The research register for this journal is available at The current issue and full text archive of this

hive of this journal is available at

http://www.emeraldinsight.com/researchregisters http://www.emeraldinsight.com/1463-7154.htm

BPMJ
8,5 Quantitative risk level
estimation of business process
reengineering efforts
490 Thomas J. Crowe
Department of Industrial and Manufacturing Systems Engineering,
University of Missouri, Columbia, Missouri, USA
Pekying Meghan Fong
Advanced Micro Devices, Inc., Sunnyvale, California, USA
Todd A. Bauman
Missouri Lottery, Jefferson City, Missouri, USA, and
Jose L. Zayas-Castro
Department of Industrial and Manufacturing Systems Engineering,
University of Missouri, Columbia, Missouri USA
Keywords Business process re-engineering, Risk, Estimating, Quantitative techniques
Abstract With risk defined as the possibility of deviation in the results from the expected goals,
business process reengineering (BPR) initiatives clearly involve risk taking. However, due to the high
expected returns of such efforts, the acceptable risk levels of BPR will tend to be greater than those
of less ambitious projects. This research reports the development of a tool to quantitatively estimate
the potential risk level of a BPR effort before an organization commits its resources to that effort.
The underlying research employed a survey of BPR-experienced organizations to collect assessment
information in order to build a BPR risk estimation model. The developed tool uses triangular fuzzy
numbers to approximate the degree of success/failure of proposed BPR initiatives. The tool can be
applied by any organization contemplating BPR, thus giving such organizations a heretofore
unavailable estimate of the risk level of proposed BPR efforts. Validation was performed based upon
an 18-month BPR project conducted at the Missouri Lottery.

Introduction
In 1992 Goll defined BPR as ``total transformation of a business, an unconstrained
reshaping of all business processes, technologies and management systems, as
well as organizational structure and values, to achieve quantum leaps in
performance throughout the business'' (Goll, 1992, p. 29). Clearly BPR is
synonymous with change, and resistance to change is human nature. The fear of
the dramatic changes from BPR, even before they take place, needs to be
anticipated and addressed to ensure the success of projects. The major element
engendering the fear is the uncertainties associated with BPR, and these
uncertainties translate into risks.
The risk of a project is the possibility of deviation in the results from the
expected goals. So risk does not equal loss as is sometimes perceived; in fact, it
Business Process Management
Journal, Vol. 8 No. 5, 2002,
can be an opportunity. Thus risk is a relative concept not an absolute one
pp. 490-511. # MCB UP Limited,
1463-7154
(Spaulding, 1997). To evaluate the risk of a BPR project it is important not only
DOI 10.1108/14637150210449148 to look at the risk involved but also at the return. Since BPR projects are
intended to deliver significant improvements, the acceptable risks may be Business process
relatively higher than with less aggressive projects. reengineering
In order to assist managers in presenting BPR projects to upper efforts
management before committing resources, a quantitative risk-assessment tool
is needed. Managers are often stranded with limited tools to assess risks and
since no formal tools are available, each employs different and often
questionable mechanisms to determine the potential benefits BPR can bring to 491
their organizations. The risk of such a practice is even greater than the BPR
projects, since the opportunity cost of not implementing BPR can outweigh all
the possible difficulties.
The tool developed in this research can be used by any organization
contemplating BPR to quantitatively estimate the risks associated with its BPR
efforts.

Literature review
Risk ± success and failure in BPR
Previous research has been reported to determine the critical success and
failure factors in BPR implementations (Lee, 1995; Crowe and Rolfes, 1998).
These success and failure factors equate to risk and will be reviewed in some
detail.
Critical success factors. Detailed studies have identified four critical success
factors in BPR (Lee, 1995). These four factors are:
(1) Egalitarian leadership.
(2) Collaborative working environment.
(3) Top management commitment.
(4) Change in management systems.
Due to the nature of BPR, which stresses radical redesign of cross-functional
business processes, organizational culture tends to be the focus of these
changes. Organizational culture defines the identity of an organization through
the common understanding of all business units about the mechanism of the
daily operation at that organization (Quinn and Cameron, 1988). This common
understanding consists of organizational purpose, performance criteria, chain
of authority, legitimate base of power, decision making mechanism, leadership
style, compliance, evaluation and motivation (Quinn and Cameron, 1988).
Egalitarian culture fosters a set of common beliefs giving value to positive
communication, the cooperative workplace, active information flow, and
empowerment of employees. Ideally with BPR employees are to be empowered
and to work cooperatively in the new system. Egalitarian culture allows such
positive changes to take place with little resistance. Egalitarian leadership is
ideologically centered in a democratic system. The interactions between chains
of command reflect the organization's ability to adapt to changes. The major
components of such leadership that have proven to affect positively BPR
efforts are (Lee, 1995):
BPMJ . shared vision/information;
8,5 . open communication;
. confidence and trust in subordinates; and
. constructive use of subordinates' ideas.

