knowledge transfer systems are labor-
intensive, although they are effective once the parties go beyond the exchange of pleas-
antries. An intranet m akes it possible to gain better access to the knowledge and
experience of the decision makers who work within it. It is a creative and empowering
tool for a company and the foundation for developing an enterprise-wide knowledge-
based interactive environment. It is a model for collaborative knowledge transfer and
knowledge sharing (see Box 10.1).
An intranet is not for every business. First, from a cost/benefit analysis view, the
investment is justified when there are 100 or more employees. One major application is
listing employee benefits, company news and views, vacation schedules, job openings,
and recognizing special employees. A company needs an intranet when:
• It has a large pool of information to share among hundreds of employees. It is an
effective way of cutting the cost of producing conventional multiple hard copies.
• Knowledge transfer must reach its destination in a hurry. Because intranets oper-
ate across platforms such as Windows, UNIX, and Mac, they are the easiest way
to get people communicating.
O ther than human resources, an intranet facilitates knowledge transfer and knowl-
edge sharing in other intelligent applications. They include detailed e-mail for inter-
office communication, internal company office circulars, bulletin board service, a daily
to-do list and assignments from a central desk to all connected desks, and a channel for

a bo x 10.1

THE MARK OF A SUCCESSFUL INTRANET SITE

BEST INTRANET SITE ployees to use the Web as the central medium of
collaboration. The Ford intranet is the repository
Bud M athaisel makes no apologies for the way
for all “corporate knowledge.” Every vehicle team
Ford M otor Co. cajoled thousands of employees
has a Web site where they post questions and
to use the Web as the engine for all kinds of busi-
results and resolve business issues. Employees in
ness activities. First, the IT department gave work-
800 facilities worldwide report that 90 percent of
ers a kick-start by seeding the intranet with tons of
the critical data they need to do their jobs is avail-
data and easy-to-use templates. Then, senior man-
able in the Web knowledge base. The intranet has
agement applied the real pressure. “G roup vice
reduced the time it takes to bring new models into
presidents ask—no, direct—their staff to publish
production, improved quality, and cut travel costs,
all of their knowledge and best practices on the
Mathaisel says. “In my career, I have seen many
Web,” says Mathaisel, Ford’s CIO. “The Web is the
promised silver bullets,” he says. “None has pro-
critical conduit for just about all activities.”
duced the value that the intranet has.”
That might explain how, 2 years after its
launch, the Ford intranet has lured 101,000 em -

SOURCE: Joachim, David. “Best Awards Business on the Internet,” InternetWeefc, April 19,1999, p. 47.

a
306 7777 PART III Knowledge Codification and System Implementation

confidential exchange of strategic decisions made by key executives. Intranet facilities

also include a chat room, where employees can post ideas for other employees in the
hope of coming up with a new product, a better policy, and the like.

EXTRANETS AND KNOWLEDGE EXCHANGE

If a company Web site links two or more trading partners, it is referred to as being business-
to-business (B2B) or an extranet. It is a B2B intranet that lets limited, controlled business
partners interact with the firm for all kinds of knowledge sharing (see Figure 10.2).
Intranets, extranets, and e-commerce have a lot in common. Intranets are more local-
ized within a firm and move inform ation m ore quickly than the more distributed
extranets. Internet (primarily Web) protocols are commonly used to connect business
users. On the intranet, Web administrators prescribe access and policy for a defined
group of users. On a B2B extranet, system designers at each participating company
must collaborate to make sure there is a common interface with the company they are
dealing with. One participating business partner might be using Microsoft Explorer;
another may be using Netscape Navigator 6.0. To collaborate via extranet, the applica-
tions have to perform consistently on all represented platforms. The same is true for
e-commerce, where trading partners may be totally unknown to one another.
Extranets are not a passing trend. They are already the backbone of the e-business
future. Extranet technology was initially adopted by corporations that had a vision.
These corporations realized that knowledge sharing contributes to productivity, prof-
itability, and growth. The obvious benefits are faster time to market, customer loyalty,
increased partner interaction, and im proved processes. The easiest way to quantify
return on investment in extranets is to identify a business unit with the company that
might benefit from an extranet. This means identifying a business goal (such as increas-
ing revenue or improving customer base) before deciding on feasibility, justification,
and return on the investment.

FIGURE 10.2: General Extranet Layout

Suppliers
 CHAPTER 10 Knowledge Transfer in the E-W orld a a a a 307
Once a business goal has been established, the next step is for IT departments to
discuss feasibility. In a vertical industry like manufacturing, the focus is on improving
operations through the existing supply chain; in vertical companies like retail chains,
the focus would be on improving revenue. Working with the IT group should bring
both technical and business information together for a master design of the extranet.
Understanding corporate business processes is the key to a successful deployment of
an extranet. By planning the deployment around a well-defined business plan, it is eas-
ier to prove how technology will help the bottom line (see Box 10.2).

1 1 1 i ■ ■ BOX 10.2

EXTRANET AT KODAK

E-COMMERCE TRENDS: KODAK’S into certain internal applications, so that, for

EXTRANET PUSH example, we can learn how products are moving
through their stores and do a better job planning
Eastman Kodak Co. is rapidly expanding its use of
for future orders from them,” says Chiazza.
extranets to cut costs and boost sales by sharing
The extranets let authorized users at other
critical information with major business partners.
com panies “tunnel” under K odak’s firewall to
Kodak has created extranet links to about 25 orga-
access specific servers and even specific applica-
nizations, including dealers, contractors, joint ven-
tions, Internet, intranet, extranets, and groupware
ture partners, and subsidiaries, and is adding new
users. Authorized outsiders can get past Kodak
extranets at the rate of two per week. It is consider-
firewalls and run applications, as they need to do
ing linking electronically to key suppliers and retail
so. Joint venture partners have access to even
chains as well, says VP and CIO John Chiazza.
more resources, such as intranets, databases, and
To boost its revenue via e-commerce, Kodak
mailboxes. Some of the networks give Kodak
is also developing virtual private networks over
access to its partners’ applications.
which customers such as major retailers can place
The extranets enable Kodak and its partners
product orders using a Web browser. A bout 40
to autom atically post product pricing, im medi-
Kodak retail custom ers accounting for about
ately update product literature, manage inven-
$1.3 billion in annual revenue already use the net-
tory, exchange information on jointly developed
works. “The initial focus is on the consumer busi-
products, access accounts, and check the status of
ness,” Chiazza says.
orders and shipments. To ensure that only autho-
The extranets are being used mainly to
rized users get under the firewall, Kodak uses an
exchange inform ation. Kodak continues to rely
extranet m anagem ent and security system that
on electronic data interchange for transactions,
includes integrated VPN (virtual private n et-
but the company is talking to some partners
work) services, as well as data encryption and
about the potential for conducting transactions
authentication for security. A dm inistrators can
over the extranets. “We’ve been involved in B2B
define privileges based on user identification, the
e-commerce for many years as a user of classic
m ethod of authentication and encryption, the
ED I, but what has been emerging recently is
inform ation resource being accessed, company
more intim ate interactions through the use of
affiliation, and day and time.
extranets, where we or our customers can reach

SOURCE: Awad, E! M. Electronic Commerce: From Vision to Fulfillment. Upper Saddle River, NJ: Prentice Hall, 2002, p. 372.

s fl a a a a
308 7777 PART III Knowledge Codification and System Implementation

Security varies with the type of user, the sensitivity of the knowledge transferred,
and the com m unication lines used. Security questions deal with access control, au-
thentication, and encryption. Access control relates to what users can and cannot
access, what can be accessed from what server(s), what accessible data are for display
only, and what accessible information can be restricted to certain times of the day. In
terms of authentication, decisions must be made regarding the level of authentication
for each user, whether passwords and user names are adequate security, and how well
other security measures complement the authentication. Is encryption required? If so,
how strong should it be? W hat type of com m unication line or inform ation should
be encrypted?
Extranets are changing how organizations share internal resources and interact
with the outside business world. Built with technology and used by people, they can
ensure lasting bonds between business partners and corporate members. The entire
commitment should be viewed as a knowledge management asset rather than a mere
networking expense to expedite business. This is where a champion becomes a critical
part of the installation.
A champion who best promotes an extranet is someone who knows the organiza-
tion’s processes, goals, and politics and has technical experience and leadership quali-
ties. This person is an advocate with ability to build company-wide support for the
extranet. It is a demanding role, requiring a detail-oriented expert who can sell top
management on the viability and potential of the extranet. The key to making his or
her case heard is to dem onstrate how an extranet can m eet the company’s revenue
goals. Specifically, there should be a convincing argument for how the extranet will
generate revenue, how the technology will solve the business problem s defined in
advance, and how the work will get done through the proposed extranet.
A champion must drive home the key advantages of an extranet:

• An extranet helps the organization ensure accountability in the way it does busi-
ness and exchanges knowledge with partners.
• An extranet promotes more effective collaboration with business partners, which
improves the potential for increased revenues.
• An extranet is a long-term investment in competitive advantage. Sooner or later,
having an early start on the competition is bound to pay off.

Think of a manufacturing organization with an e-business environment that allows

its business partners, distributors, contractors, and suppliers to access extranet resources
through some preestablished interface. Two of the applications could be an e-commerce
storefront for suppliers (or an integrated application) and a procurement system. The
extranet can be used to manage and tie together all these applications into one inte-
grated system for deriving real value from the company’s entire business relationships.
The extranet is bound to be the technical community that will eventually generate rev-
enue and ensure competitive advantage.

GROUPWARE
One of the most popular technologies that facilitate intermediation is groupware. By
interm ediation, we mean connections between people and how knowledge is trans-
ferred between knowledge seekers and knowledge providers. Groupware is software
that helps people work together, especially for organizations that are geographically
distributed and less likely to encounter face-to-face exchange among knowledge work-
ers. Through this specific class of technology, people rely on groupware to communi-
 CHAPTER 10 Knowledge Transfer in the E-World i«ai 309
cate ideas, cooperate in problem-solving, coordinate work flow, or negotiate solutions.
It seems logical to invest in an intranet that can perform locally within the company
and link globally via e-mail and the Internet. Groupware applications on a company’s
intranet can help the organization do more for less (see Box 10.3).
G roupw are is categorized according to (1) whether users of the groupware are
working together in the same place (face to face) or in different locations (distance)
and (2) whether they are working together at the same time or different times (see
Figure 10.3).
Groupware goes beyond cooperation or work by teams. In designing the technol-
ogy, the prime consideration is learning about group concepts and how group members
behave in a group setting. The goal is to identify the tasks, understand how the group
communicates, and decide on group structure and member roles. Once understood, the
next step is to have a feel for networking technology and how it can assure a produc-
tive user’s positive experience.

-------------------------- a i a a a a box 10.3 a a a a a a

GROUPWARE IN ACTION
What attorney Bill Wright was facing was a fail- work and started using it to discuss business.
ure to communicate. First, one of his three part- When Wright had to produce a draft of a pro-
ners suddenly up and left, taking 17 employees of posed settlement, for example, he no longer had
the Bellmawr, New Jersey, law firm and lots of to rummage through notes from previous dis-
clients with him. Then, within days of his depar- cussions and interrupt his partners with time-
ture, the prodigal lawyer sued his former partners. consuming questions. Instead he simply looked in
Of course, Wright’s firm turned around and filed the program ’s database for a record of earlier
a countersuit. W hat with im prom ptu hallway memos and messages relating to the topic.
discussions, emergency meetings, news flashes, A year after the suits were settled, groupware
urgent requests for background, and the rest of had become the interactive glue holding Wright’s
their caseload, the remaining staff at Farr, Burke, firm together. Their application now has 14 users
Gambacorta & Wright barely had time to breathe. (partners, lawyers, paralegals, and the office man-
“We had to find a way to help us handle the flood ager) and covers 32 topics, from legal procedures
of information,” says Wright. “And we had to find to new cases. If, for example, a bankruptcy judge
it fast.” Simple e-mail, he knew, would not suffice. makes an idiosyncratic procedural ruling about
He needed something that would organize and foreclosures, an associate adds that tidbit to the
catalog information, not just zap bulletins around pertinent topic. Some topics are available to every-
the office. one; others, like the one that disseminates confi-
W hat W right and his colleagues did was to dential information about management, are open
turn to groupware, a category of software used, as only to partners. A pregnant attorney created a
the name suggests, to help people work in groups. subject area related to her cases so that others
Instead of holding one meeting after another, could track her cases while she was on maternity
they bought and installed a program called leave and so that she could get up to speed
TeamTalk (from Trax Softworks) on their net- quickly when she returned to work.

SOURCE: Excerpted from Field, Anne. “Group Think,” www.inc.com/articles/hr/manage_emp/productivity/1995-6.html, Date

accessed November 23,2002.

a a a a a a
310 7777 PART III Knowledge Codification and System Implementation

Same Place Different Place

(face to face) (distance)

Voting, Chat Room,

Same Time
Presentation Support Teleconferencing

Peer-to-Peer, Work-Flow Process,

Different Time
Shared Computers E-Mail

FIGURE 10.3: Groupware Categories

SOURCE: Adapted from Brinck,Tom. “Groupware: Introduction,” www.usabilityfirst.com/groupware/
intro.txl, 1998, Date accessed August 2002.

Companies want to use groupware for several reasons:

• Groupware works well with groups having common interests or work and where
it would not be practical for the same number of people to meet face to face.
• Certain problems are best solved by groups, not individuals.
• Groups can bring multiple opinions and expertise to a work setting.
• Groupware facilitates telecommuting, which can be quite an effective time-saver.
• Joint meetings are often faster and more effective than face-to-face meetings.

Prerequisites for Success

Most groupware systems cannot succeed unless the group of users decides to
accept the system. Having a teleconferencing system is useless if only one side has it.
The two most critical prerequisites for system success are compatibility of software
and perceived benefit to every m em ber of the group. In the early 1980s, IBM and
UNISYS introduced workstations, but their systems could not communicate with each
other. This meant that anyone who wanted to communicate with others in the organi-
zation or elsewhere had to make sure that the other party had the same make. The
problem continued until a decade later, when open system standards made it possible
for anyone using any make to communicate with any other make, regardless of dis-
tance or location.
The second prerequisite for groupware success is the benefit perceived by group
members. Each group member must be convinced that the system will generate explicit
benefits in the way it transacts ideas across the board. Otherwise, there is the danger of
defection. For example, in a chat room, where a common problem is presented for dis-
cussion, if a member perceives that his contributions are not fairly represented for oth-
ers to consider, there is a chance that such a member will minimize or simply refuse to
contribute to the problem solution. Of course, the group may apply social pressure to
 CHAPTER 10 Knowledge Transfer in the E-W orld aaaa 311
encourage groupware use. Any disparity of individual and group benefit should be
hashed out before the final groupware system design.

Communication Structure
In a typical face-to-face session, the protocol is standard. Communication is highly
structured. Someone asks a question or explains a situation while the others listen.
There is usually a question, then a response, and so on. When a communication struc-
ture is known, a groupware system takes advantage of it to speed up communication
and improve the perform ance of the exchange. This communication environment is
referred to as a technologically mediated communication structure. The expectation is
order, discipline, and following a known protocol of question-response.
An alternative communication structure is what we call a socially mediated com -
munication structure , where individuals send a request through technology like e-mail,
with no control over how soon or whether the recipient will respond. This communica-
tion exchange is good for certain problem-solving situations, but not for others. The
im portant point is for the designer to set up a groupware system to fit the specific
requirements of the group that will use it. Unless the prerequisites are identified, there
is the risk that any groupware technology might backfire.

Session Control
Groupware systems require that sessions be conducted within the framework of
protocols designed to ensure privacy, integrity, and successful completion of each ses-
sion. A session is a situation where a group of people agrees to get together to conduct
a meeting in person, over the telephone, or in a chat room. Session control is more like
a bouncer standing at the door of a private club and checking identification, determ in-
ing whether the person is wearing the proper attire, and so forth. Session control deter-
mines who can enter and exit the session, when they can enter, and how. Once inside,
the rest of the group should be aware of the newcomer and his or her role. Some of the
rules used in session control include the following:
• Make sure users do not impose a session on others. Like a telem arketer call with
no filter such as caller ID or an answering machine, a resident feels vulnerable
not knowing who the caller is.
• Conversational group members should be identified before being allowed into a
session. This way, other participants have an idea who is included in the session
and the role they play.
• The system should control unnecessary interruptions or simultaneous transmis-
sions that might result in confusion or chaos.
• A group member should be allowed to enter and leave at any time. However, par-
ticipants should know when such a situation occurs.
• Make sure there is a ceiling on how many people can participate and how long
the session should last.
• Assign a moderator who controls access to a shared whiteboard or a repository
and how long access should last.
• Ensure anonymity, privacy, and accountability at all times. Each user should have
control over the information to be shared; other communications should be pri-
vate or confidential.
When people use technology like groupware to communicate, various types of
information are communicated through channels, implicitly and explicitly. Awareness
inform ation such as docum ents and location and tasks of coworkers should all be
312 7777 PART III Knowledge Codification and System Implementation

considered in groupware system design. This suggests that a groupware system

designer should be well versed in human interaction and communication skills and in
the technology that emulates human exchange. To find such a person is not easy and
requires special training and experience in the domain.

GROUPWARE APPLICATIONS
E-Mail and Knowledge Transfer
Besides the traditional telephone, e-mail is the most common groupware applica-
tion. Unlike the telephone, which involves two persons exchanging information via a
dedicated cable, an e-mail system provides much more. It can send and receive mes-
sages, forward messages, file messages, and create mailing groups. It can also attach
files to a message. Most e-mail systems provide features for sorting and processing
messages. A message can easily be routed to any number of destinations.
The intranet and e-mail is a marriage made in “cyberheaven.” E-mail is what a
company’s intranet is best known for; it is “the N et’s killer app” (Dell 2000). Almost
90 percent of Net users report e-mail as the most frequently used online contact
between knowledge workers. It is a major communication platform in business and
government. Scott McNealy, chief executive officer of Sun Microsystems, Inc., once
commented: “You can take out every one of the 300 to 400 computer applications that
we run our company on and we could continue —but if you took out our e-mail system,
Sun would grind to an immediate halt.”
Over 200 million in-boxes are active worldwide. Frequent e-mailers already recog-
nize that their in-box is as much a database of appointments and news as it is a place to
store messages. With e-commerce volumes on the rise, this knowledge transfer and
communication tool is becoming part of e-marketing and sales.
E-mail is also becoming smarter: It can now manage data and documents with
ease. It can direct specific messages to defined folders and be a place to check voice,
text, and fax messages. This is called content management or unified messaging services.
Managing information with e-mail is more efficient than dealing with the flood of let-
ters, faxes, and bills we handle today. As e-mail becomes the standard for content dis-
semination of all kinds, it should attract more and more users and become as popular
as the cell phone.

Newsgroups and Work-Flow Systems

Similar to e-mail in concept, newsgroups cater to large groups receiving the same
message on demand. That is, messages are available to a participant only when he or
she requests it. In contrast, a work-flow system is a smart technology for cognition. It is
designed to route information (documents, messages, reports) to specific people in the
organization and convey the business rules that govern the use of the knowledge. For
example, an employee returning from a 2-day conference submits an expense report
electronically to her manager, who approves it and forwards it to the accounting
departm ent for processing. A nother unique feature is the system’s capability to ana-
lyze a work-flow process and assess the status of docum ents (where they are, who
should have them next, and so on).

Chat Rooms
This groupware system allows many people to participate in a session, where they
write messages that appear to all participants on the bottom s of their respective
screens. In one chat session, the author was invited to give a talk on e-commerce appli-
 CHAPTER 10 Knowledge Transfer in the E-W orld i i i i 313
cations to an Association for Computing Marketing (ACM) chapter 300 miles away via
the host university’s chat room. The monitor displayed a text message that seven partici-
pants were present. It began with a message from the moderator: “Now the meeting will
come to order. I am pleased to present our speaker for the evening.” A message was dis-
played every time someone connected to the chat room or exited. As the presentation
was entered electronically in text, the presenter was informed of what was going on by
participants through the chat room. For example, when a question was entered, it was
displayed on one corner of the monitor. A t the end, an “applaud, applaud” message was
displayed with a thank-you note, which concluded the chat session.
One thing worth noting about the text version of chat is that there is a transcript of
the presentation, which makes it easier for participants to react at any time. As a result,
in most chat sessions, participants are engrossed in listening, thinking, and responding
to the ongoing exchange. It simulates a live session in a human group.

Video Communication
Another groupware application that is gaining in popularity is video communica-
tion. This system is essentially a phone system with video hardware. This application is
ideal in discussions requiring a video display or when visual information is being dis-
cussed. It also projects a view of activities going on at a remote site. Video communica-
tion systems allow live video for two-way or m ultiple locations, depending on the
requirements of the group session.

Knowledge Sharing
Knowledge-sharing groupware is unique in small business situations; it allows you
to store and share information through a resident knowledge base. The attraction of
such a system is that it involves a few steps that are easy to learn and apply; everything
resides in the knowledge base and is easily accessible (Field 1996).

Group Calendaring and Scheduling

This application provides unique features related to scheduling, project manage-
ment, and coordination of schedules among several people involved in a given project.
One of the intelligent features is detecting when a schedule is in conflict and finding a
meeting time that will accommodate each participant. This includes locating people
and synchronizing their availability.

E-Business
E-business brings the universal access of the Internet to the core business process of
exchanging information between businesses, between people within a business, and
between a business and its many clients. From an interface perspective, e-business
involves various knowledge exchanges: business-to-business, business-to-consumer,
and consumer-to-consumer. Each type is explained later in the chapter.
The focus of e-business is on knowledge transfer and knowledge sharing. It means
connecting critical business systems directly to critical constituencies—customers, ven-
dors, and suppliers—via the Internet, extranets, and intranets. The technology-based
environment provides electronic information to boost performance and create value
by forming new relationships among businesses and customers. E-business is more
than a Web site, in that it affects all aspects of business, from strategy and process to
314 7777 PART III Knowledge Codification and System Implementation

trading partners and the ultimate consumer. It combines the resources of traditional
knowledge-based systems with the global reach of the Web.
From a knowledge management viewpoint, e-business enables learning organiza-
tions to accomplish the following goals:
1. Create new products or services and, consequently, new market knowledge
2. Build customer loyalty through knowledge exchange and knowledge sharing
3. Enrich human capital by more direct and immediate knowledge transfer
4. Make use of existing technologies for research and development and creation of
new knowledge for new and more advanced products and services
5. Achieve market leadership and competitive advantage
E-business means exchange of inform ation with anyone, anywhere, anytime. It
breaks two traditional rules of knowledge sharing: Companies do not share knowledge
with competitors, and suppliers do not share knowledge with buyers, especially knowl-
edge ihat determines pricing. One of the drivers that promotes e-business is change in
organizations. Today’s learning organizations empower front-line workers to do the
kind of work once performed by junior management. There is also a trend toward part-
nering owners and managers across departm ents to develop a chain of relationships
and knowledge sharing that adds value to the enterprise. Downsizing of larger organi-
zations, outsourcing of specialized tasks, and encouraging cross-functional business
processes all require better communication between the departm ents that perform
these functions. E-business, which makes communication easy, is an ideal method of
making these connections (see Figure 10.4).

FIGURE 10.4: Changes in Organizational Makeup

• Empowerment of Front-Line Workers

• Informating of Key Business Activities
• Outsourcing and Downsizing of Large Organizations
• Partnering Between and Among Knowledge Workers
• Cross-Functional Business Processes
 CHAPTER 10 Knowledge Transfer in the E-W orld aaaa 315
E-business helps create knowledge markets. Small groups inside big firms can be
funded with seed money to develop new ideas. For example, a Boston firm formed a
team of specialists to do consumer research on electric cars and car comfort.
Electronic marketplaces improve knowledge sharing between merchants and cus-
tomers and prom ote quick, just-in-time deliveries. Convenience for the consumer is a
major driver for changes in various industries: Customers and merchants save money;
are online 24 hours a day, 7 days a week; experience no traffic jams and no crowds; and
do not have to carry heavy shopping bags. Control is another major driving factor. For
example, instead of banks controlling the relationship with the customer, customers
today can have more control over their banking needs via Internet Web sites. Banks
like Bank of America, Wells Fargo, and hundreds of others now give customers access
to their accounts via the Web.
Even though we can generate a long list of advantages and benefits, there are still
problems and drawbacks to consider before using the Web business for knowledge
exchange or knowledge transfer. The main problem is system and knowledge integrity.
Com puter viruses are ram pant, and new viruses are discovered every day. Viruses
cause unnecessary delays, file backups, storage problems, and the like. The danger of
hackers accessing files and corrupting accounts adds more stress to an already com-
plex operation.
A nother lim itation is corporate vulnerability. The availability of knowledge
regarding products, design specifics, and other information about a business through its
Web site makes it vulnerable to access by the competition. The idea of extracting busi-
ness intelligence from the com petition’s Web pages is called Web farm ing , a term
coined by Richard Hackathorn.

THE VALUE CHAIN

In 1985, Michael Porter wrote a book entitled Com petitive Advantage, in which he
introduced the concept of a value chain. A value chain is simply a way of organizing the
primary and secondary activities of a business so that each activity provides “value
added” or productivity to the total operation of the business. It is a way for businesses
to receive raw materials as input, add value to them through various processes, and sell
the finished product as output to consumers. Competitive advantage is achieved when
an organization links the activities in its value chain more cheaply and more effectively
than do its competitors. For example, the purchasing function assists the production
activity to ensure that raw materials and other supplies are available on time and meet
the requirements of the products to be manufactured. The manufacturing function, in
turn, has the responsibility to produce quality products that the sales staff can depend
on for “sale without pain.” To do the job, the human resource function must hire, retain,
and develop the right talent personnel to ensure continuity in manufacturing, sales, and
other areas in the business. Bringing in qualified people is value added to stability, con-
tinuity, and integrity of operations throughout the firm.
The knowledge-based value chain is a useful way of looking at a corporation’s
knowledge activities and at how various knowledge exchange adds value to adjacent
activities and to the company in general. Everywhere value is added is where knowl-
edge is created, transferred, or shared. By examining the elements of the value chain,
corporate executives can also look at ways of incorporating information technology
and telecommunications to improve the overall productivity of the firm. Companies
that do their homework early and well ensure themselves a competitive advantage in
the marketplace (see Figure 10.5).
316 7777 PART III Knowledge Codification and System Implementation

Purchase Materials Operations Outbound Logistics

Support Activities

Corporate Infrastructure Human Technology Procurement

(planning, finance, accounting) Resources Development

4^ ?
9 9 I fl 9 I FIGURE 10.5: A Knowledge-Based Value Chain

The trend in e-business is to integrate the knowledge management life cycle, from
knowledge creation (such as seeking inform ation from the custom er, vendor, or
employees) to knowledge distribution (such as providing reports, analysis, or forecast-
ing to top managem ent) via three major applications: business-to-consumer (B2C),
business-to-business (B2B), and business-within-business. These are also known as the
Internet, extranet, and intranet, respectively (see Table 10.1).

Business to Consumer (Internet)

The focus of this e-business application is on the use of a m erchant’s Web store-
front or Web site. Consumers anywhere can browse, learn about products, and order
goods or services online anytime. This approach is modeled on the traditional shopping
experience found in stores like Safeway, Kroger, and Kmart. For example, a shopping
cart is used to hold goods until the customer is ready to check out. Checkout is order
and payment processing. B2C is the electronic equivalent of the traditional mail order
or telephone-based ordering system. This system involves knowledge exchange
between the consumer and the e-merchant through the Web site.
 CHAPTER 10 Knowledge Transfer in the E-W orld ■ ■ ■■ 317

TABLE 10.1 Key Elements of Internet, Extranet, and Intranet E-Commerce

Element Internet Extranet Intranet
E-commerce type Business-to-consumer (such Business-to-business (procurement Internal procurement and
as mail order via the Web) and shipment) processing
Access Unrestricted (anyone can Restricted to internal company Restricted to company
access a URL address) employees, staff, and business customers, employees,
partners and staff
Security Generally minimal, except Firewalls and restricted access Firewalls to eliminate non-
for verifying credit cards to data and applications for company employees
and financial transaction partners
integrity
Payment method Credit card or electronic cash Predefined credit agreement Within-business charges
between businesses
fl m a m m a

Business to Business (Internet and Extranet)

Having a Web storefront is not a big deal anymore. Everyone has one. The real
power of knowledge management in e-commerce lies not in the direct sale of products
to customers but in the integration of relationships among merchants and suppliers for
prompt, quality customer service. Business-to-business is industrial marketing: As soon
as an online purchase is entered and payment is approved through a credit card clear-
ance procedure, a message is generally displayed that says, “Thank you for your order.
The amount of $000 will be charged to your credit card. The product should reach you
within 5 to 7 working days.” The moment the message is displayed on the customer’s
monitor, an electronic order is sent to the vendor to fill the order and ship it directly to
the customer. Doing this electronically means reduced inventory and quicker service.

SUPPLY CHAIN MANAGEMENT (SCM) AND KNOWLEDGE EXCHANGE

The concept of supply chain m anagement (SCM) means having the right product in
the right place, at the right time, at the right price, and in the right condition. This is an
integral part of the business-to-business framework. SCM cuts across application infra-
structures and business relationships. It transforms the way companies exchange infor-
mation with suppliers, partners, and even customers. The goal is to improve efficiency
and profitability, but it also means creating new opportunities through knowledge
sharing for everyone involved.
Supply chain m anagem ent employs powerful tools that allow companies to
exchange and update inform ation (such as inventory levels and sales trends) in an
effort to reduce cycle times, to have quicker delivery of orders, to minimize excess
inventory, and to improve customer service. This communication is done quickly from
one knowledge base or database to another. According to an Inform ationW eek
research survey of 300 IT executives using supply chain systems, the majority of the
respondents said the most strategic advantages of supply chain systems are better col-
laboration with business partners, lower operational costs, and reduced cycle times. See
Box 10.4 (Stein and Sweat 2000).
In SCM, the name of the game is collaboration among business partners, knowl-
edge sharing of logistics for timely delivery of goods or products, and consensus among
318 7 7 7 7 PART III Knowledge Codification and System Implementation

BOX 10.4

SUPPLY CHAIN MANAGEMENT

KILLER SUPPLY CHAINS Depot offers its partners recognition incentives to
get them on board.
For most retailers, one of the trickiest links in the
Short-term forecasting is handled locally; with
supply chain is moving goods from the supplier to
up to 65 weeks of data at the store level, store
the warehouse and then on to the store. The Home
managers are given latitude to adjust for demand
Depot, Inc., has found a simple way around that
based on merchandising programs. Home Depot
problem: Remove it. The Atlanta-based build-
prepares long-range forecasts of 3 to 5 years on a
ing supplies retailer now moves 85 percent of its
national level for its suppliers; they contain product-
merchandise —nearly all of its domestic goods—
volume data, of course, as well as where growth is
directly from the manufacturer to the storefront.
expected and where Home Depot plans to build
Product no longer languishes in warehouses, saving
new stores. That helps suppliers decide where to
both suppliers and Home Depot money. “We’re
build new plants and distribution centers, and it
treating each of our stores as if it were a distribu-
puts Home D epot in the position of helping to
tion center, says CIO Ron Griffin. Because of
determine facility location instead of simply work-
Home D epot’s high volume —its stores average
ing around it. “Rather than assume fixed capacity,
$44 million in sales and 5V2 full inventory turns a
we help shape it,” Griffin says.
year—the products frequently ship in full truck-
Home Depot opens up even more data to its
loads, making the system even more cost-effective.
biggest partners. Electric-tool manufacturer Black &
Associates walk store aisles, watching for
Decker is Home D epot’s largest supplier, and
goods that need replenishm ent. As they enter
Home Depot is its largest customer, so it benefits
orders directly into mobile com puting devices,
both companies to share information. Home Depot
called the Mobile Ordering Platform, the requests
passes point-of-sale data to Black & Decker, which
can go almost instantly via ED I connections to
helps the Baltimore company analyze sales and
more than 80 percent of Home D epot’s manufac-
determine future manufacturing volume.
turers, which can respond immediately. Home

SOURCE: Stein,Tom, and Jeff Sweat. “Killer Supply Chains ” InformationWeek, January 16,2000, pp. 1-3.

knowledge workers and suppliers to make sure orders and inquiries are filled correctly.
More and more, companies are extending their focus from internal operations like
scheduling and enterprise resource planning to relationships with external customers
and suppliers. They are looking for the perfect virtual enterprise that will link their sup-
plier’s suppliers to their custom er’s customers to operate together under one umbrella
with seamless connections among knowledge bases, databases, manufacturing, inven-
tory systems, and Web servers (see Box 10.5).
As you can see, B2B exchanges pave the way for a new model of the digital econ-
omy. It is a distinct network of suppliers, distributors, Internet service providers, and
customers that use the Internet for communications and transaction handling. As com-
m unication tools get better and cheaper, transaction costs should drop. With the
Internet, many transaction costs are approaching zero. People around the world can
now quickly and cheaply access the information they need almost instantly. Companies
can also add value to a product or service from any location, at anytime, day or night.
 CHAPTER 10 Knowledge Transfer in the E-W orld a a ■a 319

a a ■ ■ a a BOX 10.5

USES OF SUPPLY-CHAIN MANAGEMENT

• Integrated Web-based systems give companies sophisticated overseas partners that lack the
a better view of their global supply chain, resources to handle EDI.
speeding delivery of products from around • R osettaN et, a group of companies in the com -
the world. puter and electronic com ponents industry, is
• Established companies still rely heavily on elec- implementing standards based on XM L to facili-
tronic data interchange (E D I) to transmit and tate information exchange over the supply chain.
receive order requests along the supply chain. • Lucent Technologies’ Microelectronics Group cre-
• Companies employing Web technologies such as ated a supply system with D H L Airways to deliver
XML can exchange data with less technologically silicon chips halfway around the world in 2 days.

SOURCE: Chadbrow, Eric. “Supply Chains Go Global,” InformationWeek, April 3,2000, p. 52.

a a a a a a

To illustrate, G eneral M otors (GM ), Ford, and Daim lerChrysler announced in

early 2000 that they were moving all their business-to-business activity, involving more
than $25 billion and 60,000 suppliers, to the Internet. The new system will replace a
mammoth procurement process built on phone calls and fax processing. For GM, the
average processing cost of a purchase order is $125. With the Internet, the cost is
expected to drop to $1. Bidding will also drive down the cost of some goods. Parts such
as tires and headlights are already purchased through online reverse auctions, where
the autom aker names the price of the part that it needs, leaving it to a supplier to
accept the price. This approach should capture millions of dollars in savings.
The serious task is overhauling the way work gets done in a company, which for
large corporations can take years and cost millions of dollars. For example, Ford wants
to revamp its manufacturing plants to begin building customized cars for consumers; a
consumer will have a new car just 2 weeks after the order has been placed. This means
major changes for employees, dealers, and suppliers worldwide. Early in 2000, General
Motors Corporation launched an SCM project with similar goals. The work is sched-
uled for completion by 2003 and will cost well over $100 million. It could mean reengi-
neering almost all of G M ’s business processes and a big investment in new technology,
but the payback is expected to be in the hundreds of millions (Violino 1999). The top
10 e-business innovators are shown in Table 10.2.

Customer Relationship Management (CRM)

Related to SCM is customer relationship management or CRM. The goal of this
concept is for a company to enhance its relationship management with customers and
supply chain partners. CRM is a business strategy used to learn more about customers’
needs and behaviors in order to develop stronger relationships with them. It relies on
technology to bring various pieces of customer information to knowledge workers. By
doing so, companies would be able to create a more holistic view of their customers
and, in turn, tailor services to their specific needs.
CRM is about changing a company’s business processes to support a new customer
focus and applying technologies to autom ate those new processes. The technology
320 7777 PART III Knowledge Codification and System Implementation

TABLE 10.2 The Top 10 Electronic Business Innovators

Rank_______ Company_____________ URL Address___________________________ E-Business Profile

1 Office D epot www.officedepot.com This office-supply company integrates e-business tech-

nology across all channels o f its business, increasing cus-
tomer self-service and order sizes while reducing trans-
action costs.
2 IBM www.ibm.com The world’s largest computer company practices what it
preaches, employing an e-business infrastructure to link
customers, employees, and business partners.
3 Cisco Systems www.cisco.com This is one o f the first companies to succeed in using the
Internet for e-commerce. Eight of 10 orders the net-
working company receives are via the Web, with online
sales topping $1 billion.
4 Army & Air Force www.aafes.com World’s eighth-largest retailer, 9 percent of catalog sales
Exchange Service come via the Web. To reach the unconnected, the
exchange for military personnel, family, and veterans
employs Web-tied kiosks at retail outlets.
5 E-Trade Group www.etrade.com This is the leading online brokerage, touted for its high
degree of customization, which lets clients tailor the ser-
vice to their own needs.
6 D ell Computer www.dell.com The granddaddy o f customization sites, the PC maker pro-
vides an easy-to-use Web site, where customers can con-
figure, purchase, confirm, and track orders for desktops,
notebooks, and servers.
7 Lockheed Martin www.lockheedmartin.com A n Internet pioneer in the days before the Web, this de-
fense contractor’s intranet is among the world’s largest,
with more than 1,100 internal sites linking 185,000 users
to 13 enterprise resource planning systems.
8 Avnet www.avnet.com The e-business structure o f this global electronic-
components distributor includes online point-of-use
replenishment systems, accounts for 20 percent of cus-
tomer transactions and more than 90 percent of sup-
plier transactions.
9 Insight Enterprises www.insight.com This supplier of computer hardware and software products
lets custom ers—mostly small and midsize businesses—
customize online catalogs, which helped boost
unassisted Web sales by 222 percent in 1999.
10 Marshall Industries www.marshall.com This electronic-components distributor uses an online
supply chain management program—combined with
data warehousing solutions—which reduces expenses
and increases responsiveness.

SOURCE: Violino, Bob. “The Leaders of E-Business,” InformationWeek, December 13,1999, p. 72.
 CHAPTER 10 Knowledge Transfer in the E-World a a i a 321
allows multiple channels of communication with customers and supply chain partners
and uses customer information stored in databases and knowledge bases to construct
predictive models for customer purchase behaviors. Companies provide a robust net-
work system in order to support a successful CRM implementation. The main benefits
include the following:

• Increased customer satisfaction

• Cross-selling products efficiently
• Making call centers more efficient
• Helping sales staff close deals faster
• Simplifying marketing and sales processes
• Discovering new customers

The two critical elements of CRM software are operational and analytical tech-
nologies. The operational com ponent uses portals that facilitate communication
between customers, employees, and supply chain partners. As shown in Figure 10.6, the
portal resides in the presentation layer of the CRM technical architecture. Some of the
basic features included in the portal product are as follows:

• Personalization services to develop a user profile based on user cached click-

stream information
• Secure services to enforce security rules of the organization for external and
internal users
• Publishing services—an interactive mechanism for users to document the location
and meaning of business content
• Access services to help users find and access portal content
• Subscription services to deliver business content on a regularly scheduled basis
via e-mail, fax, or other media

In contrast, the analytical component uses data-mining technologies to predict cus-

tom er purchase patterns, which enables the company to apply proactive marketing
toward its target customers. Data kept at different parts of the organization, such as
sales, customer services, and billing, is organized for efficient retrieval and update.
All factors considered, the architectural imperative for CRM is to do the following:

• Allow the capture of a very large volume of data and transform it into analysis
formats to support enterprise-wide analytical requirements
• Deploy knowledge —an intuitive, integrated system; rapidly enables processing of
the intelligence gathered from analytical environments
• Calculate metrics by the deployed business rules; identifies pockets of activity by
a consumer in real time, enabling truly strategic targeting

Once operational, communications within the organization allow different parts of

the organization to share knowledge with one another. For instance, a salesperson
might retrieve a historical telephone conversation between one customer and a service
representative and then develop a sales pitch based on that particular customer’s inter-
ests. A nother example would be a customer service representative who creates a notice
to postpone the invoice for a customer due to an error in the billing cycle. The billing
departm ent must then recognize the notice and process the changes made by the cus-
tom er service representative. Local area network technologies are implemented at
most small businesses or at branch offices of larger organizations. Wide area networks
322 ■■ i a PART III Knowledge Codification and System Implementation

The portal serves as a window to the wealth o f information contained within the organization. It is the key to
a successful CRM architecture.

Customers, Partners, Employees, Suppliers

Portal
t
Presentation Layer

Personalization Application Services Secure Services

Services

Publishing Services Subscription Services Collaboration Services

Access/Search Download Center Work-Flow Services

Services

Registration Wizard Bulk Import Services Digital Dashboard

Data Marts

Customer Product Sales Orders

Access Layer

9 9 9 ■ 9 1 FIGURE 10.6: Operational and Analytical Components of CRM in Its Technical Configuration
SOURCE: Adapted from Zurell, Nancy. “Operational and Analytical Technologies for Optimizing Your
CRM System,” D M Review, February 2002, p. 82.

(also known as intranets) operate over the Internet and connect departments that are
located at rem ote areas. The overall CRM infrastructure that provides communica-
tions within the organization and with outside entities is illustrated in Figure 10.7.

■■■ ■ Implications for Knowledge Management

The Web has changed the way business and information technology work. E-commerce
is transforming the Internet from a “browse-and-surf” environment into a mammoth
knowledge exchange. Given the challenges posed by today’s e-business and its poten-
tial, when it comes to success in this emerging field, it is knowledgeable people and
managerial talent that matter. It is people with vision of the future who know how to
 CHAPTER 10 Knowledge Transfer in the E-W orld aaaa 323

Telephone Call
Customer IP Phone Supply Chain
Partners

Call Center

Purchase Order

Returns

□ Corporate
Database
Customer Information

Web-Based
Portal

FIGURE 10.7: Typical Diagram of CRM Network

handle the speed of change. The person who figures out how to harness the collective
genius of the organization will blow the competition away.
Intranets are tools to manage corporate intelligence. They offer unique leverage
and a competitive advantage at all levels of the organization. Among the key success
factors are strong leadership, a focus on users, and effective m anagem ent of the
intranet project. From a managerial view, change may not be immediate and should
be nurtured with care. This author spent 7 weeks designing and im plementing a
$700,000 intranet site for a foreign central bank with 950 employees. After 3 months
of training, coaching, selling, and demonstrating the uses of the new intranet, less than
20 percent of the employees made a habit of using the new system. The rest of the
staff continued to deliver memos, documents, reports, and messages the old-fashioned
way—in person.
Change is closely related to employee satisfaction, and the effect of the intranet on
the way employees do their jobs is important. Those who are forced to use a new sys-
tem will find a way to get back at the company. For example, gripe sites are available on
the Internet that employees use to state their dissatisfaction with the employer. For
example, www.vault.com and www.aolsucks.com are gripe sites that accept such
complaints. They are normally started by competitors, disgruntled employees, whistle-
blowers, or activists. The motive ranges from sheer anger to venting frustration or get-
ting revenge by spreading damaging inform ation. A nyone with an ax to grind can
smear the employer for the public to read. IT and company recruiters should review
such sites and check what is posted about their company.
A nother implication for management is the strategy for recruiting qualified tech-
nical personnel. The trend used to be offering significantly higher salaries than the
industry average for that job title. However, most organizations today look for appli-
cants with stability, loyalty, and com m itm ent to the work ethic. They offer bonuses
based on performance rather than raises, because they do not have to repeat them in
later years.
324 . . . . PART III Knowledge Codification and System Implementation

Extranets are career enhancers for many IT professionals. Those who work on a
successful extranet project usually end up having the biggest impact on their employer.
In one case, an extranet designer deployed an extranet with the goal of driving down
costs. The extranet met the goal by automating processes, improving overall efficien-
cies, and decentralizing functions for faster and better quality decision making. In addi-
tion to knowledge of the company’s business processes, her skills included client/server
technology, data communication and networking, and HTTR She saw a way of securely
linking customers, suppliers, and vendors to the corporate network. When the extranet
was implemented, the company recognized the change in revenue, which translated
into a hefty raise for the 23-year-old newcomer.
With change being accepted as a way of life, the human resources department has
the option of changing the staff or changing the people already on staff. There is a new
focus for building productive organizational culture, managing change and results,
building intellectual capital, creating future leaders, managing organizational learning,
and pushing for growth and innovation. As someone said, “If you are not the lead ele-
phant, you’ll never charge.”
In terms of success in today’s digital economy, the real asset is not money; money
is just a commodity. Our real value is knowledge sharing and how knowledge is used
to create value for the customer. More than half of doing business is on the soft side —
the core personnel of the firm and the customer. Having employees be part of the orga-
nization and improving their skill sets adds value and contributes to the success of
the firm.
A ttracting qualified technical people is a challenge, but finding ways to retain
them is a full-time job. In its annual job satisfaction survey, involving 575 respondents,
C om puterw orld reported that overall job satisfaction had decreased in 1999. More
than two-thirds of the IT workers said they were not working to their potential and
more than 25 percent said they planned to look for a job elsewhere. With the economic
downturn in 2001, the situation has not improved much. The ability of IT team mem-
bers to work well together contributes to their retention and productivity. Recognition
and flexible hours were found to be extremely im portant in technology-based work.
Stress stemming from heavy workloads was cited as the main drawback.
Finally, the top challenge in managing knowledge through e-commerce is in under-
standing the customer. Most of the successful companies form a 360-degree customer
view by gathering data from every possible source and analyzing it to shed light on the
kinds of details that mark the way consumers shop and buy. Obviously, companies that
better understand their customers’ preferences can sell more. They know which cus-
tomers are most important, most profitable, and most loyal.

Su m m a r y 1,11
• The Web is the fastest growing, most user friendly, and most commercially popu-
lar technology to date. Internet service providers link commercial traffic to its
destination. This involves paying for transactions, managing networks, and linking
consumers and businesses to the Internet. As the enabler of e-commerce, the
Internet has many uses; it also has many limitations. It can be a tool for gathering
opinions, exchanging knowledge, and trying out new ideas. It can also be a vehicle
for inexpensive, easy mass distribution of information, products, and services.
There are limitations as well: security and privacy issues, hackers, worms, Trojan
horses, viruses, and customer relations problems.
 CHAPTER 10 Knowledge Transfer in the E-W orld aaaa 325
• An intranet is a network connecting a set of company clients by using standard
Internet protocols, especially TCP/IP and HTTP. Intranets can handle all kinds of
communication with ease.
• An extranet or company Web site links two or more trading partners. When con-
templating an extranet installation, there are five key factors to consider: identify-
ing the user, listing the technology components, specifying the security require-
ments, setting up the administration, and understanding the usability.
• There are several benefits to an intranet: It links employees and managers around
the clock; companies gain access to their primary resources; and it is the founda-
tion for developing an enterprise-wide information system and a model for inter-
nal information management and collaborative computing. In addition, there are
cost advantages and ease of access, plus portability and scalability.
• Electronic business connects critical business systems directly to key constituents—
customers, vendors, and suppliers—via the Internet, intranets, and extranets. One
of the main advantages of electronic commerce is knowledge sharing and the
ability to swap information within organizations, between organizations, and
between organizations and their suppliers and customers worldwide. The main
limitations are concerns about knowledge protection and the integrity of the sys-
tem that handles the knowledge.
• Groupware is essentially people using technology to communicate various types
of information implicitly or explicitly.
• There are several groupware applications worth noting: e-mail, newsgroups and
mailing lists, work-flow systems, calendaring and scheduling, video communica-
tion, and knowledge sharing.
• A value chain is a way of organizing the activities of a business so that each activ-
ity provides added value or productivity to the total operation of the business.
• Supply chain management (SCM) means having the right product, in the right
place, at the right time, and in the right condition. The goal is to improve effi-
ciency and profitability.
• CRM allows companies to establish multiple channels of communication with cus-
tomers and supply chain partners and uses customer information stored in databases
and knowledge bases to construct predictive models for customer purchase behaviors.
• The knowledge transfer and knowledge-sharing life cycle includes three major
e-commerce applications: business-to-consumer (B2C), business-to-business
(B2B), and business-within-business.
• An intranet wires the company for knowledge exchange and links a company’s
knowledge network electronically. E-mail replaces paper.

Terms t o K n o w " * * '

Architecture: The hierarchical physical structure of the Electronic business (e-business): Connecting critical busi-
Internet. ness systems and constituencies directly via the
Browser: A program designed to search for and display Internet, extranets, and intranets.
Internet resources. Electronic commerce (e-commerce): Ability to conduct
Business-to-business (B2B): The networking of busi- business via electronic networks such as the Internet
nesses and their respective suppliers to exchange infor- and the World Wide Web.
mation, execute orders, and obtain supplies. Extranet: Extended intranets that share information with
Business-to-consumer (B2C): The interface where con- business partners via the Internet.
sumers access a merchant’s Web site for the purpose of Groupware: A way to use computer technology to
buying merchandise or requesting service. enhance the group work process and decision making
326 7777 PART III Knowledge Codification and System Implementation

through support in the time and place dimensions; tech- there is the assurance of having the right product in the
nology that facilitates the work of groups. right place, at the right time, at the right price, and in
Internet: An infrastructure that links thousands of net- the right condition.
works to one another; the information highway that Support activity: The activities of a business that support
makes the information stored on thousands of com- its primary activities.
puters worldwide available to millions of people Value chain: A way of organizing the activities of a busi-
everywhere. ness so that each activity provides added value or pro-
Intranet: A private network based on the same technol- ductivity to the total operation of the business.
ogy as the Internet, but restricted to an organization, its Web farming: Systematically refining information
employees, and select customers. resources on the Web for business intelligence capture.
Primary activity: The key functions of a business, includ- World Wide Web: Also known as WWW or the Web; the
ing inputs, operations, outputs, marketing and sales, whole constellation of resources that can be accessed
and service. by using tools such as HTTP; an organization of files
Supply chain management (SCM): Integrating the designed around a group of servers on the Internet,
networking and communication infrastructure programmed to handle requests from browser software
between businesses and their suppliers so that that resides on users’ PCs.

Te s t Y o u r U n d e r s t a n d in g 1 1 a a
1. What is e-commerce? How does it relate to knowledge management, knowl-
edge transfer, or knowledge sharing?
2. Contrast e-commerce with e-business.
3. Do you think the advantages of e-commerce outweigh the limitations?
Explain.
4. What is groupware? How does it differ from e-mail or the telephone?
5. What guidelines are involved in controlling chat room sessions?
6. In what way is security a limitation of e-commerce?
7. Distinguish between:
a. value chain and supply chain management
b. intranet and extranet
c. e-commerce and Internet
8. Cite the key benefits of an intranet. Is it beneficial in every type of knowl-
edge exchange? Why or why not?
9. Briefly describe the uses of the Web.
10. What is the relationship between an intranet and groupware?

Kn o w l edg e Ex e r c is e s 1111
1. How does the Web fit with knowledge sharing? Discuss.
2. “Technically, there is no difference between the Internet and an intranet,
except that only select people are allowed to connect to the intranet.”
Evaluate this statement in the light of the way intranets are designed.
3. Do you think having an intranet environment is the best way to communi-
cate within the firm? Assess alternative modes of communication, and report
your findings in class.
4. If intranets offer so many benefits, why do you think some companies resist
having them? Is it the size of the firm? The nature of the product? The caliber
of personnel?
5. How does the Web affect our traditional sales channels, partners, and
suppliers?
6. How can a company best prepare to use the Web as a channel for knowl-
edge sharing?
7. Find a company that chooses not to use e-business in its business. What fac-
tors or problems does it consider in staying away from e-commerce?
 CHAPTER 10 Knowledge Transfer in the E-W orld aai a 327
8. Go on the Internet and find a tutorial on the World Wide Web. Review the
tutorial and explain why you think it is easy (or difficult) as a learning tool.
9. Is the Internet different from other media for knowledge transfer?
10. Give an example of how e-business can help a firm reach its customers in a
very low-cost fashion.
11. Access a chat room of your choice on the Internet, and report your experience.
12. Check the following Web sites to learn more about practices in electronic
commerce:
a. Let customers help themselves: www.got.com and www.edmunds.com
b. Nurture customer relationships: www.amazon.com
c. Target markets of one: www.wsi.com
d. Build a community of interest: www.cnet.com
13. Discuss the value chain by visiting the FedEx Web site at www.fedex.com.
How does it relate to knowledge transfer? Knowledge sharing?
14. Internet participation will alter the traditional form of knowledge sharing.
Do you agree? Locate inform ation on the Internet that might support or
negate this statement.
15. Interview a businessperson or a technical person who is involved with e-business.
What has been his or her experience in incorporating the technology into the
company’s day-to-day media for knowledge transfer? What performance cri-
teria are used to judge the success (or failure) of e-business in the business?
Write a short news release for the college or university newspaper to share
your knowledge.
16. You are an Internet consultant to a company that wants you to do a 1-hour
presentation to top m anagem ent about the im portance and potential of
the Internet in the company’s electronic knowledge creation and exchange.
What information do you need before you prepare the presentation? Write a
3-page report detailing the content of the speech.
17. Visit a large firm on an intranet site. Identify the technology that operates the site.
18. Identify a large retailer in your area and determ ine whether it is ready to
adopt an intranet and an extranet. Interview the head of the IT division and
learn about the technology in use. R eport your findings to class.
19. Design an intranet (on paper) for a small bank of 65 employees. Explain the
details of the infrastructure to a local IT specialist. What did he or she find
right and wrong with your design? Write a 4-page report summarizing your
experience.

Re f e r e n c e s * 1 ' 1
Anton, Kathleen. “Effective Intranet Publishing: Getting Chadbrow, Eric. “Supply Chains G o Global,”
Critical Knowledge to Any Employee, Anywhere,” InformationW eek, April 3,2000, p. 52.
Intranet Design M agazine, August 12,2000, pp. 1-5. D eck, Stewart. “The A BC s of CRM,” CIO .com , Febru-
Awad, E. M. Electronic Commerce: From Vision to ary 23,2001, pp. 1-4.
Fulfillment. Upper Saddle River, NJ: Prentice Hall, D eck, Stewart. What Is CRM ? www.cio.com/research/
2002, p. 372. crm/edit/crmabc.html, D ate accessed D ecem -
Brandel, Mary. “D em ise of the Skill Premium,” ber 2002.
Com puterw orld, July 31,2000, p. 62. D ell, B. J. “The N et’s Killer Application,” Knowledge
Brinck, Tom. “Groupware: Introduction,” www.usabilityfirst. M anagement, D ecem ber 2000, p. 34.
com/groupware/intro.txl, 1998, Date accessed August 2002. Downes, Larry and Mui, Chunka, Unleashing the Killer
Brown, Eric, and Candler, James W. “The Elements of App. Harvard Business School Press, 1998.
Intranet Style,” Intranet Design M agazine, August 12, Elbel, Fred. “General Guidelines and Tips (H ow to Get
2000, pp. 1-5. Rid of Junk Mail, Spam, and Telemarketers),”
328 7777 PART III Knowledge Codification and System Implementation

www.ecofuture.org/jmnews.html, Date accessed Spence, Rob. “Considering an Extranet? Consider This,”

May 26,2001, pp. 1-3. Extranet Strategist, Spring 2000, p. 1.
Ewalt, David M. “Strong Growth Predicted for CRM Stein, Tom, and Sweat, Jeff. “Killer Supply Chains,”
Services Market,” InformationW eek.com, April 9,2002. InformationW eek, January 16,2000, pp. 1-3.
Field, Anne. “Group Think,” www.inc.com/articles/hr/ Tao, Paul I. “Roadmap to a Successful Intranet,” Intranet
manage_emp/productivity/1995-6.html, D ate accessed Design M agazine, August 12,2000, pp. 1-13.
October 15,2002. Violino, Bob. “The Leaders of E-Business,”
Joachim, David. “Best Awards Business on the Internet,” InformationW eek, D ecem ber 13,1999, p. 72.
InternetWeek, April 19,1999, p. 47. Violino, Bob. “Kodak’s Extranet Push,” Extranet
Koulopoulos, T. M., and Frappaolo, Carl. Smart Things to Strategist, Spring 2000, p. 17.
Know A bout. Milford, CN: Capstone Publishing, Inc., Zurell, Nancy. “Operational and Analytical Technologies
1999, pp. 67-98. for Optimizing Your CRM System,” D M R eview ,
Mandell, Janette. “E-Mail Etiquette,” Software M agazine, February 2002, p. 82.
July 1998, p. 20.
Porter, Michael. Competitive Advantage. Brookline,-M A:
The Free Press, 1985.
 Learning from
Data
Contents

In a Nutshell
The “Learning” Concept
Data Visualization
Neural Networks As a Learning Model
The Basics
Supervised and Unsupervised Learning
Business Applications
Relative Fit with Knowledge Management
Association Rules
Market Basket Analysis: PETCO
Classification Trees
Definition
Decision for Granting a Loan
Tree Construction
Implications for Knowledge Management
Summary
Terms to Know
Test Your Understanding
Knowledge Exercises
References
Web Sites

“ Unlearning must often take place

before learning can begin. ”
—G a r y Ha mmel
C o m p e tin g f o r t h e F u tu r e
330 a m PART IV K M System Tools and Portals

■■■■ In a Nutshell
Learning is an iterative process, where the final model results from a combination of
prior knowledge and newly discovered information. The improved model gives a busi-
ness an important competitive advantage. Learning tools are critical for developing a
knowledge-management environment. The approach is to identify whether such tools
are theory-driven or data-driven. Theory-driven learning tools require the user to spec-
ify a model based on prior knowledge and to test that model for validity. Data-driven
learning tools automatically create the model based on patterns found in the data.
However, the newly discovered model must be tested before it can be considered valid.
Inadequate or incomplete knowledge for decision making often plagues managers
at all levels. In this chapter, we demonstrate how certain processes coordinate fragments
of knowledge to enable learning organizations to function effectively. It is im portant to
understand how certain mechanisms, procedures, and processes mobilize the knowl-
edge needed to generate quality decision making. We start with the concept of learning,
and learning from data, as one aspect of the learning process in an organization.
In this chapter, we cover three major techniques employed in data mining: neural
networks, association rules, and classification trees. The data-mining process is covered
in Chapter 12.

■■■■ The “Learning” Concept

The collaborative intelligence layer of KM infrastructure relies on several technologies,
including artificial intelligence, expert systems, data warehousing, case-based reasoning
(CBR), intelligent agents, and neural networks. Case-based Reasoning was covered ear-
lier. Data warehousing is a unifier of databases and will be covered in Chapter 12. In this
chapter, we focus on neural networks and their role in knowledge management systems.
Using “learning” technology tools provides a collaborative environm ent for
humans to enhance the processes that are im portant to them, such as new product
development or other tasks that are regularly performed across industries. The goal is
to improve the quality of com m unication and decision m aking within the firm.
Learning tools and procedures provide new knowledge about a product or service, or
produce specific knowledge in the way of a solution, a remedy, or an outcome.
When testing learning tools, the competitive market is fertile ground for knowl-
edge validation, which is a prerequisite for knowledge use. The unifying concept of
learning is the specific mechanism that helps companies determine the kind of knowl-
edge required for decision making.
In a knowledge autom ation environment, learning is a process of filtering ideas
and transform ing them into valid knowledge, having the force to guide decisions.
Knowledge validation is a two-step process: model validation and consensual approval.
Model validation involves testing the logical structure of a conceptual or operational
model for internal consistency and testing the results for external consistency with the
observable facts of the real world. Consensual approval means approval of a special
reference group or the user of the results.
Learning from data and data patterns produces knowledge that can be used in
business decision making. In general, the goals of the learning process are as follows:
• Discovering new patterns in the data,
• Verifying hypotheses formed from previously accumulated real-world knowledge, and
• Predicting future values, trends, and behavior.
 CHAPTER 11 Learning from Data a ■a * 331
Learning from data requires models that automatically improve with experience.
There are two approaches to building learning models: top-down or bottom-up. In the
top-down model, one starts with a hypothesis derived from observation, intuition, or
prior knowledge. For example, a hypothesis could be, “Tourists visiting Egypt earn an
annual income of at least $50,000.” The hypothesis is then tested by querying the data-
bases and performing some statistics. If the tests do not support the hypothesis, then
the hypothesis must be revised and tested again. The strategy is straightforward: gener-
ate ideas, develop models, and evaluate them for validation.
In the bottom-up approach, there is no hypothesis to test. Learning techniques are
used to discover new patterns by finding key relationships in the data. An example of
pattern discovery is analysis of the shopping basket, whereby purchasing transactions
are collected and stored in the database. The goal is to identify products that are pur-
chased together. Extraction of unknown buying patterns can be used to under stand the
buying habits of customers and perform actions that may increase sales, such as adjust-
ing inventions or modifying floor layouts. For example, in one study, patterns of sales
data in a supermarket chain showed that married males between the age of 21 and 27
who went shopping for diapers also bought beer. As a result, the store began to stack
beer cases next to the diaper shelf.

■■■■ Data Visualization

Before building effective models, we must become familiar with the stored data struc-
ture, relationships, and meaning. The first step is to gather a variety of numerical sum-
maries and perform some descriptive statistics in order to get averages, standard devi-
ations, and so forth. For example, we might find meaning in learning about the average
sale of a leader product or the rate of the sale of one product, such as suits to the sale
of another, such as neckties. Exploring the data means searching visually or numeri-
cally for groups or trends. This includes the following:
• Distribution of key attributes, for example, the target attribute of a prediction task,
• Identification of outlier points that are significantly outside the expected range
of the results,
• Identification of initial hypotheses and predictive measures, and
• Extraction of interesting grouping data subsets for further investigation.
The problem in using visualization stems from the fact that models have many dimen-
sions or variables, but we are restricted to showing these variables on a 2-dimensional
computer screen or paper. This means visualization tools must use clever representations
to collapse a number of dimensions into just two.
The role of data visualization in understanding phenom ena and solving real prob-
lems is illustrated in the historical experience of Florence Nightingale and John Snow
(see Box 11.1).

■ Neural Networks As a Learning Model

As an evolving technology, neural networks have been m odeled after the human
brain’s network. The technology attempts to simulate biological information process-
ing via massive networks of processing elements called neurons. Neural nets and digi-
tal computers are not the same. Neural nets are neither digital nor serial; they are ana-
log and parallel. They learn by example, not by program m ed rules or instructions.
Digital computers do not evolve as such.
 BOX 11.1

LEARNING TO SAVE LIVES IN THE NINETEENTH CENTURY VIA DATA VISUALIZATION

FLORENCE NIGHTINGALE: The report included a urose-shaped,” polar-area

diagram, which compared the number of monthly
The Crim ean War in M arch 1854 saw the allies
deaths in two hospitals: one in Crimea, the other in
(Britain, France, and Turkey) declaring war on the
Manchester, England. Each segment, which corre-
Russians. Many medical facilities were put up, and
sponds to a time period of 1 month, has an area pro-
their conditions were criticized as being beyond
portional to the number of deaths. In the left-hand
appalling. The heroic Italian-born Florence
corner, the segments after the new sanitary im-
Nightingale was asked to assist in Scutari as a
provements can be seen, and the diagram shows
nurse. Initially, the doctors did not ask for the help
substantial decreases in the number of deaths.
of nurses, but conditions forced things to change.
More importantly, the visualization of the data
In 1858, Florence Nightingale returned to Eng-
showed the medical body that the deaths were
land and wrote a report addressed to the British
mainly due to hygiene problems in the medical
government describing the unsanitary conditions
facilities, not from wounds or general sickness.
and offering recommendations for improvement.

Proportion o f Deaths Due to:

□ Death from Wounds

□ Death from Preventable

Contagious Diseases

Death from All Other

Causes

Crimea
 CHAPTER 11 Learning from Data ■«■■ 333

Nightingale’s use of visual statistics furthered cholera deaths. He was able to locate a large cluster
the progress of the nursing profession as an inte- of deaths (over 500 in 11 days) around a water
gral and essential part of every medical facility. pump on Broad Street. The pumps were immedi-
ately disabled, and the number of deaths decreased
substantially. Surprisingly, John Snow also noted
JOHN SNOW:
that workers at a nearby brewery on the same street
In 1845, an outbreak of cholera hit the London/ were relatively unaffected by cholera. This was due
Soho area (see figure on facing page). The medical to the workers’ drinking beer instead of water.
officer commissioned to investigate the disease was Snow’s use of spatial analysis was revolution-
Dr. John Snow. He attempted to locate the source of ary at the time. Today, spatial mapping is used in
the cholera by drawing the map shown here and medical geography to locate, corner, and identify
pinpointing in black those areas with reported deaths related to all forms of diseases.

Yards
50 0 50 100 150 200
i______________i______________i______________i______________ i______________i

X Pump • Deaths from cholera

SOURCES: Tobler, W. “John Snow’s Original 1854 Map on the Location of 578 Deaths from Cholera,” www.ncgia.ucsb.edu/pubs/
snow/snow.html, Date Accessed November 2002; Evans, D. H. “The Story of Florence Nightingale,” www.florence-nightingale.
co.uk/biography.htm, 1933, Date accessed December 2002.
334 a a a a PART IV K M System Tools and Portals

One of the areas where neural nets make a definite contribution to decision mak-
ing is in knowledge automation systems. Neural nets show superior performance over
that of human experts for eliciting functional relationships between input and output
values. In a comparison of the two technologies, neural nets prove to be better at learn-
ing by example rather than by rules, and they continue to learn as the problem envi-
ronm ent changes. This capability makes them well suited to deal with unstructured
problems and inconsistent information. Neural nets can also handle fuzzy data without
losing accuracy. Knowledge autom ation systems largely depend on complete data
before they offer a final solution. In summary, knowledge autom ation systems offer
structure, explanatory capability, and validity, whereas neural nets offer the creative
part of problem-solving. They learn from experience in much the same way that a
human expert develops decision-making skills.

THE BASICS
The human brain consists of 10 billion or so neurons; each neuron interacts directly
with 1,000 to 10,000 other neurons. A neuron fires or does not fire. The rate of firing
determines the magnitude of information. When the brain accepts inputs, it generally
responds to them. The responses are a combination of learning and what has been
genetically programmed over time.
A neural network (NN) is an information system modeled after the human brain’s
network of electronically interconnected basic processing elements called neurons. It is
viewed as a self-programming system based on its inputs and outputs. Each neuron has
a transfer function that computes the output signal from the input signal. A typical
neuron building block, or node, is shown in Figure 11.1. The neuron evaluates the
inputs, determines their strengths or weights, “sums,” the combined inputs, and com-
pares the total to a threshold (transfer function) level. The threshold could be some-
thing as simple as zero or could mirror the input within a given range (0-1). If the sum
is greater than the threshold, the neuron fires. That is, it sends output from that particu-
lar neuron. Otherwise, it generates no signal.
Interconnecting, or combining neurons with other neurons, forms a layer of nodes
or a neural network. The interconnection is analogous to the synapse-neuron junction
in a biological network.

■ ■ ■ ■ ■ fl FIGURE 11.1: A Single Node with Weighted Inputs and Summation Function

Inputs Synaptic Strengths

Stimulation
Level

n
 CHAPTER 11 Learning from Data ■■■■ 335
In a real-world problem, inputs represent variables, and the connecting weights are
relationships between variables. Once input values are received by a trained network,
the output value of each unit is computed by the output function. In Figure 11.2, the
neural network model predicts a firm’s status; solvent versus bankrupt. Five input vari-
ables are fed into the system with two possible outcomes as outputs. The hidden units
adjust the weights to the right threshold and are used to support the transformation
from input to output.

SUPERVISED AND UNSUPERVISED LEARNING

The most interesting feature of artificial neural networks is their ability to learn.
Learning may be supervised or unsupervised. In supervised learning, the neural net-
work needs a teacher, with a training set of examples of input and output. Each ele-
ment in a training set is paired with an acceptable response. That is, the actual output of
a neural network is compared to the desired output. The network makes successive
passes through the examples, and the weights adjust toward the goal state. When the
weights represent the passes without error, the network has learned to associate a set
of input patterns with a specific output. This is more like learning by reinforcement.
In unsupervised (also called s e lf supervised) learning, no external factors influence
the adjustment of the input’s weights. The neural network has no advanced indication
of correct or incorrect answers. It adjusts solely through direct confrontation with new
experiences. This process is referred to as self-organization.

FIGURE 11.2 : Neural Network Model Predicting a Firm’s Bankruptcy

Five Input Two Output

Variables Variables

Net Working Capital/

Total Assets

Retained Earnings/
Total Assets
Solvent Firms

EBIT*/Total Assets

Market Value Bankrupt Firms

of Common and
Preferred Stock

Sales/Total Assets

Input
Units

* Earnings Before Interest and Taxes

SOURCE: Awad, E. M. Building Expert Systems. Minneapolis, MN: West Publishing, 1996, p. 437.
336 iii fl PART IV K M System Tools and Portals

BUSINESS APPLICATIONS
Neural networks are best applied in situations having a need for pattern recognition,
where data are dynamic. Even though this technology has been applied in virtually
every industry, the business sector has experienced significant successes with neural
networks. Financial institutions, for example, are using neural networks to simulate
cash management, asset and personnel risk management, and capital investments. In
the capital investment arena, neural networks are used to simulate the reaction of
investors to changes in organizational concerns such as capital structure, dividend pol-
icy, and reported earnings.
The following applications demonstrate neural networks’ contributions to solving
business problems in various industries.

Risk Management
A major New York bank uses a neural network to appraise commercial loan appli-
cations. The network was trained on thousands of applications, half of which were
approved and the other half rejected by the bank’s loan officers. From this experience,
the neural net learned to pick risks that constitute a bad loan. It identifies loan appli-
cants who are likely to default on their payments.

CitiBank London
CitiBank London needed to develop a successful system to forecast the fluctua-
tions of the foreign exchange market. They needed a very complex nonlinear system
suitable for pattern recognition and neural networks.
FX trader is a hybrid system that combines the strong searching ability of genetic
algorithms with the detailed pattern recognition abilities of neural nets. Genetic algo-
rithms (GA) were used to identify combinations of technical indicators that were most
accurate at forecasting FX exchange rate changes by screening and selecting those that
offered the most forecasting power. The values produced by these combinations of
indicators were then used as inputs to a neural network. Following this, the system
advised on a buy or sell decision. The system was trained for the exchange rates of the
U.S. dollar against the Swiss franc and the Japanese yen, using data from the first 6 months
of 1990. Then it was tested over an 8- to 11-week period.
The results revealed a return on capital of about 20 percent, an attractive record,
given that the best human traders could make 15 percent per year. Moreover, the sys-
tem correctly forecasted the market direction 57 percent of the time, far surpassing the
best human traders. Citibank plans to develop this system for other markets, including
U.S. interest rate markets and the Nikkei Index. It also hopes to license the forecasting
technology to fund managers and preferred customers.

Fidelity Investment
Fidelity Investment needed both to save time and to maintain high standards in
the screening of stocks for its portfolio. A universe of 2000 stocks was scanned to select
about 50 stocks for its portfolios. The scan was based on objective criteria and data
about the studied companies, such as stock prices, dividends, historical earnings, and
balance sheet information. This data was fed to the neural network, which was able to
issue a signal indicating which stock was suitable for further consideration.
The neural network system is simple and could easily be transferred to other popu-
lations of stocks given the data and the skills indicated. The performance of the mutual
 CHAPTER 11 Learning from Data ai aa 337
fund is quite impressive. The portfolio has five stars, and only 18 funds in the United
States have that number of stars. Since December 1988, the stock market has gone up
71 percent, while the Fidelity Funds portfolio has gone up 99.8 percent and has out-
performed the market every year since then.

Mortgage Appraisals
One of the most promising applications is a neural network system called AREAS,
which reviews mortgage appraisals. The program uses the data in the mortgage loan
application with a sophisticated statistical model of real estate valuations for the
immediate neighborhood, the city, and the country where the property is located. Then,
the system comes up with a valuation for the property and a risk analysis for the loan.
For each valuation, the cost has been less than one-half of the manual method, which
was also time-consuming and subject to various errors.
Innumerable neural network applications pop up each year. In business, applica-
tions identify qualified candidates for specific positions, optimize airline seating, and
determ ine fee schedules. In finance, applications identifying forgeries, interpreting
handw ritten forms, and assessing credit risks have already justified the necessary
investment for these neural networks. In manufacturing, neural network applications
control quality and production line processes, select parts for an assembly line, and per-
form quality inspection of finished products.
A common thread running through these applications is pattern recognition. The
neural network looks for a pattern in a set of cases, classifies and evaluates patterns,
and reconstructs the correct pattern with a high degree of accuracy. This technology
offers businesses a chance to assist decision makers in the kind of work that was once
relegated to human perception and judgment.
Perhaps the best a neural network can do is to advise rather than replace humans
in problem-solving. The fact that humans can still veto a neural netw ork’s decisions
attests to the higher-level of learning capacity and intelligence of humans.

RELATIVE FIT WITH KNOWLEDGE MANAGEMENT

There are several characteristics of neural networks that fit in a knowledge manage-
ment system. The main characteristics are as follows:
• A neural net exhibits high accuracy and response speed. After a period of training,
a neural net can be quite accurate, although both accuracy and response speed
tend to degrade with increased complexity of the problem being solved.
• High input preprocessed data is often required for building a neural net.
Unfortunately, this is what requires so much data, although the system has high
tolerance for “bad” data and usually ends up with accurate solutions.
• A neural net must start all over with every new application. It has to be retrained for
each specific problem undergoing analysis. A critical condition for using neural
nets in KM is the level of knowledge needed to apply the technology. Some tech-
nology requires seasoned knowledge from the user, while others assume a novice
user with no specific knowledge to contribute at any time. Another factor is
the time dimension expected before receiving a satisfactory solution from the
KM technology. For example, case-based reasoning requires a high level of user
knowledge, but the solution can be arrived at in a short time period. In contrast,
expert and rule-based systems require a low level of user knowledge, but it takes
much longer to arrive at a solution.
338 a a a a PART IV K M System Tools and Portals

A nother knowledge-based tool is the association rule technique, which generates a set
of rules to help understand relationships that may exist in data. The main types of asso-
ciations are as follows:
• Boolean rule: If a rule consists of examining the presence or the absence of items,
it is a Boolean rule. For example, if a customer buys a PC and a 17-inch monitor,
then he will buy a printer. The presence of items (computer and 17-inch monitor)
implies the presence of the printer on the customer’s buying list.
• Quantitative rule: In this rule, instead of considering the presence or absence of
items, we consider the quantitative values of items. For example, if a customer
earns between $30,000 and $50,000 and owns an apartm ent worth between
$250,000 and $500,000, he will buy a 4-door automobile.
• Multidimensional rule: This rule refers to several dimensions. For example, if a
customer buys a PC, then he will buy a browser. This rule is a single dimensional
rule, because it refers to a single attribute, that of buying. If a customer lives in
a big city and earns more than $35,000, then he will buy a cellular phone. This
rule involves three attributes: living, earning, and buying. Therefore, it is a multi-
dimensional rule.
• Multilevel association rule: A transaction can refer to items with various levels of
abstraction.
For example,
If a customer earns $50,000, he will buy a laptop.
If a customer earns $50,000, he will buy a computer.
These rules refer to different levels of abstraction. The first one is at the lower level
of abstraction, because buying a laptop also means buying a computer, and buying a
computer could also mean buying a laptop or PC, see Figure 11.3.
An association rule assumes a larger number of items, such as computers, moni-
tors, and printers, appearing together in a buying list or in a market basket. This con-
cept could be extended to text analysis. In this context, a market basket is the sentence,
and the items are the words making up the sentence. Similarly, one can consider the
document as the basket and the sentences forming the document as the items.

■ ■ ■ ■ ■ a FIGURE 11.3: Items Categorized According to Various Levels of Abstraction

Electronic Device First Level o f Abstraction

Computer Printer Monitor Second Level o f Abstraction

15 17
PC Laptop Jacket Laser Inches Inches Third Level o f Abstraction
 CHAPTER 11 Learning from Data aaaa 339
Association rules are statements of the form: When a customer buys a computer, in
70 percent of the cases, he or she will also buy a printer. This happens in 14 percent of
all purchases. This statem ent shows an association rule consisting of four elements:
Rule body: When a customer buys a computer
A confidence level: In 70 percent of the cases
A rule head: He will buy a printer.
A support: It happens in 14 percent of all purchases.
The rule body is the condition of the rule, and the rule head is the result.
Other terminologies exist to define a rule. The left-hand side (antecedent) may be
used to replace the rule body. The right-hand side (consequent) may be used to replace
the rule head. The confidence level is the ratio of the total number of transactions, with
the items of the rule body, to the num ber of transactions, with the items of the rule
body, plus the items in the result.
Consider the previous rule: A confidence level of 70 percent means that when buying
a computer occurs in a transaction, there is a 70 percent chance that the item, “printer,”
occurs at the same time. More formally, let us consider the rule: If A and B, then C. If this
rule has a confidence level of 30 percent, it means that when A and B occur in a trans-
action, there is a chance (likelihood) of 0.3 that C occurs in the same transaction.
The second measure of a rule is called support. The first rule states that it has a
support of 14 percent. This measures how often items in the rule body occur together
as a percentage of the total transactions.
Here is another example:
11,000 transactions
2.000 transactions contain a PC
500 transactions contain a laptop
1,500 transactions contain a printer
1.000 transactions contain a PC and a printer
150 transactions contain a laptop and a printer
PCs and printers appear together in 1,000 transactions out of 11,000. The support
of the rule, “When a customer buys a PC, he will buy also a printer,” is equal to 1,000/
11,000 = 0.09. Therefore, 9 percent is the support of this rule. The confidence of this
rule is 2,000/1,000 or 50 percent. It is equal to the num ber of transactions containing
PCs and printers divided by the num ber of transactions containing PCs. To find an
association rule linking computer rules to printer rules, we can say:
“A customer buying a computer in 57 percent of the cases will buy a printer. It hap-
pens in 1.5 percent of all the purchases.”
Here, we have considered a new type of item —namely, computers with a higher
level of abstraction than PCs and laptops.
The confidence is calculated as follows:
Number of transactions containing computers (PCs or laptops) and printers _
Number of transactions containing computers
1.000 + 150
2.000 + 500
The support is calculated as follows:
Number of transactions containing a computer and a printer _ 1,150 _ ^
Total number of transactions 11,000
340 a ■ i a PART IV K M System Tools and Portals

Sometimes the rules are stated in a different way, for example, when people buy
PCs, they also buy a printer half of the time. The confidence level is computed as fol-
lows: 1,100/2,000 = 0.5; the support is equal to 1,100/11,000 = 0.09.

MARKET BASKET ANALYSIS: PETCO

Founded in 1965 and incorporated in the state of Delaware, PETCO is a leading spe-
cialty retailer of premium pet food and supply items. PETCO stores carry more than
11,000 items, including premium foods, vitamins, toys, and other supplies for dogs, cats,
birds, fish, rabbits, reptiles, and other small companion animals. PETCO also provides
veterinary services, obedience training, and grooming services.

The Challenge
PETCO needed a way to make it possible for them to obtain company information
in a quick and accurate manner. They needed a store analysis to get an in-depth look at
the sales and margins section, and the base stores’ perform ance for the last 2-year
period. They also wanted to have access to inventory levels at all times, in order to
accommodate customers with the appropriate products and services. Identification of
store uniqueness was needed to allow for pricing and store assortm ent for proper
product placement and sale. A suitable solution needed to be put into practice in order
to properly target their m arket and track down customers to evaluate their specific
buying habits. Promotional, product, and vendor divisions also needed to be analyzed
to gain full interpretation of the business.

The Solution
Systech delivered an enterprise business intelligence solution, which aided ad hoc
detail merchandising to allow for reporting of data based on specific merchandising
units and also aided in overall store analysis across all reporting components. The pri-
mary platform and framework built were necessary for the data warehouse to function
completely.
A nother important component was the development of the market basket analy-
sis. PETCO wanted to be able to totally understand their customers’ buying habits and
to be able to cross-sell and up-sell to their customers. In order for the implementation
of the business intelligence solution to be successful, Systech had to carefully craft
seven crucial stages and understand each one to the fullest extent, to give PETCO the
solution they needed.

Results
Because of the business intelligence solution, PETC O now has the ability to
accomplish store and merchandising analysis. This results in better tracking of products
and services. The evaluation of vendor performances can also be obtained with the
new solution. This helps to analyze which vendors give better and faster service, allow-
ing for PETCO to have all the necessary products in stock to provide better customer
service. Prom otional analysis is now effectively generated, which helps PETCO
enhance their sales and improve their margins. Systech delivered an enterprise solu-
tion, which aided ad hoc detail merchandising. This is the reporting of data based on
specific m erchandising units. Inventory reports drawn out by the solution are effi-
ciently used to address key issues, such as the fill ratio. This effectively shows the quan-
tity sold in a particular store to quantity received from the distribution center. Some of
 CHAPTER 11 Learning from Data i a a a 341
the key reports, which are done by executives and are available through the data ware-
house, are active store comparisons, a store M erchandising Perform ance R eport,
(MPR), daily sales (showing the distribution of sales for a particular store based on
each sku, subclass, or departm ent by m onth, week, or day), the daily margin, and
departm ent sales comparisons. A m arket basket analysis was also developed to fully
understand customer-buying habits and to allow for cross-sell and up-sell. Targeted
marketing helped PETCO properly m arket their products to specific customers to give
them the full customer experience. Systech’s implementation of the business intelli-
gence solution will aid PETCO in delivering better analysis across all reporting com-
ponents (Systech 2002).

■■ Classification Trees
Classification trees are powerful and popular tools for classification and predic-
tion. The attractiveness of the classification trees m ethod is basically due to the fact
that, in contrast to NN, classification trees represent tools. Rules can be explained
so hum an beings can understand them or they can be created in a database access
language like SQL, so that records falling into a certain category can be retrieved
easily.
This point is crucial for situations where interaction is involved. For example, it is
equally im portant for the loan officer and the bank customer to understand the reason
behind denying a credit application. For the banker, it is im portant to find a reasonable
argument, even if it is generated by a computer, to justify the decision.

DEFINITION
The concept of “tree” is derived from graph theory. A tree is a network of nodes con-
nected by areas, sometimes called branches, so that there are no loops in the networks.
In general, there is a root node that is considered to be the starting node of the tree. The
ending nodes of the tree are called the leaf nodes. The root and leaf nodes are separated
by a certain num ber of interm ediate node organizations in layers, sometimes called lev-
els. Figure 11.4 shows a tree with two interm ediate levels.

FIGURE 11.4: A Binary TVee with TVo Levels

Leaf Node
i a ■ fl PART IV K M System Tools and Portals

W hen two nodes branch out from each node, this tree is called a binary tree.
Adding concepts to the nodes and to the branches transforms the tree from a theoreti-
cal network to a decision-learning tool. The decision tree is a tree where nodes repre-
sent actions and areas represent events that have a certain probability to occur.

DECISION FOR GRANTING A LOAN

Let us consider the following rule:

If the savings account > $a and the assets owned > $/?, then grant the loan;
otherwise, do not grant the loan.

The problem lies in how to reach such a rule, in other words, how to construct such
a tree. The bank has to start by collecting data about previous customers, relying on its
historical data stored in a customer database. In the historical database, the bank has
the needed information about previous customers both who failed to honor their debt
and those who were able to honor their debt.
In such a simple case, the two classes of customers are clearly separated. In addi-
tion to the rule governing the classification, it is easy to induce what would happen in
more complex situations; a systematic technique for tree construction is needed.

TREE CONSTRUCTION
Many elements should be considered when constructing a classification tree. At each
level, nodes will split data into groups until they reach the ending node, where the
classes are identified. The first question is how to assign attributes to nodes. More
specifically, if we consider the root node, one of the attributes has to be assigned to this
node. For example, we added the attribute, “savings account,” to the root node in the
bank loan cases. We could have selected the attribute, “assets owned,” for that node.
Assigning an attribute to a node that will split the data is the first task to be achieved
by classification tree algorithms. The second task is to decide when a classification tree
is complete. The third task is to assign classes to the ending nodes. Finally, the fourth
task is to convert the classification tree into a set of rules.

Implications for Knowledge Management

Neural networks represent a wide range of applications in knowledge management,
knowledge autom ation systems, and business in general. They enter as stand-alone
applications or they can be integrated with knowledge automation systems. Despite
obvious successes, they are not a cure-all. They involve several considerations worth
discussing: cost/benefit analysis, the nature of neural learning and artificial intelligence,
and maintenance.
The benefits of neural networks include improving speed, quality, and “good
enough” judgm ent in the areas in which human judgm ent might deteriorate due to
fatigue, lack of motivation, or problems of recall. The benefits are both tangible and
intangible. For example, if a neural network designed to evaluate paint quality and fin-
ish of a new car is to replace the human expert inspector, it will likely return tangible
financial savings. Likewise, a neural network that “decides” w hether a credit card
transaction is a reasonable purchase based on the credit card holder’s buying habits
can save the issuing bank millions of dollars each year by detecting fraud. Of course,
 CHAPTER 11 Learning from Data ■■«■ 343
intangible benefits include better corporate image when quality and reliability of the
product resulting from a neural network contribute to customer satisfaction. Indirectly,
it can also contribute to image building in the long run.
The other side is the cost factor. The question is how much is a corporation willing
to spend to gear up the technology? Tangible costs involve user training, hardware-
software, backup, support, and maintenance. Intangible costs are user resistance and
the time it takes to sell change. This is especially if neural networks are trained as arti-
ficial learners trying to compete with long-term specialists or experts in the area cov-
ered by the neural network. M anagement must be prepared for these problems and
take the appropriate steps to alleviate them.
A nother issue relates to the concept of neural networks as adaptive learning sys-
tems. Given the iterative process of training a neural network, is artificial learning a
“good enough” replacement for humans who learn “on the fly” by using intelligence
and common sense? Human decision making and problem-solving are also based on
their perceptions, attitudes, and beliefs in addition to methodology, use of tools, and the
following of procedures. Hum an cognitions are modified by experience, which con-
tributes to a m aturation that artificial learning lacks.
Perhaps the best function a neural network can serve is to advise, rather than replace,
humans in problem-solving. The fact that humans can still veto a neural network’s deci-
sions attests to the higher-level of learning capacity and intelligence of humans.
Finally, organizations must address the issue of maintenance. For a neural net-
work to stay up to date in a changing environment, it has to be retrained, which means
modifications in the training set, retraining, and ongoing evaluation of the neural net-
work. A firm that relies heavily on applications supported by this technology would
need to consider maintenance and enhancement as part of the total commitment to
quality assurance.
The level and frequency of maintenance often reflect the adequacy of the initial
design of the neural network. Like conventional systems, a poorly thought-out prob-
lem definition or scope requires greater and more frequent changes in the system than
a well-designed system. This suggests that, for a neural network to succeed, proper
attention should be given to the early phases of the developm ent life cycle. This
includes problem definition, problem justification, and gathering the right training data
for the job. Effective management of the front end of the development life cycle could
minimize unnecessary maintenance.

Su m m a r y a 1 a ■
• Learning tools are divided into two groups: theory-driven and data-driven.
Theory-driven learning, often called hypothesis testing , attempts to substantiate or
disprove preconceived ideas. Data-driven learning tools automatically create the
model based on patterns found in the data.
• There are several types of associations:
• Boolean rule. If the rule considered consists of examining the presence or
the absence of items, it is a Boolean rule.
• Q uantitative rule. In these rules, instead of considering the presence or
absence of items, we consider the quantitative values of items.
• Multidimensional rule. These are rules referring to several dimensions.
• M ultilevel association rule. A transaction can refer to items with various
levels of abstraction.
344 ■ ■ ■ ■ PART IV K M System Tools and Portals

• A neural network is a knowledge-based tool of electronically interconnected

basic processing elements called neurons. Neural networks are ideal for solving
cumbersome problems that traditional information systems have difficulty track-
ing, but they are terrible at problems that computers can do well.
• Neural networks can be contrasted with knowledge automation systems as follows:
• Knowledge automation systems learn by rules, and neural nets learn by
example.
• Knowledge automation systems are sequential in nature and explain their
decisions, where neural nets do not have an explanatory facility.
• Neural nets continue learning as the problem environment changes, and
knowledge automation systems depend on complete data before offering
a final solution.
• A neural network is more effective than knowledge automation systems
at developing a knowledge base with experimental data.
• A neural network consists of neurons, connected by axons that receive inputs
from other neurons. Communication between neurons is carried out by these
input-output pathways. A neuron evaluates the inputs, determines their weights,
sums the weighted input, and compares the total to a threshold. If the sum is
greater than the threshold, the neuron fires. Otherwise, it generates no signal.
• Learning occurs when the network changes weights and modifies its activity in
response to its inputs. Making weight adjustments by backing up from the output
is called back propagation. Learning may be supervised or unsupervised.
• Classification trees are powerful and popular tools for classification and predic-
tion. The attractiveness of the classification trees method is basically due to the
fact that in contrast to NN, classification trees represent tools.
• Many elements should be considered when constructing a classification tree. At
each level, nodes will split data into groups until the ending node is reached,
where classes are identified.

Terms t o K n o w * • * 1
Antecedent: When an association between two variables Neural network: A knowledge-based technology modeled
is defined, the first item (or left-hand side) is called the after the human brain’s network of electrically inter-
antecedent. connected processing elem ents called neurons; a self-
Associations: An association algorithm creates rules that programming system that creates a model based on its
describe how often events have occurred together. For inputs and outputs.
example, “When prospectors buy picks, they also buy Supervised learning: A learning strategy that specifies the
shovels 14 percent of the time.” Such relationships are correct output for a certain input.
typically expressed with a confidence interval. Support: The measure of how often the collection of
Classification tree: A decision tree that places categorical items in an association occurs together as a percentage
variables into classes. of all the transactions. For example, “In 2 percent of the
Confidence: Confidence of rule, “B given A,” is a measure of purchases at the hardware store, both a pick and a
how much more likely it is that B occurs when A has oc- shovel were bought.”
curred. A 95 percent confidence interval for the mean has Transfer function: A function that transforms the neuron
a probability of .95 of covering the true value of the mean. input into an output value.
Consequent: When an association between two variables Unsupervised learning: A learning strategy in which the
is defined, the second item (or right-hand side) is called correct output is not specified.
the consequent. Weight: An adjustable value associated with a connection
Input: Stimulation level in a neural network. between neurons or nodes in a network.
 CHAPTER 11 Learning from Data ■■■■ 345

Te s t Yo u r U n d e r s t a n d in g Mi , <
1. Define a neural network in your own words. W hat does the technology
attem pt to do?
2. “The interesting aspect of a neural net is its known contributions in solving
cum bersome problem s that traditional com puters have found difficult to
tra c k ” Do you agree? Why or why not?
3. Explain how a neural network functions. Give an example of your own.
4. In your opinion, how severe are the limitations of a neural network?
5. How are neural networks different from knowledge automation systems?
6. Explain in some detail how neural networks learn.
7. How are inputs and outputs used to contribute to a solution?

Kn o w l edge Ex e r c is e s 1 , 1 1
1. Search on the Internet or in related literature and write an essay detailing a
neural network application in business. What did you learn from this exercise?
2. Suppose you have a small kiosk database of purchased items.
The available food items in the kiosk are as follows:
• Coca-Cola
• Pepsi-Cola
• Sprite
• Budweiser beer
• Guinness beer
• Estrella chips
• Pringles chips
• Taffel chips
The database contains the following purchase transactions:

Item Item s Bought

1 Coca-Cola, Budweiser beer, Pringles chips
2 Coca-Cola, Taffel chips
3 Budweiser beer, Pringles chips
4 Pepsi-Cola, Budweiser beer, Guinness beer, Estrella chips
5 Sprite, Estrella chips
6 Pepsi-Cola, Budweiser beer, Estrella chips

Item Item s Bought

7 Sprite
8 Budweiser beer, Guinness beer, Estrella chips
9 Pepsi-Cola, Estrella chips
10 Coca-Cola, Pringles chips

What kind of rules do you get with a confidence threshold of 0.0 and a sup-
port threshold of 0.2?
Create a hierarchy for the food items in the kiosk, and try to determine if
you could get more meaningful information by using multilevel association rules.
Is it possible to set different support thresholds on different hierarchy levels?
Source : Han and Kamber 2000
3. Using a realistic example, discuss in detail how neural networks can be used
to solve real-life problems.
346 a « a a PART IV K M System Tools and Portals

4. An insurance company is planning to launch the sale of a newly developed

insurance product. As a promotional exercise, the company wishes to identify
several potential customers from the local area and send them details of this
new product, together with special discount offers. In the past, the company
has conducted similar promotional exercises, and the results were not good.
The promotion cost was high due to an extremely low response rate. Very
few people who responded to the promotions actually purchased the prod-
ucts. This time, the company aims to reduce prom otion cost and have an
increase in response rate. The company possesses about 2,000 historical
records on the previous promotions. The data set stores details of the
prospective customers, such as age groups, income bands, occupations, and so
forth. The records can be categorized into “response with purchase,”
“response without purchase,” or “no response.”
If the company wants to use a classification technique to identify the
people who are likely to purchase the new product, describe the life cycle of
such a data-mining project.
5. Review the neural network literature and cite one application of supervised
learning and another of self-supervised learning. What do you conclude is the
difference between the two forms of learning?
6. Identify whether each of the following applications is a candidate for knowl-
edge automation or a neural network:
a. Prediction of weather conditions,
b. Diagnosis of a diabetic condition by a specialist,
c. Verification of check signature, and
d. Recognition of characters on an invoice.
7. You have a set of six alphabetic characters: A, S, T, U, T, and E. Like a
Scrabble game, your job is to generate as many words as possible from these
characters. If you were to do it manually, you would have to think of a word,
verify it with the dictionary, and move on to the next word. How would a con-
ventional computer go through the same process? Is this kind of job a candi-
date for neural network?
8. If the company chooses to use the decision tree induction technique to build
a classification model, describe the main activities that need to be performed
during the project development.

Re f e r e n c e s 1 1 1 1
Awad, E. M. Building Expert Systems. Minneapolis, MN: and Applications. New York: John Wiley & Sons,
West Publishing, 1996, p. 437. Inc., 1998.
Cherkassky,Vladimir, and Mulier, Filip. Learning from Mitchell, Tom M. Machine Learning. New York: McGraw
Data: Concepts, Theory and Methods. New York: John Hill, Inc., 1997.
Wiley & Sons, Inc., 1998. Smith, Reid G., and Farquhar, Adam. “The Road Ahead
Han, J., and Kamber, M. Data Mining: Concepts and for Knowledge Management: A n AI Perspective,” A l
Techniques, section 6.3. N ew York: Morgan Kaufmann M agazine, vol. 21, no. 4, Winter 2000, pp. 17^40.
Publishers, 2000. Systech, PETCO case study, www.systechusa.com/
Michalski, Ryszard S., Bratko, Ivan, and Kubat, Miroslav. clients/pop_pet.htm, Date accessed D ecem ber 2002.
Machine Learning and Data Mining: M ethods

W e b S it e s X 1 1 1
www.dsi.unifi.it/neural This site covers recent topics in psychology, neurobiology, mathematics, physics, com-
neural networks and machine learning. puter science, and engineering.
www.inns.org/nn.html Neural Networks is an international www.ncaf.co.uk The Natural Computing Applications
journal that serves as a central, interdisciplinary publi- Forum runs meetings (with attendees from industry, com-
cation. Its editors represent a range of fields including merce, and academia) on applications of neural networks.
 CHAPTER 11 Learning from Data 1 i i « 347
www.ewh.ieee.org/tc/nnc/index.html The IEEE Neural graphics, neuroscience, artificial intelligence, and
Networks Society advances and coordinates work in many other areas.
neural networks. It is exclusively scientific, literary, and wwwl.ics.uci.edu/~mlearn/MLRepository.html This is a
educational in character. repository of databases, domain theories, and data gen-
www.ime.usp.br/-cesar/revision/neural.htm This erators that are used by the machine learning commu-
source covers a series of issues related to vision, in- nity for the empirical analysis of machine learning
cluding computer vision, image processing, computer algorithms.
 Data M ining-
Knowing the
Unknown
Contents

In a Nutshell
What Is Data Mining?
Definitions
Data Mining and Business Intelligence
Business Drivers
Technical Drivers
Role of Statistics
Machine Learning
Data Warehouses
OLAP
Evolution of the Decision-Making Architecture
DM Virtuous Cycle
Business Understanding
Develop the DM Application
Data Management
D ata Sources
Taxonomy of Data
Data Preparation
Model Building
Param eter Settings and Tuning
Model Testing and Analysis of Results
Taking Action and Deployment
Postdeployment Phase
DM in Practice
Role of DM in Customer Relationship Management
Customer Acquisition
Campaign Optimization
Customer Scoring
 CHAPTER 12 Data M ining—Knowing the Unknown a i a i 349
Direct Marketing
Integrating DM, CRM, and E-Business
Implications for Knowledge M anagement
Summary
Terms to Know
Test Your Understanding
Knowledge Exercises
References
Selected U RL Listings

When knowledge stops evolving,

it turns into opinion or dogma.
—Th oma s D a v en po r t
a n d La u r en ce Pr u sa k

In a Nutshell
In the previous chapters, we discussed knowledge management and building a KM sys-
tem from capturing tacit knowledge to codifying, testing, transferring, and sharing it. One
important issue discussed in Chapter 4 was how to create knowledge. Our perspective is
to look at explicit and tacit knowledge as the two primary types of knowledge. Tacit
knowledge is created and captured through teamwork and from one’s experience.
Explicit knowledge is knowledge codified in books, documents, reports, and the like.
D ata mining (DM) is becoming a fundamental component of the global business
infrastructure that assists firms in the decision-making process and helps them capture the
multifaceted aspects of the new economy. In this chapter, we consider data as a source of
undiscovered knowledge. We discuss the process of extracting knowledge from data and
information stored in databases, data warehouses, and other repositories. This process is
called data mining. We also discuss the role of DM in the decision-making process.
The challenge for KM is to find a way to exploit huge amounts of transactional data
that might add value and intelligence to decision making and operational busi-
ness processes. It is estimated that the amount of information in the world doubles every
20 months. Clearly, little can be done to take advantage of this flood of raw data, unless
an autom ated approach is elaborated. This is where DM provides a viable source of
knowledge discovery and an efficient way to deliver new added value for business. From
this perspective, DM is an essential component of the corporate knowledge architecture.

What Is Data Mining?

Gaining insight to improve business functions remains a major concern of any man-
ager or decision maker. The past 2 decades have seen a wide diffusion of information
technologies in all business activities. The intensive use of new electronic devices such
as point of sale, remote-sensing devices, ATMs (autom ated teller machines), and the
Internet has contributed to the explosion of available data. It is similar to the gold rush:
A new source of wealth was available to all kinds of miners. Suddenly, managers and
decision makers became conscious of this wealth of data. It was natural to think about
the best way to benefit from it. This is where data mining enters (see Box 12.1).
From a business point of view, DM has gained appeal; as a field, it has yet to
reach maturity. The many definitions of data mining show lack of consensus among the
350 i « « ■ PART IV K M System Tools and Portals

1 i I 1 1 1 BOX 12.1 1 1 1 1 1 i

SAMPLE USES OF DATA MINING

DATA MINING IN THE FEDERAL AGENCIES it to scour plane crash data to find common
causes so future failures can be avoided.
From predicting how many Marines will leave the
corps to rooting out fraudulent health-care bills MAKING PREDICTIONS
for the H ealth Care Financing A dm inistration,
data mining is becoming a popular technological Data mining is more than slicing and dicing data; it
tool for agency m anagers trying to make sense is doing predictive work as well. The Marine Corps
and b etter use of mounds of governm ent data. is using data mining to predict which types of offi-
D ata mining is accomplished with commercial cers and enlisted members will stay in the corps
off-the-shelf software applications that use so- and which will bail out. “We want to use historical
phisticated statistical analysis and advanced mod- data to predict our loss rates in specific areas,”
eling to turn volumes of data into usable informa- says M ajor Joe Van Steenbergen, a m anpower
tion. By uncovering subtle trends or patterns in manager for the Marine Corps. By mining a data
data, the application can enable users to draw warehouse containing career and biographical
conclusions or make predictions. “D ata mining is information for every officer and enlisted member
getting the ‘aha’ out of data,” says Adam Crafton, in the past decade, the Marine Corps expects to
a data-m ining m anager for IB M ’s public-sector find answers to questions such as “Are married or
operations in Bethesda, Maryland. single Marines more or less likely to leave the ser-
Commercial retailers first developed data- vice?” “A re m embers in high-skill career fields
mining software nearly 30 years ago to track con- like information technology more likely to leave
sumer buying habits. By combining those early than members with jobs that require less training?”
models with advances in artificial intelligence over By answering those questions and many oth-
the past decade, data-m ining firms have moved ers, the Marine Corps can create a profile of those
beyond helping retailers to assisting many busi- likely to stay in various positions and use it to
nesses. D ata mining’s uses now range from pre- make better management decisions when recruit-
dicting production breakdowns for large manufac- ing, assigning, and promoting personnel.
turers to helping financial institutions uncover
IMPROVING SERVICE
patterns of money laundering.
Many agencies are already using data mining The Centers for Disease Control and Prevention’s
to m ake better m anagem ent decisions and im- National Im m unization Program in A tlanta is
prove services. The Justice D epartm ent has used installing data-mining software that allows better
it to find crime patterns so it can focus its money tracking of reactions to vaccines. The program has
and resources on the most pressing issues. The a huge database of adverse reactions to vaccines
Veterans Affairs D epartm ent has used it to pre- reported by physicians, clinics and hospitals, pa-
dict demographic changes among its 3.6 million tients, and pharmaceutical companies across the
patients so it can prepare more accurate budgets. nation. Federal researchers and statisticians moni-
The Internal Revenue Service uses the technol- tor the data regularly to find problems caused by a
ogy at its custom er service center to track calls single vaccine or vaccine combinations. Data min-
in order to pinpoint the most common customer ing has helped the agency recover millions of dol-
needs. The Federal Aviation Adm inistration uses lars in fraud cases.

SOURCE: Excerpted from Cahlink, George. “Data Mining Taps the Trends,” www.govexec.com/features/1000/1000managetech.htm,
September 18,2000, Date accessed October 16,2002.
 CHAPTER 12 Data M ining—Knowing the Unknown i ■i i 351
business community. Even the term data mining is subject to controversy. Other than
data mining , the literature uses knowledge discovery from databases (KDD), informa-
tion discovery, information harvesting, data archeology, and data pattern processing.
The term K D D appeared in 1989 to refer to extracting knowledge from databases. It
has mostly been used by artificial intelligence and machine learning researchers. In
contrast, statisticians, data analysts, and the MIS community have used the term DM.
There is actually no difference between the two, although DM is not restricted to the
mechanism of discovering knowledge (Fayyad et al. 1996).

DEFINITIONS
DM definitions are derived from the interaction of three entities (Figure 12.1):
1. A body of scientific knowledge accumulated through decades of forming well-
established disciplines, such as statistics, machine learning, and artificial intelligence
2. A technology evolving from high volume transaction systems, data warehouses,
and the Internet
3. A business community forced by an intensive competitive environment to innovate
and integrate new ideas, concepts, and tools to improve operations and DM quality
A review of the literature reveals a wide variety of definitions:
• The search for relationships and global patterns that exist in large databases but
are hidden among the vast amount of data (Holsheimer and Kersyen 1994)
• A set of techniques used in an automated approach to exhaustively explore and
bring to the surface complex relationships in very large data sets (Moxon 1996)
• The process of finding previously unknown and potentially interesting patterns
and relations in large databases (Fayyad et al. 1996)

Data Mining and Business Intelligence

D a ta mining is producing knowledge and discovering new patterns to describe the
data. DM is also predicting future values and business behavior. Sending queries to the
transactional databases used to be the only way to transform data into information.
The m ajor lim itation of this approach was the incapability of extracting meaning.
“Mining” the data was relegated to the human expert. With the increasing volume of
data and the shrinking time for timely decisions, intelligent autom ated systems such as
DM became a necessity to mine the data. Box 12.2 illustrates how data mining and KM
can be integrated.

Scientific Business
Knowledge Community
i .a i a ■ ■
Information
Technology
FIGURE 12.1: DM and the
Three Bodies of Knowledge
352 «i a i PART IV K M System Tools and Portals

BOX 12.2

THE INTEGRATION OF DM AND KM

Knowledge management and data mining devel- edge is extracted. Machine learning emerged pre-
oped independently of each other, and their rela- cisely as a way of alleviating difficulties raised by
tionship has not yet been fully recognized, much knowledge elicitation from humans. In addition,
less exploited. Born of the understanding that the exponential growth of process-generated data
knowledge is one —if not the most —im portant has spurred the development for more scalable—
asset of an organization, both knowledge manage- hence, more thoroughly autom ated—ways of gen-
m ent (KM) and data mining (DM ) grew and erating useful knowledge from data.
flourished at the convergence of information tech- A second issue is knowledge refinement and
nologies (machine learning, knowledge-based sys- revision. The typical KM approach is again manual:
tems, databases), statistics and data analysis, and The knowledge engineer readjusts domain ontolo-
the business and management sciences. gies, rewrites rules in collaboration with domain
KM and DM embody distinct if not opposite experts, and so forth. Although DM research has
perspectives on different knowledge-related issues. yielded a promising harvest of automated tech-
One is knowledge capture in the broadest sense niques for knowledge revision and theory refine-
of the term. The KM community has traditionally ment, these have been demonstrated on highly cir-
focused on knowledge acquisition from humans, cumscribed domain theories and have yet to be
either directly (through manual ontology con- validated on medium and large-scale applications.
struction, expert interviews, and authoring tools) or What is needed, then, is an integrative approach
indirectly, as when human know-how is mimicked that exploits synergies between knowledge man-
by a program that observes a hum an expert in agement and data mining in order to monitor and
action (such as learning apprentices or program- manage the full life cycle of knowledge—its capture
ming by demonstration). and discovery, representation, storage, retrieval,
In data mining, databases and data ware- revision or refinement, and reuse—in an organiza-
houses are the ultimate sources from which knowl- tion or community of practice.

SOURCE: First Conference on Integrating Data Mining and Knowledge Management, Geneva 2002.

Business intelligence (B I) is a global term for all processes, techniques, and tools
that support business decision making based on inform ation technology. The
approaches can range from a simple spreadsheet to an advanced decision support sys-
tem. Data mining is a component of BI. Figure 12.2 shows the positioning of different
BI technologies used at different levels of m anagement and for different purposes,
including tactical, operational, and strategic decisions.

■■■» Business Drivers

There are many reasons why data mining has grown substantially in the last few years.
• Competition. Successfully competing in today’s economy requires an understand-
ing of customer needs and behavior, and a great flexibility to respond to market
demands and competitors’ challenges. To achieve this, data must be transformed
 CHAPTER 12 Data M ining—Knowing the Unknown i i «« 353

Decision Maker

Business Analyst

Managers

Data Architects

Database Administrator

a 9 a fl i fl FIGURE 12.2: Data Mining and Business Intelligence

into knowledge that is leveraged in the business processes. To survive, companies

must find quick answers to questions such as “Who is likely to remain a loyal cus-
tomer, and who is likely to churn?” “What service makes the most profitable cus-
tomers happy?” “What products should be prom oted to which prospects?”
Finding the right answers to these questions will differentiate a business from its
competitors.
• Information glut. Most companies have a glut of information. In fact, many are
drowning in information. Today’s manager is confronted with increasingly large
volumes of data, which has been collected and stored in various databases and
constitutes a challenge for decision making. The proliferation of data is a feature
of the e-economy. To illustrate, a customer places an order on the phone to buy
some items. Such a seemingly simple transaction involves a large number of trans-
actions in various systems at different companies:
• The merchant knows the items ordered, the shipment method, the credit card
used, and so on.
• The telephone company knows who called, from what location, for how
long, and so on.
• The credit card company knows the amount charged, the approval number,
the vendor code, and the available credit.
The information systems of the three companies must exchange information,
update records, and store data to perform each transaction properly. Each com-
pany has the opportunity to learn from the collected data. The consumer goods
company can apply DM to improve its sales process. The telephone company can
build churn models to focus on existing customers and select a promotional strat-
egy. The credit card company can leverage its vast warehouse of customer trans-
action data to identify customers most likely to be interested in a new product.
• The need to serve the knowledge workers efficiently. The databases provide the
firm with transactional memory. However, memory is of little use without intelli-
gence. Intelligence is the capacity to acquire and apply knowledge. DM incorpo-
rates intelligence in its techniques, but DM is worthless if it is not used to make
354 a ■ a a PART IV K M System Tools and Portals

better decisions. Every organization has different types of knowledge workers.

These workers include the following:
• Some knowledge analysts are researchers or experts. They demand sophisti-
cated tools for their investigations and research analysis.
• Some knowledge users are executives and decision makers. They need tools
that help them review and analyze data. They rely on easy-to-use and power-
ful systems.
• Other knowledge consumers might be account managers, customer service
representatives, or members of the sales force. They require easy to access
information.
• DM can provide the right environment to satisfy the requirements of all
these people.

■■■■ Technical Drivers

New developm ents in analytical m ethods provide advanced learning capabilities.
Answering the question “W hy?” means identifying the factors that cause a certain
result. The manager can use his know-how and expertise in making decisions and draw-
ing sound conclusions. Analytical methods allow managers to draw hypotheses and test
them, predict events and adapt decisions accordingly, find unexpected patterns and
exploit them, and discover hidden associations and undertake appropriate actions.
Without analytical methods, there is no effective analysis. Without analysis, there is no
business intelligence. Without business intelligence, there is no hope to assimilate giga-
bytes of data and consistently make decisions that keep one’s business ahead of the
competition. The objective of DM, then, is to optimize the use of available data and
reduce the risk of making wrong decisions. Statistics and machine learning are consid-
ered to be the analytical foundations upon which DM was developed.

ROLE OF STATISTICS
Statistics embraces concepts, such as hypothesis testing and correlation. The statistician
has to start her analysis by assuming a hypothesis about the relationships among the
data attributes. With databases organizing data records of hundreds of attributes, the
statistics methodology of the hypothesize-and-test paradigm becomes a time-consuming
process. DM helps by automating the formulation of new hypotheses.

MACHINE LEARNING
As a subfield of Al, machine learning (ML) has focused on making computers learn
things for themselves. Over time, attention has focused on the technological applica-
tions of machine learning. Machine learning is the automation of the learning process
that is a crucial function in any intelligent system. Its methodology includes learning
from examples, reinforcem ent learning, and supervised or unsupervised learning.
Machine learning is a scientific discipline considered to be a subfield of artificial intel-
ligence. In contrast, data mining is a business process concerned with finding under-
standable knowledge from very large real-world databases.