492 Closely related to the egalitarian leadership, collaborative working

environment is also identified as one of the BPR critical success factors.
Working environment refers to the interactions between coworkers and the
level of cooperation shown at the work place. To be considered a collaborative
working environment, the following conditions have to be maintained in the
workplace (Lee, 1995):
. friendly interactions;
. confidence and trust;
. teamwork performance;
. cooperative environment; and
. recognition among employees.
Top management commitment is significant throughout the course of the BPR
project. In order to achieve maximum potential in a BPR effort changes have to
be made to align with the organizational strategic direction, so it is vital to
secure commitment from upper management to steer the project. At the same
time resistance confronted during the course of the project can be handled most
expeditiously with clear top management commitment. In order to have a
successful BPR effort upper management must communicate to the affected
business units, motivate movement and step in to resolve differences (Stanton
et al., 1993). Therefore the following circumstances have to be displayed:
. sufficient knowledge about the BPR project;
. realistic expectation of BPR results; and
. frequent communication with BPR team and users.
In an effort that radically changes the organizational structure, management
systems have to be modified to support the newly redesigned processes.
Effective management systems monitor behavioral change in BPR efforts by
assisting employees in the transition period to the new working environment.
Such an effective change management system is characterized by the following
managerial support (Lee, 1995):
. new reward systems;
. up-to-date communication methods;
. performance measurement;
. employee empowerment; and
. timely training and education.
Critical failure factors. The seemingly high failure rate of BPR projects has Business process
always been one of the major roadblocks in convincing organizations to reengineering
commit to BPR effort. Based on a 1995 study by the Standish Group efforts
International (USA) consisting of 8,380 BPR projects at 365 companies, 84
percent of all projects failed or at least experienced some major problems
(Valimaki and Tissari, 1997). Given this unusually high failure rate compared
to other types of improvement efforts, finding critical failure factors becomes 493
an important topic for research. Past research shows that there are two primary
reasons for such incidents: employee resistance to change (Stanton et al., 1992);
and lack of resources for the BPR effort (Bashein et al., 1994).
BPR efforts have always stressed the importance of introducing radical
changes to obtain quantum leap improvements. On top of such merit
objectives, the effort also demands radical changes in employees' behavior at
work. Since it is human nature to develop inertia, it is their instinctive reaction
to resist changes instituted by BPR efforts. The resistance is especially high
among employees who are directly affected by these changes. In fact if no
resistance is detected, the BPR effort probably is not being done right.
Usually BPR projects require serious commitment of resources from the
start until the end. Restructuring an organization is an expensive operation,
and on many occasions heavy investment is indispensable to upgrade the
information technology system to support newly redesigned business
processes. At the same time BPR efforts often call for full-time dedication of
personnel directly involved, causing their normal jobs and responsibilities to
have to be delegated during the course of the project.
Of the two reasons for BPR failure, quantitative research has concluded that
only employee resistance is statistically qualified (Lee, 1995). Employee
resistance comes from fear of uncertainties that BPR changes will bring. The
most common fear is the fear of downsizing, which has been stereotypically
associated with the BPR effort. Even though the objectives of BPR efforts do
not include laying off employees, downsizing has been a result of many BPR
efforts. Below are the components of this critical failure factor that measure the
level of employee resistance (Lee, 1995):
. middle management fear of losing authority;
. employees fear of losing job;
. skepticism about project results; and
. feeling uncomfortable with new working environment.

Definition and assessment of risk

Risk is the potential, or probability, of an adverse event. We deal with risk in
our daily lives, simply driving to work and eating our dinner. Even though
driving nowadays translates to higher risk than flying due to the higher car
accident rate, people tend to believe that flying is riskier because the impact of
a flight accident is bigger. People tend to overestimate the likelihood of low
probability events (death by tornado) and underestimate higher risk levels
BPMJ (death by heart disease) (Burgman et al., 1993). Similarly BPR efforts have been
8,5 continually viewed as risky business since the changes involved are dramatic.
As with any kind of investment BPR projects introduce uncertainties. These
uncertainties can result from changes in the working environment, job duties,
and organizational structure. It is human nature to resist changes, even if the
changes are for the better. From the employees' perspective this resistance is
494 reinforced by the fear of losing direction in the new system or even not being
included in the new system. From the management point of view it is
heightened by the uncertainty in estimating the risk and return on the
investment due to the limited knowledge.
However, an investment should never be judged based only on the risk, but
also on the return. There are two elements of risk:
(1) the expected risk of the investment; and
(2) the expected return relative to the risk (Spaulding, 1997).
There is no doubt that the potential benefits of BPR effort are huge, thus
relatively the risk involved can be high.
Risk assessment has been researched in various areas; for example, finance/
investment, conservation biology, insurance business, and so on. Rowe (1977)
defines risk assessment as the process of obtaining quantitative or qualitative
measures of risk levels. Many risk-assessing methods have been developed in
different research areas, especially in the finance/investment domain. These
methods range from simple statistical measurements to sophisticated
estimating models, including standard deviation, semi-variance, value-at-risk
(VAR), and the sharpe ratio (Spaulding, 1997).
Standard deviation. Standard deviation is a very commonly used statistical
measurement due to its relative ease of calculation. It is a measure of uncertainties
and is often considered a tool to measure risk. This measure of dispersion is based
on the normal distribution, and the generic formula for calculating standard
deviation in terms of investment portfolio is (Spaulding, 1997):
v
#
u"P
u fRi R x†g 2
Sc ˆ t ;
n