DATA WAREHOUSES
For years, companies recognized that the volume of data generated by operations such
as online transaction processing (OLTP) systems, point of sale (POS) systems, auto-
mated teller machine (ATM) systems, and the Internet could be a volatile asset. The
 CHAPTER 12 Data M ining—Knowing the Unknown m■ ■ ■ 355
challenge is to make data available anytime and anywhere for the decision-making
process. For example, Midwest Card Services (MCS) is a rapidly growing financial
products services company, with more than 1,000 employees and 200,000 clients. Very
quickly, MCS developed the technology infrastructure to m eet business needs and pro-
vide clients with real-time information. Unfortunately, the company had the technol-
ogy in place, but lacked the integration of a multitude of stand-alone systems. They had
a lot of data available but could not cross-reference among systems. Bringing together
data from disparate transactional systems and providing consistent decision-making
capabilities across the various functional systems was a real challenge.
As can be seen, data warehousing (DW) is extracting and transform ing opera-
tional data into informational or analytical data and loading it into a central data ware-
house. The major features of DW are as follows:

1. Subject oriented. D ata warehouses are organized around subjects such as cus-
tom ers’ vendor products. In contrast, operational systems are organized around
business functions such as loans, accounts, and transactions.
2. Integration. D ata loaded from different operational systems may be inconsistent.
The date in one application may be “yym m dd” and “m m ddyy” in another.
Integration shows up in consistent naming of variables, in some measurements of
variables, and so forth.
3. Time variant. Data may be stored but not updated for several years. These are
considered historical data and used for prediction, forecasting, and trend analy-
sis purposes.
4. Nonvolatile. There are two kinds of operations perform ed on data warehouses:
the initial loading and the access of data. In the operational environment, insert-
ing, deleting, and changing are done regularly. These DW features lead to an envi-
ronm ent that is very different from the operational environment. Hence, there is
no redundancy between operational data houses and data warehouses.

OLAP
The need for nonstatic reporting systems has led to the development of online analyti-
cal processing (OLAP). D ata is now used in a proactive way that provides value for the
firm. This approach uses computing power and graphical interfaces to manipulate data
easily and quickly at the convenience of the user. The focus is showing data along sev-
eral dimensions. The manager should be able to drill down into the ultimate detail of a
transaction and zoom up for a general view.
OLAP has several strengths:

• It is a powerful visualization tool.

• It is an easy-to-use interactive tool.
• It can be used as a first step in understanding the data.

It has also some limitations:

• It does not find patterns automatically.

• It does not have powerful analytical techniques.

These lim itations are actually the main strengths of DM. It appears from this
analysis that data warehousing, OLAP, DM, and operational systems can be integrated
in a single framework to serve the needs of knowledge workers. This model is shown in
Figure 12.3.
356 a a ■ « PART IV K M System Tools and Portals

a fl ■ a I fl FIGURE 12.3: Data Warehouse and DM Technological Framework

EVOLUTION OF THE DECISION-MAKING ARCHITECTURE

DM is a com ponent of the business intelligence architecture. Business intelligence
makes it possible to take action that will ultim ately result in m easurable benefits.
Traditionally, it includes querying and reporting, online analytical processing (OLAP),
data visualization, and data analysis. DM com plem ents these tools and provides a
global approach that integrates the conventional tools in a whole process that leads to
new insights and actionable knowledge for managers and decision makers. It allows
users to access information from different sources through client/server or Web-based
query systems to visualize models and interact with the results of those queries to gain
new business knowledge. Figure 12.4 shows the traditional approach for the decision-
making process in contrast with the new approach, which is shown in Figure 12.5.
The integration of decision processing into the overall business process is achieved
by building a closed-loop system where the output of decision processing applications
is delivered to business users in the form of recom m ended actions (product pricing
changes, for example) for addressing business issues. In the e-business environment,
many companies are looking to extend this closed-loop processing to automatically
adjust business operations based on messages generated by a decision engine.

■ DM Virtuous Cycle
The goal of DM is to allow a corporation to improve its internal operations and its
external relationships with customers, suppliers, and other entities. H ie virtuous cycle
of DM is about harnessing the power of data and transforming it into added value for
the entire organization. It incorporates DM into the context of corporate business
processes and provides the following capabilities:
• Response to extracted patterns
• Selection of the right action
 CHAPTER 12 Data M ining—Knowing the Unknown aaaa 357

Reporting

a fl a fl 9 ■ FIGURE 12.4: Conventional Approach to Corporate Decision Making

• Learning from past actions

• Turning action into business value

Although the mechanism for extracting patterns from data is im portant, DM is

more than a set of intelligent techniques. What is more im portant is to apply DM in the
right area and deliver the right data to the right people at the right time. The virtuous
cycle incorporates DM into other business processes. Its main focus is to extract maxi-
mum benefit from data and to gain better understanding of customers, suppliers, and
the market. It helps transform an organization from reactive to proactive.
The virtuous cycle is formed by the four steps shown in Figure 12.6.

BUSINESS UNDERSTANDING
The first step in the virtuous DM cycle is identifying the business opportunity, which
means defining the problem s faced by the firm. The goal is to identify the areas
where data can provide values. This stage is the most crucial and the most delicate
stage in conducting a DM project. Finding interesting answers for a meaningless busi-
ness problem is of no use to the company. That is why defining the problem should
involve the technical people and the business experts. The DM project m anager is
expected to possess a combination of skills, including a good command of technical and
business cultures. For example, one of the insurance industry’s m ajor challenges is
fraud, which tends to take different form s—claim fraud, premium fraud, and indemnity
fraud. O ther im portant challenges are customer retention or loyalty enhancement pro-
grams. Table 12.1 lists business challenges facing different industries and the corre-
sponding DM goals. Identifying these goals is actually the outcom e of the business
understanding phase.
358 a ■ a ■ PART IV K M System Tools and Portals

Front-Office CRM
Application

Real-Time
Data Feeds

E-Business
Applications

Back-Office ERP,
SCM Applications

.pptications

Recommendations

FIGURE 12.5: DM in the Context of the New Corporate Business Model

DEVELOP THE DM APPLICATION

Once the business understanding phase is completed, the next step is to develop the
application. The goal is to extract knowledge from data and make the mechanism of
discovering a new pattern operational. This stage includes four activities:

• Define the adequate data-mining tasks.

• Organize data for analysis.
• Use the right DM technique to build the data model.
• Validate the model.
 CHAPTER 12 Data M ining—Knowing the Unknown aaaa 359

fl fl fl fl fl fl FIGURE 12.6: The Virtuous DM Cycle

The development process begins with identification of the outcomes expected from
the DM application. The outcomes are called DM tasks and consist of the following:

• Clustering. Clustering is a tool that divides a database into different groups. The
tool finds groups that are very different from each other, but whose members are
similar to each other. Unlike classification, we do not know what the clusters will
be when we start or by which attributes the data will be clustered. Consequently,
someone knowledgeable in the business must interpret the clusters. After the
clusters that reasonably segment the database are found, they may be used to
classify new data.
• Classification. The classification function identifies the characteristics of the
group to which each case belongs. This pattern helps us understand the existing
data and predict how new instances will behave. For example, we may want to
predict whether individuals can be classified as likely to respond to a direct mail
solicitation, as vulnerable to switching over to a competing long-distance phone
service, or to be a good candidate for a surgical procedure.
• A ffinity grouping. Affinity grouping is a descriptive approach to exploring data
that can help identify relationships among values in a database. The two most
common approaches to affinity grouping are association discovery and sequence
discovery. Association discovery finds rules about items that appear together in
an event, such as a purchase transaction. M arket-basket analysis is a well-known
example of association discovery. Sequence discovery is similar to association dis-
covery, in that a sequence is an association related over time.

Once the expected outcomes are identified, the next step is to collect and gather
the data needed to start the analysis and match the results to the expected outcomes
before releasing the results for decision making.
360 a a i i PART IV K M System Tools and Portals

TABLE 12.1 Business Challenges Faced by Different Industries and the DM Goals
Industry Business Challenge DM Goals
Customer services The main challenge is understanding Customer acquisition profile
that the individual preferences of
the customers is the key to Customer retention profiling
satisfying them. What one customer
may consider as being “attentive” Customer-centric selling
customer service, another may Inquiry routing
deem “oppressive ” That is why it is Online shopping
extremely crucial to understand Scenario notification
these differences and know how to Staffing level prediction
address them in the proper and Targeting market
suitable way. Web mining for prospects

Financial services Retaining customer loyalty is of the The financial services industry
industry utmost importance to this industry, is a leader in applying DM
and it is currently the key business applications. Very focused
challenge. statistical and DM appli-
cations are prevalent. Risk
management for all types
o f credit and fraud detec-
tion are among the DM
applications in financial
services companies.

Health-care business Keeping pace with the rate of Early DM activities have
technological and medical focused on financially
advancement provides a significant oriented applications. Pre-
challenge. Cost is a constant issue dictive models have been
in this everchanging market. applied to predict length
of stay, total charges, and
even mortality.

Telecommunications Keeping pace with the rate of In addition to customer

industry technological change provides profiling, subscription
a significant challenge to businesses fraud and credit applica-
throughout the telecommunications tions are utilized through-
industry. In addition to this, out the industry. Concerns
deregulation is changing the about privacy and security
business landscape, resulting in are likely to result in DM
competition from a wide range applications targeted to
of service providers. Finding these areas.
and retaining customers is
important to telecommunications
providers.

«■«« Data Management

The most challenging part of a DM project is the data management stage. It is impor-
tant to know the sources from which data are gathered, the types of data involved in
the application, and how to prepare them for the DM techniques. At least 40 percent of
a DM project is spent on this stage.
 CHAPTER 12 Data M ining—Knowing the Unknown aaa* 361
DATA SOURCES
Before data is mined, it must be accessible in an easy format. Here are some examples
of data sources:
• Flat files. These are the most common source of data. They have a simple struc-
ture and are easy to manipulate by the data-mining algorithms. Various types of
data can be found in such files.
• Relational databases. Relational databases are a set of tables containing rows and
columns representing values of entity attributes.
• Data warehouses. D ata warehouses are repositories of data collected from vari-
ous data sources.
• Geographical databases. Geographical databases combine the relational data-
bases with geographical information from maps.
• Time series databases. These databases contain time-related data such as stock
m arket and financial data. Time series databases are continuously fed new data
and require real-time analysis.
• World Wide Web. The World Wide Web is a dynamic repository with thousands of
authors and documents. Mining the Web is a challenging task, which can be helpful in
tracing the profile of the Web user or in finding a relevant document to a specific search.

TAXONOMY OF DATA
Various types of data can be collected. Numerical measurements and text documents
are simple types; geographic data and hypertext documents are complex types. They
can be found in several forms:
• Business transactions. Every business transaction, whether it is related to internal
business functions like purchasing or external business operations such as rela-
tionships with clients, is recorded in large databases. For example, supermarkets,
through the bar-coding system, store millions of transactions daily; this helps
managers reduce inventory costs.
• Scientific data. Scientists collect huge amounts of data from their experiments or
observations; this data may include astronomical data and weather data for later
analysis and better predictive outcomes.
• M edical data. Hospitals keep patients’ medical records for years. This type of data
is continuously updated based on the health situation of the patient. It is
extremely helpful to physicians in diagnosis or prescribing treatments.
• Personal data. Government and other institutions collect information about indi-
viduals, groups, and communities for security or planning purposes. This has
become especially noticeable and critical since the September 11,2001, attacks.
• Text and documents. Most of the communications between companies are based
on reports and written documents that are exchanged, communicated, and stored
for later use.
• Web repositories. For the last decade, the Internet has been a huge vehicle for
data stored and organized in repositories.

DATA PREPARATION
The purpose of this step is to change or modify data in order to fit exploration findings.
A t this point, the following operations are performed on the data:
• Evaluating data quality
• Handling missing data
362 a a a i PART IV K M System Tools and Portal?

• Processing outliers
• Normalizing data
• Quantifying data
During the data preparation phase, we begin to understand the importance of some
variables and the irrelevance of others. The result is more like a filter, eliminating use-
less variables and focusing on the significant ones. New variables could be introduced in
order to understand more accurately groupings of customers that appeared unexpect-
edly during the exploration of the initial set of data. This task is called feature extraction
and enhancement. It consists of selecting the field to use from each record. The chosen
fields are called features. Various methods are used to ensure the selection of the best
set of features. This task is not easy; many tools have been developed for this purpose

MODEL BUILDING
In this phase, various modeling techniques are selected and applied, and their param e-
ters are calibrated to optimal values. Typically, for the same data-mining problem type,
there are several techniques, some of which have specific requirements on the form of
data. Therefore, stepping back to the data preparation phase is often needed.
Once data is prepared, a model is built to explain patterns in data. The first step is
to select the modeling technique. The most popular techniques are as follows:
• Association rules
• Classification trees
• Neural networks
These techniques were covered in Chapter 11.
There are situations where more than one technique is applied. Each technique is
applied separately. The results are then compared in terms of performance, business
advantage, and gain. Sometimes, there are constraints and requirements that limit the
choice of possible techniques. Among the reasons that restrict the use of a certain tech-
nique are the assumptions that the technique makes about the data. The assumptions
must be compared to the description of the data. If the assumptions do not agree with
the data, the designer may have to step back to the data preparation phase and readapt
the data. If this cannot be done, the technique then is rejected.

PARAMETER SETTINGS AND TUNING

With any modeling technique, there are several parameters to be adjusted. Most of the
time, they are adjusted empirically. The set of initial and intermediate parameters must
be recorded for eventual use and comparison. The selection of the param eters must
also be explained, and the process of reaching the final values must be documented.
The testing and validating samples are used for this task.

MODEL TESTING AND ANALYSIS OF RESULTS

At this point, the prepared data is used, and the model is put to the test using test cri-
teria. If it fails the test, the entire model is rejected or its param eters are adjusted for
further testing. After the test phase, the model is validated. Before proceeding to final
deploym ent, the m odel is thoroughly reviewed to ensure integrity, reliability, and
usability.
A key objective is to determine if there are issues that have not been sufficiently
considered.
 CHAPTER 12 Data M ining—Knowing the Unknown ii i i 363
In summary, the tasks of this phase are as follows:
• Reviewing the business objectives and success criteria
• Assessing the success of the DM project to ensure that all business objectives
have been incorporated and identifying any factors that have been overlooked
• Understanding the data-mining results
• Interpreting the results
• Comparing the results with common sense and the knowledge base to see if there
are any new or worthwhile discoveries

TAKING ACTION AND DEPLOYMENT

At this point, the results obtained from the data-mining application need to be acted
upon. Rem em ber that a model is not the reality. It is a representation of the real situa-
tion. Like a map, it shows the roads, but it is not the actual road. The model is designed
to manipulate data and help in selecting the right action. Unless an action is taken, no
transformation is achieved.
Creation of the model is not the end of the project. Knowledge of the data must be
organized and presented in a way that the manager can use. Depending on the require-
ments, the deployment phase can be as simple as generating a report or as complex as
implementing a repeatable data-mining process. In many cases, it is the manager, not the
data analyst, who will carry out the deployment steps. It means the manager must under-
stand up front what actions will be carried out in order to make use of the created model.
The deployment phase involves several tasks:
• Summarizing the deployable results
• Identifying the users of the discovered knowledge and finding out how to deliver
and propagate it
• Defining a performance measure to monitor benefits obtained by the implemen-
tation of the DM results (This task is of extreme importance, because it can
reduce the risk of misusing the data-mining results. To that end, a monitoring
strategy has to be put in place.)
Regarding the procedure or intent, deployment means incorporating the results
obtained into the business operation. For example, a cellular telephone company dis-
covered from a data-mining project that profitable customers are the most likely to use
their call center. In response, the executive manager decided to improve the services of
their call center to keep their customers happy.

POSTDEPLOYMENT PHASE
The purpose of this stage is to provide feedback resulting from the deployment of the
newly created model. M easurements are often used to provide feedback. The question
is what to measure and how to do it. One im portant measure is the return on invest-
m ent (RO I). It m easures the financial impact, the response rate, and other impacts
resulting from the DM project.

• • • • DM in Practice
DM can be used in two ways: to improve the internal business functions and to address
the company’s relationship with its customers. Most efforts are being made to under-
stand custom er behavior and adjust a com pany’s products and services accordingly.
Over 70 percent of marketing departments currently conduct DM projects (Vee Box 12.3).
364 a a a i PART IV K M System Tools and Portals

a i a a a i b ox 12.3 a a a a a a

LESSONS LEARNED FROM DM APPLICATION IN SUPERMARKETS

INFORMATION JACKPOT WAS IN THE CARDS HOW DO YOU GET THEM TO USE THEIR CARDS
AND WHAT KIND OF RESPONSE RATES ARE
Dick’s Supermarkets uses data-mining tools from
YOU CURRENTLY GETTING?
Datasage, Inc., in Reading, M assachusetts, to
gather purchasing-history information from shop- We developed several incentive programs for cus-
pers’ scan cards. The company then uses this data tomers using the card. We have given away prizes
to identify product relationships and custom er such as lawn mowers and computers. Currently,
buying patterns. 90 percent of our total store [sales are] captured
K enneth L. Robb, senior vice president of by these cards.
m arketing at the Platteville, W isconsin-based
Brodbeck Enterprises, Inc., which operates the ARE THERE SOME GENERAL THINGS YOU’VE
eight-store supermarket chain, recently talked to LEARNED ABOUT YOUR SALES?
Computerworld about the project.
We discovered that 45 percent of our customers
represent close to 90 percent of our volume.
WHAT HAS DATA MINING DONE FOR DICK’S
Using this information, we can offer the best dis-
SUPERMARKETS?
counts to the best custom ers so that we deliver
It has m ade us sm arter about our customers, value to those customers who represent the bulk
sm arter m arketers —and made us more efficient of our business.
in our marketing and merchandising investments.
WHAT ELSE HAVE YOU LEARNED?
WHAT’S THE BASIS OF A GOOD DATA-MINING
We looked over IV2 years’ worth of data and dis-
PROGRAM?
played the top product correlations. In our stores,
You have to establish the integrity of your data we found a high correlation between yogurt and
because th a t’s im portant to the decisions you’ll granola bars and also pie filling and canned milk.
make. For us, that means getting our customers to We placed a display of granola bars adjacent to
use their scan cards with each purchase so that we the yogurt and measured a 60 percent difference
have good and thorough data about what goes on in sales between that type of display and a regular
in our stores. display in the store.

SOURCE: Deck, S. “Data Mining,” Computerworld, March 29,1999, www.computerworld.com/hardwaretopics/hardware/

desktops/story/0,10801,43509,OO.html, Date accessed August 2002.

a a a a a a

This is explained by the evolution of a company’s culture. Being customer-centric rather

than service- or product-centric, most companies are driven by the necessity of getting
closer to their customers and gaining their loyalty. Customers are extremely volatile;
they have access to a huge variety of products and brands through the Internet, regular
mail, and other means. These ideas and concepts are illustrated in Box 12.4. It shows
how HSBC Bank is using data mining to improve its relationship with the customers in
terms of proximity and efficiency.
Building customer loyalty is a challenging task for most companies. DM is among
the tools that provide insight about customers and ways to increase sales. A successful
 CHAPTER 12 Data M ining—Knowing the Unknown ■i i « 365

i 1 i 1 1 i BOX 12.4 1 1 1 1 1 1

DATA MINING IN THE BANKING SECTOR

HSBC Bank USA, a member of the global HSBC customer. “It was just a m atter of mining the data
Group, serves more than 1.4 million retail banking and analyzing the patterns to learn more about
customers, providing a variety of checking, invest- who needs what and when. This predictive analysis
ment, loan, and other financial products through helps us contact the right people at the right time
380 New York banking locations. HSBC Bank with the right offer,” said Somma.
USA, with $35 billion in assets, also serves business According to him, the key is using DM to find
and commercial customers. customers that “look like” those that have demon-
strated a particular buying behavior in the past.
IDENTIFYING THE BUSINESS PROBLEM The previous tools such as OLAP were not able to
satisfy his demand. “OLAP is a nice way to get a
Retail banking is a highly competitive business. A
sense of the characteristics of a data set, but I can’t
typical branch of a neighborhood bank might
find the strength of the associations and I can’t do
have many competitors in close proximity. It cre-
predictive modeling, and that’s the meat of what I
ates a continuous com petition to attract —and
need,” he said. “O LA P is a good reporting tool,
keep—potential customers in the surrounding area.
but without a statistical engine, it can only tell me
R ecent deregulation has also implied an addi-
where I’ve been, not where I need to go.”
tional competition coming from financial services
companies of all kinds. With so many organiza-
THE RESULTS
tions working the same customer base, the value
of custom er retention is greater than ever. The Somma and his colleagues in the bank’s individ-
issue of how to maximize sales and minimize m ar- ual product departm ents have created successful
keting costs has become critical for HSBC Bank marketing strategies based upon the data-mining
USA. This is a typical challenge arising in the predictive models. In just 3 years, he reports that
retail industry at large. sales have risen by up to 50 percent across the
b ank’s num erous product lines, and m arketing
APPLYING DATA MINING costs have been reduced by 30 percent. Somma
notes that by targeting custom ers more accu-
In the past, HSBC Bank USA often prom oted its
rately, he has been able to uncover the most
products to both new and existing customers by
promising prospects for a particular product and
using lifestyle segm entation inform ation pu r-
to save money by not contacting those customers
chased from market research companies, such as
who do not fit the predictive profile.
studies that predict income and buying behavior
based upon the neighborhood or subdivision.
THE BENEFITS
“That kind of external segmentation scheme has
its place and can be valuable when soliciting new “If I’m doing a direct mail campaign, for example,
customers. However, we realized that we already had I can mail out fewer pieces in a more targeted
much more specific and potentially valuable in- manner, get a higher percentage response rate,
formation on the buying habits and needs of our and bring in about the same amount of money,” he
1.4 million customers already locked in our data- said. “Avoiding the traditional ‘shotgun’ approach,
base,” explained Joe Somma, HSBC Bank USA’s in one recent promotion, we reduced a mail pro-
manager of customer acquisition and research. gram by more than a third, saving thousands
In response to this situation, the bank decided in postage and printing costs, but still brought in
to evolve from a mass-m arketing approach in 95 percent of the revenues of the previous mailing.
which selected services could be suggested to So our ROI was enormously improved.”
reflect more closely the real needs of a particular (<continued)
366 7777 PART IV K M System Tools and Portals

(<continued) to be inundated with messages about products

they don’t want or need. Thus, by data mining, we
“If you run just one good model, you can eas- are less likely to annoy our customers with what
ily pay for the software.” In addition, Somma they might consider junk calls or mail,” he said.
notes there is also a strong customer care “bonus” “And that, I believe, pays off in customer loyalty
to his predictive modeling efforts. “No one likes for the long term.”

SOURCE: Adapted from SPSS case studies, www.spss.com/success/pdf/HSB-0200.pdf, Date accessed August 2002.

DM application must bring data and information from various sources and business
departments. DM is a required component for various business applications, as shown
in Figure 12.7.
The most popular DM applications are as follows:

• M arket management applications: M arket segmentation, target marketing, churn

prediction
• Sales applications: Trends analysis, forecasting, pricing strategies
• Risk management applications: Fraud detection, customer retention, detecting
inappropriate medical treatments
• Web applications: Web analytics where DM is applied to the activity logs of Web
servers to gain insights about Web surfers’ behavior
• Text mining: Mining customers’ letters to reveal major complaints, mining mainte-
nance documents to match with a specific failure, mining customers’ e-mails to
automate responses

a i a FIGURE 12.7: DM as a Required Component for Various Applications in Different Departments

Marketing Sales Department

Department • Sales Forecasting
• Segmentation • Sales Trends
 CHAPTER 12 Data M ining—Knowing the Unknown ■■ ■ ■ 367

«* Role of DM in Customer Relationship Management

Most m arketers understand the value of collecting customer data, but they also realize
the challenges of leveraging this knowledge to create intelligent, proactive pathways
back to the customer. D ata m ining—technologies and techniques for recognizing and
tracking patterns within d ata—helps businesses sift through layers of seemingly un-
related data for meaningful relationships, where they can anticipate, rather than simply
react to, customer needs.
Figure 12.8 shows knowledge integration from disparate sources. Knowledge foi
m arketing decision support can come from custom er knowledge, retailers, m arket
research, or market knowledge from third parties. Among the many data-mining appli-
cations that enhance an organization’s customer services are the following.

CUSTOMER ACQUISITION
This type of application involves finding customers who previously were not aware of
the product, who were not candidates for purchasing the product (for example, baby
diapers for new parents), or who in the past have bought from competitors. Data min-
ing segments prospective customers and increases the response rate that an acquisition
marketing campaign can achieve.
Customer churn is another significant problem in many industries. It is generally
far more cost-effective to retain existing customers than to secure new ones,. It is desir-
able to identify customers who are at risk of changing service to a new provider and
offer them incentives to retain their existing service.
Customer historical data contains information that can be extremely valuable to
retention specialists. Historical data holds usage patterns and other important customer

FIGURE 12.8: Integrated Knowledge Management System for Marketing

SOURCE: Shaw, M. J., et al., “Knowledge Management and Data Mining for Marketing, Decision Support
Systems, vol. 31,2001, pp. 127-137.
368 ■ • ■ ■ PART IV K M System Tools and Portals

characteristics that can be used to identify satisfied and dissatisfied customers. By iden-
tifying which customer renewed their service, which did not renew their service, and
what incentives were offered to both groups, a predictive model can be built to predict
customers who will not renew their service and to make recommendations as to the
most effective incentives to offer.

CAMPAIGN OPTIMIZATION
In most marketing organizations, there are many ways to interact with customers and
prospects. Besides the offers made, there are multiple communication channels (direct
mail, telemarketing, e-mail, the Web) that can be used. The process of marketing cam-
paign optimization takes a set of offers and a set of customers, along with the charac-
teristics and constraints of the campaign, and determines which offers should go to
which customers, over which channels, and at what time. The key to satisfying cus-
tomers, optimizing their experience, and thereby quite possibly their loyalty lies in
understanding their individual preferences. Twenty-first century custom er relation-
ships are based on mass customization rather than mass production. What one cus-
tom er may consider “attentive” custom er service another may deem “oppressive.”
Being able to recognize the differences between these two types of customers and
making effective use of such information remains a challenge. It is a challenge that
data-mining techniques are uniquely well suited to address.

CUSTOMER SCORING
Once a model has been created by a data-mining application, the model can then be
used to make predictions for new data. The process of using the model is distinct from
the process that created it. Typically, a model is used many times after it is created to
score different databases. Identifying customers who are at risk of changing their ser-
vice and subsequently determining the appropriate actions to take that will encourage
custom ers to retain their service presents a significant challenge to organizations
responsible for customer care and customer retention. It is often unclear exactly what
characteristics dissatisfied customers exhibit and, therefore, difficult to predict exactly
which customers will not renew their service. In addition, even when it is known that a
specific custom er is dissatisfied, it is difficult to determ ine what incentives can be
offered that will encourage that customer to retain their service.

DIRECT MARKETING
Direct marketing is one of the most popular applications of data mining, because the
results can be readily applied to obtain a better campaign response and a higher return on
investment. For example, a trained statistician or consultant can develop a response
model by using data from a past direct marketing campaign to predict those most likely to
respond to the next campaign. By contacting only those most likely to respond, that mar-
keter can substantially increase the percentage of responses and generate higher profits.
Direct marketing campaigns extend beyond the retail merchandising industry to
other consum er and business services. The convergence of direct m arketing and
Internet-based electronic commerce provides a rich resource for future applications of
data mining.

INTEGRATING DM, CRM, AND E-BUSINESS

Figure 12.9 shows how DM applications for CRM are integrated with e-sales functions
in order to create the custom er-centric firm. DM applications are the first line in
understanding the customer and an integral key to segmenting the market. An intelli-
 CHAPTER 12 Data M ining—Knowing the Unknown aaaa 369
CRM E-Saies

Customer- Valued
Customer Centric Data Customer
Loyalty Warehouse Experience

Growing
Data Asset
4. Feedback ^
f o r Traffic InfovO'f
■Purchase P attei*4
Visitor Profited
tor Preference

FIGURE 12.9: Customer-Centric Data Warehousing Enables Next Generation E-Commerce

SOURCE: Microstrategy, Inc. “E-Business Elevating Customer Relationships,” p. 6, a white paper,

August 2002.

gent e-business system enhances CRM by enabling a level of responsiveness and

proactive customer care not achievable through other channels. The age of mass cus-
tomization is now at hand. Through personalization, corporations can build successful
1-to-l relationships with customers. Customers are now able to receive relevant infor-
mation derived from complex factors and inputs and enjoy a personalized purchasing
experience that meets their needs.
The ability to reach customers is fundamental to a successful e-business, because
customers have so many options to buy when and where they wish. Businesses must
bring the point of sale to a larger customer base and track buying patterns through a
variety of electronic and data channels.
Finally, feedback is an integral process that measures the overall effectiveness of
CRM. Assessing performance is required to meet growing challenges from the envi-
ronment, competitors, and stakeholders.

■■■ ■ Implications for Knowledge Management

A DM project is definitely not a straightforward project. While conducting such a proj-
ect, companies may face many problems, obstacles, and pitfalls that prevent them from
gaining returns on investing in a DM project. The main issues and challenges faced by
a company are illustrated in Figure 12.10.
What is often forgotten in the scramble to put DM into place is that the technique
chosen should reflect what the company wants to find out and the value it puts on
acquiring that information. The more powerful techniques have a cost both in project
effort and in the difficulty of interpreting results. An organization is wise to take a con-
servative approach to DM, investigating w hether the inform ation required can be
extracted by using a less costly approach before launching a project to implement a
 PART IV K M System Tools and Portals

/ Modeling \ ( Problems \
/ Mistakes \ / • No Commitment \
• Inadequate Tools • Culture
\ • Bad Sampling J \ • Structure /
J
• Inadequate Testing \ • No Deployment
\ / \ Efforts

9 9 9 9 9 9 FIGURE 12.10: Data-M ining Challenges

neural network when they cannot be sure it will pay off. Conversely, if the potential
payoff is sufficient, these techniques may be fully justified or they may indeed be the
only ones with the power to provide the required results.
Many DM projects falter because they are championed by the technical depart-
ment, which understandably wants to try out a new and exciting technique on some
interesting data that it holds. The results the project delivers may intrigue the user
departm ent to which they are offered. Ultimately, however, because the user has no
real business justification for the results and cannot act upon them in a timely manner,
there is no impetus to follow up on the project.
Problems that most data-mining project managers stumble across are listed here:
• Insufficient understanding o f business needs. Many organizations are leaping onto
the DM bandwagon; they know who their competitors are. The “me too” approach
to the technique is a mistake: Companies should have a business justification for
DM, just as they would for any other new technology. The organization will only
reap the benefits of DM if there is a real business case to answer because only then
will the project be bound tightly into a business process It should not be an experi-
ment for experiment’s sake. In other words, a business needs to resist the technol-
ogy push of tools, vendors, and technical departm ents—which often argue for a par-
ticular technique regardless of how appropriate it is—and look for user pull.
• Careless handling o f data. Data that are carelessly handled and conditioned can
mislead data miners by deceptively revealing irrelevant patterns and even
destroying existing patterns. D ata mishandling errors include the following:
a. Overquantifying data. This can lead to the “hiding” of relationships.
b. Miscoding data. This can lead to the creation of pseudopatterns.
c. A nalyzing without taking precautions against sampling errors. For example,
some data miners use the first x number of records as the training data and
 CHAPTER 12 Data M ining—Knowing the Unknown aiaa 371
blindly assume that the total population of records is randomly distributed
and independent.
d. Loss o f precision due to im proper rounding o f data values.
e. Incorrectly handling missing values. For example, a blank field replaced with
a u0” should not be used if a mean average of all the fields is to be taken.
D ata visualization and reporting tools should always be used to “see” and statisti-
cally analyze data sets to determine if the information is damaged. Additionally, a
domain expert, such as a data owner, should check to determine if obvious pat-
terns are still existent in the postconditioned data. This can be done by question-
ing the postconditioned data set by asking, “Are the relationships still there?” “If
this was to replace the original data set, can I still make sense out of it?” “Is it pre-
cise, and does it still answer what we already know?”
• Invalidly validating the data-mining model. Validating the data-mining model is of
utmost importance. This involves testing that a model is properly running as
expected by feeding it a large representative data set not used previously in train-
ing. Additionally, auditing and using multiple samples from a population can be
used to assure the validity of the model.
Most data miners will face either abundant amounts of testing data or, as
often seen, extremely scarce amounts. It is important to never ignore suspicious
findings, even if we are hastily trying to meet deadlines. Haste can lead to believ-
ing everything the data owners tell us about the data and, even worse, believing
everything our own analysis tells us. It is important to question the irrational-
sounding voice that asks, “Can this be truly right? I don’t think so.”
A good habit is to go through individual records before mining data to get a
feel for the information. Many times, a simple examination of the records will
raise the awareness of data miners and even bring to their attention to new fields
and relationships. Remember: The devil is in the details.
• Believing in alchemy. Many data owners believe that data mining is a form of an
alchemic process that will magically transform their straw databases into golden
knowledge. Data-mining efforts that find nothing are seen as “fool’s gold” by data
owners. However, they may not necessarily be failures. All data-mining efforts are
motivated by the human-induced suspicions that something exists; therefore, the
failure may be in the suspicion, not in the data to be mined or the techniques
used. Additionally, it could be that the data given to data miners are not suited for
the purpose. The technology used might also be inadequate to handle the mass of
data available or it may simply have been misapplied. The overall expectations of
the project may also play a role in making the gold seem to be fake. For example,
a company that wants to reduce all customer churn might drop a project that has
reduced it only 10 percent, even though this could be a first step.

Challenging and rechallenging the outcomes is the best way to guarantee that real
ROI is being achieved. In a rapid prototyping methodology, attacking and analyzing
the system as a whole led to finding that the data itself was the problem. In addition,
while developing a DM project, managers should consider the following issues:

• Select the right application.

• In-house development or outsourcing: The attraction of in-house DM is to gain skills
enhancement, reduce costs, and retain critical information. Outsourcing can reduce
the time to develop the project and assure the use of the most advanced technologies.
372 » ■ « « PART IV K M System Tools and Portals

• Assessing vendor selection: This task is quite difficult because the DM market is
still immature and lacks industry standards.
• Team building: Select the right people to participate in the project.

Su mma r y * 1 * 1
• D ata mining is the process of extracting useful patterns and regularities from large
bodies of data (that is, bodies of data that are too large to be analyzed by hand).
• Data mining is an umbrella term that covers the tools and techniques used in the
analysis of large databases. D ata mining is the process that is transforming raw
food into blood that nourishes the various functions of the corporation.
• Business intelligence tools are software products that provide the opportunity to
take action that will ultimately result in a measurable business benefit. They
include a range of tools such as reporting, OLAP, statistical analysis, data visual-
ization, and data mining. Business intelligence tools encompass all tools that could
be useful for supporting end users in their analysis and decision-making processes.
• Historically, the term data mining appeared in the 1990s, after a long maturation
process involving statistics, artificial intelligence, machine learning, and data ware-
housing. Tracing history to its roots, DM is a double-path discipline originating
from the birth of data organization and analytical techniques. On the data organi-
zation side, before DM can be applied, data must be properly collected and
stored. One such method is data warehousing. D ata warehousing collects data
from various operational systems, organizes it, transforms it, and makes it avail-
able for further DM investigation.
• The primary goals of data mining are prediction (predicting unknown or future
values) and description (finding human-interpretable patterns to data). The two
goals can be used to carry out primary DM tasks such as classification, regression,
clustering, summarization, and dependency modeling.
• The DM process is best illustrated by a seven-step methodology: business under-
standing, data understanding, data preparation, data modeling, analysis of the
results, knowledge assimilation, and deployment evaluation.

Te r m s to Kn o w i i i b
Categorical data: Categorical data is either nonordered claims experience. It divides a data set so that records
(nominal) such as gender or city or ordered (ordinal) with similar content are in the same group, and groups
such as high, medium, or low temperatures. are as different as possible from each other.
Classification: refers to the data-mining problem of Cross validation: A method of estimating the accuracy of
attempting to predict the category of data by building a a classification or regression model. The data set is
model based on som e predictor variables. divided into several parts, with each part in turn used to
Classification tree: A decision tree that places categorical test a model fitted to the remaining parts.
variables into classes. Data: Values collected through record keeping or by
Cleaning (cleansing): Refers to a step in preparing data polling, observing, or measuring; typically organized for
for a data-mining activity. Obvious data errors (such as analysis or decision making. More simply, data are facts,
improbable dates) are detected and corrected, and transactions, and figures.
missing data is replaced. Data accuracy: Refers to the rate of correct values in
Clustering: Clustering algorithms find groups of items the data.
that are similar. For example, clustering could be used Data format: Data items can exist in many formats such
by an insurance company to group customers according as text, integer, and floating-point decimal; refers to the
to income, age, types of policies purchased, and prior form of the data in the database.
 CHAPTER 12 Data M ining—Knowing the Unknown saaa 373
Data mining: An information extraction activity whose OLAP: Online analytical processing tools give the user
goal is to discover hidden facts contained in databases. the capability to perform multidimensional analysis of
Using a combination of machine learning, statistical the data.
analysis, modeling techniques, and database technology, Outliers: Outliers are data that fall outside the boundaries
data mining finds patterns and subtle relationships in that enclose most other data items in the data set.
data and infers rules that allow the prediction of future Pattern: A pattern is a relationship between two variables.
results. Typical applications include market segm enta- Range: The range of the data is the difference between
tion, customer profiling, fraud detection, evaluation of the maximum value and the minimum value.
retail promotions, and credit risk analysis. Alternatively, range can include the minimum and
Data-mining techniques: Procedures and algorithms maximum, as in “The value ranges from 2 to 8.”
designed to analyze the data in databases. Sampling: Creating a subset of data from the whole.
Deployment: After the m odel is trained and validated, it Random sampling attempts to represent the whole by
is used to analyze new data and make predictions. choosing the sample through a random mechanism.
Discrete: A data item that has a finite set of values. Sensitivity analysis: Varying the parameters of a model to
Discrete is the opposite of continuous. assess the change in its output.
External data: Data not collected by the organization, Sequence discovery: The same as an association rule for
such as data available from a reference book, a govern- the market basket analysis except that the time
ment source, or a proprietary database; effect on the sequence of events is also considered. For example,
dependent variable of the other. “Twenty percent of the people who buy a VCR buy a
Internal data: Data collected by an organization, such as camcorder within 4 months.”
operating and customer data. Test data: A data set independent of the training data
Missing data: Data values can be missing because they set, used to fine-tune the estimates of the model
were not measured, not answered, were unknown, or parameters.
were lost. Training: Another term for estimating a m odel’s parame-
Model: A model is a road map using data to help understand ters based on the data set at hand.
what is happening or predicting what might happen. Training data: A data set used to estimate or train a model.
Model accuracy: Refers to the degree of fit between the Transformation: Reexpression of the data such as aggre-
model and the data. This measures how error-free the gating it, normalizing it, changing its unit of measure, or
m odel’s representations are. Because accuracy does not taking the logarithm of each data item.
include cost information, it is possible for a less accu- Validation: The process of testing the models with a data
rate m odel to be more cost effective. set different from the training data set.

Te s t Yo u r U n d e r s t a n d in g 11,1
1. Why is DM a process and not an end in itself? Explain.
2. Describe the differences and similarities among DM, machine learning, and
business intelligence. How are they related?
3. “DM can be thought of as a form of advanced statistical techniques.” Do you
agree with this statement? Why or why not?
4. “DM is a tool to develop intelligent systems.” Define intelligence, explain
how systems could have intelligent behavior, and discuss this statement.
5. Describe the differences betw een O LA P and DM. When would you use
each tool?
6. What are the limitations of OLAP? How is DM able to overcome them?
7. What is the role of DM in e-business?
8. Describe, with examples, when you would use predictive DM and when you
would use descriptive DM.
9. Explain how DM is used in the health sector and in the telecommunications
industry.
10. Explain how com panies are using DM to understand their custom ers’
behavior and predict their intentions.
11. D escribe the m ajor pitfalls faced by com panies when im plementing DM
solutions.
374 a a a « PART IV K M System Tools and Portals

Kn o w l edge Ex e r c is e s a M a a
1. Discuss what types of industries can best benefit from DM. Which ones can-
not? Hint: Think of the ones having the most transactions and accessible data.
2. Statistical and DM applications both produce different results for manage-
ment, even though they might use the same historical data. Discuss the simi-
larities and differences in reporting capabilities.
3. A large online bank needs to mine data coming from many sources, including
marketing, accounting, and customer databases. Discuss the best way to col-
lect and prepare multisource data.
4. Minetise.com is an Internet company specializing in online banner ads. The
company is developing an application that customizes a banner according to
a custom er’s historic profile. Discuss how DM can be used to develop such an
application.
5. Your manager is extremely worried about integration problems that might
arise from implementing a DM application on your com pany’s SQL data-
base. Some of the questions bothering him include the following: How will it
integrate into the current computing environment? Will it work on our exist-
ing SQL database, or do we need anything else? How easily will the system
work on our intranet? Discuss the problems and possible solutions to these
questions. What other problems might your company face?
6. Finance Trance is a stock brokerage firm. They are thinking of using DM in
their customer services department. Suggest some uses and services they can
offer. Also, discuss the DM tasks that are to be used.
7. An online bookstore has asked your company to develop a DM application
to recom m end books to customers. Your m anager wants you to analyze
how the company works and see what data you can pull from their data
warehouses. How would you go about understanding the business and data
available before starting the project? What part does this fulfill in the over-
all project?
8. How could a mobile phone company use DM to lower customer churn? Can
it use DM to increase variables such as product development speed, m arket-
ing effort, or even customer retention?
9. During the data preparation stage, a supermarket omitted certain data fields
that were later shown to have significant adverse effects on the overall DM
application. Which stages of the DM process will be affected? A t which
stage could this problem have been detected? How do you think the prob-
lem was detected?
10. Design a survey to glean trends from several companies that are planning to
develop DM applications. This survey should help clarify the role of execu-
tive managers, the characteristics of the planned project, and the return
expected from it.
11. Conduct an in-depth case study with a company that has implemented a DM
solution to identify the best practices and common pitfalls.
12. Carrier Corp. is using data mining to profile online customers and offer them
cool deals on air conditioners and related products. By using services from
WebMiner, Inc., the air-conditioning, heating, and refrigeration equipment
m aker has turned more Web visitors into buyers, increasing per-visitor reve-
nue from $1.47 to $37.42.
Carrier, part of $26 billion United Technologies Corp., began selling air
conditioners, air purifiers, and other products to consumers via the Web in
1999. However, it sold only about 3,500 units that year, says Paul Berman,
global e-business m anager at the Farmington, Connecticut, company. Not
knowing just who its customers were and what they wanted was a big part of
 CHAPTER 12 Data M ining—Knowing the Unknown ■ ■■■ 375
the problem. “We were looking for ways to raise awareness [of Carrier’s Web
store] and convert Internet traffic to sales,” Berman says.
Last year, Carrier gave WebMiner a year’s worth of online sales data,
plus a database of Web surfers who had signed up for an online sweepstakes
the company ran in 1999. WebMiner combined that with third-party demo-
graphic data to develop profiles of C arrier’s online customers. The typical
customer is young (30 to 37), Hispanic, and lives in an apartm ent in an East
Coast urban area.
WebMiner matched the profiles to ZIP codes and developed predictive
models. Since May of 2002 C arrier has enticed visitors to its Web site
(www.buy.carrier.com/), with discounts. When they type in their ZIP codes,
WebMiner establishes a customer profile and pops up a window that offers
appropriate products, such as multiroom air conditioners for suburbanites or
compact models for apartm ent dwellers. “It’s the first time we’ve intelligently
delivered data-driven promotions,” Berman says.
Online sales have exceeded 7,000 units this year, Berman says, compared
with 10,000 units for a ll of last year. C arrier chose the W ebMiner service
because it was quick to implement and is relatively inexpensive —$10,000
for installation and a $5 fee to WebMiner for each unit sold, compared with
6-figure alternatives.
a. The DM application used by Carrier was one that was predictive in
nature. Could a descriptive model also be used? How would you use it,
and what outputs would you expect? Would they be of any use to Carrier?
b. What other data-driven promotions could Carrier come up with using
other data mining techniques?
c. What manufacturing-driven applications can Carrier implement using
data mining? Hint: How can it be used to forecast manufacturing defects?
d. What finance-driven applications can Carrier implement using data min-
ing? Hint: How can Carrier use DM to distinguish on-time paying cus-
tomers from doubtful ones?
SO URCE: Whiting 2001.
13. IAURIF, a French regional studies organization, needed to predict what
mode of transportation Parisians would use —and why they would use it—
from a large data set not originally collected for data mining.
With Clementine’s rule-induction algorithms, IA U RIF uncovered unex-
pected insights and proved the group’s first assumption, which was based
solely on experience, to be untrue. Instead, C lem entine’s rapid modeling
environment revealed the most im portant travel factors and derived accu-
rate results based on fact.
Results were as follows:
• Accomplished more accurate traffic forecasting
• Improved transportation planning
Analyzing and predicting traffic flows and growth is a complex process.
For IAURIF, this process started with an existing database of 400,000
records. These previously collected datia, from a detailed Parisian transport
survey, were not originally intended for data mining. That m eant a more
complex task right from the start, because IA U RIF had to complete exten-
sive preprocessing before it could begin data mining.
A rm ed with C lem entine’s data m anipulation capabilities, IA U R IF
began by grouping the 200 original fields under general headings, such as
place of residence and socioeconomic class. Then, analysts selected a repre-
sentative variable for each group of fields and ensured the groups were in-
dependent of their effect on transport mode. This im portant preprocessing
376 a a a a PART IV K M System Tools and Portals

enabled IA U RIF to pinpoint 26 fields, a core set of relevant variables that

would simplify and significantly help the group’s data-mining efforts.
IA U R IF analysts then used C lem entine’s rule-induction algorithms,
which predicted a three-way variable—whether someone would walk, drive,
or take public transportation for a specific journey. With Clementine’s power-
ful modeling techniques, analysts identified the factors behind each choice.
Based on experience, IA U RIF had first thought sociological factors, such as
income and class, would combine with the journey’s purpose to be the most
im portant causal factors.
However, Clementine uncovered a very significant, and surprising, find-
ing. The most im portant factors proved to be journey distance and trip
tim e—not factors the group had predicted on experience alone. To be sure,
IAU RIF proved Clementine’s high accuracy by testing results with a valida-
tion data set. In the end, Clem entine and this new modeling process
increased IA U R IF’s ability to plan future transport.
a. Describe the case from the perspective of the 7-step DM process. Which
parts were not covered? What recommendations do you have?
b. What difference would using online transaction data, instead of survey data,
have on the overall IAURIF project? What suggestions would you give?
c. What other DM projects could you implement for IAURIF? Which DM
techniques would you use? Why?

Re f e r e n c e s 1 , 1 1
Adriaans, Pieter, and Zantinge, Dolf. Data Mining. Harvard Business Review, vol. 77, no. 2,1999, pp. 106-116,
Boston, MA: Addison-Wesley, 1996. harvardbusinessonline.hbsp.harvard.edu/b02/en/
Cabena, Peter, et al. Discovering Data Mining: From com m on/item_detail.jhtml?id=4347, Date accessed
Concept to Implementation. Upper Saddle River, NJ: October 19,2002.
Prentice Hall, 1998. Holsheimer, M., and Kersten, M. L. Architectural Support
Cahlink, George. “Data Mining Taps the Trends,” www. for Data Mining. K D D Workshop 1994, pp. 217-228.
govexec.com/features/1000/1000managetech.htm, Hunt, E., Marin, J., and Stone, P. Experiments in
September 18,2000, D ate accessed October 16,2002. Induction. N ew York: Academic Press, 1966.
Deck, S. “Data Mining,” Computerw orld, March 29,1999, Inmon, Bill. Managing the Data Warehouse. New York:
www.computerworld.com/hardwaretopics/hardware/ John Wiley & Sons, Inc., 1996.
esktops/story/0,10801,43509,OO.html, Data accessed 2002. Kaplan, Abraham. The Conduct o f Inquiry: M ethodology
Delmater, Rhonda, and Hancock, Monte. Data Mining fo r Behavioral Science. Chicago: Chandler Publishing
Explained. Boston, MA: Digital Press, 2001. Company, 1964.
de Ville, Barry. M icrosoft Data Mining: Integrated Kaplan, Robert S., and Norton, David P. The Balanced
Business Intelligence fo r E-Com m erce and Knowledge Scorecard: Translating Strategy into Action. Boston,
Management. Boston, MA: Digital Press, 2001. MA: Harvard Business School Press, 1996.
Dhar, Vasant, and Stein, Roger. Seven M ethods fo r Kass, G. V. “Significance Testing in Automatic Interaction
Transforming Corporate Data into Business Intelligence. D etection,” A pplied Statistics, vol. 24, no. 2,1976,
Upper Saddle River, NJ: Prentice Hall, 1997. pp. 178-189.
Fayyad, Usama M., et al. Advances in Knowledge Discovery Kass, G. V. “An Exploratory Technique for Investigating
and Data Mining. Menlo Park, CA: A A A I Press, 1996. Large Quantities of Categorical D a ta ” A pplied
Groth, Robert. Data Mining: A Hands-On Approach fo r Statistics, vol. 29, no. 2,1980, pp. 119-127.
Business Professionals. Upper Saddle River, NJ: Kluge, Jurgen, Stein, Wolfram, and Licht, Thomas.
Prentice Hall, 1998. K nowledge Unplugged: The M cKinsey & Company
Groth, Robert. Data Mining: Building Competitive G lobal Survey on Knowledge Management. New York:
Advantage. Upper Saddle River, NJ: Prentice Hall, 2000. Palgrave, 2001.
Han, Jiawei, and Kamber, Micheline. Data Mining: Kuhn, Thomas. The Structure o f Scientific R evolutions,
Concepts and Techniques. San Francisco, CA: Morgan 3d ed. Chicago: University of Chicago Press, 1996.
Kaufmann, 2001. Loshin, David. Enterprise Knowledge Management: The
Hansen, Morten T., Nohria, Nitin, and Tierney, Thomas. Data Quality Approach. San Francisco, CA: Morgan
“What’s Your Strategy for Managing Knowledge?” Kaufmann, 2001.
 CHAPTER 12 Data M ining—Knowing the Unknown ■a ■a 377
Mena, Jesus. Data Mining Your Website. Boston, MA: Quinlan, R. “Discovering Rules by Induction from Large
Butterworth-Heinemann, 1999. Collections of Examples,” in D. Michie (ed.), Expert
Michael, J. A., Linoff, Berry, and Linoff, Gordon. Data Systems in the Microelectronic Age. Edinburgh, 1979,
Mining Techniques: For Marketing, Sales, and Customer pp. 168-201.
Support. New York: John Wiley & Sons, Inc., 1997. Rud, Olivia Parr. Data Mining C ookbook. New York:
Michael, J. A., Linoff, Berry, and Linoff, Gordon. John Wiley & Sons, Inc., 2001.
Mastering Data Mining: The A rt and Science o f Shaw, M. J., Subramaniam, C., Wan Tan, G., and Weldge,
Customer Relationship Management. New York: John M. “Knowledge M anagement and Data Mining for
Wiley & Sons, Inc., 2000. Marketing,” Decision Support System s, vol. 31,2001,
Michalski, Ryszard S., Bratko, Ivan, and Kubat, Miroslav. pp. 127-137.
Machine Learning and Data Mining: Methods and Smith, Reid G., and Farquhar, Adam. “The Road Ahead
Applications. New York: John Wiley & Sons, Inc., 1998. for Knowledge Management: An A l Perspective,” A l
Michie, D. “M ethodologies from Machine Learning in M agazine, vol. 21, no. 4, Winter 2000, pp. 17-40.
Data Analysis and Software,” The Com puter Journal, Sonquist, J. A., Baker, E., and Morgan, J. Searching fo r Struc-
vol. 34, no. 6,1991, pp. 559-565. ture. Ann Arbor, MI: Institute for Social Research, 1973.
Microstrategy, Inc., “E-Business Elevating Customer Spence, Robert. Information Visualization^ Boston:
Relationships,” white paper, p. 6,2001. Addison-W esley, 2001.
Millaer, William L., and Morris, Langdon. Fourth Genera- SPSS, HSBC Bank U S A , www.spss.com/success/pdf/HSB-
tion R&D: Managing Knowledge, Technology, and 0200.pdf, Date accessed August 2002.
Innovation. New York: John Wiley & Sons, Inc., 1999. Stewart,Tomas A. Intellectual Capital— The New Wealth o f
Mizuno, Shigeru. Management fo r Quality Improvement: Organizations. New York: Doubleday-Currency, 1997.
The Seven New Q C Tools. Cambridge, MA: Produc- Sturm, Jake. Data Warehousing with M icrosoft S Q L
tivity Press, 1979. Server 7.0 Technical Reference. Redmont, WA:
Morgan, J. N , and Sonquist, J. A. “Problems in the Analysis Microsoft Press, 2000.
of Survey Data, and a Proposal,” Journal o f the American Thuraisingham, Bhavani. Data Mining: Technologies,
Statistical Association, vol. 58, June 1963, p. 415. Techniques, Tools, and Trends. Ft. Lauderdale, FL: CRC
Moxon, B. “Defining Data Mining,” D B M S Online, Data Press, 1999.
Warehouse Supplem ent, www.dbmsmag.com/9608d53. Tiwana, Amrit. The K nowledge Management Toolkit.
html, August 1996, Date accessed 2002. Upper Saddle River, NJ: Prentice Hall, 2000.
O ’D ell, C., Hasanali, F., Hubert, C., Lopez, K., and Westphal, Christopher, and Blaxton, Teresa. Data Mining
Raybourn, C. Stages o f Implementation: A Guide fo r Solutions: Methods and Tools fo r Solving Real-W orld
Your Journey to Knowledge Management Best Problems. New York: John Wiley & Sons, Inc., 1998.
Practices. Houston,TX: APQC, 2000. Whiting, R. “Carrier Fans Sales With Data Mining,”
Payne, L. W., and Elliot, S. “Knowledge Sharing at Texas InformationW eek, August 6,2001, www.information
Instruments: Turning Best Practices Inside Out,” week.com/story/IWK20010802S0005, Date accessed
Knowledge Management in Practice, vol. 6,1997. November 2002.
www.apqc.org/prodserv/km-06.pdf, Date accessed Witten, Ian H., and Frank, Eibe. Data Mining: Practical
August 2002. Machine Learning Tools and Techniques with
Pyle, Dorian. Data Preparation fo r Data Mining. San Java Implantations. San Francisco, CA: Morgan
Francisco, CA: Morgan Kaufmann, 1999. Kaufmann, 2000.

S e le c te d W e b S it e s 1 1 , 1
www.dmg.org The Data Mining Group is a consortium a nonprofit international consortium that creates inter-
of industry and academics formed to facilitate the crea- operable industry specifications based on public stan-
tion of useful standards for the data-mining community. dards such as XM L and SGML. OASIS members
The site provides a member area and a software reposi- include organizations and individuals who provide, use,
tory, and it also provides news and announcements. and specialize in implementing the technologies that
www.kdnuggets.com KD Nuggets is a leading electronic make these standards work in practice. OASIS provides
newsletter on data mining and Web mining. Its monthly information on such emerging standards as predictive
release provides up-to-date news items on develop- model markup language (PMML) in the separate XML
ments in data mining and knowledge discovery. cover pages site www.oasis-open.org/cover.
www.oasis-open.org The Organization for the Advance- . www.xml.org This is an independent resource for news,
ment of Structured Information Standards (O ASIS) is education, and information about the application of
378 a a i i PART IV KM System Tools and Portals

XML in industrial and commercial settings. Hosted by also has a list of job offerings if you are looking for a
OASIS and funded by organizations that are commit- new opportunity within the field.
ted to product-independent data exchange, this Web www.mdcinfo.com This site provides information on the
site offers valuable tools, such as a catalog to help Meta Data Coalition, an organization originally set up
you make critical decisions about whether and by Microsoft to provide metadata solutions in data
how to employ XML in your business. For business- warehousing, business intelligence, and data mining.
people and technologists alike, this Web site offers www.icpsr.umich.edu/DDI The Data Documentation
a uniquely independent view of what’s happening Initiative (D D I) is an effort to establish an interna-
in the XML industry. tional criterion and methodology for the content, pre-
www.microsoft.com/sql This is the homepage for sentation, transport, and preservation of metadata
Microsoft SQL Server. This particular U R L provides a (data about data) in the social and behavioral sciences.
list of Microsoft online businesses. You can learn how Metadata constitute the information that enables the
to get real-time access to the most powerful data in effective, efficient, and accurate use of those data sets.
business intelligence, how to manage your business The site is hosted by the ICPSR (Interuniversity
partnerships more effectively, or how to share informa- Consortium for Political and Social Research) at the
tion within your organization through knowledge man- University of Michigan.
agement. You may also read how companies plan to use www.dhutton.com David Hutton A ssociates are consul-
wireless and other mobile technologies. tants in quality management. They are specialists in the
www.mlnet.org This site is dedicated to the field of Baldrige-style business excellence assessment as a tool
machine learning, knowledge discovery, case-based rea- to drive organizational change and improvement.
soning, knowledge acquisition, and data mining. This www.salford-systems.com Salford Systems are develop-
site provides information about research groups and ers of CART and M ARS data-mining, decision tree,
persons within the community. You may browse and regression modeling products. The site contains
through the list of software and data sets and locate the information about these products, white papers, and
events page for the latest calls for papers. This Web site other technical reports.
 Knowledge
Management
Tools and
Knowledge Portals
Contents

In a Nutshell
Portals: The Basics
What A re Portals?
Evolution of Portals
Key Characteristics
Illustration
The Business Challenge
Portals and the Business Transformation
M arket Potential
Knowledge Portal Technologies
Key Functionality
Collaboration
Content Management
Collaboration Versus Categorization: The Case of the World Bank
Intelligent Agents
Implications for Knowledge Management
Who Is Building Enterprise Portals?
Who Sponsors Enterprise Portals?
Implementation Issues
Bandwidth
Portal Product Selection
Summary
Terms to Know
Test Your Understanding
380 a a i i PART IV K M System Tools and Portals

Knowledge Exercises
References

What gets measured is what gets done.

—A no n

In a Nutshell
This chapter is concerned with the identification of the right tools to set up the infrastruc-
ture, services, and applications needed by the enterprise to manage effectively the flow of
information and the various activities related to the acquisition, production, and dissemi-
nation of knowledge. The goal is to provide an in-depth analysis of the technology trends
and issues related to the implementation of knowledge management solutions.
In previous chapters, we focused on knowledge management consisting of three
main processes: capture, codification, and integration. A variety of tools are emerging
to provide an enterprise with needed capabilities. The capacity to create knowledge
begins with collecting and organizing information. The value of information depends
on accuracy and timely accessibility. The more accurate and timely information is, the
more valuable it becomes for analysis and the creation of insight and new knowledge.
To be useful in creating knowledge, information must reach knowledge workers in the
most convenient, complete, and accurate way.
Portals are among the latest and most powerful tools to help achieve these goals. A
KM solution must employ a tool for capturing knowledge wherever it exits (in documents,
in experts’ minds, in databases, or as historical data). Another important tool is a user
interface that makes knowledge available to a larger community of knowledge workers.
By providing an integrated framework for linking people, processes, and knowledge,
portals can play a central role in simplifying managerial complexity, increasing opera-
tional productivity, and adding value to a company's business operations. Portals employ
distribution channels such as the Internet, intranets, and extranets. These distribution
channels enable companies to grasp potential advantages that are lying dormant in a com-
pany’s information and knowledge-based systems. Portals evolved from pure information
providers to sophisticated interfaces containing knowledge management features such as
content management for knowledge categorization, collaboration tools for knowledge
sharing, and personalization capabilities to facilitate the search function. Box 13.1 illus-
trates the impact of knowledge management portals on performance.

■■■■ Portals: The Basics

WHAT ARE PORTALS?
Portals are considered to be virtual workplaces that:
• Promote knowledge sharing among different categories of end users such as cus-
tomers, partners, and employees
• Provide access to structured data stored in data warehouses, database systems,
and transactional systems
• Organize unstructured data such as electronic documents, paper documents,
lessons learned, stories, and the like
 CHAPTER 13 Knowledge Management Tools and Knowledge Portals aaaa 381

- ■ 1 1 1 1 1 BOX 13.1 1 1 1 1 1 1 -

PERFORMANCE OF KNOWLEDGE MANAGEMENT PORTALS

FRITO-LAY: THE KNOWLEDGE CRUNCH m ation about the team ’s custom ers —including
sales, analysis, and the latest news —the portal
THE PROBLEM: UNLOCKING KNOWLEDGE
would contain profiles on who’s who in the corpo-
Corporate executives knew that capturing best prac- ration, making finding an internal expert a snap.
tices and corporate information would give employ-
ees something they could sink their teeth into. How- THE IMPLEMENTATION: BUILT FROM SCRATCH
ever, information was scattered around the company
M arino’s group established three goals for the
in disparate systems, and there was no easy way for
Frito-Lay portal: to streamline knowledge, exploit
the geographically dispersed sales force to get at it.
customer-specific data, and foster team collabora-
“We had knowledge trapped in files everywhere,”
tion. He brought in Navigator Systems, a consul-
says Mike Marino, vice president of customer devel-
tancy based in Dallas, which had worked with Frito-
opment at Frito-Lay, an $8.5 billion division of
Lay in the past and had some experience building
PepsiCo in Plano, Texas. Marino says that he knew if
knowledge management portals. Navigator built a
the 15-member sales team could only access the same
prototype in about 3 months using technologies
information, it would solve its ongoing problems
previously approved by Frito-Lay’s IS department.
within information sharing and communication.
Marino and Navigator essentially had to start
Additionally, much valuable knowledge was
from scratch when it came to populating the portal.
squirreled away on each salesperson’s system.
“Never before at Frito-Lay had they tried to cap-
There were many idiosyncratic methods of cap-
ture expertise systematically in one place,” notes
turing information, “none of which were terribly
Todd Price, a consultant at Navigator. Marino and
efficient,” Marino says. The sales team also lacked
Price did an audit within the company and then
a place for brainstorming and collaboration online.
created expertise profiles on the portal so that
If somebody got a piece of research and wanted
sales staff in the field would have an easy way to
to get input from account executives in Baltimore
learn who’s who at headquarters in Plano. That
and Los Angeles, the ability to collaborate [online]
way, people who have expertise in areas such as
just was not there.
promotion planning, activity planning, costing, or
new product announcem ents can be readily
THE SOLUTION: A KNOWLEDGE
tracked down and contacted for information.
MANAGEMENT PORTAL
The portal went live in January 2000. Since
The answer, M arino’s group realized, was to build then, three additional sales teams, or c u s to m e r
a knowledge m anagem ent portal on the corpo- c o m m u n itie s as they are called internally, have
rate intranet. A KM portal is a single point of been given access to the portal with different
access to multiple sources of information, and it co n ten t—including research abstracts and what
provides personalized access. Com panies are Marino calls performance scorecards, which evalu-
starting to pay attention to portals, because they ate account performance. “If somebody in sales or
offer an efficient way to capture information, says m arket research did a study in a particular area
Carl Frappaolo, executive vice president and like private-label trends, [the user] would be able
cofounder of the Delphi Group, a consultancy in to click to that abstract and get a summary of that
Boston. A KM portal at Frito-Lay would give the study.” Users access the portal, known as the
sales departm ent a central location for all sales- C u s to m e r C o m m u n ity P o r ta l ( C C P ) y through a
related customer and corporate information and Netscape Navigator browser and enter their
cut down on the time it took to find and share name and password on the Frito-Lay intranet.
research. In addition to different types of infor- ((c o n tin u e d )
382 ■ a « a PART IV K M System Tools and Portals

{c o n tin u e d ) catalog his or her strengths and areas of expertise.

“As a team leader, it helps me analyze where
THE RESULTS
people’s gaps might be without having to travel to
The CCP has paid off with increased sales. “What another m em ber’s location,” he says.
we expected to see was that the pilot team would The portal has also helped boost employee
outperform others in terms of sales and profitabil- retention rates, says Ackerman. Turnover used to be
ity,” Marino says. While he declined to give figures, terrible, he says, because salespeople felt pressured
he says the test team doubled the growth rate of to find vital information and communicate with the
the customer’s business in the salty snack category. rest of the team. Marino adds that salespeople felt
It also made the sales team happier. For example, frustrated and disconnected because there was no
the pilot team members reside in 10 different way to efficiently collaborate with the rest of their
cities, so the tool has become extremely valuable group unless they flew into a central location.
for communication and helps cut down on travel. Since the portal has been in place, not one per-
A year after implementing the portal, the pilot son on the 15-member team has left. Part of that
group has been able to share documents concur- can directly be attributed to the portal, says
rently instead of having to send faxes around the Ackerman, because it helps build the connection.
country to different offices. “It’s almost a distance In company surveys, salespeople previously com-
learning tool as much as anything else,” he says. plained about geographic constraints and how they
The CCP has also helped foster a sense of did not feel connected and part of a team, he says.
camaraderie and relationship building. For example, The portal has proven so successful that its use
the portal homepage lists the team members’ birth- has now become a PepsiCo initiative, says Marino.
days. People can also share best practices—on any- That means it will soon have added functionality so
thing under the sun. If someone developed an ef- that employees across all three divisions—including
fective sales presentation for a potential customer Tropicana—can take advantage of product perfor-
in Boston, a salesperson in San Francisco could mance information on a jointly shared customer
co-opt the information. Salespeople can also find like a supermarket, he says. Marino says the differ-
the latest news about their customers, and theirs is ent PepsiCo divisions will have the ability to co-
an autom atic messaging feature that tells team prom ote and co-merchandise multiple products
members who is online. that are consumed together—such as carbonated
For A ckerm an, the portal has also been an beverages and salty snacks—to drive greater sales
invaluable tool for helping him assess employee internally, naturally and for its customer. T hat’s
skill sets, because each salesperson is required to talking more than just peanuts.

SOURCE: Shein, E., “ The Knowledge Crunch,” CIO, May 1,2001, www.cio.com/archive/050101/crunch.html, Date accessed
August 2002.

Portals are Web-based applications providing a single point of access to online

information. They assure secure and reliable interface to participants in a business
process and collaborate with users through the integration of external Web-based
applications or internal back-office systems.
Portals are emerging as the most promising tool that could:
• Simplify access to data stored in various application systems
• Facilitate collaboration among employees
• Assist the company in reaching its customers
From a business perspective, portals provide the company’s employees with task-
relevant information. They can also quickly supply partners and customers with knowl-
 CHAPTER 13 Knowledge Management Tools and Knowledge Portals i i ai 383
edge. The goal of such a portal is the transparent enterprise, reducing the complexity
for reaching needed information.

EVOLUTION OF PORTALS
Initially, portals were mere search engines. They employed simple search technology to
find information on the Web. The next phase transformed the portals to “navigation
sites” to describe the functions available at sites like Quicken, MSN, Lycos, and Yahoo!.
Those portals categorize personal interests in groups (such as news, sports, finance,
education, science, and others). An example of the logical hierarchy of groups is shown
in Figure 13.1. They are referred to as Internet public portals.
Because of the increasing amount of information accessible internally and exter-
nally, the need for m ore personalized inform ation is becoming more critical and
urgent. To facilitate access to large accumulations of information, portals evolved to
include advanced search capabilities and taxonomies. Because emphasis was on infor-
mation, they were called information portals.
The evolution of the portal concept is shown in Figure 13.2.
Information portals provide the next step in the portal technology evolution. In
November 1998, Christopher Shilakes and Julie Tylman of Merrill Lynch’s Enterprise
Software team introduced the concept of “enterprise portal.” The concept was
described as “applications that enable companies to unlock internally and externally
stored inform ation and provide users a single gateway to personalized information
needed to make informed business decisions. They are an amalgamation of software
applications that consolidate, manage, analyze, and distribute information across and
outside an enterprise, including business intelligence, content management, data ware-
house, data mart and data management applications” (Shilakes and Tylman 1998).
Companies are becoming more aware of the opportunities for using and adding
value to the information lying dorm ant in scattered information systems. Companies
need to use both “publish” (pull) and “subscribe” (push) media to ensure that the right
information is available or distributed to the right people at the right time. The main

FIGURE 13.1: Yahoo! Home Page

j Fio Edt View Favorites Toob Help
j Back » - (& [£) Ja} | '^ S e a rc h [jJ Favorite* ^ H itt a iy | Gji)-* gfr {
JAddress [^Jhttp://wwwyahoocom/

^
P«itonjl fm jn c t
^
Shop
" V i« L H o o f #Emlil e> i
M±!r

Yahoo! Travel - Reserve: Flights. Hotels. Cars. Cruises Gel Deals: Web Fares from American Airlines

11A v FANTASYFOOTBALL 1
Season starts Tonight!
Piav Yahoo! Sports Fantasv Football
Yahoo! PayOirect - The safe, quick way to move money

Shop Auctions. Autos. Classifieds.—

-------------------------------- Real Estate. Shopping.
------------- Travel
------ ---------
sfcjUn
Find Careers. Maps. People Search. Personals. Yellow Pages Sign up to personalize Yahool - learn more j
Connect Chat. GeoCities. Greetings. Groups. Mail. Messenger. Mobile
10.53«m ET. Thu S«p 3
Organize Addresses. Briefcase. Calendar. Mv Yahool. PayDirect. Photos
Fun Games. Horoscopes Kids. Movies. Music. TV > Bomb rocks Kabul market, at least 1Q dead
Info Finance. News. Sports. Weather More Yahoo! .. - Israeli police detonate 1 300-pound car bomb
> M a n .YflthJS.guns in car arresled_in.Q.C.
Yahoo! Shopping Computer* - fitctronlct - Apparel - Mort...
> A 's 'n 2Qlh win. braaL&Liscflrd ■ U.S. Qcan

[ great deals on
IfrHI •GHome
ear up *or
Essentials
19
> Markets, sap soo ♦ . % • N**daq ♦ . % 27
r*H o ;C
p u c l 3
m\
^ 1*57
'. s<?ny
DVDP lavers under*1QQ
..Sbaw£i)se - Ho! Products
Newa - W n f t t r - Sports - Stock Ouoto

I ___________ u r m r t ■ Gifts for All Occasions

search I More...
• Labtec webcam only S24,95 with free shipping
D*pt>: Booki. Spoilt. Horn* &<3<rd«r>. F lo w rf. more... - Share your world todayl jj

SOURCE: www.yahoo.com, Date accessed September 5,2002.

384 ■ • i a PART IV KM System Tools and Portals

Search Indices Search Indices Search Indices

Categorized Content Collaboration

Personalization

Content Management

Examples

AltaVista Yahoo!, MSN Amazon.com

FIGURE 13.2: Evolution of the Portal Concept

benefits for companies include lowered costs, increased sales, and better deployment of
resources. Integration is the key feature of enterprise portals. Portals integrate applica-
tions by combining, standardizing, analyzing, and distributing relevant information and
knowledge to end users, whether they are customers, employees, or partners.
The effectiveness of information portals can be enhanced by building applications
that combine the search, analysis, and dissem ination of inform ation. This goal is
achieved through knowledge portals.
In the knowledge portal, the focus is not on the content of information but on the
way knowledge workers will use it. The knowledge portal is a key component in the
knowledge management architecture. It is the central piece, allowing producers and
users of knowledge to interact. A major trend in knowledge management systems is
the move to online knowledge portals.
Knowledge portals provide information on various topics, and they can be cus-
tomized to meet a user’s individual needs. Portals make it easy to access knowledge
because of their universal interface —a Web browser. Online portal systems let IT
organizations access a variety of back-end systems (such as process management soft-
ware and methodology databases).
Knowledge portals provide two kinds of interfaces:
• The knowledge producer interface. It facilitates the knowledge worker’s job of
gathering and analyzing information, collaborating with peers or colleagues, and
finally generating new knowledge.
• The knowledge consumer interface. It facilitates the dissemination of knowledge
across the enterprise. A key feature of knowledge portals is a sophisticated per-
sonalization facility that takes into account the consumer profile.

KEY CHARACTERISTICS
Knowledge portals provide information about all business activities and also supply
m etadata for decision making. Enterprise knowledge portals (EKP), therefore, distin-
guish knowledge from information. They provide a facility for producing knowledge
from data and information. They also provide a better basis for making decisions than do
enterprise intelligence portals. Those who have knowledge gain competitive advantage
 CHAPTER 13 Knowledge Management Tools and Knowledge Portals ■«** 385

a a i a « a bo x 13.2 a a a a a a

THE NEXT GENERATION OF PORTALS

THE SECOND WAVE OF EIP agement applications. ECPPs are characterized by

“virtual project areas” or communities.
The second wave is shifting the primary focus of
• Enterprise mission management portals (EMMPs)
the enterprise intelligence portal (EIP). Whereas
provide a “digital expertise-oriented workplace” —a
EIP did emphasize broadly based and generalized highly specialized and personalized Web site where
decision processing and mass dissem ination of everything a user team needs (such as access to ERP
corporate information, it now targets collabora- applications, productivity and analysis tools, and rel-
tion and highly targeted and personalized distribu- evant internal and external content) to effectively
tion of content, bundled with multiple types of spe- manage mission-critical management activities such
cialized expertise-oriented services. These trends, as customer relationship management (CRM) is
happening over the past year or so, indicate that consolidated and made accessible via the Web.
the EIP concept is on the verge of an explosion • Enterprise extended services portals (EESPs) do
along four key directions: everything the first three types do, but focus on
providing comprehensive job support from the
• Enterprise business intelligence portals (EBIPs) standpoint of “virtual enterprises” by creating
provide the central launching point for corporate communities and “virtual service spaces” of chan-
decision-processing and content-management nel partners, suppliers, distributors, and customers.
applications. Their primary focus is to connect
users with structured and unstructured content The convergence of the first and second EIP
relevant to them. waves will occur within IV2 years.This time will be
• Enterprise collaborative processing portals (ECPPs) spent on extending architectural frameworks that
connect users not only with all the information guide technology from the search-based, first-
they need, but also with everyone they need. wave portals to a fully functional architecture
ECPPs consolidate groupware, e-mail, work-flow, capable of enabling expertise- and service-based,
and critical desktop applications under the same second-wave portals.
gateway as decision-processing and content-man-

SOURCE: Excerpted from Davydov, Mark M. Intelligent Enterprise, vol. 3, no. 4, March 1,2000. Accessed from www.intelligent
enterprise.com/000301/supplychain.html. August 2002.

over those who have mere information. The essential characteristics of enterprise intelli-
gent portals and enterprise knowledge portals are shown in Table 13.1.

ILLUSTRATION
Let us illustrate using the case of the army’s knowledge online portal. The objective of
this portal is to transform the army into a networked organization that leverages its
intellectual capital to better organize, train, equip, and maintain a strategic land com-
bat force. More specifically, the army needs to do the following:
• Allow its enterprise information to be accessed more quickly and easily for less cost
• Use information technology to leverage army-wide innovation in services,
processes, and knowledge creation
Figures 13.3,13.4, and 13.5 show the capabilities of the army knowledge portal, the
communities of practice that were created, and their features.
■i n PART IV K M System Tools and Portals

TABLE 13.1 Knowledge Portals Versus Information Portals

Enterprise Information Portals Enterprise Knowledge Portals
U se both “push” and “pull” technologies to Are goal-directed toward knowledge
transmit information to users through a production, knowledge acquisition,
standardized Web-based interface knowledge transmission, and
Integrate disparate applications including knowledge management
content management, business intelligence, Are focused on enterprise business
data warehouse/data mart, data management, processes such as sales, marketing, and
and other data external to these applications risk management
into a single system that can “share, manage, Provide, produce, and manage
and maintain information from one central information about the validity of the
user interface” information they supply
Have the ability to access both external and Include all EIPs functionalities
internal sources of data and information, and
the ability to support a bidirectional exchange
of information with these sources
'fl ■ « ■ ■ ■ -
SOURCE: Firestone, J. “Enterprise Knowledge Portals,” White Paper 8, www.dkms.com, Date accessed
August 2002.

" The Business Challenge

Businesses from all sectors of the economy confront new market realities daily. At the same
time, there is inherent pressure to optimize the performance of operational processes in
order to reduce costs and enhance quality. Customer-centric systems allow companies to
understand the customer, predict customer behavior, and offer the right product at the right
time. Companies also need to commercialize their products at the lowest price possible.

FIGURE 13.3: The Army Knowledge Portal

J Fie Ed* Browse Go To Favorites Help

1^ Back. ® ® &1$ SM
reh <3 ^History |
jAddeM,jg j7 //C :/P o c w ^ «agQan(0^e>ting«/uharwky^X5ag2DSettir^enwaiy^2af<eme<X20FtetA^ntentlE5/4DY1AB67/38Q.6.Stde6jj fr>Go |

AKO Knowledge Communities lW S B

Knowledge Center

I®'U nknown Zone

SOURCE: E-Gov Knowledge Management Conference. “Using Army Knowledge Online for Mission
Success,” April 11,2000, www.army.mil/ako/webbriefing.ppt, Date accessed August 2002.
 CHAPTER 13 Knowledge Management Tools and Knowledge Portals aaaa 387
j F fe : E d f B ro m c G o to F a v o r ite ,' H * ' ■ : ’
J £»B ack » »» -
hannky/locaK20Se<lingj/Tefnpo<a»y5:2C(n<e(ne<5J2CiFtes/Conten( IEfvMDY1A867/356.7.0ff>c*f C«e« Management Knowledge Center

^fTri®“uSno«nZonT
■ ■ ■ I ■ I FIGURE 13.4: The Army Communities of Practice
SOURCE: E-Gov Knowledge Management Conference. “Using Army Knowledge Online for Mission
Success,” April 11,2000, www.army.mil/ako/webbriefing.ppt, Date accessed August 2002.

PORTALS AND THE BUSINESS TRANSFORMATION

The problems arise from two fundamental aspects underlying the current computing
environment. First, the explosion in the quantity of key business information already
captured in electronic documents has left many organizations literally losing their grip
on information as they transform into new systems and process upgrades. Second, the
speed by which the quantity and kinds of content is growing m eans that what is

FIGURE 13.5: Features of Officer Career Knowledge Management Center

SOURCE: E-Gov Knowledge Management Conference. “Using Army Knowledge Online for Mission
Success,” April 11,2000, www.army.mil/ako/webbriefing.ppt, Date accessed August 2002.
388 ■ « ■ ■ PART IV K M System Tools and Portals

required to meet these challenges is a rigorous internal discipline to expose and inte-
grate the sources of enterprise knowledge. As a result, organizations are facing trem en-
dous challenges. Consider these pressures faced by just about all organizations:

• Shorter time to m arket New products and services have to be conceived, devel-
oped, and delivered in months, or even weeks.
• Knowledge worker turnover. When a pivotal person leaves, the pain is widely
and quickly felt. “It’s becoming increasingly difficult to acquire and retain
employees, and a company’s strongest asset is its people,” says Chris Moore,
chief technology officer at Training Server, Inc. “Organizations that do not
tap into their mind share and take advantage of the knowledge within will
quickly fall behind.”
• More demanding customers and investors. For virtually every organization, the
squeeze is on customers wanting to pay less while investors want more value
from their portfolios. This means that all the resources to which an organiza-
tion can lay claim, including its intellectual resources, must be managed for
the best result.

To sustain competitiveness, organizations carefully craft processes to efficiently

m anufacture products and care for their custom ers —all in an effort to effectively
exploit valuable resources. However, Ernst &Young estimates that up to 80 percent of
a company’s intellectual resources—the knowledge residing within it—is not systemati-
cally applied to business processes.
Today, more companies realize that they must develop strategies and processes
designed to best utilize intellectual resources at both the strategic and operational
levels. Ten years ago, companies began using groupware (such as e-mail, discussion
forums, and document libraries) for coordinating activities. Now, they are inundated
with new tools for communicating, sharing knowledge, and interacting electronically.
They are deploying next-generation information, application platforms (such as enter-
prise portals), and real-tim e tools (such as instant messaging, Web conferencing,
streaming audio/video) while struggling to manage process engineering across partners
and suppliers as another aspect of collaboration.
Research from International Data Corp., for example, indicates that 50 percent of
companies adopting data warehousing are planning or already implementing knowl-
edge management. According to a survey by Cambridge Information Network (a divi-
sion of Cambridge Technology Partners) of its 3,500 member CIOs, 85 percent believe
that knowledge m anagem ent generates competitive advantage. Figure 13.6 reports
reasons for launching KM projects.
Organizations are looking for solutions to support their new e-business models. As
a result, the demand for tools to more effectively negotiate, plan, decide, and collabo-
rate has dramatically increased. However, most organizations are meeting collabora-
tion requirements on a piecemeal basis, fulfilling requests as they emerge from busi-
ness units or partners without an overall blueprint. The result is a hodgepodge of
overlapping and redundant technologies (Meta Group 2000).
The benefits companies are expecting from their EKP initiatives are shown in
Figure 13.7.
In the late 1990s, enterprises started to implement intranets to satisfy the enter-
prise requirements for information access, sharing, and dissemination. However, they
are still poorly equipped to take advantage of the rapid developm ent of intranets,
enterprise applications, and electronic business.
 CHAPTER 13 Knowledge Management Tools and Knowledge Portals iii « 389
Why Organizations Launch Their KM Programs

Increase Profits or Revenues 67%

Retain Key Talent and Expertise 54%

Improve Customer Retention and/or Satisfaction 52%

Defend Market Share Against New Entrants 44%

]
Accelerate Time to Market with Products 39%

I
Penetrate New Market Segments 39%
I 9 ■ fl I I
] FIGURE 13.6: Main Reasons
Reduce Costs 38% for Launching KM Projects

SOURCE: International Data

Corporation, 1999 Knowledge
Develop New Products and Services 35% Management Survey, 1999.

MARKET POTENTIAL
Knowledge portals have emerged as a key tool for supporting the knowledge work-
place. According to David Folger, a senior program director at WCS, “Portals are big
business. More than 85 percent of organizations plan to invest in portals during the
next 5 years. The median expenditure is $500,000. We believe that as the world becomes
more networked, these estimates are bound to climb. Why? Portals can provide easier,

a a a a a fl FIGURE 13.7: The Benefits Knowledge Portals

Benefits of Knowledge Portals

E-mail Traffic
Bandwidth Use
Productivity Time in Meetings
Locating Documents Phone Calls
Collaboration Response Times
Better Decisions Redundant Efforts
Quality of Data
Sharing Knowledge
Identifying Experts
390 a i a « PART IV K M System Tools and Portals

* ! * ■ ■ ■ BOX 13.3

THE ROLE OF PORTALS IN FACING BUSINESS PRESSURES

Business integration versus information integration Information with workflow, categorization, and
or application integration. Integrating information taxonomy equals context.
and applications is the first step to supporting • A pplication and information integration. This is a
business integration. A portal can support applica- strict reference to source systems from which
tion and information integration, but the weak- information is generated. It also includes the data
ness at this point can be a lack of providing busi- administration layers, such as cleansing, transfor-
ness context, or meaning, to the information. mation, and extraction.
Business integration is the integration and trans-
• Enterprise metadata repository. Metadata is vital
lation through the metadata layer of information
as the mechanism for context. The repository
and applications to provide context.
should not include every piece of information that
Process integration. In addition to providing con- exists, only the information that will be utilized to
text, the portal solution should support process provide meaning. The repository then becomes
integration. Process integration includes work- the “single version of understanding,” not just the
flow, categorization, and taxonomy services. These “single version of the truth,”
services provide the foundation for context.

SOURCE: Bolds, Robert. “Enterprise Information Portals: Portals in Puberty,” a white paper KMWorld, vol. 2, July/August 2001.

a a a a a a

unified access to business information, and better communications among customers

and employees”. See Box 13.3 for sample pressures facing portals.
The EIP market is comprised of the following infrastructure components:
• Content management
• Business intelligence
• Data warehouses and data marts
• Data management
An example of a portal in action is explained in Box 13.4.

BOX 13.4

FORD MOTOR COMPANY BUSINESS-TO-EMPLOYEE PORTAL

A BUSINESS-TO-EMPLOYEE PORTAL AT FORD study example won D C I’s Annual Portal Ex-
MOTOR COMPANY cellence Award for the Internet portal category,
announced in January 2001.
Ford Motor Company’s Internet portal, using the
Plum tree portal, is an example of a business- (continued)
to-em ployee (B2E) Internet portal. This case
 CHAPTER 13 Knowledge Management Tools and Knowledge Portals aaaa 391

(<continued ) drop file sharing, multithreaded discussions, real-

time messaging, and polling that Ford is deploying
Ford used Plumtree to implement an expan- enterprise-wide. For example, paint shop workers
sive fram ework for its am bitious business-to- from manufacturing departments on different con-
employee e-business strategy. The world’s second tinents can share skills and ideas easily, shortening
largest autom aker chose Plum tree to create a development cycles and increasing product quality.
single, simple, Web destination for 200,000 em - Com m unity o f practice. To foster collabora-
ployees enterprise-wide to find and share the con- tion, project mangers at Ford create community
tent and services they need to support customers pages of content and services shared by entire
and speed products to market. Ford is deploying business units.
the Plumtree corporate portal as part of an initia-
tive to m odernize the w orld’s largest intranet,
hub.ford.com, which spans 800 Ford facilities and BENEFITS TO FORD
150 manufacturing plants worldwide. The portal
Increased R O I on inform ation technology. The
enables Ford to integrate the hundreds of thou-
Plum tree-powered hub.ford.com organizes scat-
sands of Web pages that comprise hub.ford.com
tered intranet sites into a framework that every-
in one enterprise-w ide Web destination. Now,
body can use, anywhere, and broadens the audience
Ford employees anywhere can draw on a com -
for applications previously limited to specialists,
mon base of best practices, m arket news, prod-
increasing return on Ford’s electronic assets. Ford
uct specifications, perform ance metrics, and pol-
will integrate multiple data sources into the Plum-
icy and procedures for the inform ation they
tree corporate portal, including Docum entum ,
need to make confident business decisions and
whose content management platform is deployed
act quickly on revenue opportunities. Information
throughout the enterprise.
is available for custom er relationship m anage-
Increased productivity. A single, personal-
m ent, order fulfillment, custom er satisfaction,
ized destination for corporate content and ser-
sales and volume tracking, economic assessments,
vices helps focus everyone on company strat-
competitive inform ation, and Ford community
egy, brands, and com petitors, increasing the
initiatives.
impact of every employee. Hub.ford.com users
know what is happening across the enterprise
and can stay on top of their customers, products,
FEATURES OF FORD’S B2E PORTAL
and markets, driving sales. The portal is also the
Personalization. Each hub.ford.com user tailors fram ework for a broad e-learning initiative to
the portal experience to his or her role. To assem- foster employee competency, leadership, and
ble a complete view of business, Ford employees advancement.
can select gadgets for em bedding e-mail, real- Close collaboration. Desktop access to digital
time news feeds, stock reports, sales histories, per- workplaces is driving business-to-employee and
sonnel directories, self-service travel booking, business-to-business collaboration across Ford’s
corporate expense reporting, and pay and bene- enterprise. Now, Ford engineers can easily com-
fits in personalized portal pages. municate with one another and with suppliers
Collaboration. The portal enables Ford em - and product designers, and executives can conduct
ployees to use a Web-based workplace for drag and secure meetings across time zones and borders.

SOURCE: Finkelstein, C. “Building Enterprise Portals Using XML,” The Data Warehouse Institute Conference, 2001.

■ i a ■ fl a
392 aaaa PART IV K M System Tools and Portals

■■■■ Knowledge Portal Technologies

In this section, we discuss the technologies needed to build portals. We proceed by
identifying the functionalities to be included in a portal solution. For each capability,
we describe the corresponding technology.

KEY FUNCTIONALITY
The main goal of a portal is to provide a single point of access to all inform ation
sources. Therefore, portals must be the ultimate tools for universal integration of all
enterprise applications. At the same, because every individual has different informa-
tion needs and knowledge uses, portals have to deliver a personalized interface. Given
the complexity of this challenge, portals must include the following functionalities:
• Gathering. Documents created by knowledge workers are stored in a variety of
locations (such as files on individual desktops, Web sites on the network, and
databases on servers). In order to be accessible, data and documents need to be
captured in a common repository.
• Categorization. This facility profiles the information in the repository arid orga-
nizes it in meaningful ways for navigating and searching. Portals should support
categorization at all levels, including the employee, partner, and customer levels.
It should also support categorizations in various dimensions, including the
process, product, and service dimensions.
• Distribution. This facility acquires knowledge, either through an active mecha-
nism (search interface) or a passive mechanism (push). This facility supports the
distribution of structured and unstructured information in the form of electronic
or paper documents.
• Collaboration. Collaboration is achieved through messaging, workflow, discussion
databases, and so forth. This facility expands the role of portals from a passive
information provider to an interface for all types of organizational interactions.
• Publish. This facility publishes information to a broader audience, including indi-
viduals outside the organization.
• Personalization. Personalization is a key component of the portal architecture
because it allows individuals to enhance their productivity. It is becoming a neces-
sity for successful portals. This is due to the proliferation of information available
through the portal. To take advantage of this facility, knowledge
workers must be able to manage and prioritize the delivery of information
on task-function or interest bases.
• Search/navigate. This component provides tools for identifying and accessing spe-
cific information. The knowledge worker can either browse or submit a query.
Figure 13.8 shows the most common features of portals. It shows also their busi-
ness benefits.
Box 13.5 shows the Microsoft portal architecture.

COLLABORATION
The goal of the collaboration tool is to create a basic, collaborative knowledge man-
agement system that supports sharing and reusing information. To support the require-
ment of capturing undocumented knowledge, the messaging and collaboration module
is a perfect tool. If writing e-mail, sending documents, or participating in discussions
are easy tasks for knowledge workers, the motivation to act with knowledge manage-
 CHAPTER 13 Knowledge Management Tools and Knowledge Portals aaaa 393

r
Common Features Business Benefits
A

^Search Quick access to hidden information to facilitate business processes ^

^Categorization Ability to organize information assets by business process, group, or job j

category thus promoting access to relevant information J

( Query, Reporting, Better decision support as well as information dissemination and sharing^
I and Analysis

[ Integration of Information Ability to access through a single interface, all applications and |
I and Applications information required for increased job throughout J

^Publish and Subscribe Maturation of business processes by collaborating with others, sharing ]
information, and improving business performance J

^Personalization Arranging the interface to meet an individual’s needs and desires for |
increased job productivity J

FIGURE 13.8: Portal Features and Their Corresponding Benefits

SOURCE: Bolds, Robert. “Enterprise Information Portals: Portals in Puberty,” KMWorld, July/August 2001,
www.kmworld.com/publications/whitepapers/portals/bolds.htm, Date accessed December 2002.

ment in mind is much higher than in infrastructures that do not support this or that
make it hard to use collaborative features. In a well-designed collaborative environ-
ment, this knowledge flow can be easily captured in e-mail, stored in document and dis-
cussion databases, and archived in the knowledge management system for later reuse.
Collaboration in the knowledge management context is the ability for two or more
people to work together in a coordinated manner over time and space using electronic
devices. One has to distinguish between two types of collaborations: asynchronous and
synchronous collaboration.
• Asynchronous collaboration is human-to-human interactions via computer sub-
systems having no time or space constraints. Queries, responses, or access occur
anytime and anyplace.
• Synchronous collaboration is computer-based, human-to-human interaction that
occurs immediately (within 5 seconds). It can use audio, video, or data technologies.
Table 13.2 describes the requirements for a successful collaboration component.
A nother im portant distinction to make is that of “push” versus “pull” technologies.
Push technology places information in a place where it is difficult to avoid seeing it.
E-mail is a classic example of a push technology. Pull technologies require you to take
specific actions to retrieve information. The World Wide Web is a good example of a
pull technology. An electronic mailing list that uses the push technology of e-mail is
extremely powerful as a collaborative tool because it requires little learning or behav-
ior change on the part of the user. In spite of its relative simplicity when compared to
sophisticated Web-based collaborative forums, electronic mailing lists are the most
popular collaborative tool.
Table 13.3 shows the advantages and disadvantages of the different technologies
available to perform asynchronous and synchronous collaboration.
394 7777 PART IV K M System Tools and Portals

1 ■ fl fl I ■ BOX 13.5 fl fl fl fl I fl

LAYERS OF THE PORTAL ARCHITECTURE

MICROSOFT’S KM PLATFORM • The knowledge services layer provides important

knowledge management services such as collabo-
The knowledge m anagem ent platform offers a ration, document management, and search and
typical, but extended, 3-layered architecture that deliver functionality, with modules for tracking,
allows a company to build a flexible, powerful work flow, and data analysis.
knowledge management solution that is capable • The system layer is a foundation that includes
of being scaled. The knowledge desktop layer administration, security, and directories for manag-
consists of familiar productivity tools, such as ing the knowledge management platform. All ser-
Microsoft Office, and integrates tightly with the vices run on the.system layer and benefit from the
knowledge services layer. integrated communication services that connect
with external solutions, platforms, and partners.

Knowledge Knowledge Portal

Desktop Knowledge Tools

Search and Deliver

Collaboration Document Data

Management Warehousing
Knowledge Business and
Services Intelligence
(data analysis)

Tracking and Work Flow

Business Logic Security Replication

System Communication Directory Administration

SOURCE: Honeycutt, Jerry. “Knowledge Management Strategies,” Redmond, WA: Microsoft, p. 167.

a a a a a a

CONTENT MANAGEMENT
Content managem ent requires directory and indexing capabilities to automatically
manage the ever-growing warehouse of enterprise data. This component addresses the
problem of searching for knowledge in all information sources in the enterprise. This
knowledge includes structured and unstructured internal information objects such as
office documents, collaborative data, MIS and ER P systems, and experts, as well as
information from outside sources. This component handles how knowledge assets get
into the knowledge management information base. To handle this new complexity of
 CHAPTER 13 Knowledge Management Tools and Knowledge Portals aaaa 395

TABLE 13.2 Requirements for Successful Collaboration Tools

Requirements fo r Successful Collaboration
• Comfortable e-mail systems that support collaborative services such as shared calendars, task
lists, contact lists, and team-based discussions
• A Web browser for browsing and presenting the documents to the user
• Simple search functionalities such as file services integrated with the operating system or
search services integrated with an application (e-mail, discussions)
• Collaboration services with a multipurpose database for capturing the collaborative data
• Web services for providing the access layer to documented knowledge
• Indexing services for full-text search of documents
• Well-organized central storage locations such as file, Web server, and documents databases
! ■ ■ ■ ■ ■ ----------------------------------------------
SOURCE: Honeycutt, Jerry. “Knowledge Management Strategies,” Redmond, WA: Microsoft, p. 176.

TABLE 13.3 Advantages and Disadvantages of Synchronous and Asynchronous

Collaboration Tools
Synchronous Collaboration Asynchronous Collaboration
Teleconferencing Electronic Mailing Lists
In use extensively by senior management and staff, Lists have been in use for a number of
conference telephone calls represent an effective years and represent an extremely cost-
(if relatively expensive) collaboration technology. effective collaboration technology.
Advantages: personal, immediate feedback Advantages: cheap
Disadvantages: expensive, often does not work well Disadvantages: limited communication
across time zones medium

Computer Video/Teleconferencing Web-Based Discussion Forums

Computer-based teleconferencing and video- There are a number of different online
conferencing is a rapidly evolving technology discussion forum applications in use.
that has tremendous potential for distributed Advantages: same as electronic mailing
organizations. lists except requires slightly faster
Internet connection
Online Chat Forums Disadvantages: cultural resistance
Allow multiple users to communicate
simultaneously by typing messages on a Lotus Notes
computer screen. Lotus Notes is a comprehensive
collaboration tool that includes
e-mail and groupware.
Advantages: comprehensive
collaborative solution employing
state-of-the-art technologies for
communication, document manage-
ment, and work flow
Disadvantages: expensive to deploy
when compared with other collabo-
ration technologies
m m m m m a ---------------------------------------------
396 a a a a PART IV K M System Tools and Portals

the knowledge management information base and to help the knowledge workers to
stay focused on solving business problems (without disappearing in technology), a
sophisticated knowledge management taxonomy needs to be built based on metadata
(data that describes other data). M etadata is needed to define types of information.
The content management component also needs to publish information in the knowl-
edge base (for example, categories and attributes).
A nother im portant issue handled by content management is the way documents
are analyzed, stored, and categorized. Once the documents have been gathered, they
must be analyzed so that their content is available for retrieval and use by the system
or end users. As documents enter the portal system, they are stored for later retrieval
and display. However, it is not useful to simply put the documents away in their raw
form. Systems typically analyze the docum ent content and store the results of that
analysis so that subsequent use of the documents by the system and users will be more
effective and efficient.
As the number of documents under management grows, it becomes increasingly
important to gather similar documents into smaller groups and to name the groups.
This operation is called categorizing. All automatic categorizing methods use features
to determine when two documents are similar enough to be put into the same cluster.
Because document collections are not static, portals must provide some form of
taxonomy maintenance. As new documents are added, they must be added to the tax-
onomy at appropriate places, using a classification technology. As the clusters grow,
and especially as the conceptual content of the new documents changes over time, it
may become necessary to subdivide clusters or to move documents from one cluster to
another. A portal administrator, using the taxonomy editor, can monitor and imple-
ment these suggestions and can periodically assess the health and appropriateness of
the current taxonomy and document assignments within it (Mack, Ravin, and Byrd
2001). In the publishing process, several things should be considered concerning the
knowledge m anagem ent taxonomy. Although tagging docum ents with m etadata is
important for the quality of content in this stage of document publishing, it is a burden
to submit information if tagging the metadata is a complex or time-consuming process.
This is where the XML language comes into play. Box 13.6 provides a brief description
of this language.

COLLABORATION VERSUS CATEGORIZATION: THE CASE OF THE WORLD BANK

The World Bank’s main activity has been that of lending. It lends some $15 billion to
$30 billion a year, receiving profits of about $1 billion a year. Knowledge management
has become a strategic thrust. In fact, the World Bank is regarded as a leader in the
knowledge management arena. In 1996, the World Bank was drowning in information,
and even though it was spending a fortune on classifying all this knowledge, there was
very little information flow within the organization (see Figure 13.9) (Sestina 2000).
An electronic document m anagement system (EDM S) represents a substantial
undertaking for any organization. The World Bank case provides a good example of
the inherent opportunities and difficulties. The bank employs a state-of-the-art, XM L-
enabled Oracle data engine to drive a document management system linked to Lotus
Notes groupware. In spite of the leading-edge nature of this technology, it is impossible
to get staff to consistently and correctly file documents. The bank has employed a staff
of 100 catalogers internationally to see that information is correctly codified. By ensur-
 CHAPTER 13 Knowledge Management Tools and Knowledge Portals aaa■ 397

a a ■ i i a box 13.6 a a a a a a

THE ROLE OF XML

WHAT IS XML? default definition for undeclared components of

the markup.
XML is a subset of the standard generalized
XML allows users to do the following:
markup language (SGML) defined in ISO standard
8879:1986 that is designed to make it easy to inter- • Bring multiple files together to form compound
change structured documents over the Internet. documents
XML files always clearly mark the start and end of • Identify where illustrations are to be incorporated
each of the logical parts (called elements) of an into text files, and the format used to encode each
interchanged document. XML restricts the use of illustration
SGML constructs to ensure that fall-back options • Provide processing control information to sup-
are available when access to certain components of porting programs, such as document validators
the document is not currently possible over the and browsers
Internet. It also defines how Internet uniform re- • Add editorial comments to a file
source locators (URLs) can be used to identify
component parts of XML data streams. It is im portant to note, however, that XML
By defining the role of each element of text in is not:
a formal model, known as a document type defini- • A predefined set of tags, of the type defined for
tion (D TD ), users of XML can check that each HTML, that can be used to mark up documents
com ponent of the docum ent occurs in a valid
• A standardized template for producing particular
place within the interchanged data stream. An types of documents
XML DTD allows computers to check, for exam-
ple, that users do not accidentally enter a third- XML was not designed to be a standardized
level heading without first having entered a sec- way of coding text. In fact, it is impossible to
ond-level heading, som ething that cannot be devise a single coding scheme that would suit all
checked using the hypertext m arkup language languages and all applications. Instead, XML is a
(HTML) previously used to code documents. formal language that can be used to pass informa-
However, unlike SGML, XML does not re- tion about the component parts of a document to
quire the presence of a DTD. If no DTD is another computer system. XML is flexible enough
available, either because all or part of it is not to describe any logical text structure, whether it be
accessible over the Internet or because the user a form, memo, letter, report, book, encyclopedia,
failed to create it, an XML system can assign a dictionary, or database.

SOURCE: Bryan, Martin. “An Introduction to the Extensible Markup Language (XML),” SGML Centre, 1997,
www.personal.u-net.com/~sgml/xmlintro.htm, Date accessed December 2002.

a a a a a a

ing rigor in classification through the use of controlled vocabulary, resource discovery
is much more efficient, both in terms of searching and browsing.
With regard to building knowledge bases (comprehensive knowledge reposito-
ries), the overwhelming message from the World Bank is that this is a difficult, ex-
tremely labor-intensive process. After 4 years of investment in knowledge management,
398 ■ ■ ■ ■ PART IV K M System Tools and Portals

Knowledge

\nnovatioii

Oracle InterMedia for Indexing

XML as
Emerging
Language of
Exchange Lotus Notes:

i
Communication, Document
Management, Work Flow

Output
Strategy Lotus Notes PeopleSoft

•Lotus Notes SAP (systems,

• Oracle EDMS (electronic document applications, and
management system) products in data
processing)

Oracle 8 Database

1 m i ■ i i FIGURE 13.9: KM Architecture at the World Bank

SOURCE: Sestina, A., Graham, H. et al. “Technology for Knowledge Management,” www.bellanet.org/km/
report/KMFinall4.doc?ois=y;template = lists_word.htm, Date accessed August 2002.

the bank reports that they have had much more success with collaboration strategies
than with codification.
Although codifying knowledge is a fundamental activity of any knowledge-based
organization, the cost of com prehensive codification technologies such as EDMS
needs to be evaluated in terms of a return on investment with respect to both technol-
ogy and people (Denning 1998).

INTELLIGENT AGENTS
Intelligent agents are tools that can be applied in numerous ways in the context of EKPs.
As a tool, intelligent agents are still in their infancy. Most applications are experimental
and have not yet reached the efficient commercial stage. However, there is no doubt that
they will play a crucial role in all aspects of EKPs, especially in intelligent search and fil-
tering the right documents according to some criteria. One application is the use of intel-
ligent agents in ERPs such as CRM portals, specifically in areas of competition.
As the relationships between companies and their customers become more com-
plex, the enterprises need more information and advice on what these relationships
mean and how to exploit them. Intelligent agent technology offers some very interest-
ing options for addressing such needs.
Customers set certain priorities while purchasing products and services. Intelligent
agents can m aster individual custom ers’ or custom er groups’ dem and priorities by
learning from experience with them, and they can quantitatively and qualitatively ana-
lyze those priorities. Agents are software entities that are able to execute a wide range
 CHAPTER 13 Knowledge Management Tools and Knowledge Portals ■■* i 399
of functional tasks (such as searching, comparing, learning, negotiating, and collaborat-
ing) in an autonomous, proactive, social, and adaptive manner. The term intelligent in
this context means only that we are dealing with entities that are able to adjust their
behavior to the environment. In other words, intelligent agents are able to learn from
previous situations and replicate the behavior of the customer to predict purchasing
patterns. There is a vast range of services customers require that intelligent agents can
address. Some of these services may include the following:
• Customized customer assistance with online services: news filtering, messaging,
scheduling, making arrangements for gatherings, ordering, and so on
• Customer profiling, including inferring information about customer behavior
based on business experiences with the particular customer
• Integrating profiles of customers into a group of marketing activities
• Predicting customer requirements
• Negotiating prices and payment schedules
• Executing financial transactions on the custom er’s behalf
These examples represent a spectrum of applications from the somewhat modest,
low-level news filtering applications to the more advanced and complicated customer
relationship management applications that focus on predicting customer requirements.
The main point is that an intelligent agent is an intermediary between the enterprise
and its customer, and a source of effective, utilitarian information encountered at dif-
ferent virtual destinations.
Figure 13.10 shows the new technology trends in im plem enting portals. The
emphasis is on collaborative technologies to create communities of practice; advanced
human computer interaction to enhance performance; and intelligent agents to auto-
mate the search function.

■■■« Implications for Knowledge Management

WHO IS BUILDING ENTERPRISE PORTALS?
META Group conducted a survey of 350 organizations (respondent organizations had at
least 500 employees) to find how widely enterprise portals are being deployed, how the
portals are being utilized, and what services and vendors organizations use for portal
deployment. More than 80 percent of the respondents knew the term portal, and one in
three currently either have a portal installed or have one under development. Large orga-
nizations (more than 10,000 employees) have a significantly higher portal installation rate.

WHO SPONSORS ENTERPRISE PORTALS?

New portal sponsorship is shifting away from IT —which drops from 63.3 percent of
sponsorship of existing portals to 46.6 percent for planned portals —and toward the
marketing/sales group (20.6 percent versus 6.7 percent), which is a strong indicator of
the influence of e-commerce on portal development. The line of business has increased
from 2.5 percent of share to 9.2 percent, while the corporate staff sponsorship has
remained the same.

IMPLEMENTATION ISSUES
Although technology issues can be categorized in many different ways, the codification
versus collaboration paradigm also provides a particularly useful structure for under-
standing current trends in information technology. For globally distributed organizations
400 * * * * PART IV K M System Tools and Portals

Portal & New Technology Directions

Global, just-in-time knowledge

sources and services

Analytic Tools User-, task-, and situation-

Intelligent Training tailored interaction
Collaborative Learning
Human Computer
Performance Support Interaction

Collaborative Filtering Multimedia

Information Brokers Multilingual
Knowledge Integration Multidocument

Knowledge Management Digital Libraries

Seamless collaboration across Intelligent agents to monitor, filter, search, extract,

geographic, temporal, organizational, translate, fuse, mine, visualize, and summarize
and mission boundaries information for a variety of operational needs

Collaborative Environments Intelligent Agents

■ ■ i l l ! FIGURE 13.10: New Trends in Portal Technologies

SOURCE: Conover, Joan. “Using Portal Technology to Fuse Corporate Information Knowledge
Management, Information Management, Data Warehousing,” New Technology Digital Library.
www.c3i.osd.mil/km/proceedings/53.ppt, Date accessed August 2002.

(that is, most international development organizations) that rely on the Internet as a
medium for the sharing of knowledge, the issue of bandwidth is fundamental. At this point
in the evolution of the Internet, bandwidth is a chief constraining factor for many applica-
tions. The determination of an organization’s overall KM strategy will provide guidance for
the implementation of appropriate technology. Hansen, Nohria, and Tierney (1999) pre-
sent a valuable model to help guide thinking about managing organizational knowledge by
distinguishing between ‘"codification” versus “personalization” strategies. This dichotomy is
useful in making the critical decisions required to ensure the right technological mix.
Codification focuses primarily on computer use, whereby knowledge is carefully coded
and stored in databases for easy access (see Chapters 7 and 9). By contrast, the personaliza-
tion KM strategy makes use of computers to help people communicate knowledge, not to
store it. The emphasis is on knowledge sharing via direct person-to-person contacts.

BANDWIDTH
Current trends point toward a steady decrease in the cost of Internet access. The rapid
and pervasive spread of Internet communication coupled with the evolution of faster
and cheaper technology is resulting in improved access to the Internet at lower costs.
 CHAPTER 13 Knowledge Management Tools and Knowledge Portals ai aa 401
This trend has been slowest at m anifesting itself in Africa. However, even their
Internet access is spreading rapidly and is becoming much less expensive, especially in
capital cities.
Given the importance of collaboration and the creation of communities of practice
as a method for knowledge sharing, it is worth investigating the costs of a significant
increase in bandwidth for regional offices that could support the following:

• Desktop videoconferencing
• Internet telephony
• Improved access to information systems based at headquarters
• Other collaborative tools
• Access to more sophisticated information resources

Most of the technological tools now available tend to help dissemination of know-how,
but they offer less assistance for knowledge use. Tools that assist in knowledge creation are
even less developed, although collaborative work spaces offer promising opportunities by
enabling participation, across time and distance, in project design or knowledge-base
development; those most knowledgeable about development problems—the people living
them on a day-to-day basis—can actively contribute to their solutions. Some of the more
user friendly technologies are the traditional ones—face-to-face discussions, the telephone,
electronic mail, and paper-based tools such as flip charts. Among the issues that need to
be considered in providing information technology for knowledge-sharing programs are
the following:

• Responsiveness to user need. Continuous efforts must be made to ensure that the
information technology in use meets the varied and changing needs of users.
• Content structure. In large systems, classification and cataloguing become impor-
tant so that items can be easily found and quickly retrieved.
• Content quality requirements. Standards for admitting new content into the system
need to be established and met to ensure operational relevance and high value.
• Integration with existing systems. Because most knowledge-sharing programs aim at
embedding knowledge sharing in the work of staff as seamlessly as possible, it is
key to integrate knowledge-related technology with preexisting technology choices.
• Scalability. Solutions that seem to work well in small groups (such as HTML
Web sites) may not be appropriate for extrapolation organization-wide or on a
global basis.
• Hardware-software compatibility. Hardware-software compatibility is important
to ensure that choices are made that are compatible with the bandwidth and com-
puting capacity available to users.
• Synchronization o f technology with the capabilities o f users. This is important so as
to take full advantage of the potential of the tools, particularly where the technol-
ogy skills of users differ widely.

PORTAL PRODUCT SELECTION

Creating an enterprise portal vendor shortlist can save time and effort typically devoted
to large-scale market scans. Once a portal vendor shortlist has been created, each ven-
dor should be evaluated through the use of scenarios that test key business functional-
ity, based on architectural fit, vendor viability, and product features. Table 13.4 shows the
features of some products and the best way to use them.
402 aaaa PART IV K M System Tools and Portals

TABLE 13.4 Portal Vendors

K M Portal
Vendor Product Feature Summary Best Uses
Lotus/IBM Lotus Raven 1.0 • Intelligent taxonomy; builder • Self-creating and refining taxonomies
(in beta) automates document clustering; learns • Personnel resources linked to
your preferences; expertise locator data sources
links authors to content • Advanced collaboration
• QuickPlace collaboration tool for • Easy portal repurposing
instant messaging, app sharing, and • Rapid application development
virtual meetings with associated KM packages
• Assigns value to data based on how
often it is used
• Portal replication
• Facilitates content management

Open Text MyLivelink • Integrated work flow and project • Integrated KM

Portal 1.0 with management with task lists and Java- • Document management and work flow
Livelink 8.5.1 based process maps • Custom collaboration spaces
KM software • Quick integration of features from (personal, project, or enterprise)
other apps via widgets
• Quick portal deployment with more
than 600 syndicated application
com ponent gadgets, which can be
freely distributed

Plumtree Plumtree • Automatic population of mapped • Easy and extensive content and
Corporate taxonomies with filtered content application integration
Portal 4.0 • E-mail, voice, and wireless notification • Scalability
of content updates • Advanced security and access control
• Integration with LDAP directories • Trainable taxonomies
• E-room tools for collaboration and • Voice and wireless data access
virtual communities • Customization and extensibility

Woolamai WebMeta • Quick integration of corporate portals • Usability

Engine 1.0 with data and other apps without • Tracking site statistics
specialized training • Content streaming to wireless devices
• Intuitive, flexible portal interface
• Knowledge taxonomy adapts to data
views; provides relevance filtering
• Data-mining functionality
• Web site statistics monitor site activity
......... m m m m m m

Su mma r y a1 1a
• A portal is a secure, Web-based interface that provides a single point of integra-
tion for and access to information, applications, and services for all people
involved in the enterprise, including employees, partners, suppliers, and customers.
• Born with search engines such as Yahoo! and Alta Vista, portals have since made
their way into enterprises, bringing together not only information from the
Internet, but in-house data as well. These portals, which are known as enterprise
 CHAPTER 13 Knowledge Management Tools and Knowledge Portals a a a a 403
knowledge portals (EKPs), aim to offer a single, uniform point from which all of
an enterprise’s data sources can be accessed.
• The term data sources encompasses structured data (such as databases, Lotus
Notes, and so forth) and unstructured data (such as e-mails, files, and archives); it
also includes the data resulting from specific processes and enterprise applica-
tions (such as ER P and CRM tools). Today, the EIP market is thriving, and many
vendors are betting big on portals’ well-founded ability to fulfill enterprise needs.
• Content management in the EKP context requires directory and indexing capa-
bilities to automatically manage the ever-growing store of structured and unstruc-
tured data residing in data warehouses, Web sites, ERP systems, legacy applica-
tions, and so forth. Using m etadata to define types of information, good content
management can serve as the backbone for a system of corporate decision mak-
ing where business intelligence tools mine data and report findings back to key
players in the enterprise. Content management may also involve going outside
the enterprise, employing crawlers that find pertinent data via the Internet, incor-
porating it into existing systems, indexing it, and delivering it to appropriate ana-
lysts, knowledge workers, or decision makers.
• The collaborative functionality of EKPs can range from tracking e-mail to devel-
oping workplace communities. Some EKPs might allow workers in different parts
of the world to easily create virtual meeting rooms where they can conference by
chat, voice, or video communication.

Terms t o K now • • • •
Bandwidth: The term used to describe the speed of a net- Enterprise knowledge portal: An electronic doorway into
work connection. A fast connection allows the user to a knowledge management system.
view images and videos, and interact with remote sites Extensible markup language (XML): A specification
as if they were local. developed by the W3C designed especially for Web
Browser: A software application used to locate and dis- documents.
play Web pages. Groupware: A generic term used to describe applications
Content management system (CMS): A system used to that support and facilitate the work of teams and
manage the content of a Web site. groups of people.
Document management system: The computerized Intelligent agents: Programs used extensively on the Web
management of electronic as well as paper-based that perform tasks such as retrieving and delivering
documents. information and automating repetitive tasks.
Electronic collaboration: A process through which pro- Intranet: A network based on TCP/IP protocols (the
ject partners can contribute jointly to works in progress Internet) belonging to an organization, usually a corpo-
via e-mail, groupware, and public networks. ration, accessible only by the organization's members,
Electronic mailing list: A list of e-mail addresses identi- employees, or others with authorization.
fied by a single name, such as [email protected]. Knowledge repository: The software system that is a col-
When an e-mail message is sent to the mailing list lection of internal and external knowledge inside the
name, it is automatically forwarded to all the addresses knowledge management system.
in the list. Knowledge repository model: Model that describes the
Enterprise application integration (E A I):T he process that 2-step knowledge transfer procedure of person-to-
integrates applications inside an organization or appli- repository and repository-to-person.
cations of many organizations in a seamless fashion. Metadata: Data about data, such as indices or summaries.
Enterprise information portals: Portals that tie together Organizational knowledge base: The collection of knowl-
multiple, heterogeneous internal repositories and appli- edge related to the operation of an organization.
cations as well as external content sources and services Personalization: Software systems that allow an Internet
into a single browser-based view that is individualized site to provide the user with a Web page that reflects
to a particular user’s task or role. the interests, needs, and actions of the user.
404 a a a a PART IV K M System Tools and Portals

Portal: A Web page that offers links to other Web sites. Video mail: Application that enables users to transmit
Portals can be broad or narrow, specific or general. photos, sounds, and video.
Search engines: Software agents whose task is to find Voice portals: Gateways that allow customers and
information by looking at keywords or by following employees to use ordinary phone lines to access and
certain guidelines or rules. request information.
Teleconferencing: Conferring wjth a number of people via Work flow: The defined series of tasks within an organiza-
telephone or computer systems. tion to produce a final outcome.
Vertical portal: Electronic exchanges that combine
upstream and downstream e-commerce activities of
specialized products or services.

Te s t Yo u r U n d e r s t a n d in g 1111
1. Why is there a need for portals? How are portals similar to the concept of
data warehouses and data marts?
2. W hat are the advantages and disadvantages of having your portal on the
Internet instead of an intranet?
3. List the differences between knowledge and information portals. Discuss the
benefits of each.
4. Discuss the strategic and technological fit required for an organization to
implement a portal.
5. Discuss the advantages and disadvantages of purchasing a portal from a ven-
dor. Make sure you explore vendor Web sites such as Viador (www.viador.com)
and Autonomy (www.autonomy.com).
6. Discuss the differences between static and dynamic portals. When would you
use each each?
7. Discuss how you can use content management to sort knowledge from exter-
nal and internal sources. Illustrate with examples.
8. Discuss the issues that can arise when implementing a portal. Focus on tech-
nology, management, corporate strategy, and end users.
9. Give examples and uses of portals for B2B, B2C, B2G, C2C, and C2G.
10. List a number of possible ways a portal can be made accessible, given current
technological trends. Focus on five of these technologies and discuss their
strengths and weaknesses.

Kn o w l edg e Ex e r c is e s 1111
1. In the past, companies used electronic data interchange (ED I) to communi-
cate with suppliers and customers. Discuss how portals can be used to replace
the functions of EDI.
2. An audit firm needs to develop a system that allows auditors and public
accountants to search accounting standards, share knowledge, communicate,
and share Word and Excel files between the head office and clients’ sites. As
a consultant, you have been asked to recommend such a system. What would
you suggest?
3. A hardware retailer wishes to offer real-time support to customers via the
Internet. Suggest how a knowledge portal, equipped with chat and CRM, can
be used to accomplish this. W hat additional support can the hardware
retailer offer? What information can he give to the manufacturer?
4. A m ultinational conglom erate has a centralized human resources depart-
ment in Cleveland, Ohio. The human resources director wishes to launch a
new set of multilingual policies to all employees, according to their function,
CHAPTER 13 Knowledge Management Tools and Knowledge Portals ■i ■a 405
category, and grade. The director also wishes to have employees interact and
give feedback on the policies. Suggest a computerized solution.
5. Discuss how synergy between different strategic business units can be har-
nessed and utilized by knowledge portals.
6. Discuss how portals can offer a solution to the centralized versus decentral-
ized information dilemma. What forms of knowledge can be collected cen-
trally, and what should be left decentralized? Why?
7. How can personalized portals use data-m ining techniques? Suggest how
knowledge management and data mining can be integrated on a portal, and
give supporting examples.
8. Challenge: Establish e-learning portals for customers and partners to help
those audiences succeed with their NCR products and to generate new reve-
nue for NCR.
Strategy:
Use THINQ e-learning solutions to launch and track courses and pro-
vide community features—such as chat rooms and message boards—around
the courses.
Results:
NCR has extended nearly 4,000 online and classroom courses to more
than 2,000 registered users and is meeting its e-learning revenue goals.
NCR has come a long way in the 117 years since it introduced the first
mechanical cash register, but the $6 billion com pany’s trem endous success
still hinges on the individual customer relationship. Today, NCR helps compa-
nies harness the vast amount of customer information they collect at the sales
counter, over the phone, at the ATM, and on the Internet. With this informa-
tion, businesses can satisfy each customer’s unique needs, often automatically,
and transform customer transactions into rich customer relationships.
From the dawn of the cash register through N C R ’s leadership in data
warehousing, the com pany’s offerings have always been sophisticated.
Accordingly, it has always been important for NCR to help customers learn
how to reap the full potential of their NCR hardware and software. It is also
important for NCR partners to completely understand the company’s prod-
ucts so they succeed in selling and implementing them.
NCR has traditionally offered classroom training for customers and part-
ners to achieve these goals. Recently the company started offering courses
over the Internet. These classes combine the incisive content of NC R ’s class-
room training programs with the reach and efficiency of the worldwide net-
work, letting students anywhere in the world take a class any time they can
access a Web browser. In addition to helping customers and partners succeed,
this e-learning program also generates new revenue for NCR.
NCR evaluated a number of e-learning tools for the job of powering its
two custom er and partner e-learning portals. One portal is the Teradata
Education Network (TEN), an e-learning Web site for the company’s data-
warehousing customers. The other is the external NCR University (NCRU),
which extends award-winning NCR employee e-learning to partners. After a
rigorous review, NCR selected THINQ to build and power the portals, which
offer customers and partners Web-based, self-paced training, course tracking,
employee-learning reporting, live virtual classes, hosted educational chats, a
reference library, message boards, and instructor-led courseware registration.
“Em ployee training is im portant, but it just scratches the surface of
e-learning’s potential,” said Janet Perdzock, program manager, Global
Learning Operations of NCR. “These initiatives improve our customers’ and
partners’ businesses, enrich our relationships with those audiences, and generate
406 ■ i ■ * PART IV K M System Tools and Portals

significant new revenue for us. THINQ provides a bridge to our courses, con-
tent, and other back-end systems and is a key ingredient in our success.”
NCR chose THINQ over other e-learning vendors because of T H IN Q ’s
long experience in the learning industry, its integrated product, reputation,
flexibility, affordability, and ability to interoperate with NCR’s existing infra-
structure. TH IN Q ’s satisfied customers also tipped N CR’s decision.
The NCRU portal for partners went live in July 2001 and offers 3,600
courses, including 859 NCR proprietary Web-based training courses, NETg
inform ation technology courses, NETg desktop computing courses, and
multilanguage courses. The portal offers CDs, books, tapes, and classroom
courses —all filtered based on each com pany’s profile. Since the launch,
NCRU has signed up 766 users at 49 companies who have completed 375
courses. They can create individual learning plans, register online, and view
their entire training histories. Counting users who registered for a previous
“interim ” site prior to July 2001, the NCRU portal for customers and part-
ners has 1,528 authorized users.
N C R ’s Teradata Education Network (TEN) learning environment went
live March 30, 2001, to its membership-based learning community. Teradata
Education Network is exceeding all projections with more than 1,300 mem-
bers, more than 1,000 user sessions every week, and more than 660 course
completions in the first 6 months. One-third of all Tbradata companies have
an associate who is a member of TEN, and TEN is on track to meet its reve-
nue goal for 2001. A 13-month membership to the site comes in many shapes
and sizes, depending on the custom er’s need. An individual membership can
cost as little as $895, and a corporate membership can cost as little as $6,795.
Not only do members receive access to the learning community, but they also
are provided with access to over 50 Teradata courses. Members get into mes-
sage boards, access white papers, take virtual classroom courses, and review
recorded virtual classroom presentations.
Unlike most online training programs, Teradata Education Network
allows students from around the globe to communicate with other students
and make direct contact with instructors. One of the most powerful aspects
of the network is the access to knowledgeable Teradata professionals world-
wide, giving students a chance to stay current in an ever-changing market.
In the future, NCR plans more curriculum-mapping/skill-building capa-
bilities and additional, customized portals for specific customers or partners.
It also plans thorough profiling of individual users and corporations to better
meet their e-learning needs.
“Customers were asking for an alternate way to train without leaving the
office that would supplement their classroom learning,” said Adam Zaller,
program manager of e-learning for Teradata Customer Education. “Our cus-
tomers say the Teradata Education Network learning community is like hav-
ing ‘Partners,’ our annual user conference, 365 days a year. It’s a community
of customers helping each other, a revenue stream for the company, and a
great way to help our technology solve critical business problems in the real
world. Our customers and our prospects are now a lot more aware of educa-
tion and what we have to offer. We anticipate usage to continue to grow, reve-
nue to expand, and the program to eventually cover the globe.”
a. Discuss the advantages of Teradata as a learning option. What advantages
does it have over conventional learning in terms of content delivery, con-
venience, and growth opportunities?
b. Discuss the possible disadvantage of having a purely electronic learning
solution. How can the human element be incorporated?
CHAPTER 13 Knowledge M anagement Tools and Knowledge Portals i i i ■ 407
c. Suggest ways for NCR to incorporate curriculum-mapping/skill-building
capabilities and customized portals for specific customers or partners, as
mentioned in the case.
d. Suggest ways in which Teradata can be used to train some of its divisions
and departments. Can Teradata also be used to train line, middle, and
upper management? If so, how can they go about this, and what content
can they use?
SOURCE: THINQ 2002
9. With the help of a corporate portal, the U.S. Postal Service (USPS) delivers
quality KM.
A t many organizations, corporate portals are seen as a convenient way
to centralize proprietary inform ation and m ake it easily accessible to
employees. As such, portals can serve as an ideal knowledge management
tool where employees can tap into a wealth of corporate know-how.
Unfortunately, many portals fall short. Instead of serving as sleek KM vehi-
cles, all too often portals resem ble black holes where inform ation gets
dumped, never again to see the light of day.
For those striving to dust off their dormant corporate portals and trans-
form them into a KM tools, John Gregory has a few sound words of advice.
Gregory, a m arket research analyst for the U.S. Postal Service in Arlington,
Virginia, is in charge of MarketTracks, a knowledge retrieval and competitive
intelligence portal used by 1,000 sales and marketing employees.
Since 1994, the postal service has offered employees centralized sales and
marketing information, first in the form of a client/server system and then,
beginning in 1997, on the Web. Over the years, the organization has honed a
practical strategy for creating useful, relevant online resources that actually work
as advertised—they help employees do their jobs rather than hinder them.
As Gregory sees it, there are a few mysteries to solve when it comes to
figuring out how to create and sustain a useful portal. When considering con-
tent, Gregory assembled focus groups of users and asked them what they
wanted to get out of a portal. In most organizations where 1,000 employees
are the target audience, not everyone needs or wants to see the same infor-
mation. As a result, personalization of content became an im portant criteria.
Once he examined the content issues, Gregory turned his attention to ven-
dors and specific technologies. Vendor selection, Gregory says, is an area hav-
ing many pitfalls, but it is one that often gets short shrift. As a result of lack-
adaisical vetting processes, many companies are saddled with software that
does not fit their needs or vendors that do not work well with them. The end
result is money down the drain. “It can be enormously expensive to roll out a
full-featured portal,” Gregory says. “But it doesn’t have to be that expensive.”
For example, Gregory ruled out Plum tree’s portal software because the
postal service did not need its entire suite and balked at the prospect of pay-
ing for add-ons as they were introduced. Instead, the postal service served as
a beta tester for Epicentric, trying out the company’s enterprise portal soft-
ware among a small group of employees.
“T here’s a lot to be said for being an early adopter,” Gregory says,
adding that the chance of having a good relationship with a vendor is greatly
enhanced by doing so.
Even with careful vendor choices, Gregory says the odds are against
most portals. “A major flaw is design,” he says.
Gregory says that good design means first figuring out what people are
going to do with a portal by focusing on function rather than content.
Essential to function is navigation that is well planned and efficient.
408 ■ ■ ■ ■ PART IV K M System Tools and Portals

“Navigation isn’t just a box marked ‘search,’” he says. “There’s got to be the
taxonomy, links, site map, and a help feature as well.”
The navigation should never become static, however. As organic entities,
portals need to change and adapt as the organization does, a lesson that the
USPS is putting to good use.
“People think of a portal as a database and end up putting up everything
that they’ve got instead of what users need,” Gregory says. “A portal really is
an organic corpus of knowledge.”
S O U R C E : Santosus 2002.
a. Suggest ways in which content can be managed so that it can be person-
ally available to an employee.
b. Discuss the advantages and disadvantages of being an “early adopter”
and how it may impact organizations such as USPS.
c. Discuss the advantages and disadvantages of designing a portal on the
basis of functionality instead of content.
d. What does the author mean by “navigation should never become static”?

Re f e r e n c e s 1 1 1 1
Bolds, Robert. “Enterprise Information Portals: Portals in Honeycutt, Jerry. “Knowledge Management Strategies,”
Puberty,” white paper, K M W orld, vol. 2, July/August Redmond WA: Microsoft, 2000, p. 167,176.
2001, www.kmworld.com/publications/whitepapers/ IBM. “Portals: An Overview,” www-l.ibm.com/services/
portals/bolds.htm, D ate accessed August 2002. kcm/cm_portal.html, Date accessed August 2002.
Bryan, Martin. “A n Introduction to the Extensible IBM. “Portals, Knowledge, and Content M anagement,”
Markup Language (XM L) ” SGML Centre, 1997, www-l.ibm.com/services/kcm/know_mngt_con.html,
www.personal.u-net.com/~sgml/xmlintro.htm, Date D ate accessed D ecem ber 2002.
accessed D ecem ber 2002. IDC, “Knowledge Management Survey,” 1999.
Conover, Joan. “Using Portal Technology to Fuse Mack, R., Ravin, Y., and Byrd, R. J., “Knowledge Portals and
Corporate Information Knowledge Management, the Emerging Digital Knowledge Workflow,” IBM
Information Management, Data Warehousing,” New Systems, vol. 40, no. 4,2001, research web. watson. ibm.com/
Technology Digital Library, www.c3i.osd.mil/km/ journal/sj/404/mack.pdf, Date accessed August 2002.
proceedings/53.ppt, D ate accessed August 2002. Meta Group, “Business Collaboration,” www.metagroup.
Davydov, Mark M. “The Second Wave of E IP ” Intelligent com/cgi-bin/inetcgi/commerce/productDetails.
Enterprise, vol. 3, no. 4, March 1,2000. jsp?oid_29277, Date accessed August 2002.
Denning, Stephen. “What Is Knowledge M anagement?” Santosus, Megan. “In the Know,” CIO.com, February 19,
World Bank Knowledge Management Board, 1998, 2002. www.cio.com/knowledge/edit/K021902_portal.html.
www,apqc.org/free/whitepapers/index.htm, Date Date accessed December 2002.
accessed August 2002. Sestina, A., Graham H., Weaver Smith, B., and Sorg, S. “Tech-
E-G ov Knowledge Management Conference. “Using nology for Knowledge Management,” www.bellanet.org/
Army Knowledge Online for Mission Success,” km/report/KMFinall4.doc?ois=y;template=lists_word.
April 11,2000, www.army.mil/ako/webbriefing.ppt, htm, Date accessed August 2002.
D ate accessed August 2002. Shein, E., “The Knowledge Crunch,” CIO, May 1,2001,
Enterprise Knowledge Portal, www.askmecorp.com, Date www.cio.com/archive/050101/crunch.html, Date
accessed August 2002. accessed August 2002.
“Enterprise Knowledge Portals to Becom e the Shared Shilakes, C. C., and Tylman, J. “Enterprise Information
Desktop of the Future,” www.itweb.co.za/office/bmi/ Portals. In: In-depth Report,” Merrill Lynch, 1998,
9903300919.htm, D ate accessed August 2002. www.sagemaker.com/company/downloads/eip_indepth.
Finkelstein, C. “Building Enterprise Portal Using XM L,” pdf, D ate accessed 2002.
The D ata Warehouse Institute Conference, 2001. T H IN Q ,“Case Study: NCR,” www.thinq.com/pages/
Firestone, J. “Enterprise Knowledge Portals,” White Paper 8, cus_case_ncr.htm, 2002, D ate accessed December
www.dkms.com, D ate accessed August 2002. 2002 .
Hansen, Morten, Nohria, Nitin, and Tierney, Thomas. www.brint.com, D ate accessed August 2002.
“What’s Your Strategy for Managing K nowledge?” Yahoo! Hom e Page, www.yahoo.com, Date accessed
Harvard Business R eview , March-April 1999, p. 106. September 5,2002.
 PARTV: E T H I C A L , L E G A L ,
A N D M A N A G E R I A L I SSUES

Who Owns
Knowledge?
Ethical and
Legal Issues
Contents

In a Nutshell
Knowledge Owners
Knowledge for Sale
Releasing Knowledge Gained on the Job
Becoming an Expert via a Corporate Knowledge-Based System
Legal Issues
The Liability Question
The Basis of Liability
Copyrights, Trademarks, and Trade Names
Warranties
Strict Liability
Legal Disputes in Knowledge Management
Web Linking and Domain Name Disputes
The Malpractice Factor
The Ethics Factor
What Is Ethics?
Ethical Decision Cycle
Major Threats to Ethics
Improving the Climate
Where to Start?
Code of Ethics
The Privacy Factor
Implications for Knowledge Management
Summary

409
410 7777 PART V Ethical, Legal, and Managerial Issues

Terms to Know
Test Your Understanding
Knowledge Exercises
References

D o n ’t ask what you can do,

ask what you should do.
— H . Je f f er so n Sm it h

■■»■ In a Nutshell
Julie, a knowledge developer of a bank, was asked by her company’s marketing depart-
ment for access to the customer mortgage information base. They wanted to use this
information, which included household income, debt history, and other personal items, in
a marketing knowledge base to help target appropriate customers for new bank services.
Although several mortgage managers complained that such information was not
supposed to be used for marketing purposes, Julie delivered the information anyway.
She justified her actions by saying that it was not her job to rule on the uses of infor-
mation, but to expedite their exchanges.
The corporation was soon contacted by a local newspaper reporter, who had heard
of the marketing campaign. He wanted some details about the marketing plan for a
story on privacy and confidentiality. The result was a potentially embarrassing public
relations disaster that had to be quickly and deftly handled by the CEO. After the inci-
dent, the CEO reprimanded Julie for not taking responsibility to ensure proper infor-
mation flow. He made it clear that it was her duty to bring such'issues to the attention
of senior executives.
In this real incident, a custodian of a company’s knowledge base was confronted
with an ethical and potentially legal dilemma: how to manage sensitive information,
how to be the corporation’s knowledge conscience, and how to handle tough corporate
questions regarding the consequences of knowledge-based applications.
By adopting a proactive approach grounded in sound ethical and legal reasoning,
knowledge developers can extend their leadership from the technological arena into
the domain of knowledge ethics. Knowledge managers who assume the moral steward-
ship of knowledge initiatives can provide long-term benefits to both their departments
and their companies.
Unethical acts raise legal and litigation issues as well. Cases involving knowledge-
based systems will depend on who owns knowledge, who is held liable, and on what
basis liability is determined. The laws also vary depending on whether a knowledge
system is considered a product or a service. Warranties and the laws enforcing such
assurances protect employees and users in various ways.
The material covered thus far has examined the principles, procedures, tools, and
methodologies of knowledge management. Each chapter discussed the managerial
implications of this evolving concept and how it might serve productivity of the firm.
However, other questions arise at each stage of development: Who owns knowledge?
What are the ethical issues involved in building and managing knowledge-based sys-
tems? How does an organization address knowledge as a property right related to
employee knowledge or expertise? W hat are the legal im plications of knowledge
management?
 CHAPTER 14 Who Owns Knowledge? Ethical and Legal Issues a ■■■ 411

■ Knowledge Owners
In an oral culture where literacy does not exist, knowledge is not owned by anyone: It
is performed. For example, the skills involved in boat building are passed from one
craftsman to another through apprenticeship and tribal arrangements. In contrast, in
literate societies the invention of writing separated text from performance and knowl-
edge from the knower. A manuscript, for example, can be stored and retrieved. By sepa-
rating knowledge from the knower, knowledge can be owned by separate individuals.
The printing press later severed the connection between the author and the transmis-
sion of knowledge.
Following this separation, copyright laws were created in the eighteenth century as
a way of preserving intellectual property. The author began to get paid for his or her
ideas and was given the right to protect the work’s integrity by owning the right to
modify, correct, or retract the content. Knowledge is not what you see, but what you
say. Your knowledge is the result of your grasp of information from observing in the
field. Your knowledge is your own, because you arrive at it by yourself and draw your
own conclusions, (see Box 14.1) If you think that you can gain knowledge by simply
rem em bering what an expert says, then ask yourself if you could learn Russian by
memorizing a Russian dictionary.
In a corporate environm ent, the three possible owners of knowledge are the
expert, the employing organization, and the user who acquires the knowledge autom a-
tion system. O ther situations in which knowledge ownership could be an issue in the
future include the following:
1. If an expert sells personal knowledge
2. If an expert is unwilling to release knowledge gained on the job

KNOWLEDGE FOR SALE

In any environment, there is a price for knowledge (see Box 14.2). Imagine that a firm
has hired a consulting group to build a knowledge automation system for a new area of
the business. Because a resident expert does not exist, the consulting group is forced to
locate and hire an outside expert to build the system. In this scenario, the expert would
be the rightful owner of the knowledge if no prior agreements were established. Both
the company and the expert are open targets for liability and ownership lawsuits. If an
agreement made before developing the system allows the expert to receive royalties,
then ownership and liability are obvious. To avoid this situation, the company should
make arrangements to “purchase” the expert’s knowledge outright. Chances are that
experts may agree to it in order to escape future liability.

RELEASING KNOWLEDGE GAINED ON THE JOB

Charles Amble worked for General Motors (GM) as a maintenance engineer for more
than 20 years. To avoid losing his know-how, GM developed a knowledge automation
system nam ed Charley in order to capture A m ble’s expertise in vibration analysis.
Because this expert was willing to cooperate with GM during the knowledge capture
phase, knowledge developers were successful in building the system. The knowledge-
based system provided GM. with quick, standardized diagnoses of complex vibration
maintenance problems and substantial savings in all manufacturing plants, because all
plant mechanics were given access to the knowledge needed to diagnose these prob-
lems. Does the company have the right to Am ble’s knowledge? Who owns the knowl-
edge in this scenario?
412 7777 PART V Ethical, Legal, and Managerial Issues

i a fl I a a box 14.1 a a a a a a

CLARIFYING KNOWLEDGE OWNERSHIP

WHO OWNS KNOWLEDGE? take the drive to verify the information. Otherwise,
what you have is not knowledge but directions.
“Knowledge is not what you see, but what I say. ”
Your knowledge is your own, because you
If I said that, your answer would likely be, “Oh arrive at it yourself by making your own observa-
yeah?” You’d wonder who gave me the right to tions and drawing your own conclusions. You need
decide things for you. However, what if I said to check out the information provided by experts.
the same thing differently? “The storehouse of Pretending that experts implant knowledge in you
knowledge is too vast for any one person to com- is tempting when it seems like a quick way to get
prehend, so we must accept what the experts tell smart. In fact, it is a quick way to get dumb. It is
us.” In other words, “Knowledge is not what you confusing information with knowledge. It is recit-
see, but what the experts say.” ing lines and doing what you are told.
An expert is one who has good knowledge in Anyone can claim to be an expert, and many
some area. It is not your knowledge, but his. He do. If knowledge is owned by experts, then there
has observed, identified, confirmed, and com - must be an expert with the knowledge of who the
bined evidence into a grasp of the subject. If he experts really are. However, anyone could claim
tells it all to you, then he has informed you of his to be that expert, too. In the end, it will always be
knowledge. You have been given information. To up to you to decide for yourself who knows what
make the information into knowledge, you must they are talking about and who does not. How do
observe, identify, confirm, and combine it into a you do that? You check it out.
grasp of the subject. To check out information, you trace it to its
“I know how to get there,” says a friend. “Take base in sensory evidence and fit it in with every-
the Central Freeway south, get off at Main, and thing else. You see where it came from and how it
turn right.” “Okay!” you say. “Now I know how to relates. If it came from the unsupported word of
get there.” If you are already familiar with the area, an expert, then it is not knowledge but opinion.
you do know, because you can mentally verify the Informed opinion tells you where to look for evi-
information by checking it against your observa- dence. It cannot tell you what to think, no m atter
tions. If you are unfamiliar with the area, you must how much you may want it to do so.

SOURCE: Excerpted from Sedgwick, James. The Certainty Site, Oxford Press, 2001, p. 5, homel.gte.net/cpqlszzy/
whoownsknowledge.htm, Date accessed August 2002.

a a a a a a

Amble developed the knowledge by using the resources made available by the
firm, but he also developed personal reasoning and deductive abilities at the same
time. Unless he had signed a preemployment contract or intellectual property agree-
ment releasing his ownership of knowledge, he is the owner of knowledge. From the
expert’s point of view, a preemployment contract or intellectual property agreement
helps to limit personal liability for the knowledge automation system. GM would also
want an intellectual property agreement so that it could have title to the savings real-
ized by the knowledge automation system.
 CHAPTER 14 Who Owns Knowledge? Ethical and Legal Issues a i a ■ 413

1 1 i 1 1 1 BOX 14.2

WHAT PRICE KNOWLEDGE

The sale of assets from failed Internet start-ups Beyond book value, a large portion of a corpo-
has captured recent headlines, focusing public ration’s remaining worth lies in the knowledge it
attention on the difficulty of putting a price tag on possesses. Companies such as Qualcomm, Inc., and
intangible assets. Intellectual property represents Texas Instruments, Inc., have found greater value in
a growing proportion of a com pany’s value, licensing their know-how than in making products
whether that company is alive or dead. of their own, observers say. Houston-based Bryan
A key in determ ining true corporate worth Benoit is responsible for the intellectual property e-
is enterprise value, which for a publicly traded business (called the Intellectual Property Exchange)
corporation is simply what is left when its of PricewaterhouseCoopers LLP. He commented,
book value is subtracted from its m arket capi- “Intellectual property is difficult to value because
talization. “That num ber expressed as a frac- it’s an asset that can be used by more than one com-
tion of m arket cap has been steadily rising for pany in more than one way simultaneously.”
20 years,” says Alex Arrow, chief financial offi- As for the big picture, Benoit observes that
cer of the Patent & License Exchange, Inc., in two complementary forces are driving the interest
Pasadena, California. “G eneral Electric, for ex- in intangibles. “Our economy is knowledge-based,
ample, now has probably less than $40 billion and the trend is for it to become even more so
in cash, buildings, and machinery, a small frac- over time,” he says. “The other force is that be-
tion of its approximately $400 million in market cause of that first force, investors are becoming
capitalization.” more interested in the value of intangibles.”

SOURCE: Rapport, Marc. “The Price of Knowledge,” Knowledge Management, August 2001, p. 14.

a a fl fl i a

BECOMING AN EXPERT VIA A CORPORATE KNOWLEDGE-BASED SYSTEM

Jim DeMong has worked as a timber bucker for Roland Pacific for the past 12 years.
DeM ong’s job requires him to make many tactical decisions when cutting trees into
logs and then allocating them to appropriate m anufacturing facilities. As an expert,
DeMong knows just how to cut each tree into various log lengths and then distribute
the logs among the different mills in the most efficient and cost-effective way. Each
mill produces a different end product such as lumber, plywood, or paper. As a result,
his knowledge is invaluable to Roland Pacific. Wrong cutting and allocation decisions
could drastically change the profit contribution from each tree.
What is unique in this situation is how DeMong acquired his knowledge. Several
years ago, Roland Pacific introduced a knowledge automation system to train and assist
its timber buckers with their jobs. The new system was installed in the office where tim-
ber buckers spent a few hours a week simulating actual tree-cutting decisions and evalu-
ating the resulting costs associated with each decision. Because the system was
portable, it was also used in the field to make real-time cutting and allocation decisions.
DeMong used this knowledge-based system extensively during the first 2 years
with the firm. However, for the past few years, most of the decisions he made were his
414 77T7 PART V Ethical, Legal, and Managerial Issues

own. They were virtually identical to the system’s decisions. Does the company have a
right to DeM ong’s knowledge? Can the company prevent him from using this knowl-
edge if he leaves the company to start his own business or work for a competitor?
Again, unless DeMong signed some type of intellectual property agreement that
would release ownership of his knowledge, he owns the knowledge and has a right to use
it as he wishes. Intellectual property agreements are also important to the expert in order
to avoid strict liability. This brings up legal issues that are covered in the next section.

«««■ Legal Issues

Every legitimate business operates in a legal environment. Contracts, taxation, and
copyrights are among the issues that fall under the legal system. Many questions that
arise from knowledge ownership are not settled, but new laws can change the rules and
plug loopholes in this emerging area. In an age of prolific litigation, employees and
employers should be aware of the legal ramifications arising from knowledge sharing.
The excitement generated by knowledge capture and knowledge sharing has also
led to substantial financial investment in knowledge ownership and the technology that
makes it available. This includes e-mail infrastructure, intranets, stand-alone knowledge
automation systems, and knowledge repositories. In an age of prolific litigation, users
and developers should be aware of the legal ramifications arising from knowledge
sharing and knowledge automation in general. Consider the following situations:
• An attorney, using a knowledge base, advises his client of the tax forms to file and
what to include in his articles of incorporation to qualify as a federally tax-
exempt corporation. The client is later sued by the Internal Revenue Service
(IRS) for back taxes.
• A physician diagnoses a patient after consultation with her medical knowledge-
based system. The patient is treated but soon dies as a result of misdiagnosis and
improper treatment.
• A large computer firm sells a local area network to a client after using a knowl-
edge-based system to identify the proper configuration. When the system is
installed, the client-user discovers that the configuration is wrong and that the
equipment, as it stands, is worthless.
• A knowledge-based system used by an architect incorrectly determines the stress
requirements of an expensive new public building. The completed structure later
collapses, killing dozens of people.
• A radiation machine guided by a knowledge-based advisory system calculates the
dosage in the treatm ent of cancer patients. Two years after treatment, hundreds of
patients die due to radiation overdose. The problem was located in a faulty pro-
gramming module.
Each of these cases is real. These unfortunate situations lead to questions such as
“Who owns knowledge?” “Who is liable in such situations and for what reasons?” This
section examines the legal aspects of knowledge ownership.

THE LIABILITY QUESTION

Regardless of where knowledge originates, when it is misused, liability will become an
issue. The blame may fall on the knowledge developer who might have tapped the
wrong knowledge or on the expert who provided the wrong knowledge. It could also be
the fault of the knowledge repository that produced the wrong solution. In day-to-day
business, tort and contract theories present challenging questions for organizations and
 CHAPTER 14 Who Owns Knowledge? Ethical and Legal Issues aaaa 415
the legal community. If a knowledge repository produces the wrong solution, which
causes losses or injury to others, the resulting damage often leads to litigation dealing
with strict liability or negligence. Each individual involved in the process (company,
knowledge developer, or expert) is potentially vulnerable and subject to legal scrutiny.

Liability of the Knowledge Developer

In knowledge design or software development, the developer is often responsible
for knowledge accuracy and reliability. In building the knowledge-based environment,
a variety of errors may become embedded in the system: Some are nontrivial and oth-
ers are out-of-bounds errors. An out-of-bounds error is one that occurs because either
the software did not have the expertise to address the particular problem or the devel-
oper improperly condensed the captured knowledge. A nontrivial error has a large
financial impact on the business, especially if the knowledge is used by employees
company-wide. The consequence is decommissioning the system or facing litigation.
Because the developer relies on his or her experience to develop the product or
software and a malfunction occurs, the developer is vulnerable to charges of personal
liability under the doctrine of respondeat superior (an employer-em ployee relation-
ship). If the designer is an employee of the organization that sells the software, the
employing firm is involved in the negligence action. In the end, the company is respon-
sible for certifying the system before it is released for company use.

Liability of the Expert

Another possible target for litigation is the expert, because his or her knowledge is
codified in the system. Should the system be faulty, liability will surely follow. The compe-
tency of the expert, however, is difficult to evaluate; because of their individual experiences,
feelings, and instincts, even experts addressing the same problem do not always agree on all
issues. Experts open up their knowledge to scrutiny, even when the resulting system is far
removed from the expert’s control. For example, in one incident where a walkway at a new
hotel collapsed, the two structural engineers in charge were found guilty of negligence.
With each situation, expert involvement and potential liability vary. For example
could a court of law hold an expert liable if the expert were an employee of the firm that
developed the knowledge-based system for resale? If the expert is an outside consul-
tant, who would be liable? As of yet, answers to these questions are unknown, because
only one case involving knowledge-based system liability has been tried in court to date.

Liability of the User

Even end users of a knowledge-based system are not immune from a lawsuit. They are
directly responsible for proper use of the system. Within the duties of their jobs, refusing to
comply with the system’s directions will increasingly come into question. By not properly
using an available resource by affirmative duty, users could be negligent by omission or
“passive negligence.” For example, the use of a knowledge automation system in medical
diagnosis could place as affirmative duty the responsibility, for using the system. It could
also be used in court as support for a diagnosis, even though the diagnosis was incorrect.

THE BASIS OF LIABILITY

Although the focus in this book has been on knowledge management principles and
procedures, the nature of knowledge and knowledge capture and use have not yet
reached a point where new legal remedies are required. Meantime, the old familiar
liability issues are still applicable. Tort law is a major area of concern, with the issue of
strict liability and negligence falling under it.
PART V Ethical, Legal, and Managerial Issues

Tort Law in Knowledge Management

Torts are wrongful acts subject to civil action. Tort law is a special area of law that
remedies wrongs between parties. In knowledge management and knowledge owner-
ship, it involves settling contract problems between the employee as an expert and the
employing organization in terms of knowledge ownership. Tort cases have already
been brought up in a court of law against businesses. Most of the cases have been
related to negligence, misrepresentation, and trademark violations.
Negligence is failing to take certain action, causing injury or m aterial losses to
another. For example, fearing lawsuits, loss of secrets, or loss of intellectual property,
many banks are now scanning employee e-mail messages at least randomly. There have
been several high-profile court cases where e-mail has turned up as evidence. A busi-
ness could conceivably be found negligent if it did not exercise due care in monitoring
and safeguarding its intellectual property.
Misrepresentation is a tort area that tags to fraud. Similar to false advertising in
intent, claiming that a certain knowledge product will perform certain functions when
in fact it could not is misrepresenting the product. Likewise, if a salesperson fails to dis-
close the negative aspects of a product when he or she knew all along about such weak-
nesses, they would be subject to prosecution.

Knowledge—A Product or a Service?

In terms of legal ownership of knowledge, the question of whether knowledge is a
product or a service attracts varied opinions. If one views knowledge as what you say,
not what you see, then knowledge can be treated as a service. One legal opinion sug-
gests that if knowledge is codified in a package and sold as a mass-marketed item, then
it is a product. If the package is custom-designed, then it is viewed as a service.
The resolution of this issue is im portant for users and knowledge developers. If a
knowledge repository is a product, proving negligence is unnecessary to hold the
developer liable. The Uniform Commercial Code (UCC), however, allows the devel-
oper to limit liability for defective knowledge bases through a disclaimer of warranties
in the contract. For these liability activities, the loss falls on the defendant (developer),
regardless of fault, as a cost of doing business.
Conversely, if a knowledge repository is a service, the contract law of the state in
question would apply, rather than the UCC. Negligence principles should be used. A
negligence cause of action is more difficult to prove, because the plaintiff must show
the aspect of the process that caused the defect and prove that failure to use sufficient
care caused the defect.
Many legal experts want knowledge-based systems to be considered as services in
order to avoid the strict liability associated with products. For example, a medical
knowledge base mass-produced to hospitals throughout the country could be classified
as a product. However, such a system should not be presumed to have the physician’s
intuitive capabilities, knowledge of the particular patient, or perception of the subjec-
tive factors. Knowledge-based systems that require the user and the system to make a
diagnosis jointly will most likely be considered a service. A summary of these relation-
ships is presented in Table 14.1.

COPYRIGHTS, TRADEMARKS, AND TRADE NAMES

In terms of copyright and trademark violations, this is an area that falls under intellec-
tual property law. It includes software, books, music, videos, trademarks, copyrights,
and Web pages. There is a growing controversy over who owns the intellectual property
of product design, creative advertisements, and the like.
 CHAPTER 14 Who Owns Knowledge? Ethical and Legal Issues i i ■■ 417

TABLE 14.1 Knowledge as a Product or a Service in Litigation Issues

Product Service
• Off-the-shelf software • Custom-designed software
• Mass-marketed software • Negligence principles used
• Custom-designed but affects a large number • Negligence cause of action more
of customers difficult for plaintiff to prove
• Proving negligence unnecessary to holding • For liability, law of state applies, rather
developer liable than UCC
• UCC liability limit allowable via disclaimer
of warranties

Copyright is ownership of an original work created by an author. It is a form of

intellectual property protection that covers the “look and feel” and content of printed
media like articles and textbooks as well as software programs and software packages.
Copyright law gives the author or creator of a tangible product the right to exclude oth-
ers from using the finished work. That is why authors and publishers place a copyright
notice after the title page, although a notice does not make the work freely available.
A copyright owner has exclusive rights to his copyrighted work. They include the
right to copy, duplicate, and distribute the work in any form. He or she has the right to
modify the work in an effort to create a new work. In addition, a copyright owner has
the right to show the work at a public place or by means of a film or television show.
This includes showing the work on the Web. Such rights are protected by law; a copy-
right owner can recover damages in a federal district court.
Several other works are protected—literary works; musical works; dramatic works;
pictorial, graphic, and sculptural works; Web sites; sound recordings; and architectural
works. Knowledge-based programs and most compilations may be registered as “liter-
ary works.” There are several categories of material, however, that are not eligible for
copyright protection. For example, works consisting entirely of information considered
common property and containing no original authorship are not copyrighted. Familiar
symbols or designs or mere listings of ingredients or contents are also not copyrighted.
A copyright is good for the life of the author plus an additional 70 years after the
author’s death. In the case of joint authorship where the authors did not work for hire,
the term lasts for 70 years after the last surviving author’s death. There are also specific
conditions and laws to protect people from copying someone else’s work without per-
mission. For example, an author can quote up to 250 words without permission, pro-
vided he or she gives recognition to the author of the quoted work. The same proce-
dure applies to copying material on other people’s Web pages. Beyond that, you need
permission from the copyright holder to quote or to copy. A magazine that copied only
300 words of form er President G erald F ord’s 200,000-word autobiography got into
trouble. It was found guilty of copyright violation (Emery 1997).
In knowledge management, a knowledge repository and the way it is organized are con-
sidered a compilation and are copyrightable. A compilation’s copyright protects all compo-
nents within it. An original compilation of names and addresses that are now in the public
domain is also copyrightable. The same applies for logos and trademarks. By international
agreements, only an expression can be copyrighted, not facts. The biggest problem is not
from the text content, but from images and programs. On the Web, for example, images are
easy to download or cut and paste into any other page; pirating images is an everyday affair.
418 7T77 PART V Ethical, Legal, and Managerial Issues

A word, a picture, or an image that identifies a product or service is intellectual

property and is protected by a trademark. The term trademark means registration of a
company’s trade name so that others cannot use it. It is also a word or a symbol that
distinguishes a good from other goods in the market. As shown in Figure 14.1, “For
Dummies” is a tradem arked name; “ID G Books” is a tradem arked logo; and the icon
on the left is a tradem ark symbol of the same firm.
Who owns a tradem ark (or a copyright) is often a contractual matter. Trademark
protection is a maze of federal and state laws that have to be reviewed carefully before
securing protection. Some trademarks might be registered in one state, but not in oth-
ers, and some states have individual laws covering trademarks. Copyright and trade-
mark examples are shown in Figure 14.2.
Related to the copyright issue, if a company hires an outside firm to develop its
Web site, the contract should clearly give the company all intellectual property rights.
The Web designer provides a “work made for hire.” If the design is done in-house, the
employment contract should also make it clear that any creative work performed by
employees belongs to the company. For a $30 fee, anyone can copyright work with the
Library of Congress. Its Web site address is www.loc.gov/copyright (see Figure 14.3).
The resolution of this issue is important for users and developers of Web sites. If a
Web site representing a body of knowledge is a product, proving negligence is unnec-
essary to holding the developer liable.

WARRANTIES
The UCC is the foundation of commercial contract law in all states, except Louisiana.
As such, it defines the concepts of product law and contains provisions for computer
contracts in the form of warranties. A warranty is an assurance made by the seller
about the goods sold. An additional safeguard to the UCC is the federal Magnuson-
M oss Consumer Product Warranty Act, enacted in 1975. It clarifies the issues relating
to warranty inform ation disclosure requirem ents and regulates the lim itation of

FIGURE 14.1: An Example of a Trademark and a Trade Name

Edit View Favories Took Help

[ »a«a . ® @ ,a| ®S.«ch a|Fav«te (gHiaoiy | Ig)-,St B - S| . .
Wtp/Aww.dummie*.com/

i& i g / Register for eTips '

J R © ’ Subscribe to our
d u m H i e S .COA/J' free newsletters.

The O nline Resource foi the Rest o f Us!

Dummies Book Search I © Browse How-To's, Articles, and Featured Titles

Bookstore o r Browse the complete D | TECHNOLOGY

Certification
LIFESTYLE
Food & Drink
Hardware Home & Garden
Internet Pets
Networking
Programming
Software

i MONEY
( Career
Personal Finance
Small Business
FPt.Ptjmmles,
Online B u iin a n
) RECREATION
3rd Edition
Outdoors
Recreational Sports
Dummies Specialty Shop: Chack out more For Dummies products Team Sports
T_
including fitness videos, sewing patterns, conversation cards, and
other fun items. J Weekly Highlights

SOURCE: www.dummies.com. Date accessed November 30,2002.

 CHAPTER 14 Who Owns Knowledge? Ethical and Legal Issues ■■■■ 419

COPYRIGHT

All content included on this site, such as text, graphics, logos, button icons, images, audio clips, digital downloads, data
compilations, and software, is the property of Amazon com or its content suppliers and protected by United States and
international copyright laws. The compilation of all content on this site is the exclusive property of Amazon com and protected
by U.S. and international copyright laws. All software used on this site is the property of Amazon com or its software
suppliers and protected by United States and international copyright laws.

TRADEMARKS

AM AZON COM, AM AZON.COM B OOKS. EARTH’S BIGGEST BOOKSTORE; IF IT S IN PRINT, IT S IN .

STOCK; MOODMATCHER, 1-CLICK; and other marks indicated on our site are registered trademarks of Amazon com.
Inc. or its subsidiaries, in the United States and other countries EARTH'S BIGGEST SELECTION, PURCHASE
CIRCLES. SHOP THE WEB, ONE-CLICK SHOPPING. AM AZON.COM ASSOCIATES, AM AZON.COM MUSIC.
AMAZON COM VIDEO, AM AZON.COM TOYS. AMAZON COM ELECTRONICS. AM AZON.COM e-CARDS.
AM AZON.COM AUCTIONS. zSHOPS, CUSTOMER BUZZ, A M A ZO N C O UK. AMAZON.DE, B ID -O iC K .
GIFT-CLICK. AM AZON.COM ANYWHERE. AM AZON.COM OUTLET. BACK TO BASICS. BACK TO BASICS
TOYS. NEW FOR YOU. AM AZON HON OR SYSTEM. PAYPAGE. UNPAY. and other Amazon com graphics, logos,
page headers, button icons, scripts, and service names are trademarks or trade dress of Amazon com. Inc. or its subsidiaries.
Amazon.corn's trademarks and trade dress may not be used in connection with any product or service that is not
Amazon.corn's, in any manner that is likely to cause confusion among customers, or in any manner that disparages or
m zr~ mumssm
■ FIGURE 14.2: Copyright and Trademark Statements

SOURCE: www.amazon.com. Date accessed August 2002.

implied warranties. Both the UCC and the warranty act identify the various types of
warranties that can exist and serve as references for further information on the subject.
Of the two types of warranties, an express warranty is offered orally or in writing
by the m aker of the product. It is usually part of a sale. The buyer purchases the goods
in part because of a statem ent by the seller with respect to the quality, capacity, or some
other characteristic of the package. An express warranty need not be an expressed
statement. It may be found in the seller’s conduct.

FIGURE 14.3: Library of Congress Copyright Web Site

J: Ffe Edft Vie* Favorite" Tod* Heb

|jAddrcn |g] http:/A>wb.loc.oov/copyrighU>

'To promote the Prvgrest of Science and useful Arts,

0 ) Copyright Office . by ucunr.jfor limited Tares (o Authors and Inventor,
'he etdusne tight to their respective Writings and Discoenrs"
— US Con3tlnitun.Attielet.Sei

•^);"liS2!uSar8BS3SflMrffMSMTiv!ZBS5KMJLLiiiLlJjLjr^ili^^ ?
. An important message about mall delivery

About Copyright How to Register a Work

Hot Topics
o Copyright Basics o Literary Works
° Fee Changes
o Frequently Asked Questions (FAQ) o Visual Arts
o Current Fees q Performing Arts
« Sound Recordings
Search Copyright Records o Serials/Pehodicals
What’s New
a N t«nN «t o Registrations and Documents
How to Record a Document
o Currant Fees
« Notices of Restored Copyrights n
o Online Service Providers o Record a Document
O Website Redesign
o Vessel Hull Designs
O 2 00 0 Annuel Law and Policy
R eport
o Copyright Law
O W ebcasting
<3 Circulars and Brochures o Federal Register Nodces
q Forms « Current Legislation

SOURCE: www.loc.gov/copyright. Date accessed December 2,2002.

PART V Ethical, Legal, and Managerial Issues

An implied warranty arises automatically from the fact that a sale has been made
and that the good will do what it is supposed to do. For example, a knowledge base
should be fit for the ordinary purposes for which it is used. This implied warranty of
merchantability indicates that the Web site should do what it is expected to do. The
other aspect of implied warranty is one of fitness. A knowledge base should be fit for
the particular use intended by the buyer. Violation of this warranty is probably not
common among knowledge developers, although it may be more common among com-
panies that do customized programming.
Disclaimers and warranties are closely related. A disclaimer is the seller’s inten-
tion to protect the business from unw anted liability. Many software packages are
labeled “as is,” meaning they are sold without warranty of any kind regarding perfor-
mance or accuracy. More pronounced disclaimers go so far as to state that neither the
developer, retailer, or anyone affiliated with the developer is liable for damages even if
the developer has been forewarned of the possibility of such damages.
Even though disclaimers are clearly stated, their legal status is fuzzy at best. The
main issue centers on whether the knowledge in question is a product or a service. In
either case, the courts are adverse toward excluding warranty disclaimers or attempts by
the knowledge developer to avoid their applications as unconscionable. Express war-
ranty disclaimers are effective, provided they are conspicuously placed and in writing.
A knowledge-base client can reasonably look toward warranties as protection if
damage is caused through the use of a product purchased through a company’s Web
site. However, showing reason why warranty exclusion should not be accepted is diffi-
cult. In fact, two states have enacted “shrinkwrap” laws, which hold that all warranties
made or disclaimed on the license found inside the shrinkwrapping are legal and final.
Cases involving warranties also require that the user show who is at fault and why,
which is a difficult task. The knowledge automation system in question must necessar-
ily be considered a product under UCC rules for warranty issues to be relevant.

STRICT LIABILITY
Tort theory is based on several issues, namely that the producer of a product is in the
position to reduce risks and insure against injuries that could result. As with warranties, a
software package must be considered a product for the tort theory of strict liability to
apply. If this criterion is met, developers, manufacturers, and distributors could all be held
liable for injuries even though reasonable care standards have been satisfied. For exam-
ple, even though no errors are found in a Web site, the Web designer could still be held
liable under the tort theory of strict liability should damages or losses result from the use
of the Web site. Imposition of this theory protects the Web visitor regardless of whether
anyone is at fault, in a strict sense. The major legal issues are summarized in Table 14.2.

LEGAL DISPUTES IN KNOWLEDGE MANAGEMENT

In knowledge management, several disputes may arise that have legal implications:
• An employee delivers a problem solution based on years of proven experience,
but the company fails to support the work or process. As a result, the product in
question fails the customer test of approval.
• A company has a quality standard for a procedure, but the employee purposely
shows reluctance to cooperate to meet the standard.
• A company engineer does not like the established design, but the company is
reluctant to make changes. As a result, the sale of the product falls behind due to
new competition.
 CHAPTER 14 Who Owns Knowledge? Ethical and Legal Issues m■ i ■ 421
TABLE 14.2 Legal Issues of Knowledge-Based Systems
• A n expert owns his or her knowledge of the work if no prior agreement was established.
• A preemployment contract or intellectual property agreement can limit the expert’s own
liability for the knowledge automation system.
• If a knowledge developer builds the system and a problem arises, he or she is subject to
charges of personal liability under the doctrine of respondeat superior. If the developer is a
company em ployee, the organization is also involved in the negligence action.
• If a knowledge-based system is a product, proving negligence is unnecessary to hold the
developer liable. The UCC allows developers to limit liability for defective work via a dis-
claimer of warranties in the contract. For these liabilities, the loss falls on the developer,
regardless of fault, as a cost of doing business.
• If a knowledge-based system is a service, the contract law of the state would apply, rather
than the UCC.
• Cases involving warranties require that the user show who is at fault and why.
• The knowledge base in question should be considered a product under UCC rules for war-
ranties to be relevant or for tort theory of strict liability to apply.
------------------------------------------------------ 1 1 1 I 1 1 -------------------------------------------------------

WEB LINKING AND DOMAIN NAME DISPUTES

In electronic commerce, unique knowledge about a product, a company, or a service
resides in Web sites. The infrastructure of the Internet is designed around hyperlinks—
text or an image whose address can be linked to another Web page for reference.
When one clicks on the link, it automatically goes to the attached location and the des-
ignated Web page. This “jumping from one site to another” raises legal issues that
include the following:
• Referencing a linked site without permission from the site owner
• Retrieving or downloading information from a linked site without referencing or
permission
• Unauthorized use of a company’s registered trademarks
• Adding a Web program to a company’s Web site without permission
Inappropriately referencing a Web site is not a clear-cut issue; it depends on the
intent of the referencing. For example, one bank’s Web site advertised online automo-
bile loans. When a visitor entered the amount she wanted to borrow (in this case, it was
$47,000 for a Mercedes), the Web site asked the visitor to click on the “referral” button,
giving the impression that her application would get special attention. Instead, she
ended up on the Web site of a large bank in another state that specializes in “jum bo”
auto loans. The visitor was unhappy about this runaround and promptly clicked away
from both sites.
One New Jersey bank Web site posted a question on its homepage: “Want to com-
pute your mortgage rate on a house? Click here.” The software module that did the cal-
culation was registered to a Chicago bank. The networking algorithm detected un-
authorized use of the package by the New Jersey bank Web site and promptly sued for
damages. The case was settled out of court; the New Jersey bank was forced to delete
the referral to the mortgage calculator.
Domain names have been viewed as representing the intellectual capital of a busi-
ness. Disputes over who owns a domain name have been with us since business hit the
Internet. In 1992, the U.S. government contracted with Network Solutions, Inc. (NSI),
which also goes by the name In terN IC (Internet N etw ork Information Center), to
422 7777 PART V Ethical', Legal, and Managerial Issues

manage the top-level domains. Initially, domain names were assigned on a first-come,
first-serve basis. However, this caused companies or individuals to register for domain
names that they had no use for and hold them for future sale at exorbitant prices. Since
1995, the policy has changed. Domain names are still issued on a first-come, first-serve
basis, but applicants are reminded in writing that such issuance does not duplicate or
replace the legal right of another party such as a registered tradem ark to use the name.
The low-cost fee of $70 for 2 years of registration has caused several individuals to
register and hold known names hostage for big money. One poacher, for example, regi-
stered for 200 domain names, including his former employer’s. However, he kept losing
cases filed against him by firms defending their right to use their tradem ark name. The
general rule for resolving domain name disputes is to determ ine the date that the
claimant first used his tradem ark or the effective date of his validated tradem ark regis-
tration. If the registered holder appears to have infringed on the registered trademark
Owner, then NSI assigns the registered holder a new domain name. The court is the
only other avenue to seek relief (see Box 14.3).
Here are some ideas regarding domain names and trademarks:
• Find out whether the proposed domain name infringes upon any trademarks. The
fact that one registers for a domain name does not in itself give the owner the
legal right to use it.
• Secure federal trademark registration of the proposed domain name. Once it
clears against possible claims of infringement, then it should be registered as a
trademark with the U.S. Patent and Trademark Office.
• Register the proposed domain name with InterNIC (Internet Network Information
Center)—the agency that represents the U.S. government in assigning domain names.

* i i * 1 1 BOX 14.3

THE COST OF POACHING

In what is thought to be the first ruling on a faith intent.” Key to the New York telephone com-
potent provision of a recently signed cybersquat- pany’s case was a provision that explicitly allows
ting legislation, a U.S. judge has empowered Bell tradem ark owners to take legal action directly
Atlantic Corp. to take possession of nearly 2 dozen against domain-name holders, without the neces-
Internet addresses. U nder the ruling, verbally sity of hauling each alleged transgressor into court.
issued by a m agistrate judge, Bell A tlantic can Although tradem ark owners have praised the
transfer the registered domain names, said Sarah legislation, critics worry that it goes too far and
Deutsch, chief intellectual property counsel for may end up snaring Internet entrepreneurs and
Bell Atlantic. others who innocently register names similar to a
In entering his order, the judge relied on the trademark. Individuals whose last name happens
Anticybersquatting Consum er Protection Act, to resemble a corporate moniker, for example, are
which gives trademark owners a powerful weapon concerned that they will be dragged into court
to combat cybersquatters, those who register vari- and forced to pay a huge fine for registering an
ous perm utations of company names with “bad address with their own last name.

SOURCE: Adapted from Plitch, Phyllis. “Court Order Lets Bell Atlantic Wrest Dom ain Nam es from Cybersquatter,” D ow
Jones & Company, February 2,2000.

fl fl fl fl fl fl
 CHAPTER 14 Who Owns Knowledge? Ethical and Legal Issues mmmm 423
• In the event of a poached domain name, bring a lawsuit to force InterNIC to reas-
sign the name.
• Get permission before linking to other Web sites. There are issues of liability in
linking to other Web sites without such permission.

THE MALPRACTICE FACTOR

Most people have heard of litigation involving doctors who have been sued for mal-
practice. As a result, surgeons and other professionals pay high premiums for malprac-
tice insurance. M alpractice in knowledge management is negligence applied to knowl-
edge developers for design defects in systems tailored specifically for professional use.
For experts to be held liable for malpractice, they must be considered professionals,
belong to a profession in which standards of care have been established, and owe a cer-
tain duty of care to those for whom they work.
A t present, neither standardization nor certification that recognizes knowledge
developers as professionals has been established; therefore, they are not yet liable for
malpractice. Some people argue that knowledge developers should be licensed as a
condition to building knowledge-based systems (see Box 14.4).
The discussion of liability and knowledge-based systems would not be complete
without bringing up the potential liability for failure to use knowledge-based systems.
Professionals may be held liable for not using such a system when it is available. One
such case is T. J. Hooper. H ooper’s tugboat lost the barges it was towing in a storm that
the captain would have known about had the boat had a radio aboard. Even though a
radio was not a required item, Hooper was found liable for the loss of the barges.
In another classic case, that occurred during the Persian Gulf war, the failure to
comply with expert systems resulted in benefits. The U.S. intelligence community, rely-
ing on knowledge-based systems and satellites, found that only 25 percent of a key
bridge in Baghdad had been destroyed. A call was m ade to G eneral H. Norm an
Schwarzkopf, asking him to send a squadron to destroy the bridge. The general refused
on the ground that the 25 percent damage had severed one span, therefore rendering
the entire bridge useless. Sending solders to destroy the bridge would expose them to
unnecessary risks or even casualties (Awad 1996).

« ■ 9 I a a BOX 14.4

THE LIABILITY OF WEATHER FORECASTING

A case involving the National W eather Service proves to be an inexact science. The fishermen
(NWS) illustrates the potential danger of hyping were confronted with an unpredicted storm. All
knowledge autom ation system technology. A c- four men died. Their families sued the U.S. gov-
cording to the evidence, four fishermen consulted ernment for $1.25 million and won. Certainly, one
a w eather forecast released by the NWS before who develops a knowledge-based system should
embarking on an expedition. Unfortunately, even be aware that the system may be held to certain
after years of study, the prediction of weather still standards.

SOURCE: Awad, E. M. Building Expert Systems. Minneapolis, MN: West Publishing Co., 1996, p. 346.
424 7777 PART V Ethical, Legal, and Managerial Issues

If knowledge-based system developers are professionals, they may be sued for

malpractice. What about knowledge-based systems that have evolved to the level of
legitimate professionals? Could they not also be held liable for malpractice? If such
systems can be held liable for anything, then perhaps malpractice claims against them
would be reasonable. For example, if a legal knowledge-based system has taken on all
of the attributes and functions of a judge, then it should be held liable for malpractice
just as a human judge would be. This scenario may be stretching the imagination a bit,
but suing a knowledge-based system for malpractice may become a reality.
Malpractice is only one aspect of the liability that may exist in a world in which
knowledge systems have reached their full potential. Others will undoubtedly be dis-
covered to fill the legal gap that has been created by this new technology. Soon, new
legal reasoning and new laws will be needed.
Anyone to whom liability for knowledge systems can reasonably be traced will bear
responsibility. Picture a company that develops and sells a knowledge system that assists
in the diagnosis of medical diseases. Over the next 10 years of its use, the program repos-
itory encounters enough new diseases or new situations that it is able not only to expand
its knowledge base but also to modify itself in such a way that the code could no longer
be attributable to the original developer. If an injury occurred in such a situation, logi-
cally, the correct “person” to hold liable would be the knowledge base. Again, holding
such a system liable for anything may seem rather far-fetched, but it may not necessarily
be out of line as laws change to reflect the influence of this technology in everyday life.
Another area that may be viewed under malpractice relates to a growing list of
corporations that export their intellectual property abroad to shelter income and
evade U.S. taxes. It is not clear how many patents or trademarks U.S. firms hold off-
shore, because the practice started in the mid-1990s. According to a source detailed in
Box 14.5, a large Bermuda law firm promotes its expertise, spelling out the advantages
of holding intellectual property offshore.

■■■« The Ethics Factor

Organizations throughout the United States are troubled by ethical problems. High
ethical standards in a learning organization strengthen its competitive position. The
CEO plays the most significant role in setting ethical standards for employees. This sec-
tion focuses on ethical issues in knowledge-based systems.

WHAT IS ETHICS?
The word ethics is not easy to define. It means different things to different people. To
begin to discuss the concept, a common definition is needed. Ethics is one or all of the
following: fairness, justice, equity, honesty, trustworthiness, and equality.
Stealing, cheating, lying, or backing out on one’s word are all descriptive of a lack
of ethics. Something is ethical when a person feels it is innately right, which is a subjec-
tive judgment. For example, “Thou shalt not steal” is a belief held by most people; how-
ever, a single parent who steals a loaf of bread to feed starving children may be for-
given for the behavior even though it is illegal.
An unethical act is not the same as an immoral or an illegal act, although one may
lead to the other (see Box 14.6). For example:
1. Cheating on o n e’s federal income tax return is more illegal than immoral,
although it is implicitly unethical.
2. Cheating on a friend is more immoral than illegal, although it is tacitly unethical.
 CHAPTER 14 Who Owns Knowledge? Ethical and Legal Issues a ■a a 425

BOX 14.5

SENDING INTELLECTUAL PROPERTY ABROAD

While Washington lawmakers are cracking down are stashing their intellectual property abroad to
on traditional offshore corporate m aneuvers to shelter income from overseas sales.
avoid taxes, another strain has sprouted that Computer companies, drug firms, and others
could be costing Uncle Sam billions: Companies that derive most of their value from research
and ideas find the tactic particularly attractive.
For example, Columbia Laboratories, Inc., a small
pharm aceutical firm in Livingston, New Jersey,
for years has been transferring patents to Bermuda,
U.S. a tiny island in the A tlantic Ocean. When two
Company
U.S.-based Columbia researchers patented a
constricted-artery treatm ent in the mid-1990s,
Buy-In Payments Intangibles Cost-Sharing they turned over the rights to subsidiary Colum-
bia Laboratories (Bermuda) Ltd.
One reason: Royalties from sales of the prod-
ucts made outside the United States would flow
to the Bermuda subsidiary and could be parked
there tax-free.
Here is how the deals typically work:
1. U.S. parent company sets up offshore holding
company (IH C o) located in a low-tax jurisdiction
such as Bermuda.
2. Offshore unit buys stake in parent’s Buy-In
Payments Intangibles Cost-Sharing existing
intangibles such as trademarks, Arrangement
patents, or other intellectual property. Future
intangibles are developed jointly by parent and
IHCo under cost-sharing arrangements.
3. Offshore unit licenses intangibles to subsidiary A,
typically in a third country, which then collects roy-
alties from foreign subsidiaries (B, C, D ) that sell
parent company’s products to foreign customers.
4. Royalties are returned by A to IHCo, which for-
wards one portion back to U.S. parent and keeps
its own portion offshore to avoid U.S. taxes.

SOURCE: Excerpted from Simpson, Glenn. “A New Twist in Tax Avoidance: Firms Send Best Ideas Abroad,” The Wall Street
Journal, June 13,2002, p. AlOff.
426 «■«« PART V Ethical Legal, and Managerial Issues

I 1 I i ■ 1 BOX 14.6 1 1 1 1 1 1

THE SAGA OF A YACHT

During the February board of directors meeting ment of bank customers, bank officers, and com-
of a commercial bank, the first agenda item was a m unity leaders on his yacht. The chairman has
review of the statem ent of condition (expenses, certain privileges.” Joan then asked, “For a bank
revenues, and so forth) of the bank for the previ- this size, how can one justify this much expense
ous month. Joan, a new m em ber of the board, each month? What kind of entertainment adds up
noticed a line item under “entertainm ent” in the to $14,000 per m onth? How long has this been
amount of $14,000. She thought to herself, “Here going o n ?” The president, his face turning red
is a local bank of 240 employees. W hat kind of with irritation, said, “I really don’t want to elabo-
entertainment is going on at the bank to add up to rate further about this. Remember, you are new
this much expense?” She had seen the same fig- on the board. I wouldn’t advise asking the chair-
ure during last m onth’s meeting. man about it. I’d let it go. The bank is making
Out of curiosity, she asked the chairm an of enough money. The chairm an has 70 percent
the board, “Mr. Chairman, I’d like to know more equity in the business. What more explanation do
about the entertainment expense item. Could this you want?”
be back from the Christmas party, reflected as a Sensing futility in the attempt, Joan later dis-
January expense?” The bank president, who was a covered that the monthly charge was a dockage
m em ber of the board snapped, “Well, as you fee for the chairman’s 190-foot yacht. As a board
know Joan, the bank incurs all kinds of entertain- member, she felt responsible for assuring integrity
ment expenses. Why don’t you stop by my office in bank managem ent. What made this situation
after the meeting and I ’d be happy to explain it more difficult for Joan was that for the past 6 years,
further.” there had not been one occasion where entertain-
A fter the meeting, Joan and the bank presi- ment on the yacht was related to the bank in any
dent had a brief chat. The president explained, measurable way. She is now in a quandary won-
“Joan, the entertainm ent item is the b ank’s dering whether she should continue on the board
monthly contribution to the chairman’s entertain- or resign.

3. Submitting a “padded” bill to a client is more unethical and possibly illegal,

although it is considered immoral.
4. Stealing an item from a store is illegal, immoral, and unethical in various ways.

Incidents like these are related to one’s value system, beliefs, and culture. Laws are
often created to combat unethical acts that threaten societal image and survival. They
are also used to reinforce existing ethics. Ethics creates a strong sense of professional-
ism. Ethical misbehavior among knowledge developers is no greater or less than it is
among the mass public (see Box 14.7).
A model of acceptable behavior, with ethics as a factor, is shown in Figure 14.4. If
a person falsely reports a donation to a charitable organization, it is considered both
illegal and immoral (quadrant A). This action implies a lack of ethical standards and
personal values. In quadrant D, rescuing hostages from a foreign country might be ille-
gal (and dangerous) by that country’s laws but be moral (and implicitly ethical) in
terms of justice under U.S. laws. Deciding what is moral, legal, or ethical creates a con-
stant struggle.
 CHAPTER 14 Who Owns Knowledge? Ethical and Legal Issues aaaa 427

■ I ■ ■ ■ I BOX 14.7

WHO DECIDES OPERATIONAL ETHICS?

A knowledge-based system built for a com m er- tem. Why don’t you talk with him?” In addressing
cial bank uses 17 criteria to determine whether an the equity question, the vice president replied, “I
applicant qualifies for a mortgage loan. For sev- make decisions on mortgage loans based on
eral years, the bank has been criticized by state guidelines from the board of directors. Our chair-
auditors for not granting enough loans to minori- person is a major stockholder and pretty much the
ties or to people living in m inority districts. A owner of the bank. She does not live in the area
review of the knowledge-based system also re - and wants to make sure we make secured loans.
vealed that most mortgage loans favor applicants Why don’t you talk to the president about it?”
who are married rather than those who are single, In a meeting with the president the next day,
separated, or divorced. the president said, “I’m surprised you’re bringing
When the discriminating knowledge rules up bank policy. We paid for a knowledge-based
were pointed out, the knowledge developer said, system to comply with our requirements. So, what
“I developed this system based on the vice presi- is the fuss all about?”
dent of mortgage loan’s thought process. He re- However, the question raised here still re-
viewed several prototypes and approved the sys- mains: Is the system unethical, illegal, or both?

a fl fl a fl a

ETHICAL DECISION CYCLE

Knowledgeable people with common sense are expected to be people of ethics. They
consider a number of elements when they make ethical decisions:
1. The nature and essence o f the act. Is the act fair, reasonable, or conscionable?
2. The consequences o f the action or inaction on the parties involved. Who gains by
the act, who loses, and by how much?
3. The far-reaching consequence o f the action or inaction on the organization, com -
munity, and society. Will the act, if left unchecked, lead to societal ills?
Figure 14.5 summarizes the three major steps of a hypothetical ethical decision
cycle. A person faces a situation and must evaluate alternative approaches that lead to
a decision to commit an act. The outcome is initially reviewed by the person. If it is
believed to be rewarding, fair, and just, chances are it will be repeated, and vice versa.
The outcome in general becomes known to others (groups, organizations, and society)
who would condone or condemn the outcome. The result of this feedback is eventual
advancement and refinement of ethics within families, the workplace, and society in
general.

MAJOR THREATS TO ETHICS

Ethics are more openly discussed as a serious concern today than they were in the past,
because the threats have steadily increased. As shown in Figure 14.6, the main threats
are as follows:
• Faster computers and PCs
• Sophisticated telecommunication and computer networks
428 rT77 PART V Ethical, L eg a la n d Managerial Issues

Legal

Example: Example:
Restricting Immigration Donating to Charity

B C

A D

Example: Example:
• Robbing a Bank Rescuing Hostages from a
• Falsely Reporting Foreign Country
Charitable Donation

Illegal

FIGURE 14.4: M odel of Acceptable Behavior

• Massive integrated knowledge bases

• Ease of access to information and knowledge bases
• The view that captured knowledge is a competitive weapon
With these threats, ethical dilemmas 01 uimensions not imagined 10 years ago are
facing today’s firm, stretching standard ethical considerations to the limit. In fact, tech-
nological advancements have resulted in the need to reevaluate ethical standards and
their implications for privacy, confidentiality, accuracy, and integrity.
Knowledge hoarding, unauthorized e-mail access, and sale of competitive knowl-
edge are serious ethical issues. Also, taking out life insurance policies on employees to
benefit the company when the employee dies is questionable ethics (see Box 14.8). In
2002, a senior vice president at a m ajor credit card company was dismissed for

9 fl fl fl fl fl FIGURE 14.5: Ethical Decision Cycle

 CHAPTER 14 Who Owns Knowledge? Ethical and Legal Issues m■ ■ i 429

9 9 ■ 9 9 ■ FIGURE 14.6: Major Threats to Ethics

allegedly using his position and company affiliation to promote his side business and
pressure company suppliers into buying products from him —all of which were carried
out using e-mail and fax. Because no one knows more about information technology
and has better access to the company’s knowledge base than the head of the IT depart-
ment, should the head of the IT«department be the one to establish and enforce ethical
standards for the company as a whole?

BOX 14.8 9 9 9 9 9 9

THE ETHICS OF EMPLOYEE INSURANCE

Felipe M. Tillman loved music. O ne of his jobs though you’re an employee of theirs, you should
was a brief stint in the early 1990s at a Camelot know w hat’s going on in your nam e,” says
Music store in Tulsa, Oklahoma. In 1992, Felipe, Anthony Tillman, who was surprised to learn that
then 29 years old, died of complications from his b ro th er’s death benefited Felipe’s one-tim e
AIDS. He never bought life insurance, so his employer. “It isn’t fair.”
family received no death benefit. However, CM Companies can use the death benefits for any-
Holdings, Inc., th en the parent company of thing they like. CM Holdings used $168,875 of the
Camelot Music, did. It received $329,302. death benefit on Mr. Felipe Tillman for executive
CM Holdings took out life insurance policies compensation, court documents show. The docu-
on thousands of its employees, with itself as the ments also show that $280 went to Star County
beneficiary. Most workers covered this way do Children’s Services to help cover child-support pay-
not know it, nor do their families. “If som eone ments for a nephew of Camelot Music’s founder,
is going to use your nam e for something, even who was working at the company at the time.

SOURCE: Excerpted from Schultz, Ellen, and Francis, Theo. “Worker Dies, Firm Profits—W hy?” The Wall Street Journal,
April 19,2002, p. A l.
430 7777 PART V Ethical, Legal, and Managerial Issues

■««« Improving the Climate

Learning organizations can take several approaches to improve ethical behavior in the
firm as a whole:
1. To ensure ethical behavior throughout the firm, top management should act as
the role model.
2. A code of ethics should be established, taking into consideration the state of
technology of the organization. The goals should be realistic, achievable, and
agreed upon by the employees. Each organizational level should create an acting
ethics program within the company’s code of ethics and one that befits the nature
of the operation.
3. Unethical behavior should be promptly dealt with based on criteria and proce-
dures set in advance.
4. The firm should set up and support a strong ethics training program for all new
employees and reinforce the code of ethics on a regular basis.
5. The organization must provide motivation to focus on honesty, integrity, fair-
ness, and justice as goals that are just as im portant as money or the so-called
“bottom line.”

WHERE TO START?
Once the meaning of ethics has been agreed upon, the next step is to decide who is
going to lead the “ethics movement.” With knowledge workers, learning organizations
have used two approaches: bottom-up and top-down. The b o tto m -u p approach incul-
cates ethical behavior at the employee level with full support of top management. As
employees follow ethical practices, the entire organization benefits. The impact creates
waves of appreciation up the chain of command and benefits the entire firm over the
long run.
The to p -d o w n approach suggests that actions of the company start with the CEO.
By virtue of personal acts, decisions, and overall behavior, the top corporate officer sets
the tone for the kind of image the corporation will adopt. Corporate obligations extend
to a variety of stakeholders (such as customers, employees, and vendors), and these
obligations are generally recognized as being central to the ethical standards of the
firm. Carrying out these obligations requires a com prehensive ethics code that is
understood and supported in every sector and at each level of the organization.
Take the case of Boeing A ircraft’s chief executive William Allen. In September
1945, he resigned from his law firm to lead Boeing Aircraft to focus on postwar prod-
ucts after World War II. Allen had served as a company attorney for 20 years and direc-
tor for 14 years. He is remembered as a man of great sincerity, honesty, and integrity. It
was a great environment for knowledge sharing. When he accepted the offer of presi-
dent, he offered the following resolution as a reflection of his personal values:
• Do not be afraid to admit that you don’t know.
• Be definite —tell it like it is.
• Try to promote honest feelings toward the company around Seattle.
• D on’t talk too much . . . let others talk.
• Be considerate of your associates’ views.
• Above all, be human —keep your sense of humor and learn to relax.
Like Allen, the firm acquired a reputation as a highly ethical firm whose employ-
ees developed a strong sense of values and integrity in the way they approached their
 CHAPTER 14 Who Owns Knowledge? Ethical and Legal Issues aaa1 431
work and shared their ideas and knowledge. His term in office is rem em bered as a
period of “uncompromising high standards and clean ethics.” Employees always knew
where they stood. It shows that an ethics program should be “rolled out” from the top
down the line, which contributes to making each manager an effective ethics leader.

CODE OF ETHICS
Business ethics are closely tied to corporate culture and values, which means that a
code of ethics should represent the entire company. The code should be all encompass-
ing and stable over time. It does not make sense, for example, to change the code for
every new situation that arises.
Once a code of ethics has been posted and approved by management and employ-
ees, it becomes a commitment to behave within its guidelines on a day-to-day basis. An
honest workplace where m anagers and employees are held accountable for their
behavior and trust one another is the best environment in which to promote ethical
corporate behavior. To keep the ethical climate healthy, an organization must stress
regular self-assessment and encourage open debate within the workplace. Such a step
can be a sure winner in knowledge exchange and knowledge sharing.
Self-assessment is a question-and-answer procedure that allows individuals to
appraise and understand their personal knowledge about a particular topic. In the
case of ethics, it is not an exercise to satisfy others. The goal is to think about ethics
and adjust one’s behavior accordingly. It should be an educational experience for the
participant.
One self-assessment procedure asks a participant to assess a scenario and judge
whether an ethics issue is involved. The response is recorded on a special form and later
compared to the judgment of a panel of experts. The following story is an example.
St. John’s Corporation manufactures computerized voting machines and has won
several contracts from southern states to install them well before the next presidential
election. Alpha Corporation programs the machines for St. John’s and carries out the
installation on the sites. An Alpha engineer, Smith, is visiting St. John’s plant one day
and learns that problems in the construction of the voting machine could result in mis-
counts as high as one in 10 under heavy volume. Smith reports this finding to her
supervisor, who informs her that it is St. John’s problem, not A lpha’s. Smith does noth-
ing further.
Question: Is an ethics issue involved?
Opinion: Participants nearly unanimously agreed that doing nothing further
would be unethical. Use of inaccurate voting machines could
invalidate elections and potentially harm the general public trust.
Responsible (ethical) behavior and good business practice are not
inconsistent. The engineer should pursue the m atter further.

THE PRIVACY FACTOR

Privacy is a basic American value. It is also one of the most pressing concerns of knowl-
edge workers today and an issue that is inadequately addressed in knowledge manage-
ment. Cyberspace, originally intended for scientists, is now dom inated by virtually
everyone who has access to a computer. What makes knowledge and information so
valuable is that most of it is gathered discreetly. Private inform ation is being docu-
m ented and sold without the knowledge owner’s permission or awareness. Obviously, a
learning organization has an ethical responsibility to inform users of what information
is being captured and how it is being used.
432 7777 PART V Ethical, Legal, and Managerial Issues

The thought of being monitored is unsettling. Hidden video cameras, phone taps,
and e-mail analysis are all examples of technologies that are considered to be unethical
(and sometimes illegal), because they allow information to be captured about individ-
uals without their knowledge.
The FTC (Federal Trade Commission) has identified five principles of privacy pro-
tection, which are widely recognized in the United States, Canada, and Europe:
1. Notice. Employees have the right to be told in advance about any personal infor-
mation being gathered.
2. Choice. Employees should have the final say regarding the use of personal infor-
mation, other than the processing of such information.
3. Access. Employees should be able to access and correct any personal information
captured in company files or databases.
4. Security/integrity i Employees’ personal information should be processed, stored,
and transmitted in a secure way so as to assure integrity at all times.
5. Enforcement. The courts should back employees if any of these principles are violated.
There are three categories of concern regarding information and knowledge pri-
vacy. The first involves the electronic knowledge that businesses store about con-
sumers, vendors, and employees. Who owns such knowledge? The second is the security
of electronic knowledge transmission. Encryption has been promoted as a secure way
to transmit information or knowledge over the Internet. The third concern is the un-
authorized reading of personal files. Public key infrastructure (PKI) and other tech-
nologies are used to control unauthorized access.

■■■■ Implications for Knowledge Management

One conclusion from the discussion on legal and ethical issues is that the legal rules that
define who owns knowledge are yet to be clarified. The questions that come up before
various technical, organizational, academic, and governm ent groups dealing with
knowledge managem ent are “W hat rules should be instituted to govern knowledge
ownership?” “Who will make and enforce those rules?” “What shape should copyright
protection take in a world of costless, instantaneous, and undetectable copying?”
If one looks at communication networks, they are essentially defined by a set of
rules —the “network protocols,” which specify the characteristics of the messages to be
transm itted, the media through which they can travel, and how the messages are
routed through the media to their destinations. Because knowledge management is not
a set of interrelationships among people or procedures, it is difficult to protect it when
it is simply in the human mind. The legal implications will continue to be problematic.
Another area of concern is the long-range effect of continued knowledge hoarding as
a way to safeguard knowledge embezzlement. So far, learning organizations have differen-
tiated themselves on the basis of what is unique to the way they put knowledge to work.
Regardless of the laws that might assure integrity and protection for the employee
with knowledge and the company that employs him or her, the ultimate goal in knowl-
edge management is to nurture, promote, and protect the type and quality of knowl-
edge that contributes to the firm ’s competitive advantage. The Internet makes it so
much easier to trade and go global, but unfortunately it is also easier to abuse pro-
tected knowledge and, therefore, violate the laws.
In developing knowledge automation systems, knowledge developers, experts, and
corporations may choose to ignore legal issues on the basis that the likelihood of litigation
is remote. On the other hand, responsible system developers can become more proactive
 CHAPTER 14 Who Owns Knowledge? Ethical and Legal Issues ■a ■■ 433
by approaching knowledge repositories and systems building with ethics and integrity.
Having integrity includes never overselling the system or making outlandish claims. The
zeal with which knowledge-based systems are built should be tempered with the realiza-
tion that they are based on human knowledge and ideas, which are less than perfect.
Finally, management must focus on legal and employee protection issues surround-
ing knowledge ownership. For example, the issue of ethics may be tangential to running
a business, but without it there could be serious erosion of the company’s customer
base. Value America had that very problem seal its fate: It went bankrupt. Also, ethics
may mean one thing in one country and another in the country where business is con-
ducted. It is important to develop a code of ethics unique to the country or region.

Sum m ary * 1 * •
• Ethics may include any of the following: fairness, justice, equity, honesty, trustwor-
thiness, and equality. Something is ethical when a person feels it is innately right,
which is a subjective judgment. An unethical act need not be immoral or illegal,
although one act may imply another.
• Several issues represent major threats to ethics. These include faster computers,
sophisticated networks, ease of access to information, and transparency of soft-
ware. Software copyright infringement, unauthorized e-mail access, and sale of
competitive information are also serious ethical issues. To improve the corporate
ethics climate, top management should act as the role model, establishing a realis-
tic code of ethics and a strong training program.
• A code of ethics is a declaration of principles and beliefs that govern how
employees of a corporation must behave. The code should be all encompassing
and stable over time. Once posted, it becomes a commitment for the organization
as a whole. Self-assessment allows periodic adjustments in the code.
• In an age of litigious proliferation, users and developers should be cognizant of
legal issues arising from knowledge ownership and knowledge-based system use
or misuse. The key issue is determining who owns the knowledge. Depending on
interpretation and the laws, it could be the user, the expert, the organization, or
the knowledge developer. With no prior agreements, the expert is the rightful
owner of knowledge and can be the target for litigation resulting from a faulty
knowledge-based system. A preemployment contract releasing ownership to the
firm would help limit the liability of the expert.
• The issue of whether a knowledge-based system is a product or a service depends
on interpretation. If the software is custom-designed, then it is a service and the
contract laws of the state in question would apply. If it is off-the-shelf software, it
is a product, which means proving negligence is unnecessary to holding the devel-
oper liable.
• The Uniform Commercial Code (UCC) offers provisions for computer contracts
in the form of warranties. Warranties may be implied or expressed. Implied war-
ranties expect the product to do what it is intended to do, which is covered under
implied warranty of merchantability or fitness. Express warranties are presented
orally or in writing by the seller.
• Knowledge developers are open to malpractice—negligence due to design defects
and resulting damages. A t present, neither standardization nor certification recog-
nizes knowledge developers as professionals. In this respect, they are not yet
liable for malpractice. Sooner or later, knowledge developers should be licensed
to build knowledge-based systems.
434 7777 PART V Ethical, Legal, and Managerial Issues

Te r m s to Kn o w 1111
Code of ethics: A declaration of principles and beliefs narily regulate human affairs, would do; lack of reason-
that governs how em ployees of a corporation are to able conduct and care.
behave. Product liability: A tort that makes a manufacturer liable
Copyright: Ownership of an original work created by an if its product has a defective condition that makes it
author; a form of intellectual property protection that unreasonably dangerous to the user or consumer.
covers the “look and feel” and content of printed media Self-assessment: A question-and-answer procedure that
like articles, textbooks, and software packages. allows individuals to appraise and understand their per-
Copyright law: A law that gives the author or creator of a sonal knowledge about a particular topic.
tangible product the right to exclude others from using Stakeholder: Customer, employee, vendor, or distributor
the finished work. who has a vested interest in a company, a product, or a
Disclaimer: Renunciation of a claim or power vested in a system.
person or product. Strict liability: A seller is liable for any defective or haz-
Ethics: Justice, equity, honesty, trustworthiness, equality, ardous products that unduly threaten a user’s safety.
and fairness; a subjective feeling of being innately right. Tort: Wrongful act, subject to civil action; a legal wrong
Express warranty: Warranty offered orally or in writing committed upon a person or a property independent of
by the maker of the system or product. a contract; a wrongful injury to a person, a person’s rep-
Eraud: An intent to deceive; for example, knowing in advance utation, or a person’s property.
of a material fact about a product that is covered up in a sale. Trademark: Registration of a company’s trade name so
Implied warranty: Presumed warranty; certain implied that others cannot use it; a word or a symbol that distin-
facts that represent the product. guishes a good from other goods in the market.
Malpractice: N egligence or professional liability of a cer- Uniform Commercial Code (UCC): A law drafted by the
tified professional related to design defects in systems National Conference of Commissioners on Uniform
tailored specifically for professional use. State Laws that governs commercial transactions.
Misrepresentation: A tort area that tags to fraud. Warranty: A promise made by the seller that assures cer-
Negligence: Omission to do something, which a reason- tain facts are truly representative of a product or ser-
able person, guided by those considerations that ordi- vice, subject to certain limitations.

Te s t Yo u r U n d e r s t a n d in g a ' ' a
1. In your own words, describe how ethics relates to knowledge.
2. Review material on the Internet, and write an essay detailing the latest views
of ethics in knowledge management.
3. In what respect is an ethical act different from an immoral or legal act? Give
examples of your own.
4. Distinguish between:
a. product liability and tort laws
b. out-of-bounds and nontrivial errors
c. code of ethics and self-assessment
d. implied and express warranties
5. Explain briefly the ethical decision cycle. In what respect is it viewed as a
cycle?
6. In your own words, discuss the major threats to ethics. Is one threat more
serious than others? Be specific.
7. How can a code of ethics be applied in an organization? What are some of
the barriers?
8. Suppose you were asked to set up a code of ethics for a learning organiza-
tion. How would you proceed? What steps would you take to do the job?
9. Visit a company in your community, and investigate whether it is a learning
organization. What ethical issues exist in the firm? Write a brief report on
your findings.
10. What is the purpose of self-assessment? How does it contribute to a corpo-
rate code of ethics?
 CHAPTER 14 Who Owns Knowledge? Ethical and Legal Issues ■a ■a 435
11. Is a knowledge-based system a product or a service? Explain the pros and
cons of this issue and its implications for litigation.
12. Describe the circumstances under which knowledge developers might be
sued for malpractice in the future.

Kn o w l edge Ex e r c is e s 1 a 81
1. The chapter suggests three possible owners of knowledge. Who are they?
Do you agree?
2. In what ways will knowledge ownership be an issue in the future?
3. Who would be held liable for a defective knowledge-based system? Discuss
in detail.
4. How different do you think a learning organization views or upholds stan-
dards of ethics compared to other organizations?
5. Do you think litigation is the best route to protect knowledge ownership?
Form a group with your peers, and brainstorm the pros and cons.
6. The Internal Revenue Service acquires demographic data about tax-paying
citizens in an effort to elicit relationships to their tax returns. In your opinion,
is this effort an unethical act? Illegal act? Immoral act? Discuss.
7. Shoppers at a national retail chain are asked for their ZIP codes as part of
the checkout process. This information is used to determ ine the pattern of
business coming from various regions in the community. As a result, the store
decides on its products, prices, and specials to maximize sales volume.
Shoppers are not told why ZIP codes are solicited. Is the store’s action ethi-
cal? How does it compare to the use of cookies in Web shopping?
8. The board of directors of a commercial bank had an attorney whose job was
handling cases of bad loans and other legal m atters affecting the bank. The
attorney’s main line of business was real estate. Cases requiring specific
expertise were referred to attorneys in tow at the recom m endation of the
bank’s attorney.
One day, a case came up in which an employee sued the bank and one of
its vice presidents for sexual harassment. The bank’s attorney decided to han-
dle the case himself. A fter a lengthy trial, the court ruled in favor of the
employee. The jury awarded her $350,000 in damages.
W hen asked why he did not have another attorney try the case, the
bank’s attorney replied that he thought he was saving the bank money by try-
ing it himself, even though he had no prior experience with sexual harass-
ment cases. A t the outset, he did not think the case had much substance. The
bank, on the other hand, was not happy with the results and blam ed the
attorney for botching the case.
a. Could any aspect of the attorney’s decision to handle the case be consid-
ered unethical, immoral, or illegal?
b. Should the attorney be fired by the bank? Why or why not?
c. Visit a local law library or talk to a local attorney and determine whether
the bank attorney violated the American Bar Association’s code of ethics.
9. A knowledge-based repository was installed in the loan departm ent of a
local commercial bank to determ ine the qualifications of auto loan appli-
cants. The knowledge captured was based on the departm ent vice president’s
experience. The bank paid the knowledge developer (an outside consultant)
in full for the knowledge database.
Four months later, the developer learned through one of the bank
employees that the bank had modified the knowledge base to develop a real
estate loan knowledge-based system on its own. The junior loan officer took
the software home and spent time learning how to build a knowledge-based
436 77TT PART V Ethical, Legal, and Managerial Issues

system. He then worked with the senior vice president of the real estate
departm ent to build the new database. Several modules of the auto loan
database that check for the applicant’s assets, salary, place of employment,
and so on were copied into the new system. The approach to the system and
the procedure were identical.
The developer filed a lawsuit against the bank, seeking damages and the
right to ownership to the system. The bank, in turn, filed a countersuit, claim-
ing it purchased the knowledge package outright and, therefore, owed the
developer nothing.
a. What went wrong in this case?
b. Was an implied or express warranty part of the auto loan knowledge base?
c. Are any ethical, moral, or legal considerations at issue here?
d. Who do you think owns the knowledge in the real estate repository?
e. Based on the material presented in the chapter, how do you think the
court would rule?
f. Who owns the procedure used to build the knowledge base? Is it
proprietary?

Re f e r e n c e s 1 1 1 1
Amazon.com, “Conditions o f U se,” www.amazon.com/ Okerson, Ann. “Who Owns Digital Works?”
exec/obidos/tg/browse/-/508088/002-2707765-6824853, www.mhhe.com/socscience/english/holeton/chap8/
D ate accessed August 2002. okerson.mhtml, Date accessed August 2002.
Awad, E. Building Expert Systems. Minneapolis, MN: Platt, Nina. “Knowledge Management: Can It Exist in a
West Publishing Co., 1996, p. 346. Law O ffice?” www.llrx.com/features/km.html, Date
Brent, Doug. “History of Ownership,” www.ucalgary.ca/ accessed Decem ber 3,2002.
ejournal/archive/rachel/vln3/vln3.html, Date accessed Plitch, Phyllis. “Court Order Lets Bell Atlantic Wrest
August 2002. Domain Names from Cybersquatter,” D ow Jones &
Emery, Vince. H ow to G row Your Business on the Company, February 2,2000.
Internet, 3rd ed. N ew York: Coriolis Group Books, 1997, Powell, Adam. “MP3: Legal and Ethical Issues,”
pp. 121-140. hotwired.lycos.com/webmonkey/99/06/indexla.html.
Kull, Michael D. “G et a Competitive Edge with Rapport, Marc. “The Price of Knowledge,” Knowledge
Knowledge M anagement,” The Law Marketing Portal, M anagement, August 2001, p. 14.
www.lawmarketingcom/news/km.cfm, D ate accessed Schultz, Ellen, and Francis, Theo. “Worker Dies,
August 2002. Firm Profits —Why?” Wall Street Journal, A pril 19,2002,
Kurland, Norman. “Ownership: Knowledge as a Capital p. A l.
A sset,” cog.kent.edu/archives/ownership/ Sedgwick, James. The Certainty Site. Oxford Press,
msg02255.html, D ate accessed August 2002. 2001, p. 5, hom el.gte.net/cpqlszzy/whoownsknowledge.
Louis-Jacques, Lyonnette. “Legal Research on International htm, D ate accessed D ecem ber 2002.
Law Issues Using the Internet,” www.lib.uchicago.edu/ Simpson, Glenn. “A New Twist in Tax Avoidance: Firms
-llou/forintlaw.html, Date accessed August 2002. Send Best Ideas Abroad,” The Wall Street Journal,
Mason, Richard O. “Four Ethical Issues of the Information June 13,2002, p. AlOff.
Age,” www.misq.org/ archivist/vol/nol 0/issue 1/vol 10/ www.loc.gov/copyright, Date accessed D ecem ber 2,2002.
nol/mason.html, D ate accessed August 2002.
 Managing
Knowledge
Workers
Contents

In a Nutshell
What Is a Knowledge Worker?
Personality and Professional Attributes
Business Roles in the Learning Organization
M anagement and Leadership
Work M anagement Tasks
Work Adjustment and the Knowledge Worker
Profile of Vocational Needs and Reinforcers of Knowledge Workers
Smart Leadership Requirements
Technology and the Knowledge Worker
Knowledge W orker’s Skills
Role of Ergonomics
Role of the CKO
Managerial Considerations
Becoming a Change Leader
Managing Knowledge Projects
The Soft Side Always Wins
Implications for Knowledge M anagement
Summary
Terms to Know
Test Your Understanding
Knowledge Exercises
References

He who receives an idea from me, receives

instruction him self without lessening mine; as
he who lights his taper at mine, receives light without darkening me.
—T h o m a s Je f f e r so n

437
438 7777 PART V Ethical, Legal, and Managerial Issues

■■■ ■ In a Nutshell
We are living in a continuously changing world fraught with uncertainty, market fluctu-
ations, and aggressive competition. Today’s emerging age of knowledge economy and
knowledge management has created a new breed of company employees, whose intel-
lectual capital is the accumulated experience, commitment, and potential for develop-
ing and maintaining the learning organization. Such a breed is referred to as the knowl-
edge worker. A knowledge worker puts people first. He or she leverages technology to
maximize efficiency and corporate success round the clock. According to one source,
knowledge workers already make up one-third of the U.S. workforce. It is the fastest
growing employment sector of wealthy countries.
The driver of success in the new knowledge economy is knowledge. It is impressive
that people can earn a living by working with something as intangible as knowledge. In
1959, m anagem ent expert Peter D rucker popularized the term know ledge worker ;
which was invented by Fritz Machlup, a Princeton economist. Although the term is
widely used, there is confusion about its m eaning—even among knowledge workers. In
this chapter, we define knowledge workers, their attributes, job qualifications, poten-
tial, and how they should be managed. This is the crux of knowledge management as an
emerging challenge for the learning organization.

■■■« What Is a Knowledge Worker?

In reviewing the literature, several views and definitions have been proposed for the
knowledge worker. Here is a representative sample:
• A knowledge worker is someone who uses IT in conducting day-to-day business
and one that has direct impact on the efficiency and productivity of the job and
the work process (Awad 1996).
• A knowledge worker is someone who follows a process requiring knowledge
from both internal and external sources to produce a product that is distinguished
by its specific information content (Kappes and Thomas 1993).
• Anyone who makes a living out of creating, manipulating, or disseminating
knowledge is a knowledge worker (Bennett 2001).
• Individuals who have seriousness of purpose and are in the upper part of the tal-
ent and knowledge curves are knowledge workers. (Packaged Business Solutions,
Inc. 2001).
• Individuals who add to a company’s products and services by applying their
knowledge are knowledge workers (Drucker 2001).
• A knowledge worker is anyone who works for a living at the tasks of developing
or using knowledge (SearchCRM.com 2002).
• A knowledge worker knows what can actually be accomplished (see Box 15.1,
Dove 1998).
• A knowledge worker is one who gathers data/information from any source; adds
value to the information; and distributes value-added products to others (Kappes
and Thomas 1993).
• Knowledge workers are people who use their heads more than their hands to
produce value (Horibe 1999).
These definitions have several threads in common. Knowledge work embodies
experience, innovation, creativity, and transform ation of experience into knowledge
for leveraging products and services. Our definition of a knowledge worker is a person
 CHAPTER 15 Managing Knowledge Workers aaaa 439

a fl i fl i ■ bo x 15.1 a ■ fl a a fl

THE KNOWLEDGE WORKER

Just exactly what is a knowledge worker, anyway? A here. Business essentially is the application of knowl-
little reading and you come away thinking it means edge to produce/provide something that someone
people who work principally with computers—that else is willing to buy. If you don’t “know” how to do
seems to be the most vocalized worldwide-need for this, it doesn’t work. If anyone else “knows” how to
special skills and talent. The Fortune article talks do it better, it won’t work for long. When someone
about a school in Canada with a telling dropout “knows” a new way to do it, everybody else is in
problem. They train video-game developers and for a shock.
lose promising students to corporate scouts. Note That last item is what puts the spotlight on
two things here: These are not college-educated today’s knowledge worker. Our collective knowl-
people, and they do not even finish their “technical” edge is changing much faster today. The value of
training. So, where does this knowledge come from corporate experience is being recognized, and
that makes a knowledge worker so valuable? companies put their best into the corporate con-
L et’s look at this knowledge worker concept sulting pool. Though these people may have other
more closely. This may be one of the latest buzz- knowledge of value, knowing what can actually be
words, but knowledge workers have always been accomplished is the leverage they offer.

SOURCE: D ove, Rick. “The Knowledge Worker,” 1998, www.parshift.com/library.htm#0ther% 20publications, D ate accessed
Decem ber 3,2002.

------------------------------------a a a a a a ---------------------------------------------------------------------

who transforms business and personal experience into knowledge through capturing,
assessing, applying, sharing, and disseminating it within the organization to solve spe-
cific problem s or to create value. He or she is the “product” of experience, values,
processes, education, and the ability to be creative, innovative, and in tune with the cul-
ture of the corporation. For example, horse trailer salespeople do not sell trailers, but
they do sell horse transportation solutions. They have to know their customers’ busi-
ness, the shows they attend, distances to haul horses, frequency of use, and the horse
trailer line before they propose the right trailer to meet a custom er’s requirements. In
this respect, the salesperson becomes a knowledge worker. Managers, lawyers, doctors,
systems analysts, and accountants are all knowledge workers (see Figure 15.1).

PERSONALITY AND PROFESSIONAL ATTRIBUTES

A knowledge worker incorporates several personality and professional attributes:
• Holds unique values and understands and adopts the culture of the organization
• Aligns personal and professional growth with corporate vision and the achieve-
ment of strategic goals
• Adopts an attitude of collaboration and sharing
• Has innovative capacity and a creative mind
• Has a clear understanding of the business in which he or she is a part
• Is willing to learn, unlearn, and adopt new ways that result in better ways of
doing a job
• Is in command of self-control and self-learning
• Is willing to tolerate uncertainties and grow with the company
440 7777 PART V E thicalL egal, and Managerial Issues

FIGURE 15.1: M akeup of the Knowledge Worker

Corporations preparing their human capital to be knowledge workers in a com-

petitive environm ent must consider several core competencies of the self-directed
knowledge worker:

• Thinking skills. A knowledge worker is expected to possess strategic thinking

skills that shed potential on the work performed or the ideas provided on the job.
Strategic thinking means having a vision of how the product can be better, how
the company can improve by the value-added contributions of it employees, and
how continuous learning contributes to a knowledge worker’s career, loyalty to
the firm, and satisfaction on the job. The organizational behavior literature cites
two ways of thinking: right-brained thinkers, who tend to be intuitive and non-
linear in their approach to problem solving, and left-brained thinkers, who are
known for logic and decision making based on facts. For either type of thinker,
the point is for the knowledge worker to stretch his or her thinking to achieve
worthwhile results.
• Continuous learning. Knowledge work implies innovation through continuous
learning on the job, professional seminars, and working in an environment con-
ducive to creativity and advancement. This implies unlearning and relearning to
be in tune with the fast-changing business. It also means the learning corporation
must provide both the support and the funding for allowing its employees to con-
tinue to learn, in the hope that the outcome is better products or better quality
service, or both. In one professional’s words, “Learning is about working and
working is about learning” (Frank 2002).
• Innovative teams and teamwork. As competition becomes more intense, the prob-
lems facing today’s corporation become more complex, requiring innovative team-
work and joint decision making for solutions. Teamwork requires collaboration,
cooperation, and coordination, based on a knowledge-sharing attitude and com-
mitment to knowledge exchange. A prerequisite for successful teamwork is man-
agement support and attractive rewards, both intrinsic and extrinsic. For a career-
oriented knowledge worker, intrinsic rewards (such as recognition) go farther than
extrinsic rewards (such as salary increases or stock options) in the long run.
 CHAPTER 15 Managing Knowledge Workers aaaa 441
• Innovation and creativity. The spirit behind innovation and creativity is for knowl-
edge workers to expand their vision and “dream ” a new or a different product or
service for the advancement of the firm. For that to happen requires open and
equal opportunities to explore, to test, and to try things out. It means a solution
focus mind-set, a passion to create knowledge, and a strong desire for idea genera-
tion and follow-up. From management’s view, it means providing an environment in
which knowledge sharing is encouraged within the framework of people’s abilities
and potential to succeed on behalf of the firm. The ultimate goal is to focus on cre-
ating tomorrow’s product or service that the knowledge worker wants for the firm.
• Risk taking and potential success. Innovation and creativity mean risk taking.
Maintaining the status quo requires minimum risk, but it also breeds no change
for the betterm ent of knowledge workers or the organization. Like investing in
stocks, the higher the risk, the greater the chance of higher return on investment.
It could also result in a greater loss. What the knowledge worker or management
must do is first share and exchange the best knowledge available, work together
as a team, and make a joint decision with a calculated risk. In risk taking, you
have to be willing to lose as much as you hope to gain. Vision, experience, and
seasoned knowledge enter the picture in risk taking for the forward-looking firm.
It is part of what a learning organization has to undergo to gain and advance its
employees’ knowledge base.
• Decisive action taking. With all these factors considered, decisive action taking
means that knowledge workers should be willing to embrace professional disci-
pline, patience, and determination. Motivation is a critical factor in keeping the
focus on a product or a service for the future. Analyzing choosing among alterna-
tives, testing, and the selling of change must come before final adoption
(Garfinkle 2002).
• A culture o f responsibility toward knowledge. This core competency means loyalty
and commitment to one’s manager or leader. Knowledge workers must consis-
tently support their leaders, their peers, and the company as a whole. When a
problem arises, a knowledge worker is expected to take the problem to a respon-
sible source, discuss or brainstorm it, settle on a “best solution” outcome, and let
it go at that. In traditional organizations, employees with problems or grievances
often gripe about them through the rumor mill, which breeds m isinterpretation or
misunderstanding. This route finds no winners.
Related to this core competency is a subculture of referrals and knowledge
exchange. By that, we mean identifying knowledge workers who have specialized
knowledge for certain problems and who are willing to provide such knowledge
on call. A network of knowledge sources and knowledge availability is a critical
component of the learning organization.

■ Business Roles in the Learning Organization

By learning organization, we mean an organization of people with ingrained commit-
ment to improve their capacity, to create, and to produce—who respond to uncertainty,
to challenges in the marketplace, and to change in general. Senge defines a learning
organization as “a group of people continually enhancing their capacity to create what
they want to create” (Senge 2002). M alhotra defines it as an “organization with an
ingrained philosophy for anticipating, reacting, and responding to change, complexity,
and uncertainty” (M alhotra 2000). In brief, the rate of learning of an organization may
become the most critical source of competitive advantage.
442 7777 PART V Ethical, Legal, and Managerial Issues

To a learning organization, data and information are givens. We have seen suc-
cesses in data processing and information processing from the 1960s to the 1980s. The
focus then was on efficiency, where computers replaced human redundant arithmetic
(algorithmic) work. There were quantitative savings and everyone benefited. In the
1990s, information was collated, processed, and converted into relevant (nonalgorith-
mic) knowledge for the decision maker. A t this level, the focus shifted from quantita-
tive to qualitative performance-oriented value-added decision making. Captured and
processed knowledge became the fundam ental building block for today’s learning
organization, where everything operates creatively in real-time as problems arise.
Personalization in electronic commerce and just-in-time (JIT) inventory management
are examples of today’s learning organization. The applications are smart enough to
personalize the inform ation based on the user’s current location and needs (see
Figure 15.2).

MANAGEMENT AND LEADERSHIP

In knowledge management, we distinguish between managers and leaders. Traditional
managers focus on the present. They are action-oriented, spending most of the time
delegating, supervising, controlling, and ensuring compliance with set procedures. They
were once workers or superemployees and were prom oted to managers. When they
manage subordinates, they know all aspects of the business, because they were once
there. The goal of management is stability on the job and meeting deadlines.
In contrast, smart managers focus on organizational learning to ensure operational
excellence. Because of continuing change and improvements in the workplace, they
cannot be expected to have mastered the work of subordinates. As a result, they take
on the role of leaders, where change is the primary goal. Clearly, the challenge is to get
the departm ent or the organization moving in the direction of the goal(s) in line with
the rate of change. Their focus is on the future, developing strategies and sharing vision
through effective communication with knowledge workers. Learning becomes the key
focus for the organization’s survival and growth.

i fl fl I 9 1 FIGURE 15.2: From Data Processing to Self-Learning —A Trend

 CHAPTER 15 Managing Knowledge Workers ■■■i 443
When it comes to leadership, the leader’s role in a learning organization is more of
a facilitator than a supervisor, a teacher than an order giver, a steward of the collective
knowledge of his or her staff than a reporter to top management, and a designer more
than the traditional role of merely seeing things done. A leader also has responsibility
for knowledge workers who are continually expanding their capabilities to mold their
future. This has human resources management implications, where the leader has a
serious commitment to learning.
Learning and teaching are two faces of the same coin. In teaching, the focus is on
knowledge transfer from the instructor to the learner. The instructor is supposed to be
the expert. The role is to deliver quality content and to communicate content with
potential. The interaction should instill serendipity and thinking about better ways of
handling problems. Learning should promote a new way of thinking, not just facts. The
key is not listening and retaining ideas or knowledge, but raising questions that might
trigger new ways of decision making or problem-solving. When this is applied in a
learning organization, the smart manager becomes the instructor, and the knowledge
workers become the learners. A smart manager provides opportunities for knowledge
workers to brainstorm ideas, exchange knowledge, and come up with new ways of
doing business. All that is carried out in a culture amenable to change. As a leader, a
smart manager personally leads discussions, poses questions, and provides constructive
feedback. The end result is sending a signal to knowledge workers that knowledge is to
be shared, not hoarded, especially among coworkers. As Ralph Waldo Emerson, the
Am erican essayist, noted over 100 years ago: “ I pay the schoolm aster, but ’tis the
schoolboys that educate my son” (Gavin 2002).

WORK MANAGEMENT TASKS

A key ingredient in management and leadership is the way managerial experience is
processed. Smart managers learn from their experience rather than being bound by
their past habits. The trick is not so much focusing on what managers learn, but how
they learn what they learn —the process of learning. With this in mind, managers in a
learning organization recognize and reward creative thinking, openness, and creativity.
Work management tasks include the following:

• Searching out, creating, sharing, and using knowledge in everyday activities

• Managing knowledge workers and cultivating their knowledge-oriented activities
to achieve effective results by using time and mental resources efficiently
• Maintaining work motivation among knowledge workers
• Ensuring readiness to work, especially during an emergency
• Allocating effort and switching control among tasks
• Managing collaboration, coordination, and concurrent activities among knowl-
edge workers
• Sharing information and integrating work among knowledge workers (Davis 2001)
• Hiring or recruiting bright, knowledge-seeking individuals that keep the intelli-
gence “pum p” primed for future achievements of the firm

It should be noted that measuring productivity of knowledge workers is not that

simple. Traditional productivity measures are inadequate and inaccurate. At best, they
are subjective and indirect. For example, how would one decide between two knowl-
edge workers who plugged the same leak in an oil well? Is it by the shortest time
taken? The least costly approach? The most durable weld? Work efficiency addresses
cost reduction in terms of time and effort. In contrast, work effectiveness emphasizes
444 7777 PART V Ethical, Legal, and Managerial Issues

increased value in meeting the requirements of the task. Both are important, but the
approach to productivity measures is bound to be different.
In managing knowledge workers, a m anager must consider several factors that
limit knowledge worker productivity and ways to get around them:

• Time constraint. Time is the enemy of successful knowledge workers. There is

always more work to do. As a result, either quality suffers or completion time
lags. Obviously, such stress can work against one’s motivation to contribute.
• Working smarter and harder and accomplishing little. This frustrating result is
often triggered by limited time, limited staff support, or financial constraints.
Management can do a lot to alleviate this type of productivity problem.
• Knowledge workers doing work that the firm did not hire them to do. The way to
get around this is for the smart manager to explore the specificity of the task or
the job, match the task to the knowledge of the worker, and eliminate the
nonessentials. Unlike manual or skilled labor, knowledge workers are hired
because they know something the firm does not. This means that a knowledge
worker is expected to open up to the organization, explain what they know and do
not know, and provide the goals that their knowledge will produce for the firm.
• Work schedule. Heavy work demands invariably affect a knowledge worker’s
attention span, motivation, and patience, regardless of pay or benefits. The man-
ager should be careful in planning work schedules and work rotation to assure
cooperation and successful completion of jobs on schedule.
• Motivation against knowledge work productivity. Knowledge workers are not all
programmed to follow the ideals proposed by management. Avoiding task uncer-
tainty or job complexity can pose productivity problems and affect the productiv-
ity of other knowledge workers if their work depends on one another’s input.
Motivation is also affected in situations where urgency supercedes motivation. If
productivity takes a nosedive, the knowledge worker can always blame it on the
time constraint, lack of adequate input, and the like.

In summary, knowledge worker productivity is a challenge, requiring tender care,

personal attention, consistent recognition, and timely rewards. Managing knowledge
workers requires expertise in handling specialists with control of corporate knowledge as
the core asset of business. Strategic planning means carefully selecting a knowledge
worker when they join the organization, matching the knowledge worker’s vocational
needs with the requirements of the job, monitoring progress and improvements made
over time, and ensuring stability and tenure on the job.

Work Adjustment and the Knowledge Worker

Smart managers strive to ensure the right match between the vocational needs of their
knowledge workers and the requirements of their jobs. The goal is to assure stability of
the workforce and continuity on the job in the interest of the corporation. One
approach to work adjustment is a model developed by a group of counseling psychol-
ogists at the University of Minnesota who were concerned with the general problem
of adjustment to work (Dawis, Lofquist, and Weiss 1968). The model is based on the
concept of “correspondence” or a match between the individual and his or her work
environment.
 CHAPTER 15 Managing Knowledge Workers aaaa 445

FIGURE 15.3: A Work Adjustment M odel for Knowledge Workers

Achieving and m aintaining correspondence with the work environm ent are
viewed as basic motives of human work behavior. As shown in Figure 15.3, correspon-
dence begins when the individual brings certain skills (abilities) that enable him or her
to “respond” to the requirem ents of the job or the work environment. On the other
hand, the work environment provides certain rewards (reinforcers such as wages, pres-
tige, personal relationships, and so forth) in response to the individual’s requirements.
W hen both the individual’s and the work environm ent’s minimal requirem ents are
mutually fulfilled, correspondence exists.
When the individual achieves minimal correspondence, he or she is allowed to
stay on the job and have an opportunity to work toward a more optimal correspon-
dence and to stabilize the correspondence relationship. For example, an experienced
teller in a commercial bank who handles customer problems effectively day after day
is rewarded with a prom otion in pay and rank to senior teller and on to head teller
and beyond. It is presum ed that the teller is satisfied on the job and is also consid-
ered a satisfactory employee (satisfactoriness) by the bank. W hen this relationship
(correspondence) continues, it contributes to job tenure and stability of the bank as
an entity.
PART V Ethical Legal, and Managerial Issues

PROFILE OF VOCATIONAL NEEDS AND REINFORCERS OF KNOWLEDGE WORKERS

In a 2002 preliminary study (Awad 2002) of select knowledge workers in the teller depart-
ment of a medium-size bank, knowledge workers reported several vocational needs:
1. Achievement or a drive to accomplish worthwhile, complex tasks and a feeling of
accomplishment
2. Use of their abilities on matters related to problem-solving and solutions rather
than problem implementation based on a stable set of predetermined, mecha-
nistic tasks
3. Authority exercised in terms of telling peers and others what to do
4. High pay and the prestige that goes with high compensation
5. The congenial atmosphere created by knowledge workers as coworkers
6. The freedom to try out their own ideas and create a new way of doing things
7. Recognition by both departm ent officials and peers for work done
8. The chance to exercise responsibility in planning one’s work and the work of others
9. The drive to do different things (variety) within the job scope from time to time
10. The social status that accompanies knowledge work where one can be important
(somebody) in the eyes of customers and peers alike
11. Creativity or a chance to try out his or her own ideas, because no two customers
or customer inquiries are ever the same
In contrast to the knowledge worker, traditional tellers’ vocational needs relate to
the following dimensions:
1. Handling a stable set of predetermined, relatively mechanistic tasks such as with-
drawals and deposits
2. Independence on the job, in terms of doing their work alone rather than with
coworkers; minimum socialization
3. The security of a steady job
4. Opportunities for advancement
With these vocational needs, the reinforcers are the efforts on the part of the
employer to meet these needs and support the motivational factor that is part of the
knowledge worker in general.

SMART LEADERSHIP REQUIREMENTS

Knowledge work requires smart leadership that can facilitate effective use of knowl-
edge, where time and timing become critical in the com petitive environm ent.
Obviously, when dealing with smart people, you need a smart manager on your side.
The critical point here is the knowledge chain , where a smart manager will focus not
only on the intellectual capital of the corporation but also on the return on time.
The knowledge chain is a series of steps that determine the potential of a learning
organization. One approach offers five key steps:
1. Assessment o f the core competency o f the organization. The key question for the
smart manager is “What are we doing (not what we make) that makes us what we
are?” This addresses the company’s current inventory of skills, competencies, and
talent. In today’s global market, products must be regularly reinvented, reshaped,
or refashioned to meet continuing changes in consumer tastes and preferences
and to deliver the next more successful product or service in time to beat the
com petition. In A ndrew C arnegie’s words “The only irreplaceable capital an
organization possesses is the knowledge and ability of its people. The productiv-
 CHAPTER 15 Managing Knowledge Workers a i a a 447
ity of that capital depends on how effectively people share their competence with
those who can use it” (Koulopoulos and Frappaolo 1999).
2. Response to the organization’s internal shortcomings. Traditional organizations
are known for controlling and often suppressing new ideas. Even when they do
not, in most cases, managers take credit for innovations rather than the em -
ployee^) in the departm ent or division. A truly learning organization encour-
ages new ideas and knowledge sharing among employees through job coordina-
tion and effecting change in time to meet or beat market demands. Implied in
this step is an assessment of how quickly core competency can be put to effec-
tive use to introduce a product to the market or respond to external company
demands. A smart manager should have a grip on where the company’s core
competencies lie and how to take advantage of and reward its return on time.
3. Vivid knowledge o f the external market and the tricky nature o f competition in the
marketplace. A classic example is Amazon.com. Soon after Amazon.com was
founded, top management began to realize that in order to keep customers they
needed to give advice to book buyers by posting book reviews and reviews of other
purchasers on the Web site. Amazon.com also decided to post information about
books with similar authors or topics. This unique approach, where human and com-
puter intelligence are coordinated for marketing strategy, paid hefty dividends.
A smart manager in a learning organization must assess the external envi-
ronm ent and introduce innovative features that dem onstrate his or her com-
pany’s superiority over the competition.
4. Online response to the com pany’s external environment. This includes the com-
pany’s customers, vendors, suppliers, community, government agencies, threats,
and opportunities. One of the main problems with e-merchants is their inability
to deliver merchandise on time or in time to be useful or satisfactory to the cus-
tomer. A smart manager knows the serious nature of this link in the knowledge
chain and will make every effort to fulfill this step effectively. To do so means set-
ting strategies and procedures, especially for emergency situations, to ensure the
organization's viability.
5. Measure the return on time. Knowledge has already become the most important
factor of production. Learning organizations have realized that having tacit
knowledge and putting such knowledge to good use is what powers today’s
knowledge economy. A smart manager recognizes the fact that the traditional
way of commanding workers to put their tacit knowledge to work for mere pay
no longer works. The approach to knowledge content is to voluntarily share it for
the organization to compete, innovate, and succeed. The magic ingredient is inno-
vation. Philip Kotler, a noted marketing specialist, said, “Every firm should work
hard to obsolete its own product line before the competitors do” (Koulopoulos
and Frappaolo 2000). Innovation and unique ideas increase returns through
product design knowledge and redesign. The return on time is like a feedback
loop that judges the effectiveness of new knowledge over time.

Technology and the Knowledge Worker

Knowledge work focuses on thinking, using information for processing, and recom-
mending value-added change. The primary activities of knowledge work are assess-
ment, decision making, monitoring, and scheduling. A knowledge worker is a manager,
a supervisor, or a clerk who is actively involved in thinking, processing information,
analyzing, creating, or recommending procedures based on experience and cumulative
PART V Ethical, Legal, and Managerial Issues

knowledge. For example, a stockroom clerk’s job may be highly structured; therefore, it
provides limited knowledge work. In contrast, a head teller job in a bank entails a sig-
nificant amount of knowledge work, including the following:
1. Assessing the amount of cash available in the teller’s drawer before the lunch break
2. Deciding on the procedure to use in handling irate customers
3. M onitoring cash supply for the drive-in tellers on Friday, when employees of a
local plant cash their paychecks
4. Scheduling lunch breaks of available tellers during the flu season
Most people have an innate ability for problem-solving. However, routine activi-
ties often make it difficult to concentrate on the creative phase of problem-solving.
This situation can be improved in an IT environment that provides the following:
1. A knowledge-based retrieval system that generates knowledge and ideas quickly
2. Interactive functions to derive feasible solutions
3. Functions to communicate a knowledge w orker’s activities to the appropriate
people at the appropriate time via technology such as e-mail or an intranet
IT plays a role in the learning organization in three key processes: knowledge capture,
information distribution, and information interpretation. In knowledge capture, a place
for IT is in market research and competitive intelligence systems. Scenario-planning
tools can be employed to generate the possible futures (Malhotra 1996). Likewise, the use
of e-mail, intranets, and bulletin boards can facilitate information distribution and inter-
pretation. The accumulated archives of such traffic can develop the basis for organiza-
tional memory.
It should be noted that the mere availability of technology does not assure a learn-
ing organization. Take the example of a multibillion-dollar bank, where the author
complained about a cash deposit that would not appear on the customer’s online bank-
ing screen until the following day, after the deposit had been processed. The bank’s
senior vice president of operations replied. “Well, everything, whether cash or checks,
must be held overnight and cannot be made available for the customer before the next
day.” When asked, because the deposit was in cash, how the customer could withdraw
some of this money during the day when the computer did not reflect a deposit. The
answer was, “We try to discourage customers from doing so. We don’t want any cus-
tomer who makes a deposit in the morning to withdraw money later in the same day.”
The contrast is a smaller bank in the same community, where a customer can view
any deposit, withdrawal, or other activity within seconds from anywhere —by the teller,
via ATM, or through the customer's personal computer using the bank’s online bank-
ing system. Obviously, true online banking favors the latter bank, based on the policy
of full customer service. In fact, the latter bank provides newer, quicker, and cheaper
services for its customers on a regular basis.
The ultimate goal of technology is to serve organizational memory and to create a
working environment that provides these conditions. There are various types of equip-
ment and software unique to a knowledge w orker's tasks. They include intelligent
workstations, e-mail, and local area networks. E-mail and local area networks have
been covered in C hapter 3. The intelligent workstation autom ates repetitive, tedi-
ous tasks, freeing the knowledge worker’s time for more creative work. Through the
keyboard or the mouse, the end user performs a host of functions, ranging from cus-
tom er tracking to creating illustrated reports and from calendaring to working with
knowledge-based systems. The single most used application for the executive work-
station, however, is the electronic spreadsheet.
 CHAPTER 15 Managing Knowledge Workers i i aa 449
The capabilities of today’s intelligent workstation fall in line with linear thinking—
processes that are carried out step by step in a specific sequence. The learning organi-
zation depends on the knowledge worker to synthesize the information and develop
conclusions. In addition to the mundane tasks, an intelligent workstation should per-
form the following functions:
1. Administrative support functions. These functions include such tasks as maintain-
ing tickler files, scheduling meetings, and maintaining electronic telephone direc-
tories. With the position of private secretary on the wane, professionals are taking
on a number of support tasks.
2. Personal computing functions. These include spreadsheets and business graphics.
3. Intelligent databases. These databases allow knowledge workers to create and
maintain information on an ad hoc basis.Today’s intelligent databases easily han-
dle files, voice messages, and message-switching.
With knowledge worker time at a premium, learning organizations are offering flexi-
ble schedules through telecommuting. The case summarized in Box 15.2 illustrates the role
of technology and networking in today’s knowledge worker’s performance and lifestyle.
A knowledge worker is expected to possess professional experience and technical
know-how in order to access, update, and disseminate information and ideas from
databases and knowledge bases. The com puter should be owned by the knowledge

1 I 1 I 1 I BOX 15.2 fl fl fl fl fl fl-------------------------------

FLEXIBLE SCHEDULING THROUGH TECHNOLOGY

THE BEST PLACE TO WORK avoid losing them is one reason why this consult-
ing firm ranks among the top 25 employers for
John Lewis, junior partner at a nationwide con-
benefits and low turnover.
sulting firm, became the father of twins. He and
Among the unique benefits consulting firms
his wife already had a boy and a girl from previ-
offers knowledge workers like John Lewis are
ous marriages. The new additions meant more
extended leaves of absence so they can pursue
demands on both parents. In response, John’s
academic degrees or a lifelong dream. In fact, such
company told him to work from home when nec-
a firm gives its employees a chance to set their
essary. This m eant scheduling meetings with
performance goals, review them with their mana-
clients by phone and planning meetings with his
gers, and periodically perform self-assessment to
staff around slack times. John happens to run a
see how well they’re meeting their goals and to
$35 million developm ent project with a new
determ ine ways to correct deviations from the
client, and working from home turned out to be as
goal. The shared responsibilities, teamwork, and
effective as working from the office.
collegiality of the firm are well known among
Without telecommuting, via e-mail and tele-
competitors, vendors, and customers. In fact, it has
conferencing technology, John might have had to
been known to generate more business and ensure
leave his job downtown (a 20-mile drive each way)
security and stability of the workforce.
and look for another job with a company closer to
his home. Accommodating employees this way to

SOURCE: Excerpted from Zetlin, Minda. “Best Places to Work,” Computerworld , May 6,2002, p. 40ff.

l fl I fl a l
aaaa PARTV Ethical, Legal, and Managerial Issues

worker rather than by the company. Knowledge workers should keep their software
and be able to take their computer from one job to another. One would expect them to
customize and handle jobs within their range of experience.
Since the late 1990s, several software packages have become available; their intent
is to em ulate a knowledge w orker’s tasks. For example, KnowledgeW orker from
Datum Consulting is designed to store the experience and skills of company employ-
ees. Through an Internet browser, you can create a secure picture of the unique devel-
opments on jobs, projects, work flows, team communications, and the like. The inte-
grated com ponents include a knowledge library that manages documents that are
easily navigable via a Web browser; a business work flow to manage business
processes; a digital dashboard to promote collaboration and knowledge sharing; and a
search engine that allows users to surf a company’s corporate information as if they
were surfing the Web (Datum Consulting 2002).

KNOWLEDGE WORKER’S SKILLS

In addition to technical skills and abilities, a knowledge worker is expected to have
both professional experience and soft traits such as a sense of cultural, political, and
personal aspects of knowledge in the business. The personal aspects include open, can-
did, and effective communication skills, a warm and congenial personality that accom-
modates knowledge sharing in a group setting, sensitivity to the political pressures in
the departm ent or organization in general, and accommodating the job as well as peers
within cultural constraints. These are people who want a challenge and to be on a win-
ning team. They employ and create knowledge, and that is a high-demand intuitive
skill. They prefer to be employed by a company that creates knowledge, not one that
employs knowledge. All of these might seem ideal, but they can be a reality from the
recruiting phase to the maintenance phase of human resources management.
Lawyers, consultants, and scientists are all knowledge workers. One difference
between knowledge workers and white-collar workers is the level of education and
specialized training. As a rule, the majority of knowledge workers have a university
degree. The question of whether a knowledge worker should have a vocational degree
versus a university degree continues to be a point of debate. Because knowledge work
is about creating, manipulating, and sharing knowledge in the workplace, it is obvious
that one needs more education than stacking products on supermarket shelves. Like it
or not, employers continue to be more impressed with paper qualifications than practi-
cal experience. A paper qualification might not assure competence, but it is a door
opener with long-term potential for the applicant.
A university degree is often not needed in a specialized area of work, such as dous-
ing oil well fires or designing and running a high-level video- and audio-oriented Web
site. More specifically, most employers expect the university degree to be in the area of
the assigned job or department. Like medicine, a patient expects the specialist, not the
paramedic, to handle serious surgery. One problem with today’s higher education,
however, is that most students prefer to take courses with a vocational (hands-on)
drift, thinking that such experience will land a good job. To sit back and think through
a concept or a strategy is a chore to many students.

Role of Ergonomics
With regular use of the computer day after day, end-user performance and comfort are
closely interrelated. No matter how sophisticated the electronic support, a knowledge
worker can achieve full productivity with proper design of the work environment.
 CHAPTER 15 Managing Knowledge Workers 1 «« i 451
Ergonomics is a key issue here. Ergonomics involves comfort, fatigue, safety, under-
standing, ease of use, and any other areas that affect welfare, satisfaction, and perfor-
mance of knowledge workers working with user-machine systems.
The list of factors that affect the ergonomics of knowledge workers falls into three
categories:
1. E nvironm ental issues that include pro p er lighting, layout, and temperature.
Indirect lighting, covering windows near the workstation with blinds or curtains
to reduce the glare, and positioning terminals at right angles to windows —all
contribute to work productivity.
2. Hardware issues that focus on furniture, com fortable seating, and well-designed
workstations. An ergonomically acceptable chair should be adjustable and sup-
portive. Proper lum bar (lower back) support eases back strain over extended
hours of work (see Figure 15.4). The workstation itself might have a built-in
swivel to tilt to the angle of the user.
3. Knowledge w orker-system interface. This addresses software, user training, and
easy-to-follow documentation.
The knowledge worker-system interface emphasizes several features:
1. M inim um w orker effort and memory. This means that inform ation should be
entered only once and nonproductive work eliminated. Documentation should
also be available online in the form of a help routine. User manuals should be
clear, complete, and easy to follow.
2. Best use o f human patterns. Workstations should consistently place similar infor-
mation on screens in the same position, by familiar screen formats, and so on.The
knowledge w orker should use a consistent approach and term inology for all
functions. Performing these functions should require minimum training.
3. P rom pt p roblem notification. The knowledge worker should be alerted to
changes taking place that might adversely affect problem solution. For example,
if a file is approaching capacity (say, 90 percent), a message should be displayed
to indicate the status of the file.
4. M axim um task support. Task docum entation should be com plete and readily
available so that the knowledge worker is not required to use other resources for
task performance.

FIGURE 15.4: Ergonomic Environment and Features—The Basics

«aii PART V Ethical, Legal, tf/id Managerial Issues

Demand for systems with good ergonomic designs continues to increase, pointing
to the growing im portance of the human factor. The benefits that ergonomics con-
tribute to productivity are not readily quantifiable, partly because it is not easy to m ea-
sure productivity.

Role of the CKO

Some experts feel that KM, by definition, should be centralized; others prescribe that a
specialized person be accountable for knowledge creation, transfer, organization, dis-
semination, and overall management. Toward that end, many large corporations such
as General Electric, Dow Chemical, Texas Instruments, and Chevron Corporation have
created a full-time position of chief knowledge officer (CKO) (see Box 15.3).
As the knowledge management czar, the CKO has the main mission of identifying
knowledge within the company and encouraging workers to share it. Specifically, his or
her main functions include the following:

• Maximize the returns on investment in knowledge —people, processes, and

technology
• Share best practices and reinforce the goodness of knowledge sharing among
employees on a regular basis
• Promote and improve company innovations and the commercialization of new ideas
• Minimize “brain drain” or knowledge loss at all levels of the organization

An effective CKO must be good at communications, interpersonal skills, technical

skills, and people management. He or she must be a thinker, enforcing the big picture for
top management. Other attributes include being an evangelist, a facilitator, and a jug-
gler (Flash 2001). Interpersonal skills instrumental in CKO work include the following:

• Teaching and selling. This includes educating the user on technology; selling and
promoting change through demos; seminars; and specialized behavioral training.
In the end, the CKO must make the user want to “buy” change.
• Communicating. The CKO must speak the language of the user, mediate, work
with management at all levels, and be politically sensitive.
• Understanding. This includes identifying problem areas and determining their
impact, grasping employee expectations, and being sensitive to the impact of the
KM system on people.

The key technical skills are as follows:

• Broad knowledge of business practice in general and the ability to translate tech-
nical information at the employee level
• Dynamic interface, making effective use of technical and nontechnical elements
in KM design, especially if the CKO is the chairperson of the KM team
• Knowledge of information technology, information systems, software, and tech-
nology in general; knowing how information works and how it is transformed
into knowledge

In real life, we expect greater emphasis on interpersonal skills during the early phase
of the KM development life cycle. Technical skills become important during knowledge
capture, knowledge storage, data mining, and knowledge organization. In addition
 CHAPTER 15 Managing Knowledge Workers ■a a i 453

I 1 fl 1 1 I BOX 15.3 fl I I fl I fl

ROLE OF THE CHIEF KNOWLEDGE OFFICER

Agencies are looking beyond chief information offi- ample, she plans to im plem ent portals through
cers to take advantage of the reams of information which users could make requests and a Web crawler
stored in a multitude of systems. To meet the chal- that would gather information overnight. She also
lenge, some agencies are creating a new post: chief said successful implementation of KM consists of
knowledge officer. The groundbreaking for the foun- five components: business purpose and measure-
dation of knowledge management within agencies ment of results; strategy; leadership; a culture of
occurred with the June 1999 appointment of Shereen sharing; and technology. However, the emphasis
Remez as the General Services Administration’s— in her office is on the human com ponent—where
and the government’s—first CKO. the knowledge worker is most important. “Knowl-
Admiral George Naccara, the Coast G uard’s edge management will continue to grow as we see
CIO, said the guard is selecting a CKO who will continued pressure on the government to per-
also serve as deputy CIO. Knowledge within the form more efficiently. It will allow us to reuse and
Coast G uard is transitory because personnel innovate ways to use knowledge.”
move from station to station, taking expertise Susan Hanley, knowledge management director
with them, he said. The CKO will help ensure that for American Management Systems, Inc., of Fairfax,
knowledge is not lost and improve critical deci- Virginia, has said that to succeed, a CKO must lever-
sion support, he said. age an organization’s technological infrastructure to
Remez said knowledge management requires meet the data requirements of users and customers.
data sharing at an enterprise-w ide level and Remez says that citizens are demanding the same
bridging local islands of inform ation. For ex- type of service from the government.

SOURCE: Excerpted from Date, Shruti. “A gencies Create CKO Posts to Get in the Know," Government Computer News ,
Novem ber 8 ,1 9 9 9 ,pp. 1-2, www.gcn.com /voll8_no36/news/950-l.htm l, Date accessed October 20.2002.

fl fl I fl I I

to these skills, academic preparation and a career in knowledge management are pre-
requisites for filling this emerging position in industry. For example, to help prepare people
for CKO work, the University of California-Berkeley’s School of Information Manage-
ment Systems offers graduate courses for knowledge managers. They include offerings in
social science, management theory, and information technology. A list of these and other
courses are available on UC Berkeley’s Web site at www.sims.berkeley.edu.
In terms of roles, the CKO performs multiple roles such as:

• Agent o f change. This requires having the style to sell the change from knowledge
hoarding to knowledge sharing. The style ranges from that of persuader to
imposer of change, depending on employees’ level in the organization and the
leverage of the CKO.
• Investigator. This includes identifying the real problem(s) in knowledge sharing
and mapping out procedures for alternative solutions.
• “Linking pin. ” This means performing the role of liaison between employees’
expectations and how a KM system must perform to meet those expectations.
• Listener. This includes reaching people, interpreting their thoughts, and drawing con-
clusions from interactions. Listening is the other side of effective communication.
454 7777 PART V Ethical, Legal, and Managerial Issues

• Politician. This means solving a problem by not creating another—an art in itself.
A successful CKO knows who to contact, what to say, and how to use his or her
knowledge to maximize the use of corporate knowledge for profitability and per-
formance. Diplomacy and finesse in dealing with employees and managers can
also improve the acceptance of a new KM system environment.

In 1999, many organizations had no staff devoted to knowledge management.

Today, the number of people dedicated to KM is growing rapidly, although the title
“CKO” remains rare (Flash 2001). Box 15.4 summarizes where CKOs come from and
why organizations decide to hire one. When it comes to the serious decision to appoint
a full-time CKO, the least the CIO should do is to support the cause. The CIO is not
only the leader of KM efforts, but also the guarantor of its success.

1 I I « 1 1 BOX 15.4 1 i i M 1 1

DESCRIBING THE JOB OF THE CKO

WHO IS THE CKO? says. He helped to identify the experts who had
worked on the B-2 and used video recordings to
Given the range of responsibilities involved, it is not
capture their knowledge. The team indexed the
surprising that the backgrounds of actual chief
interviews on the com pany’s Web sites to make
knowledge officers vary. David Owens, CKO of the
them available to others.
St. Paul Cos., an insurance company in St. Paul,
In 1999, Hallmark Cards, Inc., realized that its
Minnesota, is a former teacher. SAICs Greenes is a
customer research division had built vast reposito-
geophysicist who moved through management posi-
ries of data and information, but no one knew
tions into human resources. Dow Chemical Co/s
what to do with it all. The company appointed Tom
Jim Allen worked as a manager in research and
Brailsford manager of knowledge leadership in
development before becoming director of KM
charge of creating a KM infrastructure to tie the
under the CIO of the Midland, Michigan, company.
fragmented sources together and convince the cor-
The career routes that knowledge leaders
porate culture to embrace this knowledge. The
take to their positions show a similar diversity.
result was an intranet site called Voice o f the Market-
Scott Shaffer, project m anager for knowledge
place, which includes customer research data that is
management at Northrop Grumman Corp. in Los
accessible throughout the company.
Angeles, rose through the ranks of the company
No matter who the knowledge leaders report
and took on a knowledge leadership role that
to, CKOs agree that to succeed they must have
evolved over time. After 17 years with the defense
support at the top. Getting executive support can
contractor, Shaffer took a position as program
be difficult because it is not easy to demonstrate
manager on the B-2 stealth bom ber program in
the explicit financial benefits of knowledge manag-
1997. He became concerned about how the com-
ement. “My CEO has to make a leap of faith and
pany would maintain its knowledge about the B-2
say T believe this is right,’ even though I can’t
over the life of the aircraft, especially after the
prove it in dollars and cents,” says Sam Saint-Onge
engineers who had designed it left.
of Clarica Life Insurance Company.
“A t the time, we didn’t know about knowl-
edge management, but we discovered it,” Shaffer

SO U RC E: Excerpted from Flash, Cynthia. “Who Is the CKO?” Knowledge Management, May 2001, pp. 37^1.
 CHAPTER 15 Managing Knowledge Workers aaaa 455

«»j » Managerial Considerations

Regardless of the product, service, size, or culture of the organization, managing knowledge
workers is unique. It requires a special breed of managers and a new set of responsibilities
and challenges. Included in the managerial phase are job definitions, job performance evalu-
ations, and an attitude to accommodate the value system and the culture of the organiza-
tion. For the organization, it means facilitating collaboration among work groups and pro-
moting the productivity potential of technology. For the knowledge workers, it means
managing task schedules; automating repetitive, labor-intensive jobs; and providing an envi-
ronment where knowledge workers can concentrate on challenging and creative work. In
this case, the managers should support intelligent work and articulate, recognize, and reward
the knowledge worker for the contributions they make at the time when they make them.

BECOMING A CHANGE LEADER

The greatest difficulty in dealing with people is not so much convincing them to accept
new ideas, but in getting them to abandon old ones. Managing knowledge workers is no
different. If one asks a knowledge worker “What do you expect your job to offer?” he
or she is likely to cite the following:
• An opportunity to do creative work
• A chance for independence of decision making
• Good and conducive interpersonal work environment
• Management support
• Intrinsic rewards like recognition and appreciation of quality work
• Potential for advancement
• Supportive company policies
• G reater responsibility and autonomy on the job
• Variety in the job
For the change leader, the end result is likely to be improvements in efficiency and
performance, improvement in focus by devoting more time to the primary functions of the
job, and reduction in the amount of work that must be done. From the organization’s side,
what could block the change are fewer knowledge workers available to do the work, lack
of up-to-date technology and support, inconsistent decision quality, and high turnover.
With that in mind, it is important that a change leader consider the following:
• Focus less on problem s and more on successes and opportunities. A negative atti-
tude with a focus on problems rather than successes is the wrong recipe for a
knowledge worker manager. Certain problems can be tabled, while other prob-
lems must be addressed as they occur. A change leader must prioritize each prob-
lem, decide on its severity, and determine who, when, and where it must be
resolved. This means that problems must be viewed in the context of the larger
picture for the product, departm ent, or organization as a whole. Likewise, suc-
cesses have to be touted, and knowledge workers behind them should be recog-
nized to promote excitement and motivation for others to follow.
• A d o p t an attitude that views challenges as opportunities. One of the unique expec-
tations of a change leader is strategic thinking and a knack for professional disci-
pline. Challenges will come up at times when they are least expected. They also
bring with them unexpected difficulties. A change leader must find a way of pre-
senting or supporting challenges as ones with potential opportunities for a new
product or better service, despite the risks that invariably accompany challenges.
• Work on creating tom orrow’s business instead o f hammering on yesterday's problems.
■■ si PART V Ethical, Legal, and Managerial Issues

Managing Knowledge Projects

Managing the hum an dimension of knowledge work is the most challenging job a
knowledge m anager can have. Knowledge managers coordinate special knowledge
projects designed to take a product or a service a step forward under competitive pres-
sures. TTie main managerial functions include drafting knowledge teams, deciding on
customer requirements, identifying project problems, and ensuring successful value-
added results. The role goes beyond management. A knowledge manager is expected
to possess psychological, technical, and business skills. He or she needs to be on top of
all these aspects, depending on the project.
Of all the challenging skills, the main challenge is assembling the knowledge team.
As described in Box 15.5, the focus is first on people, processes, and cultural aspects of
the job. The use of technology to facilitate knowledge sharing comes later in the team ’s
operations. Overall, a knowledge manager must leverage the strength of each team
member by the assignments made, the leadership provided, and the support given.

THE SOFT SIDE ALWAYS WINS

People will always be people. A knowledge manager must recognize certain challenges
that have to be faced throughout a knowledge project. Convincing people to share indi-
vidual experiences has been an elusive idea for most organizations. In a learning orga-
nization, where knowledge work reigns supreme, imagine the benefits of the following:

• The newest recruit in a customer service departm ent addressing customer com-
plaints as effectively as the most senior representative
• Any salesperson in the field counting on the collective knowledge of the experi-
enced staff anytime, anywhere

Managing knowledge projects, where knowledge workers are involved, requires

the knowledge manager to focus on the following:

• Encourage every team member to create new knowledge in the interest of the
project. New knowledge creates wealth, which is the backbone of a learning
organization.
• Help knowledge workers do their jobs. A knowledge manager must help them
understand the company’s business environment and its strategic orientation, and
how they can apply them to the project at hand.
• Allow knowledge workers to participate in major company decisions, which can
pay off in intrinsic and extrinsic benefits for the company and employees alike.
This includes company image building, improved quality products, and stronger
identification with a successful employer.
• Encourage knowledge workers and employees in general to learn as they earn a
living on a day-to-day basis. This is made possible through on-the-job learning, job
rotation, training sessions, and self-training (see Box 15.6).

An organization that manages to connect employees to employees and employees

to m anagers to collectively solve problem s makes the m om entum unstoppable.
However, we know from experience that once an organization grows larger, it tends to
become more impersonal, which impedes knowledge sharing. This is where the chal-
lenge becomes obvious for the knowledge manager; that is, the challenge is managing
the connection among people, processes, and technology on a full-time basis. It is also
focusing not only on knowledge availability and dissemination like what one finds in a
 CHAPTER 15 Managing Knowledge Workers aaaa 457

fl ■ 1 I fl fl BOX 15.5 I fl fl fl fl fl

DRAFTING KM TEAMS
KNOWLEDGE TEAM FORMATION Inc., in Fort Lauderdale, Florida, has derived six key
roles, which he calls “hats,” that one team member
By definition, it takes at least two people to share
or another must play. In principle, one person might
something. When that something is knowledge
wear more than one hat if the team is small.
applied to the goals of a larger enterprise, the shar-
ing multiplies and the number of people involved 1. Business leads specialize in understanding the
must increase. Therefore, most organizations make measurable results to be obtained.
establishing a project team an early priority in their 2. Organizational leads specialize in communication
knowledge management initiatives. plans, motivational incentives, and other organiza-
Choosing and recruiting the right people for tional needs.
the team is a vital task that will have enduring 3. Superusers can influence the user community and
consequences. Accomplishing corporate business provide practical feedback.
objectives requires a mix of skills, experience, and 4. Interface design leads specialize in
stakeholders interests, so the team members usu- computer/human interaction and integrating con-
ally will come from multiple disciplines. Team tent into work processes.
developers also must consider issues of trust 5. Training leads specialize in cognitive needs, learn-
across departm ents and business units as well as ing strategies, and repurposing legacy content.
between individuals. 6. Technical architects specialize in technology
When helping clients to implement knowledge installation and administration.
management projects, consultants at Chicago-based
Andersen suggest staffing teams with individuals After you have selected the right team mem-
from the business process in question and from the bers, you may have to negotiate their availability
human resources and information technology for the project. Several KM experts advocate full-
departments. A team needs members from all time devotion from members, at least in the early
three sectors. “You don’t find all these skills in one stages. A team member replacement plan must be
person, so you have to ensure that all are repre- worked out at an early stage so it is ready to acti-
sented,” says Christina Schultz, manager of global vate, if necessary. The team also must capture the
knowledge services at Andersen. knowledge that it generates in the process of work-
A team m em ber’s primary duty within the ing on the project, to guide future projects, and for
group is as important as his or her qualifications. transfer to managers who will later take on respon-
From his KM project experience, Fred Merrill, a sibility for the company’s KM efforts. This, after all,
consultant at Ariel Performance Centered Systems, is the purpose of knowledge management.

SOURCE: Excerpted from Robb, Drew. “Draft Your Dream Team,” Knowledge Management, August 2001, pp. 44—50.

knowledge-based system, but also on creating a way to reach out to and learn from one
another within the organization and with others in the know.
The bottom line is that an employee has to be convinced of the goodness of knowl-
edge sharing before the creative process becomes a reality. Woodrow Wilson once said,
“I use not only the brains I have, but all that I can borrow.” For the knowledge man-
ager, the crux of knowledge sharing is externalization or transforming captured tacit
knowledge to explicit knowledge for others to learn from and use anytime, anywhere.
458 7777 PART V Ethical, Legal, and Managerial Issues

■ ■ fl I fl 9 BOX 15.6 fl fl fl fl fl fl

IMPORTANCE OF SELF-TRAINING

SELF-TRAINING The company released the brainX Digital

Learning System, which is a personal knowledge
W hat happens to all the money that com pa-
management system. The software mimics two be-
nies spend on pum ping their employees full of
haviors: browsing the Web and highlighting printed
new knowledge? On average, U. S. companies put
materials. When a user has collected material,
78.6 percent of their employees through training
brainX automatically formulates a study plan and
programs. Companies spend an average of $677
begins quizzing. Then, the system shows the cor-
per worker. All in all, training costs U. S. busi-
rect answer and asks the user whether he or she
nesses approximately $50 billion annually.
got it right with ease, with difficulty, by guessing,
It is questionable whether that money is well
or not at all.
spent. For example, M otorola, Inc., found that
This type of system has applications beyond
employees tend to retain only 15 percent of what
training, such as refreshing employees’ memories for
they learned within 3 weeks after taking a corpo-
periodic compliance of safety testing and harvesting
rate education course. To address the retention
knowledge from seasoned employees and making it
problem , brainX.com hired experts in learning
available throughout the organization. “Companies
theory, curriculum design, and educational psy-
need an easy way to capture the knowledge in ex-
chology and incorporated their ideas into a new
perts’ heads in a way that makes it easy for others to
software design. The company adopted a premise
learn,” Dan Lewolt, founder of brainX.com says.
put forth in studies on education that the most
“We are not storing documents that no one else can
powerful way to learn new information is to cre-
use but capturing the key pieces of knowledge in
ate questions based on the material and practice
those documents, as identified by the experts.”
answering those questions.

SOURCE: Excerpted from Barth, Steve. “Worker, Teach Thyself,” Knowledge Management, August 2001, pp. 65-66.

I fl fl fl fl fl

It is also the question of culture apd trust that must be dom inant in the decision-
making environment and the factors that help build community in the organization.
In one medium-size bank, a strong community culture was discovered. Early one
week, one-third of the bank tellers submitted their resignations. They were all going to
a new bank across the street for higher salaries, bonuses, benefits, opportunities for
advancement, and the like. To survive the “gash” in the teller line, four of the remaining
tellers got together and devised an emergency schedule that handled the normal cus-
tomer traffic without noticeable delay. For example, they staggered the lunch schedule
so that only one teller would be absent for no more than 15 minutes. They also bor-
rowed two former tellers from the loan departm ent to help out. The security guards
were also asked to escort customers with basic deposits or withdrawals to a special ser-
vice window to expedite the traffic. As a result, the “fast checkout” window was
adopted as a great idea for customer service.
The instinct for survival is a good test in any organization, where intelligent, caring,
and trusting employees band together in the interest of the entity. They often come up
with creative ways to repair the wounds that otherwise could leave the organization
bleeding. They share what they know to survive and move on for a better tomorrow.
 CHAPTER 15 Managing Knowledge Workers i aaa 459
The bottom line is sharing, cooperating, coordinating, and collaborating for a common
goal (see Box 15.7).
Of these elements, sharing what you know is the most challenging for the knowl-
edge manager. To maintain knowledge sharing, he or she must reinforce and continue
to instill the goodness of sharing and how every knowledge worker benefits with this
new approach to survival and growth. The knowledge manager must deliver consistent
signals that knowledge sharing is what the organization supports and rewards. There
must also be a dem onstration of the value of sharing through unique incentives,
bonuses, and so forth. It means the hujnan resources departm ent must work in a

■ ■ 1 1 1 1 BOX 15.7 1 1 1 1 1 1

IMPACT OF KNOWLEDGE SHARING

A CASE FOR KNOWLEDGE SHARING the system. Not only does that arrangem ent
impede the flow of data outward, it also makes it
The Executive Office of Health and Human Ser-
difficult to share internally.
vices (EOHHS) is the Commonwealth of Massa-
Things came to a head in 1993, when seven
chusetts’ largest secretarial, consisting of 15 agen-
youths were murdered on the streets of Boston by
cies and a budget of roughly $9 billion. EOHHS
other juveniles. Investigation revealed that although
acts as the policy and steering arm to departments
various agencies addressed the problem, they were
covering functions as diverse as children, benefits,
incapable of solving the problem individually.
disabilities, and public health.
“These people knew each other but didn’t nor-
Those agencies operate in close proximity, but
mally work closely together,” says then Commis-
in terms of technology and knowledge collabora-
sioner of Youth Services William O ’Leary, a prime
tion, they may as well be situated on opposite
mover in M assCARES —a joint technology/KM
sides of the Atlantic. Henry Swiniarski, the state’s
initiative designed to centrally coordinate state
assistant secretary for EOHHS, gives examples of
health and human services information, share it
some of the many agencies involved in home visit-
among multiple agencies, and use it to provide bet-
ing programs. “Most of these agencies are entirely
ter care across the state. “There was a commitment
disconnected,” says Swiniarski. “We didn’t know
to collaborate. We got everyone on the same page
who was servicing whom, and multiple interviews
in terms of how we could get data, how we could
were going on covering the same inform ation
share information, and how we could develop pre-
about the same people. As a result, the knowledge
vention strategies to reduce violence.”
of one agency was largely unavailable to others.”
The results were surprising and immediate. They
Sharing inform ation does not necessarily
found out something that individual agencies had no
come easily in government. A hierarchical struc-
way of ever discovering—a handful of individuals
ture and silo-like IT organization mean that each
were responsible for most of the city’s violence. By
department can operate like an information fief-
identifying those responsible and organizing better
dom. The leader runs his or her own staff, intera-
targeted and more closely coordinated actions, the
gency collaboration is typically weak, and data
problem was resolved. For the next 2 years, the city
resides in huge repositories only accessible to one
experienced no further youth shootings.
or two technical staff with the know-how to run

SOURCE: Excerpted from Robb, Drew. “M assCARES Seeks to U nite Disparate State and Local A gencies,” Knowledge
Management, June 2002, pp. 24-25.

fl fl ■ ■ 1 fl
460 7777 PART V Ethical, Legal, and Managerial Issues

creative way to assure the extrinsic and intrinsic returns on knowledge sharing. The
ultimate goal is to practice knowledge sharing as a two-way, reciprocal type of interface
between the knower and the one who wants to know. Because knowledge has value,
the company should subject it to the same economic factors and the resulting rewards
that cement relationships and contribute to tenure and stability on the job.
In terms of incentives, various methods have been tried; some succeeded and oth-
ers failed. Much of what ends up being the right set of incentives depends on company
culture, company policies, and financial considerations. Based on our experience in the
field, the following points should be considered:
• Link incentives to a team approach, where the performance of the team on the
project will determine the size and nature of the incentive. It is assuming that the
team is working together and sharing their knowledge in the interest of the proj-
ect, where their collective strengths represent the quality of the results.
• Use awards or symbols of recognition for teams as well as individuals for unique
contributions, whether it is a new method that saved time and money or a new
idea that led to the creation of a new product or service. Awards tend to be moti-
vational, because recognition is an important human need. They carry little mon-
etary value, but when recognized through the company newsletter or Web site, an
employee of the month program, or special parking space to recognize that
special employee, they have far-reaching effects toward employee satisfaction
and stability on the job.
• Flextime is an approach where the company allows the team to decide on when
to work, when to quit, and so forth. The move away from the traditional 8 a .m . to
5 p .m . schedule implies trust in the team ’s judgment and also freedom for deciding
the time frame for the particular problem under discussion.
• Monetary rewards, bonuses, and special prizes can be a hit with the winning team,
especially when publicized throughout the company. This can be an effective
management tool to jump-start a new knowledge management initiative.
From the knowledge m anager’s view, the ultimate goal is personal advancement of
the knowledge worker’s career path. The incentives provided should be heavily based
on sharing and reciprocity, not individual contributions. If compensation is based solely
on the individual’s contributions, it could easily erode the sharing spirit and revert
team spirit to knowledge hoarding and disintegration.

■■■■ Implications for Knowledge Management

It should be obvious by now that we are moving into an age where business is m ea-
sured not in terms of historical performance, but in terms of how it makes use of the
knowledge and intellectual capital of its knowledge workers. New human management
challenges arise from the increasing dem and for knowledge workers. This means
encouraging productive knowledge to come forward and encouraging employees to
learn. This is where mutual trust and employee loyalty become critical.
Knowledge management is about companies realizing anew the assets that have
never appeared in the balance sheet but are vital to their future success. These hidden
assets are the people with know-how. In the 1990s, the geometric growth of companies
like Microsoft was based on intellectual property and the outcomes created in the way
of products and services.
How a company performs depends not only on its financial capital but also on how
it invests its intellectual capital. Its valuation is the sum of the talent, knowledge,
 CHAPTER 15 Managing Knowledge Workers i i i ■ 461
patents, image, and brands that far outweigh money, brick, or mortar. Intellectual capi-
tal is measured by capturing, consolidating, sharing, and managing knowledge workers’
tacit knowledge, knowledge repositories, and intelligent databases. Companies that
make effective use of their knowledge workers will prosper, and those that do not will
struggle. The ultimate goal is to harness employee knowledge and effectively turn it
into intellectual capital that can increase and assure a company’s competitive edge.
From the knowledge worker’s view, as they begin to recognize the value of their
knowledge, they will change their relationship with their employer from em ployer-
employee to custom er and partner. They become the source of an organization’s
competencies.
For a company employee to become a knowledge worker, it is im portant that he or
she become special to the employing organization. This means that the company will
have to pay a high price to replace a person’s skills and knowledge. In fact, training and
development can no longer be viewed as a benefit at the discretion of the employer,
but as a part of employment that will increase the worth of the knowledge worker over
time. By the same token, as knowledge work becomes an increasingly costly item in a
company’s budget, learning organizations will be under continued pressure to look for
more creative ways of making every bit of knowledge count.

Su m m a r y 1111
• The driver of success in the new economy is knowledge. Knowledge embodies
experience, innovation, and creativity.
• A knowledge worker holds unique values, aligns personal and professional
growth with corporate vision, adopts an attitude of collaboration and sharing, has
innovative capacity and a creative mind, is willing to learn, is in command of self-
control, and is willing to tolerate uncertainties and grow with the company.
• There are several core competencies of the self-directed knowledge worker:
thinking skills, continuous learning, innovative teamwork, creativity, risk taking,
decisive action taking, and a culture of responsibility toward knowledge.
• When discussing business roles in the learning organization, management and
leadership become important. Smart managers focus on organizational learn-
ing to ensure operational excellence. In contrast, the leader’s role is more
of a facilitator, a teacher, a steward of the collective knowledge of the staff,
and a designer.
• A key ingredient in management and leadership is the way managerial experi-
ence is processed. Work management tasks include searching out, creating, shar-
ing, and using knowledge in everyday activities; maintaining work motivation;
ensuring readiness to work; allocating effort and control-switching among tasks;
managing collaboration; sharing information; and hiring bright, knowledge-
seeking individuals.
• In managing knowledge workers, a manager must keep in mind time constraints,
working smarter and harder but achieving little, knowledge workers doing the
wrong job, work schedule and deadlines, and motivational forces against knowl-
edge work productivity.
• Smart managers strive to ensure the right match between the vocational needs of
their knowledge workers and the requirements of their jobs. A proper match
results in worker satisfaction and hopefully a satisfactory employee as viewed by
management. Vocational needs are many; they include achievement, ability uti-
lization, recognition, and the like.
462 7777 PART V Ethical, Legal, <
2/7d Managerial Issues

• Knowledge is a series of steps that determine the potential of a learning organiza-

tion. The ultimate goal is to encourage knowledge workers to focus on thinking,
using information for processing, and recommending value-added change.
• Technology plays a major role in knowledge work. It facilitates information avail-
ability and information processing for creative thinking and intuitive planning. The
usfc of e-mail, intranets, and bulletin boards can facilitate information distribution
and interpretation. With technology in mind, a knowledge worker is expected to
possess professional experience and technical know-how to access, update, and
disseminate information and ideas from databases and knowledge bases.
• There are several environmental, hardware, and knowledge worker-system inter-
face factors that affect the ergonomics of knowledge workers. The goal of
ergonomics is to provide minimum worker effort, make best use of human pat-
terns, and provide maximum task support.
• To be a change leader, several factors should be considered: a focus on successes
and opportunities, adopting an attitude that views challenges as opportunities,
and creating tom orrow’s business instead of hammering on yesterday’s problems.
• When it comes to managing knowledge projects, the human factor becomes
important. Convincing people to share individual experiences continues to be an
elusive item in most organizations. A knowledge manager must build trust to dis-
courage people’s tendency to hoard knowledge, allow knowledge workers to par-
ticipate in major company decisions, recognize and respect the experts, ensure
that the technology meshes with the knowledge worker’s culture and company
strategy, and keep knowledge up to date.
• In terms of incentives, a knowledge manager must try various methods that work
best for the knowledge worker, especially in a team environment. Incentives
should be linked to a team approach, and awards should be used to recognize
value-added work. Monetary rewards, bonuses, and special prizes can be a hit
with the winning team, especially when publicized throughout the company.
• A major implication for knowledge management is that knowledge management
is about companies realizing anew the assets that have never appeared in the
balance sheet but are vital to their future success. These hidden assets are the
people with know-how. How a company performs in the future depends not only
on its financial capital but also on how it invests its intellectual capital.

Terms t o K now • * ■ 1
Achievement: A drive to accomplish worthwhile, complex Leader: In a learning organization, a facilitator, a teacher,
tasks and a feeling of accomplishment. a steward of the collective knowledge of his or her staff,
Creativity: Trying out on e’s own ideas in an effort to and a designer.
come up with new ones. Smart manager: A person whose main job is to focus on
Ergonomics: The science of providing a physical environ- organizational learning to ensure operational excellence.
ment (such as lighting or swivel chairs) for human com - Traditional manager: An action-oriented person who
fort at work. focuses on the present and spends most of the time
Knowledge worker: A person who transforms business delegating, supervising, controlling, and ensuring com-
and personal experience into knowledge through cap- pliance with set procedures.
turing, assessing, applying, sharing, and disseminating it Work adjustment: An individualized model that prescribes
within the organization to solve specific problems or to achieving and monitoring correspondence between an em-
create value. ployee’s vocational needs and the reinforcers of the job.
 CHAPTER 15 Managing Knowledge Workers aaaa 463

Te s t Yo u r U n d e r s t a n d in g 1111
1. What is a knowledge worker? Do you agree with any of the definitions in the
chapter? Why or why not?
2. List and briefly explain personality and professional attributes of the knowl-
edge worker.
3. The self-directed knowledge worker must consider several core competen-
cies. Explain three core competencies of your choice. Why are they called
core competencies?
4. Elaborate on the business roles in the learning organization.
5. In what ways are data and information considered as givens?
6. W hat is the difference between m anagem ent and leadership? Traditional
managers and smart managers?
7. How are learning and teaching related?
8. What do work management tasks focus on?
9. Explain the main factors that limit knowledge worker productivity and ways
to get around them.
10. Explain the work adjustm ent model. How does it relate to the knowledge
worker?
11. Briefly list the vocational needs and reinforcers of knowledge workers.
12. How does creativity relate to achievement?
13. Elaborate on smart leadership requirements.
14. What is meant by return on time? How does it relate to the knowledge chain?
15. Briefly explain the key steps in the knowledge chain.
16. How does technology assist the knowledge worker?
17. What should an intelligent workstation do?
18. List the knowledge w orker’s key skills. Do you agree with them?
19. In what way does ergonomics relate to the knowledge worker?
20. What does it take to become a change leader?
21. “Managing knowledge projects requires a set of qualifications and responsi-
bilities.” Elaborate.
22. Explain the various incentives that might be successful for the knowledge
worker.
23. What implications can you draw from the chapter for knowledge management?

Kn o w l edg e Ex e r c is e s 1 1 , 1
1. Discuss the similarities and differences between the traditional manager and
the knowledge manager.
2. Try to verify the personality and work attributes of the knowledge manager
by doing research on the Internet. Report your findings to the class.
3. Is a college degree im portant for knowledge work? Discuss your beliefs
with the class.
4. Cite three companies that qualify as learning organizations. W hat makes
them unique?
5. Does a leader’s job include management? If so, why do we need managers?

Re f e r e n c e s 1 1 1 1
Apfel, Ira. “Hiring Changes,” www.potomactechjoumal.com/ Awad, E. M. Building Expert Systems. Minneapolis, MN:
displayarticledetail.aspl, Date accessed October 20,2002. West Publishing, 1996, p. 471.
Appleton, L., and Gavin, C. Intellectual A sset Valuation. Awad, E. M. “Select Attributes of the Knowledge Worker
Boston, MA: Harvard Business School Publishing, 2000. in Banking.” Unpublished manuscript, 2002, pp. 1-18.
464 7777 PART V Ethical, Legal, a/zd Managerial Issues

Baltazar, Henry. “Overcoming KM’s Obstacles,” Gravallese, Julie. “Knowledge M anagement,” www.mitre.
Knowledge Management, February 18,2002, pp. 4 3 ^ 4 . org/pubs/edge/april_00/index.htm, D ate accessed
Barth. Steve. “Worker,Teach Thyself,” Knowledge August 2002.
Management, August 2001, pp. 65-66. Hall, Mark. “Build a Long-Term Work Relationship,”
Bennet, Bill. “Just Who Are Knowledge Workers?” C om puterw orld, May 6,2002, p. 46ff.
November 11,2001, Australia.internet.com/r/article/ Hildebrand, Carol. “D oes KM = IT?” www.cio.com/
jsp/sid/11300, Date accessed August 2002. archive/enterprise/091599ic.html, D ate accessed
Bhushan, Navneet. “Creating Next-Generation August 2002.
Enterprises,” E AI Journal, May 2002, p. lOff. Horibe, F. Managing K nowledge Workers, New York: John
Chaleff, Ira. “Process Improvement for Knowledge Wiley & Sons, Inc., 1999.
Workers,” www.ibt-pep.com/afsmart.htm, Date Horvath, Dean. “Knowledge Worker,” searchcrm.
accessed August 2002. techtarget.com/sDefinition/0,290660,sidll_gci212450,00.
Date, Shruti. “A gencies Create CKO Posts to Get in the html, Date accessed August 2002.
Know.” Government Computing N ews, November 8, Kappes, Sandra, and Thomas, Beverly. “A M odel for
1999, www.gcn.com /voll8_no36/news/950-l.htm l, Date Knowledge Worker Information Support,” Knowledge
accessed October 20,2002, p. 1-2. Worker Information Management, September 1993,
Datum Consulting. “Knowledge Workers,” www.datum pp. 1-4, www.cecer.army.mil/kws/kstoc.htm, D ate
consulting.com/solutions_kw_architecture.asp, Date accessed August 2002.
accessed August 2002. Koulopoulos, Thomas, and Frappaolo, Carl. Smart Things
Davis, Gordon B. “An Emerging Issue: Knowledge Worker to K now A b o u t Knowledge Management. Oxford:
Productivity and Information Technology,” Information Capstone Publishing Limited,. 1999.
Science Conference, Krakow, Poland, June 20,2001. Maccoby, Michael. “Knowledge Workers N eed New
Dawis, Rene V., Lofquist, Lloyd H., and Weiss, David. Structures,” www.maccoby.com/Articles/
“The Theory of Work Adjustment.” Minneapolis, MN: KnowledgeWorkers.html, D ate accessed August 2002.
University of Minnesota, 1968. Malhotra, Yogesh. “Knowledge Management, Knowledge
Dove, Rick. “The Knowledge Worker,” www.parshift.com/ Organizations, and Knowledge Workers: A View From
library.htm#0ther%20publications, Date accessed the Front Lines,” April 2001, www.brint.com/
December 3,2002. interview/maeil.htm, Date accessed August 2002.
Drucker, Peter. “Who Is the Knowledge Worker?” Robb, Drew. “Draft Your Dream Team,” Knowledge
www.pbsilink.com/knowledgeworkers.htm, Date Management, August 2001, pp. 44-50.
accessed August 2002. Robb, Drew. “M assCARES Seeks to U nite Disparate
Dvorak, John C. “Know-Nothing Knowledge Workers Must State and Local A gencies,” K nowledge M anagement,
Go!” PC Magazine, May 20,2002, pp. 12-15, www.pemag. June 2002, pp. 24-25.
com/article/0,2997, s%3D1500%26a%3D22201,00.asp, Russom, Philip. “An Eye for the N eedle,” Intelligent
Date accessed August 2002. Enterprise, January 14,2002, pp. 27-30.
Flash, Cynthia. “Who Is the CKO?” Knowledge Manage- Sebra, Bill. “Electrifying Your H R Management,”
m ent, May 2001, pp. 3 7 ^ 1 . www.knowledgeworkers.com/news/news_258.cfm,
Flash,T. “Personal Chemistry” Knowledge Management, D ate accessed August 2002.
August 2001, pp. 36-43. Wesley, Doug. “Retaining Workers in the Knowledge
Frank, J. T. “Strengths of K-Professionals,” www.kwx.com. Economy,” www.findarticles.com/cf_dls/mOFWE/
my/kwx/asp/articles02/articles02/articles0207.asp, Date 10_4/66276585/pl/article.jhtml, D ate accessed
accessed D ecem ber 3,2002. D ecem ber 3,2002.
Garfinkle, Joel. “Steps to Becoming a Change Leader,” Zetlin, Minda. “Best Places to Work,” C om puterw orld,
www.kwx.com.my/kwx/asp/articles02/articles_0203.asp, May 6,2002, p. 40ff.
Date accessed August 2002.
 What More Do We
Need to Know?
Contents

The Shareability Factor

The Human Element
There Is More to Know
The Social Factor
One Final Note
References

Today’s knowledge manager wants to know the reality of tomorrow, not just the reality
of today. In a strategic sense, knowledge management in the next 5 to 10 years is likely
to be shaped by many initiatives that we do not know for certain. Before we talk about
future KM initiatives, our view of today’s KM foundation is shown in Figure E .l.

I I I i fl fl FIGURE E.l: Today’s Knowledge Foundation

465
466 7777 What More D o We Need to Know?

The first step is to autom ate and routinize basic tasks that can be handled with
minimum effort by technology. This allows the knowledge manager to concentrate on
higher-order decision making or decide on the knowledge needed to do creative work.
Creativity is the driver of innovation, which should result in a new product or
improved service over time. Knowledge for creative work comes from knowledge
repositories and other sources.
What we now know is this: Knowledge management is a prerequisite for competition.
It is the critical element for innovation. Consider Japan’s industrial growth in the 1960s as
compared to today’s production of Lexus and Camry vehicles, among others. Technology
alone did not do the trick. Products were initially not differentiated from competitors.
When Toyota, for example, began to focus on quality, reliability, and dependability at a
competitive price, it took on the U.S. market. In fact, American car manufacturers began
to rethink their way of building cars to compete with a new way of doing business.

■■ ■« The Shareability Factor

The greatest factor in knowledge management is sharing tacit knowledge. The chal-
lenge is not so much availability but willingness to share with others. We have known
for years that tacit knowledge can be captured and stored before it is shared.
Unfortunately, because knowledge tends to become richer, not leaner, over time, it
must be kept up to date, which can be a problem. Also, the key is not availability of
tacit knowledge, but how it is used. This is a major parting point from knowledge
automation systems.
A nother component of the shareability factor is organizational control. A tradi-
tional organization assigns certain authority, power, influence, and control to each
m anagerial level. Sharing knowledge will require decentralized intelligence to
empower knowledge workers to function more quickly and intelligently. Doing so
requires transplanting vast volumes of decision making to lower levels. The problem is
that the farther down the line intelligence is transported, the more “diluted” it
becomes. When it reaches the intended level, only content is transported, not context.
It is like going to a museum to view the paintings. They are nothing more than artifacts
containing basic information about their composition or age. It is left for the visitor’s
imagination to draw conclusions about their prominence.
One conclusion is that shareability is more than mere availability of repositories
or warehouses where the burden is on the knowledge worker to find the relevant
expertise and decide how it should be applied. To achieve context, knowledge man-
agement must shift emphasis to cultivating expertise. The shift is fraught with two key
misconceptions.
M isco n cep tio n 1. Generalizing expertise will reach a wide audience; this is often not
true. Consider this scenario:
One o f the authors recalls a consulting assignment, involving a $200,000 hard-
ware installation fo r a small bank. The vendor was known fo r giving no dis-
count on the product, service, or support. Sensing that the order could go to a
competing bidder, the sales representative came up with the offer o f writing the
installation under a 6-month testing, which meant no charge. Instead o f giving
the client the 20 percent discount that was refused by headquarters, the value o f
the 6-month period o f use was worth the discount when considering the interest
saved on the $200,000 at 10 percent fo r 6 months, plus savings on maintenance
and support, to be worth $20,000. The sales rep drafted this indirect discount so
 What More D o We Need to Know? a a * i 467
that other sales reps could apply the same, with no reference to the customer or
to the plan as a direct discount. The list was then posted on the K M portal
available only to sales reps under the heading “one-time discount procedure
fo r sales exceeding $200,000

Misconception 2. The more knowledge is available, the more likely individuals will
find what they need to solve a business problem. Humans in general do not have the
patience to surf or scan hundreds of documents to locate the desired item. Also, as
more docum ents are stored in a repository, the num ber of dated docum ents also
increases. This means finding a way to filter useless documents would be required at a
high cost, which is difficult to justify.
Another problem with knowledge search is that knowledge workers do not always
know how to search effectively for the right document to meet specific needs. With an
ever-increasing num ber of knowledge repositories, there is an inevitable increase in
the number of “hits.” This is where frustration sets in, and people want to find an easy
and a quick way out.
One way to address this problem is for a select team of company experts to iden-
tify repository content that provides the most value to the company business in gen-
eral. They could apply the 20/80 rule, where the top 20 percent of the content will sat-
isfy 80 percent of the queries or problems. In addition, an online environment would
continuously purge old documents through filtering. Technology may be employed to
filter out the waste and organize existing resources based on value, frequency of use,
and so forth. This helps ensure targeted knowledge to authorized users.

■■■■ The Human Element

People do not devote much time or show much interest in contributing to a knowledge
base. This is especially true with top performers. They perceive their own knowledge as
something they own. Why would they want to share it? In addition, most people do not
know enough to distinguish between what should and should not be shared. Even if
they did, not everything people share is necessarily valuable. However, if only a few
contributors dominate a knowledge repository, it can discourage the participation of a
majority of a company’s experts.
To encourage experts to contribute to the knowledge management initiative, it is
important for experts to be treated as experts. This means recognition and appreciation
of the contributions they make, for the way they think, and how they explain things.
Experts are not the most humble people or the easiest to understand. In fact, many
experts with whom we have worked with over the years tend to be moody and get rest-
less when asked to extend sessions or repeat something they have explained. However,
there is a powerful emotion attached to being recognized as an expert.
This brings up im portant issues in knowledge management. King et al. (2002) con-
ducted a study to identify and evaluate im portant knowledge management issues and
how they should be managed. Table E .l summarizes the top 20 KM issues as reported
by survey respondents. The issues were factor analyzed into four grouping issues:
Executive and strategic managem ent; operational managem ent; costs, benefits, and
risks issues; and standards in KM technology and communication networks. The pre-
diction is that these issues and their respective grouping represent the basis for effec-
tive management in the future.
Knowledge organizations face an investment choice as they consider knowledge
management applications. The ingredients in knowledge management applications are
468 7777 What More D o We N eed to Know?

TABLE E.l. KM Issues

Executive and Strategic Management Operational Management
1. Providing for strategic advantage 3. Knowledge currency
2. Top management support 5. Identification of
4. M otivation to participate organizational knowledge
9. Sustainability 7. Verification of knowledge
15. Evaluating the CKO contributions
16. Creativity and innovation 8. System design
17. Effect on organizational processes 11. M ethodologies
18. Organizational responsibility

K M Issues
Management
Program

Costs, Benefits, and Risks Standards

6. Financial cost and benefit 12. D evelopm ent of technical
10. Security standards
13. Nonfinancial cost and benefit 14. Operational definition
18. Investment 20. Fit with IT infrastructure

Note: Numbers reflect priority ranking of each issue

SOURCE: King, W. R., Marks, P. V., and McCoy, Scott. “The Most Important Issues In Knowledge
M anagem ent,” Communications o f the ACM, September 2002, p. 95.

fl

technology and content. Our preference is to employ technology to provide knowl-

edge managers direct access to a lower volume of high quality content that is tailored
to the exact needs of the solicitor. Sophisticated indexing, for example, goes a long
way toward providing the knowledge needed at the time needed, and in the proper
format. Experts should also be interviewed, and their new or updated knowledge
should be captured on a regular basis to ensure quality and dependability of the
knowledge repository. This means that knowledge management needs to be managed
and leveraged; this can only be achieved when paying attention to the best source of
knowledge—people.

■ There Is More to Know

Knowledge managers think ahead. They want to know how to strategize for tom or-
row’s business and how to continue using technology to leverage resources. Based on
the authors’ years of experience and existing literature, several trends are worth noting:
• Knowledge is productive only when it is captured in people’s minds. Because
most knowledge is acquired on the job, through training and in interaction with
people, it behooves every organization to nurture and invest in the skill level of
employees so that they become more productive in present and future jobs. With
proper incentives and benefits, this is the kind of investment that promotes loy-
alty and discourages defection.
 What M ore D o We N eed to Know? aaaa 469
• More and more companies are discovering that leveraging knowledge is best
served through decentralized intelligence. This means empowerment of knowl-
edge workers and funneling tacit and explicit knowledge to the level where the
actual work is performed. This type of decentralization is expected to make orga-
nizations more responsive to a changing environment, where the time for decision
making is measured in seconds, not hours. Companies that have human knowl-
edge updated regularly will find greater opportunities for creativity and innova-
tion through the unique ideas that inevitably result.
• It is worth repeating that the driving force of knowledge work is creativity—an
essential driver for innovation. The key is not so much the amount of knowledge
available, but how knowledge is used. This is a major parting point from knowl-
edge automation systems. Creativity is breaking loose from the traditional way of
doing business. Recruitment and compensation strategies are being revised to
address the creativity imperative. Aspiring candidates with potential will be
highly sought. Technology will continue to take over routine and redundant work,
freeing knowledge workers to concentrate on human insight and foresight.
• We believe that knowledge management will be an entry requirement for com-
petitive advantage and that the future will belong to innovators, not copycats.
This, of course, will be expensive: Recruiting intelligence is not cheap or easy. The
pressure will increase on human resource managers and recruiters to scout for the
right sources of talent and secure a continuing stream of candidates to support
the long-term needs of the corporation.
• For years, we have been working on the assumption that corporate capital is the
enabler of the enterprise. Capital continues to be important, especially when it
comes to funding new projects and investing in equipment and technology. How-
ever, the trend is toward using corporate knowledge in a global competitive market.
Knowledge management requires action, foresight, and focused strategic planning.

The Social Factor

Knowledge management is seen as being passive, in that it is composed of facts and
heuristics that can be stored, retrieved, and transferred with little concern for the con-
text in which it will be used. Most of the writings and practices in knowledge manage-
ment seem to focus on the content of knowledge management’ systems, overlooking
how knowledge is presented or communicated. We believe that knowledge manage-
ment is more than getting the right information to the right person at the right time.
Managing knowledge occurs within a complex structured social context. That is, there
must be social and hum an factors in the creation and exchange of knowledge
(Thomas et al. 2002).
In the field of computer-supported cooperative work (CSCW), knowledge work
was found to involve communication among communities of people and the social
practices that occur in a particular context. For example, O rr studied the practices of
photocopier technicians and found that their technical knowledge was socially distrib-
uted across the technicians of the firm and disseminated by storytelling, by memos, and
by graphic sketches (O rr 1996). A nother study brought up a variety of social factors
that affect communication of knowledge among coworkers and illustrated how greater
shared background helps in establishing common grounds for knowledge sharing. The
rate of motivation also happens to be a factor in knowledge sharing. Failure of motiva-
tion was reported to be a major factor adversely affecting the adoption of groupware in
most cases.
470 7777 What More D o We Need to Know?

It is im portant to think of knowledge management as not just a bunch of isolated

facts stored in documents, knowledge bases, or repositories; knowledge management is
shaped by social and human factors that require the involvement of the contributor
and the recipient of knowledge. Communication occurs when individuals or groups are
motivated by personal and social motivators that make it convenient and appropriate
to share experiences and arrive at consensus in a problem-solving context.
Given the role of technology in transferring and disseminating knowledge, the
true picture of knowledge is one where people voluntarily explore, use, and adopt
knowledge for the good of the project and in the best interest of the firm. Such a
knowledge community includes people who know one another. They use storytelling,
war stories, and other forms of narrative to enrich professional relationships, using spe-
cific techniques. This type of exchange is bound to lead to building social capital, coop-
eration, coordination, trust, and solidarity among professional peers. Cultural, organi-
zational, and political factors must be considered in knowledge transfer and
knowledge exchange.
Viewing knowledge m anagem ent continues to be a problem. Some emphasize
intellectual capital; some focus on technology; others associate different views of
knowledge with personality types. Dueck did a major study in which he found that a
person’s temperament is a major influence on how that person views knowledge man-
agement (Dueck 2001). Briefly, he found that “corporate guardians” (caretakers, care-
givers) are concerned with security of knowledge to ensure that no knowledge is lost.
Everyone in the organization uses the same technology and speaks the same language
for solidarity and unity in front of the customer. The other group consists of the “utili-
tarian rationals,” who live for their work and focus on competence and the need to see
daily improvements. They are not always sensitive to interpersonal exchange. They
need databases and hate soft knowledge.
For knowledge m anagem ent to be established in an organization, the so-called
guardians will try to set standards; however, some professionals are likely to resist,
because they view the “right” knowledge management approach as a function of a per-
son’s temperament.
In the final analysis, we feel the right approach to knowledge management is a unified
approach, involving technological, social, human, and organizational elements that lead to
economic value. Each element and temperament has to be part of the mix, the exchange,
or the approach. The behaviorists have to accept the use of technology for storing and dis-
seminating knowledge; the technologists have to understand that “tacit knowledge” and
expertise are the foundations of knowledge management. The inner personalities of the
experts must blend with those in technology to make things happen. It is the future of the
organization, not its past or present, that is the goal of knowledge management.

■■ ■■ One Final Note

We have been undergoing dynamic and somewhat turbulent changes in the business
environment. M alhotra (2002) referred to it as the world of “re-everything,” tanta-
mount to performing maintenance on a passenger train while the train is moving at top
speed. What used to be predictable and programmable is becoming less relevant to sur-
vival. The question is how does a company manage a business that has the capability of
questioning its executives’ logic and revising their heuristics in real time as the business
senses dynamic changes in the product or the environment? The concept of strategy
and advance planning has shifted from predicting the future to what to do in the event
of a surprise on a daily basis.
 What More Do We Need to Know? aaa■ 471
With the increasing return on intangible assets becoming more prominent, there
is a pronounced shift from the “brick and m ortar” to the “click and mortar,” as seen in
e-commerce, e-business, supply chain management, customer relation management,
and the value chain for most companies worldwide. One may speculate that tom or-
row’s firms may far exceed the current logic of e-technologies as business becomes
interlocked with learning and with digital storefronts along the Internet—the infor-
mation highway.

R eferen ces • • • •
Davenport, Tom. “From Data to Knowledge,” Malhotra, Yogesh. “Knowledge Management and
www.cio.com/archive/040199_think_content.html, Date E-Business in the New Millennium,” January 2000,
accessed August 2002. www.brint.com/ advisor/aOl 1700.htm, Date accessed
Davenport,Tom. “Knowledge Roles: The CKO and November 29,2002.
Beyond,” www.cio.com/archive/040196/davenport_ Olsen, Bjorn. “Mismanagement of Tacit Knowledge:
content.html, Date accessed November 30,2002. Knowledge Management, the Danger of Information
Dueck, Gunter “Views of Knowledge Are Human Technology, and What to D o About It,” www.program.
Views,” IBM Systems Journal, vol. 40, no. 4,2001, forskningsradet.no/skikt/johannessen.php3, Date
pp. 885-88. accessed August 2002.
Dyson, Esther. “Intellectual Value ” www.wired.com/ Thomas, J. C., Kellog, W. A., and Erickson,T. “The
wired/archive/3.07/dyson.html, Date accessed Knowledge Management Puzzle: Human and Social
November 30,2002. Factors in Knowledge M anagement,” Communications
King, W. R., Marks, E. V., and McCoy, Scott. “The Most o f the A C M , October 2002, pp. 863-84.
Important Issues in Knowledge M anagement,” VisionCor. “The Human Element: Knowledge M anage-
Communications o f the ACM , September 2002, m ent’s Secret Ingredient,” www.visioncor.com, Date
pp. 93-97. accessed October 2002.
This page is intentionally left blank.
 Index

A Boolean rule, 338

Bottom -up approach
Churchman, C. W., 116
Circular errors, 246, 248
Action clause, of a rule, 226-227 to building learning models, 331 Circular triad, 170
Action protocol, 182 to improving ethical behavior, 430-431 Cisco Systems, 296
Adair, Red, 292-293 Bounded rationality, 274 CitiBank London, 336
Adaptive learning systems, 342 Brailsford,Tom, 454 Classification, 359
Affinity grouping, 359 Brain drain, 35 Classification trees, 341-342
A ge effect, 169 Brainstorming, 183 Clustering, 359
A ggression, 108, 262 electronic, 184-185, 186, 187 CM Holdings, Inc., 429
A llen, Jim ,454 brainX Digital Learning System, 458 Code of ethics, 431
A llen, William, 430-431 Branches, 341 Codification, 400
Alpha Corporation, 431 Break-even analysis, 95 See also K nowledge codification
Ambiguity, 106 Buckman, Robert, 302 Cognitive science (psychology),
Am ble, Charles, 4 1 1 -4 1 2 Business drivers of data mining, 352-354 59-60, 232
Andersen Consulting (A ccenture), 36 Business intelligence (B I), 351-352 Collaboration, 41 ^ 2
Anticvbersquatting Consumer Protection Business process reengineering (B P R ), 44 versus categorization, 396-398
A c t,422 Business-to-business (B 2B ),306 knowledge portals and, 392-393, 395
Application layer, 103-105 Business to business (Internet and Collaborative agents and filtering, 103
Architects, of KM systems, 102 extranet). 317 Collaborative intelligence and filtering
Arcs, 197 Business to consum er (Internet), 316-317 layer, 129-133
A R E A S, 337 Business transactions, 361 C ollective sequential transfer, 288-291
Aristotle, 44 Columbia Laboratories, Inc., 425
Arrow, A lex, 413 Com mon sense, 58-59, 68
Artificial intelligence, 131-132 Com m unication, 120
Association discovery, 359 C difficulties, 171
interpersonal, 232
Association rules, 338-341
Asynchronous collaboration, 393, 395 Calogero, Barry,42 training, 263
Atlantic Corp.,422 Cambridge Information Network, 388 Com petition, 352-353
Authentication/security layer, 103 Campaign optim ization, 368 Com petitive advantage, 44
Authorized access layer, 127-129 Capturing phase of KM life cycle, 48 Com pilation, 72
Avoidance. 108, 262 Case, 68, 228 Computer-supported cooperative work
Case-based reasoning (C B R ), 67-68, 131, (C SC W ),469
132, 228-230, 231 Concept mapping, 195
(B ox 2.3), 69 procedure, 196-197
B (Figure 4.13), 133 semantic nets and, 197-198
Cases, 188 Consensual approval, 330
Bandwidth, 400-401 Categorizing, 396 Consensus, 183, 190
BAN K O R , 9 0 -9 1 , 92, 9 4 -9 5 , 98, 102, 105 Certainty factor, 225 Consensus decision making, 190-191
Barker, Joel Arthur, 65 Champion, 260-261, 308 Consistency, 125
Behavioral feasibility, 94 Change, 323, 324 Contam ination effect, 167, 170
Beliefs, 57 Change leader, 455 Content m anagem ent, 312
Bell. D an iel,44 Chase, W. G., 72 of know ledge portals, 394, 396, 397
Bennet, Alex, 29 Chat rooms, 312-313 Continuous learning, 440
Best Buy Co., 39 Chiazza, John, 307 Control m echanism , 200
Biometrics, 129 Chief knowledge officer, 4 52-454 Conversion, 106
Blackboard, 199 Child frames, 224 Copyright, 417
Blackboarding, 198-200 Chrysler Corp., 42 statem ent (Figure 14.2),419
Boeing Aircraft, 430 Chunking, 72 Copyright and trademark violations, 416
Boland, R. J., 116 Chunks, 72, 151 Copyright laws, 411, 417

473
474 7777 Index

Corney, Paul, 42 implications for knowledge m anage- Disclaimer, 420

CO RVID, 91 ment, 369-372 Discovery, 56
Cost/benefit analysis, 94 integration of, with knowledge man- DM tasks, 359
Crafton, A dam , 350 agem ent (B ox 12.2), 352 D ocum ented knowledge, 71
Creativity, 440 in practice, 363 Docum ents, 361
Cultural values, 279 banking exam ple (Box 12.4), 365-366 D ocum ent type definition (D T D ),397
Culture, 84, 97 supermarket exam ple (Box 12.3), 364 Dom ain name and Web linking disputes,
Cupello, J. M., 115 problems, types of, 370-371 421-423
Customer acquisition, 367-368 sample uses of (Box 12.1), 350 Drivers for knowledge management
Customer com munities, 381 technical drivers financial, 38-39
Customer know ledge, 30 data warehouses, 354-355 knowledge-related, 38
Customer relationship m anagem ent decision-m aking architecture, 356, personnel-specific, 38
(C R M ), 319, 321-322 357,358 process, 38, 39
See also D ata mining (D M ) machine learning, 354 technology, 37-38
Custom er scoring, 368 OLAP, 355-356 Drucker, Peter, 44, 438
Customer, understanding the, 324 statistics, 354
Customer users, 107 U R L listings, 377-378
Cybersquatting, 422 virtuous cycle, 356, 359 E
business understanding, 357, 360
DM application, developing, 358-359 Ease of use, 131
D Data preparation, 361-362 Eastman Kodak Co., 307
D ata sources, 361 E-business, integrating with DM and
Daim ler-Benz AG, 42 D ata visualization, 331, 332-333 CRM, 368-369
Daim lerChrysler, 319 D ata warehouses, 361 See also Knowledge transfer
Data, 32 D ata warehousing, 45-46, 354-355 Econom ic feasibility, 94
information, knowledge, and, 60-65 D avenport, Thomas, 349 Electronic brainstorming, 184—185, 186, 187
taxonom y of, 361 Decision-m aking architecture, 356, 357 Elements, 191, 397
Data, learning from, 329 Decision-m aking flow integration, 258 E-mail and knowledge transfer, 312
association rules, 338-340 Decision support systems (D SS), 64 Em bodied knowledge, 71
market basket analysis, 340-341 D ecision tables, 221, 222, 223 Emerson, Ralph Waldo, 443
classification trees, 341-342 D ecision trees, 221, 222, 223 Em ployee insurance, ethics of, 429
data visualization, 331, 332-333 Declarative knowledge, 68-69, 70 Encryption, 432
“learning” concept, 330-331 Decom piling.chunked knowledge, 72 Enhancem ent,108, 109
neural networks as a learning D eductive reasoning, 66-67 Enterprise business intelligence portals
m odel, 331 D eep knowledge, 66 (E B IP s),385
basics, 334-335 Default, 223 Enterprise collaborative processing por-
business applications, 336-337 D ell Computers, 296 tals (ECPPs), 385
fit with know ledge m anagem ent, 337 Delphi m ethod, 194-195 Enterprise extended services portals
supervised and unsupervised learn- D eM ong, Jim, 413-414 (E E S P s),385
ing, 335 Dem ons, 223 Enterprise intelligence portal (EIP), 385
Web sites, 346-347 D eploym ent, 363 Enterprise mission management portals
Data m anagem ent issues related to (EM M Ps), 385
See D ata mining (D M ) champion, role of, 260-261 Enterprise portals, 383-384
Data mining (D M ), 48, 348 ease of understanding the system, Enterprise-wide integrated knowledge
business drivers, 3 52-354 255-256 m anagement system
business intelligence and, 351-352 integration alternatives, 257-258 (Figure 4.14), 135
custom er relationship m anagem ent knowledge base problem, selection Episodic knowledge, 68-71
(CR M ) of, 254-255 Ergonomics, 450-452
campaign optim ization. 368 knowledge transfer, 256-257 Ernst & Young, 388
custom er acquisition, 367-368 maintenance, 258-259 ERP (enterprise resource planning), 275
custom er scoring, 368 organizational factors, 259-260 Errors in knowledge base design, 106
direct marketing, 368 quality information factor, 260 Ethical issues of knowledge, 409-410
integrating DM , CRM , and return on investment (RO I) em ployee insurance example
e-business, 368-369 factor, 260 (Box 14.8),429
data m anagem ent, 360 of the KM system, 254 ethical decision cycle, 427, 428
action and deploym ent, 363 postim plem entation review, 263-264 ethics defined, 424, 426-427
data preparation, 3 61-362 user training and implications for knowledge manage-
data sources, 361 preparing for KM system training, ment, 432-433
m odel building, 362 261-262 improving
m odel testing and analysis o f results, resistance to change, combating, code of ethics, 431
362-363 262-263 privacy, 4 31-432
parameter settings and tuning, 362 See also System testing and deployment where to start, 430-431
postdeploym ent phase, 363 Descartes, Rene, 56 threats to eth ics,427-429
taxonom y of data, 361 Deutsch, Sarah, 422 E-world
defined, 349, 351 Diagnostic KM system, 212 See Knowledge transfer
faltering of, reasons for, 370 D ichotom ous(yes/no) questions, 163, 164 Exact reasoning, 66-67
global business infrastructure and, 349 Direct marketing, 368 Example, learning by, 76
 Index i aaa 475
Expandability, 137 Formal reasoning, 66-67 If-changed, 223
Experience, 58, 75 -7 6 Frames, 223-224 If-needed, 223
understanding, 160-161 Frappaolo, Carl, 381 If-then programming statements, 225
Expertise, 7 2 -7 4 Friendly type of expert, 165 Im plem entation, 106
Expert knowledge, 7 2 -7 4 , 153 Frito-Lay, 381-382 Implied warranty, 420
Experts, 58, 92, 106, 153 Fuzzy reasoning Incom pleteness, 106
assigning to know ledge areas, 122-124 See Tacit knowledge capture Inconsistency, 170
becoming, via a corporate knowledge- FX trader, 336 Inconsistent knowledge, 249
based system, 4 1 3 ^ 1 4 Inductive reasoning, 67
(Box 2.4), 74 Inference engine, 131, 228
in knowledge capture, 101-102 G Inferencing, 131
liability o f,415 in knowledge-based codification, 228
reliability from, 168-169 G ender effect, 169 Information, 32
resistors to change, 107 General Electric (G E ), 36 analogies and uncertainties in, 160
types of, 157, 165 General Motors (G M ). 319, 411-412 data, knowledge, and, 60-65
See also Knowledge architecture; Tacit Geographical databases, 361 quality, 260
knowledge capture Gill,T. G .,116 Information glut, 353
Expert systems, 131-132 G lobalization, 40 Information portals, 383
Explicit interteam transfer, 291-292 Global K nowledge Exchange, 85 Information revolution, 40
Explicit knowledge, 27, 44, 71 G odfather type of expert, 157 Information value chains, 32
knowledge codification and, 211-212 G odwin, Judith, 29 Inheritance concept, 224
N onaka’s m odel of know ledge creation Griffin, Ron, 33, 318 Innovation,4 1 ^ 2 , 440
and transformation, 119-121 Gripe sites, 323 Innovative teams, 440
technology support of, 125 Group calendaring and scheduling, 313 Inputs, 334
Explicit to explicit com m unication, 120 Groupware, 181, 308 Instantiated frame, 224
Explicit to explicit knowledge, 215 in action (B ox 10.3), 309 Instruction/training KM system, 212-213
Explicit to tacit com m unication, 120-121 applications Integration, 131
Explicit to tacit knowledge, 215 chat rooms, 312-313 decision-m aking flow, 258
Express warranty, 419 e-mail and knowledge transfer, 312 knowledge-sharing, 258
EXSYS Corporation, 91 group calendaring and scheduling, 313 technical, 257-258
Externalization, 120, 215 knowledge sharing, 313 workflow reengineering, 258
Extranets, 128, 324 newsgroups and work-flow systems, 312 Integrity, 57
at Eastman Kodak Co. (B ox 10.2), 307 video com m unication, 313 Intellectual capital (IC), 26, 32
knowledge exchange and, 306-308 communication structure, 311 (Box 1.2), 33
prerequisites for success, 310-311 Intellectual property,416
session control, 311-312 sending abroad (B ox 14.5), 425
F Intellectual Property E xchange,413
Intelligence, 57-58
Facet. 223 H Intelligent agents, 103, 129-130, 398-399
Face validation, 251 Internalization, 120-121
Fact, 57 H am mel, Gary, 329 International Data Corp., 388-389
False negative, 106 Hammer, Kay, 42 Internet,45, 354
False positive, 106 Hanley, Susan, 74, 453 knowledge transfer and, 293
False representation, 106 Hemingway, Ernest, 210 benefits of, 296-297
FAQs (frequently asked questions), 219 H eterogeneous operation, 130 Internet service providers (ISPs),
Fault-tolerance, 130 Heuristic, 57 294-295
Feasibility, 9 4-95 Heuristics, 71 limitations, 297
(Box 3 .5),96 reasoning and, 66-68 stability and reliability of the
Feature extraction and enhancem ent, 362 H om e D epot, Inc., 318 Web, 295
Federal Trade Com m ission (FT'C), princi- H ooper, T.J., 423 onset of, 44
ples of privacy, 432 H ostile attitude, 171 sharing of knowledge, 34
Fidelity In vestm en t,336-337 Human brain, 333 See also World Wide Web
Filtering, 103 Human elem ent, in knowledge m anage- Internet protocol (IP), 105
Filtering layer, 129-133 m ent, 467-468 Internet public portals, 383
Financial drivers, 38 -3 9 Human learning, 75 Internet service providers (ISPs),
Financial knowledge, 30 See also Learning 294-295
Firewalls, 129 Human thinking InterNIC, 422
Fischer, Bobby, 72 See Learning Interpersonal com m unication, 232
Flat files, 361 Hyperlinks, 421 Interpretive codified knowledge
Flexibility, 131, 137 H ypertext transport protocol (H TTP), systems, 213
Flexible scheduling, 449 104-105 Interviews, as tool for knowledge cap-
Flohr, Salo, 72 ture, 162
Folger, David, 389 managerial implications, 175
Force-fail testing, 244 successful, guide to, 175
Ford Motor Co., 305, 319 ending the interview, 171-172
business-to-em ployee portal (B ox 13.4), IBM, 297, 310 errors made by the knowledge
390-391 Ideawriting, 194 developer, 169-170
476 7777 Index

issues to address, 172-173 knowledge centers, identifying decision tables, 221, 222, 223
phrasing the questions, 166 experts, assigning to each knowledge decision trees, 221, 222, 223
problem s encountered during the area, 122-124 frames, 223-224
interview, 170-171 knowledge content satellites, activat- knowledge-based agents, 230-231
question construction, 166-167 ing, 122 knowledge maps, 217-221
rapid prototyping, 173-174 people core, 121-122 production rules, 224-228
reliability of information from technical core, 124 See also Production rules for tacit
experts, 168-169 authorized access layer, 127-129 knowledge codification
setting the stage and establishing collaborative intelligence and filter- K nowledge consum er interface, 384
rapport, 165-166 ing layer, 129-133 K nowledge conversion, modes of, 214-215
things to avoid, 167-168 knowledge-enabling application K nowledge creation, 115-118
types of, 163-165 layer, 133 transformation and, N onaka’s
Intranet,43, 127-128, 303-306, 323 middleware layer, 133-134 m odel, 119-121
Iterative approach in know ledge repositories layer, 134, 135 K nowledge developer
capture, 101 transport layer, 133 knowledge requirements
user-interface layer, 125-127 cognitive psychology, 232
K nowledge autom ation systems, 333 com puter technology, 231
J K nowledge-based agents, 230-231 dom ain-specific knowledge, 231-232
K nowledge-based rules, 225 knowledge repositories and data
Jefferson, Thomas, 437
K nowledge capital (B ox 3.5), 96 mining, 232
Justifier, 131
K nowledge capture, 33, 41, 60, 9 9-100 liability o f,415
1 8 0-181,414 skills requirements
K blackboarding, 198-200
brainstorming, 183-185, 186, 187
ability to articulate the project’s
rationale, 232
Kelly, Thomas, 147 challenge in (B ox 5.1), 147 interpersonal com munication, 232
KM architecture, layers of, 103-105 concept mapping, 195 job roles, 234
KM blueprint, 102-105 procedure, 196-197 personality attributes, 233, 234
KM system design, 102-105 semantic nets and, 197-198 rapid prototyping, 232-233
Know-how, 56, 66, 71 consensus decision making, 190-191 See also K nowledge capture;
licensing, 413 defined, 146-150 Knowledge management system
Knowing, 277 Delphi m ethod, 194-195 life cycle (KMSLC);Tacit
K nowing-doing gap, 276-277 experts for, 101-102 knowledge capture
K nowledge, 55 -5 6 iterative approach in, 101 Knowledge dim ensions and bottlenecks
cognition and know ledge m anage- nominal group technique (N G T), (Figure 7.3), 214
m ent, 5 9 -6 0 193-194 K nowledge discovery from databases
as a corporate asset, 44 on-site observation, 182-183 (K D D ), 351
data, information, and, 60-65 protocol analysis, 185, 187 K nowledge em bezzlem ent, 432
defined, 57 procedure, exam ple of, 189-190 Knowledge-enabling application layer, 133
definitions scenario, 188-189 K nowledge evaluation, 84, 85
com m on sense, 5 8-59 repertory grid, 191-193 Knowledge hoarders, 262
experience, 58 “smash and grab” approach, 101 K nowledge im plem entation, 86
intelligence, 57-58 techniques, choosing, im plications for Knowledge infrastructure, of a corpora-
knowledge, 57 knowledge management, 200 tion (B ox 9.2), 286
human thinking and learning, 75-76 See also Tacit knowledge capture Knowledge m anagement
im plications for know ledge m anage- Knowledge centers data mining and (B ox 12.2), 352
ment, 76 -7 7 See Knowledge architecture data mining implications for, 369-372
product or service?,416, 417 K nowledge chain, 446-447 fit with neural networks, 337
select characteristics o f (B ox 2.1), 63 Knowledge codification, 209 foundation of (Figure E .l ) ,465
types of defined, 210-212 future, 470-471
com m on sense, 68 how to codify knowledge, 215-216 human elem ent, 467-468
explicit and tacit, 71 im plications for knowledge m anage- knowledge architecture implications
know-how, 66 ment, 234-235 for, 138-139
procedural to episodic, 68-71 knowledge conversion, modes of, 214-215 knowledge m anagement system life
reasoning and heuristics, knowledge developer’s skill set cycle (K M SLC) implications for
6 6 -6 8 , 69 knowledge requirements, 231-232 KM, 109-110
shallow and deep, 66 skills requirements, 232-234 knowledge transfer implications for, 298
vision, 64 reasons for knowledge workers implications for,
(B ox 2.2), 65 diagnosis, 212 460-461
See also Ethical issues o f knowl- instruction/training, 212-213 legal disputes in, 420
edge; Expert knowledge; interpretation, 213 neural networks implications for, 342-343
K nowledge owners; Legal issues planning, role of, 214 shareability factor, 466-467
o f know ledge planning/scheduling, 213 social factor, 4 69-470
K nowledge architecture prediction, 213 system testing and deploym ent impli-
build in-house, buy, or outsource?, things to remember, 213-214 cations for, 264-265
135-138 tools and procedures tacit knowledge capturing implications
im plications for know ledge m anage- case-based reasoning (C B R ), for, 174-175
m ent, 138-139 228-230, 231 team formation (B ox 15.5), 457
 Index i a a i 477
techniques, choosing, im plications user training, 106 e-world, 302
for, 200 manage change and reward struc- extranets and know ledge exchange,
trends, 46 8 -4 6 9 ture, 107-108 306-308
Knowledge m anagem ent (KM ) postsystem evaluation, 108 groupware, 308-313
acceptance of, in industries, 40 Knowledge m anagem ent tools im plications for knowledge m anage-
benefits of, 34 -3 9 See K nowledge portals ment, 322-324
challenges, 4 1 -4 4 K nowledge maps, 217-221 intranets, 303-306
cognition and know ledge m anage- K nowledge networking, 286 explicit interteam transfer, 291-292
m ent, 5 9 -8 0 K nowledge organization framework for (B ox 9.1), 275
defined, 2 6-27 See Knowledge m anagem ent (KM ) im plications for know ledge m anage-
drivers, 3 7 -3 9 K nowledge owners, 409-410 m ent, 298
exam ple (B ox 1.1), 29 clarifying (B ox 14.1), 412 Internet, role of, 293
failure to em bed an operation, 36-37 expert, becom ing, via a corporate benefits of, 296-297
historical overview, 44^-5 knowledge-based system, Internet service providers (ISPs),
im plications for, 50-51 413-414 294-295
intellectual capital (B ox 1.2), 33 know ledge gained on the job, 4 1 1 ^ 1 2 limitations, 297
justification (Table 1.2), 36 price for knowledge, 41 1 ,4 1 3 stability and reliability of the Web, 295
knowledge codification im plications K now ledge portals, 379 know ledge infrastructure, wiring a cor-
for, 234-235 architecture, layers of (B ox 13.5), 394 poration’s (B ox 9.2), 286
knowledge organization, 27-31 business challenge, 386 m ethods
knowledge sharing market potential, 389-391 how knowledge is transferred, 288-293
(B ox 1.4), 42 portals and business transformation, inhibitors o f know ledge transfer,
(B ox 1.3), 39 387-389 287-288
life cycle, 4 8 -5 0 characteristics, 384-385, 386 nature o f the problem , 284-285
myths, 45 -4 7 defined, 380, 382-383 strategies, 285-287
origins of, 39-41 evolution of, 383-384 as a step in a process, 273-275
what it is not, 31 -3 4 illustration, 385, 386, 387 knowing-doing gap, 276-277
See also Drivers for knowledge implications for knowledge management prerequisites for transfer, 277-284
m anagem ent bandwidth, 400-401 tacit know ledge transfer, 292-293
K nowledge m anagem ent system life im plem entation issues, 399-400 See also Groupware; Knowledge shar-
cycle (K M SLC), 82-83 portal product selection, 401-402 ing; Supply chain m anagem ent
B A N K O R exam ple (B ox 3.4), 90-91 who is building enterprise portals?, 399 (SCM ) and knowledge
challenges who sponsors enterprise portals?, 399 exchange; Value chain
knowledge, estim ating worth of next generation of (B ox 13.2), 385 K nowledge validation, 330
(B ox 3.2), 85 perform ance of (B ox 13.1), 380-381 K now ledge value chains, 32
knowledge, processing of portal vendors (Table 13.4),402 K nowledge work, 44
(B ox 3.3), 85-86 technologies K nowledge workers, 44, 353-354, 437
knowledge sharing (B ox 3.1), 84 collaboration, 392-393, 395 business roles in the learning organiza-
challenges in building KM collaboration versus categorization, tion, 441
systems, 8 4-86 396-398 m anagement and leadership, 442-443
versus conventional information systems content m anagem ent, 394, 396, 397 work m anagem ent tasks, 4 43-444
differences, 86-88 functionality, 392, 393, 394 chief knowledge officer (CKO), 452-454
similarities, 88 intelligent agents, 398-399, 400 core com petencies, 440-441
users versus knowledge XM L, role of (B ox 13.6), 397 d efined, 438
workers, 88-89 K nowledge processing, 85-86 (B ox 15.1), 439
design errors to look for, 106 K nowledge producer interface, 384 personality and professional attrib-
evaluating existing infrastructure, 92 K nowledge-related drivers, 38 utes, 439-441
expert, selecting an, 101-102 K nowledge sharing,73, 313, 414 ergonom ics and, 450-452
im plications for know ledge m anage- (B ox 3.1), 84 flexible scheduling through technology
m ent, 109-110 cultural differences’ impact on (B ox 15.2), 449
know ledge developer, role of (B ox 1.4), 42 im plications for knowledge m anage-
design the KM blueprint, 102-105 exam ple (B ox 1.3), 39 m ent, 460-461
im plem ent the KM system, 106 impact of (B ox 15.7), 459 know ledge projects, managing, 456, 460
test the KM system, 105 im pedim ents to (Figure 4.2), 119 KM teams, drafting (B ox 15.5), 457
know ledge m anagem ent developm ent See also K nowledge transfer know ledge sharing (B ox 15.7), 459
(Table 3.2), 8 9 -9 0 K nowledge sharing core, 116 self-training (B ox 15.6), 458
managerial factors, 109-110 Knowledge-sharing integration, 258 knowledge-worker system interface, 451
most critical phase, 83 K nowledge source, 199 managerial considerations, 455
quality assurance, role of, 106 K nowledge testing productivity limitations on, 444
rapid prototyping, 101 See System, testing and deploym ent skills requirements, 450
system justification, 92 K nowledge transfer, 256-257, 272 technology and, 447-450
feasibility question, 9 4 -9 5 , 96, 97 collective sequential transfer, versus users, 88-89
form the KM team, 9 8 -9 9 288-291 work adjustment and, 444-445
knowledge capture, 9 9 -1 0 0 e-business, 313-314 smart leadership requirements,
scope factor, 9 3 -9 4 supply chain management (SCM) and 446-447
strategic planning, 97 -9 8 knowledge exchange, 317-322 vocational needs and reinforcers of
user support, 9 5-97 value chain, 315-217 know ledge workers, 446
478 7777 Index

Koch, Ed, 157

KPMG Peat Marwick LLP, 85, 86
Marino, Mike, 381-382
Market basket analysis (PETCO ex-
O
Krishna, Vish, 33 am ple), 340-341 Office of the Comptroller of the
Kworld, 85 Market potential, knowledge portals and, Currency (O C C ), 90-91
389-390 O ’Leary, William, 459
Master plan, 95 Online analytical processing (O L A P ), 355
L Matching, 224 Online transaction processing (OLTP), 354
Mathaisel, Bud, 305 On-site observation, 182-183
Language problems, 161-162 Maturation, 117 Organizing phase of KM life cycle, 48
Language understanding, 58 M eans-end decision strategies, 73 O ut-of-bounds error, 415
L aoT su, 25 Medical data, 361 Output, 334
Leadership, 442-443 Memory, 56, 58 Owens, David, 454
Leaf nodes, 340 Merchantability, implied warranty of, 420
Learning, 58 Merger, 42
by discovery, 76 Merrill Lynch. 35-36 P
by exam ple, 76 META Group, 399
by experience, 7 5 -7 6 Meyer, A llen, 39 Parameter settings and tuning, 362
See also Data, learning from Microsoft, 34 Parent frames, 224
“Learning" concept, 330-331 Expedia Travel Service, 296 Payback analysis, 95
Learning organization KM platform (B ox 13.5), 394 People, 121
See K nowledge workers Middleware layer, 133-134 People core, 116, 121-122
Legacy systems, 133-134 Midwest Card Services (M CS), 355 PepsiCo, 381-382
Legal issues o f know ledge,40 9-410 Mind mapping, 195 Perceptual slant, 168
copyrights, trademarks, and trade Minsky, Marvin, 223 Performance. 137
names, 416^1-18, 419 M isrepresentation,416 Personal data, 361
disputes, 420 Mission scenarios, 188-189 Personality
im plications for know ledge m anage- Model building, 362 of a knowledge developer, 233, 234
ment, 432-433 M odel of acceptable behavior, 426, 428 in knowledge sharing, 119
intellectual property, sending abroad Model testing, 362-363 of a knowledge worker, 439^441
(B ox 14.5),425 M odel validation, 330 Personnel practices knowledge, 30
liability, 41 8 -4 2 0 Modularity, 137 Personnel-specific drivers, 38
malpractice, 4 2 3 -4 2 4 M otorola, Inc., 458 PETCO, 340-341
poaching (B ox 14.3),422 M ultidimensional rule, 338 Physical layer, 105
strict liability,420 M ultilevel association rule, 338 Planning
warranties,418-4 2 0 Multiple-choice questions, 163, 164 in production rules, 227-228
Web linking and domain name dis- Multiple expert aggregations (Box 5.6), 155 role of, in a KM system, 214
putes, 421—423 Planning KM system, 213
Levels, 341 Plato, 44
Liability, 414 N Poaching, 422
basis of, 4 1 5 -416 Point of presence (POP) application layer, 105
expert, 415 Naccara, G eorge, 453 Point of sale (PO S) systems. 354
knowledge develop er,415 Nader, Abraham, 33 Polanyi, M ichael,44, 71
strict, 420 Narrow-minded “superstars,” 107, 262 Politics, 259
user,415 Navigation, 126 Portability, 131, 137
See also Malpractice Negligence, 415, 416 Portals
Library of Congress, 418 in testing, 246 See also K nowledge portals
copyright Web site (Figure 14.3), 419 Network protocols, 432 Porter, Michael, 315
Logical hierarchy of groups, 383 Network Solutions, Inc. (NSI), Predictive KM systems, 213
Logical testing, 243 421-422 Premise, 225-226
See also System testing and deployment Neural networks, 132-133 Presession plan, 184
Lotus, 34 implications for knowledge m anage- PricewaterhouseCoopers, 4 1 ,4 1 3
ment, 342-343 Privacy, 431^4-32
Neural networks (N N ) Procedural knowledge, 68-7 1 , 73
M See also Data, learning from Procedural rule, 57
Neurons, 331 Procedure type of expert, 157
M achiavelli, 82 Newsgroups and work-flow Process drivers, 38, 39
Machine learning (M L), 354 systems, 312 Production rules for tacit knowledge
Machlup, Fritz, 438 Nightingale, Florence, 332-333 codification, 224
M agnuson-M oss Consum er Product N ode, 334 action clause, 226-227
Warranty Act (1 9 7 5 ),418-419 Nom inal group technique (N G T ), illustration, 227
M aintenance, 108, 258-259, 343 193-194 inferencing, role of, 228
Malhotra, Yogesh, 116, 470 Nonaka, Ikujiro, 44 planning, role of, 227-228
M alpractice,4 2 3-424, 425 model of knowledge creation and premise, 225-226
Management and leadership,442-443 transformation, 119-121 Product knowledge, 30
Managerial factors in know ledge man- Nontrivial error, 415 Profiling, 121-122
agem ent system life cycle Novartis, 286 Projection, 108, 262
(K M SLC), 109-110 Novice, 106 Prosumer, 34
 Index i l i a 479
Protocol analysis, 185, 187 Scalability, 131 System testing and deploym ent, 240-241
procedure, exam ple of, 189-190 Scenarios, 71, 188-189 knowledge testing
scenario, 188-189 Scheduler, 131 definitions, 243
Prototype, 101 Schema-driven orientation, 151 issues, 244-246
Prusak, Larry, 84 Schultz, Christina, 457 logical testing, approaches to
Prusak, Laurence, 349 Schwarzkopf, H. Norm an, 423 circular errors, 246-248
Public key architecture (P K I),432 Scientific data, 361 inconsistent knowledge, 249
Pull technology, 393 Scoping, 9 3-94 redundancy errors, 248
Pupil users, 107 Sears, R oebuck & Co., 45 -4 6 subsumption errors, 248-249
Push technology, 393 Seasoned experience, 117 unusable knowledge, 248
Second-w ave portals, 385 quality and quality assurance, 242-243
Security, 131 testing phase, managing, 253
Q safeguarding the KM system , 264 user acceptance testing, approaches to
users and, 308 techniques, 251-253
Qualcomm , Inc.,413 Self-assessm ent, 431 test criteria, 251
Quality, 242 Self-organization, 335 test team and plan, 249-250
Quality assurance, 106, 242 Self-supervised learning, 335
Quantitative rule, 338 Self-training, 458
Questions, for an interview, 163-165
construction, 166-167
Sem antic knowledge, 69, 70 T
Sem antic nets, 197-198
length of, 171 Sequence discovery, 359 Tacit knowledge, 27, 43, 44, 57, 71
phrasing, 166 Serviceability, 137 codifying, 216
standardized, 171 S e s s io n a l 1 know ledge codification and, 211
Session control, 311-312 N onaka’s m odel o f knowledge creation
R Shaffer, Scott, 454
Shallow knowledge, 66
and transformation, 119-121
subjective nature of, 245
Share, 277 uncaptured (Figure 1.2), 30
Race effect, 169
Range, 223 Shareability factor, in knowledge m an- Tacit know ledge capture, 144-145
Ranking scale questions, 163, 164 agem ent, 466-467 experts, developing a relationship with,
Rapid prototyping, 101 Sharing, 47 156-159
in interviews, 173-174, 175 Shilakes, Christopher, 383 experts, evaluating, 150-151
skills for knowledge developer, 232-233 Short-term memory, 68 aggregating judgm ent of experts
Raven (know ledge m anagem ent strat- Shuttle process, 183 (B ox 5.6), 155
egy), 34 Simon, H. A ., 72 characteristics and capabilities of
Reasoning, 228, 2 7 9 -280 Simon, H erbert, 274 (B ox 5.4), 153
by analogy, 66 Simple mail transfer protocol levels of expertise, 152, 153
heuristics and, 6 6 -6 8 (SM TP), 105 m ultiple experts, pros and cons,
Reasoning errors, 106 Slot, 223 154-156
Redundancy errors, 248 Smart leadership, 446-447 single experts, advantages and draw-
“R e - e v e r y t h in g 470-471 Smart managers, 442-443 backs, 153-154
Refining phase of KM life cycle, 48 “Smash and grab” approach in knowl- single versus multiple experts’
Relational databases, 361 edge capture, 101 knowledge, 152-153
Relevancy, 125 Smith, H. Jefferson, 410 implications for know ledge m anage-
Smith, Reid, 74 ment, 174-175
Reliability, 241
Snow, John, 333 interviews, 162
R em ez, Shereen, 453
Repertory grid, 191-193 Social factor, in know ledge m anagement, guide for, 165-173
R epositories, 105 469-470 rapid prototyping in, 173-174, 175
Repositories layer, 134-135 Socialization, 120-215 types of, 163-165
Resident expert type, 165 Social m ediated com m unication struc- quality of, fuzzy reasoning and, 159
Resistance to change, 107-108, 262-263 ture, 311 analogies and uncertainties in infor-
Resources, 97 Somm a, Joe, 365-366 m ation, 160
Respondent superior,415 Sopheon PLC, 42 experience, understanding, 160-161
Response bias, 170 Spatial mapping, 333 language problem , 161-162
Return for know ledge m anagem ent ini- Stakeholders, 430 See also K nowledge capture
tiatives, 85 Standardized questions, 171 Tacit know ledge transfer, 292-293
Return on investm ent (R O I), 260, 363 St. John’s Corporation, 431 Tacit to explicit com m unication, 120
Return on time, 446 Storyteller type o f expert, 157 Tacit to explicit knowledge, 215
Rickard, Norm an, 241 Strategic planning, 97-98 Tacit to tacit com m unication, 120
Robb, Kenneth L., 364 Strict liability,415, 420 Tacit to tacit knowledge, 215
Subsum ption errors, 248-249 Taxonomy o f data, 361
Role bias, 169
Role negotiation, 263 Supervised learning, 335 TCP/IP, 105
Supply chain m anagem ent (SCM ), 44 Teams
knowledge exchange and, 317-322 KM, developing (B ox 4.2), 138-139
S Support, 338 KM, drafting (B ox 15.5), 457
Surface knowledge, 66 KM, role and formation of (B ox 4.1), 118
Salesperson type o f expert, 157 Swiniarski, Henry, 459 knowledge capture and transfer
Satisficing, 274 Synchronous collaboration, 393, 395 through (Figure 3.6), 100
480 7777 Index

knowledge m anagem ent,9 8-99 Trust, 278, 287 Values, 57

Teamwork, 44, 1 1 7 -1 1 8 ,4 4 0 role of, in the KM life cycle, 49 -5 0 Van Steenbergen, Joe, 350
Technical core Truth,56 Vasella, Dan, 286
See Knowledge architecture Turillo, Michael J.,Jr.,85 Verification, 105
Technical drivers of data mining Tutor user, 107 V ideo com munication, 313
See Data mining (D M ) Tylman, Julie, 383 V ideoconferencing, 43
Technical feasibility, 94 Type I error, 253 Virtual private network (V PN ), 307
Technical integration, 257-258 Type II error, 253 Virtuous cycle
Technologically m ediated com m unica- See Data mining (D M )
tion structure, 311 Vision, 64, 97
Technology, 31, 32
backbone of K M ,40
u (B ox 2.2), 65
Visual clarity, 125
focus of, in the 1990s, 45 Unified messaging services, 312 Vocational reinforcers, 119
intranet, 43 Uniform Commercial Code (U C C ), 416
lack of com m on sense, 59 UNISYS, 310
videoconferencing system, 43 Universal, 125 W
See also Knowledge workers Unstructured (open-ended) questions,
Technology core, 116 163 Warranty
Technology drivers, 37-38 Unsupervised learning, 335 defined, 418
Teltech Resource Network Corp., 42 Unusable knowledge, 248 disclaim er,420
Test cases, 252-253 Usability, 126, 127 express, 419
Test plan, 249-250 User acceptance testing, 243-244, 245 implied, 420
Test team, 249-250 See also System testing and deploy- Web linking and domain name disputes,
Texas Instruments (TI), 147, 413 ment 421-423
Text, 361 User-attitude survey, 263 Web repositories, 361
Think-aloud m ethod, 188 U ser interface, 103 W eighted evaluation m atrix,95, 97
Thinking skills,440 User-interface layer, 125-127 Weights, 334
Threshold level, 334 Users Welch, Jack, 36
Tillman, Anthony, 429 versus knowledge workers, 88-89 Wells, Charles, 29
Tillman, Felipe M., 429 liability of,415 Whispers in the ears, 71
Time series databases, 361 manage change and reward structure, Wilson, Woodrow, 457
Timid type of expert, 165 107-108 Wisdom, 64
Toffler, A lvin, 44 postsystem evaluation, 108 Work adjustment and the knowledge
Top-down approach resistors to change, 107 worker
to building learning models, 331 support of, importance of, 95-96 See K nowledge workers
to im proving ethical behavior, 430-431 training, 106 Workflow reengineering
Tort law, 416 User training and deploym ent integration, 258
Torts, 416 See System testing and deploym ent Work-flow systems and
Total quality m anagement (TQ M ), 44 newsgroups, 312
Trademark,418 Working smarter, not harder, 25-51
statem ent (Figure 14.2),419 V Work m anagem ent tasks,
Trade name, 418 443^44
Training, KM system , 212-213 Vagelos, Roy, 26 WorldCom, Inc., 84
Training sessions, 263 Valid, 56 World W ide Web, 3 6 1 ,363
Training set, 335 Validation, 105 changes to KM through, 37
Transfer function, 334 Validity, 56 See also Internet
Transfer phase o f KM life cycle, 49 Value-added layer, 133 Wright, Bill, 309
Transfer protocol, 289-291 Value Am erica, 433 Wyse, Lois, 145
Transmission control protocol (TCP), 105 Value chain, 315
Transport/Internet layer, 105 business to business (Internet and
Transport layer, 133 extranet), 317 X
Tree, 341-342 business to consum er (Internet),
Troublemakers, 107, 262 316-317 XML, role of (B ox 13.6), 397