where:
Ri = the return of the ith portfolio;
R(x) = the arithmetic average of all portfolios;
n = the number of portfolios.
Semi-variance. Semi-variance was introduced to the investment world as a better
way to measure risk than standard deviation. Unlike standard deviation, which
does not take into consideration that ``over-performers'' should not be discounted
like the ``under-achievers'', the semi-variance will only measure the ``downside
risk''. The semi-variance is simply half the variance of the differences between all
possible points spaced a constant distance apart (Hoskins, 1973). Thus, the Business process
formula for the semi-variance for portfolio evaluation is (Spaulding, 1997): reengineering
X
SV t† ˆ 2
Ri t pi t Ri † ; efforts

where
Ri = a possible return; 495
t = standard deviation;
pi = the probability of the actual return being Ri;
t = the target return.
Value at risk. Value at risk (VAR) is a new concept in financial risk
measurement, which is used extensively for measuring the market risk of
portfolios of assets and/or liabilities (Jorion, 1996a, b). VAR summarizes the
worst expected loss over a target horizon within a given confidence interval.
The one advantage of this risk measurement tool is that its unit is currency,
which is more easily understood by common investors.
The two most frequently used VARs are the closed-form VAR and the
Monte-Carlo VAR. The closed-form VAR (which is also referred to as the linear-
or parametric-VAR) is used to estimate financial risks for simple portfolios by
assuming the portfolio's profitability to be normally distributed. The formula
for the closed-form VAR is (Jorion, 1996a, b):
VAR t† ˆ PF  K%

where:
t = observation time period;
PF = portfolio;
P
K% = R(PRL), all occurrences of returns less than K% add up to PRL
of the total number of t;
PRL = 100% ± desirable confidence interval.
On the other hand, the Monte-Carlo VAR applies Monte-Carlo simulation to
randomly construct a histogram of possible profits or losses for a portfolio over
a specific time period. The analysis uses the same formula, but the data are the
result of the simulation instead of historical information. Relative to the closed-
form VAR, Monte-Carlo VAR is difficult to implement in terms of the modeling
and the time to run the model (Glasserman et al., 2000).
Sharpe ratio. The Sharpe ratio was first introduced in 1966 as a fresh model
for financial portfolio management, and was then widely accepted as a common
measure of risk. This ratio is a direct measure of reward-to-risk by measuring
the excess return that a portfolio provides over the cash return, divided by the
standard deviation of the portfolio's return. The formula for the Sharpe ratio is
(Sharpe, 1994):
BPMJ S x† ˆ ‰R x† R$ Š= x†;
8,5
where:
x = amount of investment;
R(x) = average annual rate of return of x;
496 R$ = best available rate of return of a ``riskless'' security (i.e. cash);
(x) = standard deviation of x.
Even though there is a wide array of risk assessment tools readily available,
these tools focus mainly on different sets of criteria, based on different sets of
factors, and under different sets of assumptions. Since there is very limited
research in the area of BPR risk assessment and even less on quantitative
methods, the goal of this research is to develop a quantitative risk-assessment
tool for managers to evaluate the risk-level of a BPR effort at their
organizations before committing themselves to the investment.

Research methodology
This research developed a tool that utilizes the critical success factors (CSFs) and
critical failure factor (CFF) to assess the risk level of a potential BPR project. In
order to develop this tool, a series of activities were performed as follows:
(1) define underlying elements;
(2) construct questionnaires;
(3) conduct survey;
(4) analyze data; and
(5) present results.
In a previous section the critical factors affecting BPR effort, as discovered by
Lee (1995), have been discussed at length. The critical success factors (CSFs)
are egalitarian leadership, collaborative working environment, top
management commitment, and change in management systems. The only
significant critical failure factor (CFF) is employee resistance. These factors
will be used to construct the questionnaires.

Fuzzy sets and fuzzy numbers

``The key elements in human thinking are not numbers, but labels of fuzzy sets,
that is, classes of objects in which the transition from membership to non-
membership is gradual rather than abrupt'' (Zadeh, 1973). In addition, the
process of human thinking often involves approximate reasoning rather than
precise determination (Kaufmann and Gupta, 1988). Thus, a fuzzy set is defined
as class A in X by a membership function A(x) which associates with each
element of X, say x, a real number in [0, 1]. A(x) represents the grade of
membership of x to A; for A(x) = 1, one has a total belonging, while for A(x) = 0,
one does not belong to class A (Nola and Sessa, 1989).
 Triangular fuzzy numbers (TFN). A TFN, as illustrated in Figure 1, can be Business process
defined by a triplet (a1, a2, a3), where the membership function of the TFN is reengineering
defined as (Kaufmann and Gupta, 1988): efforts
A x† ˆ 0; x < a1 ;
ˆ x a1 †= a2 a1 †; a1  x  a2 ;
ˆ a3 x†= a3 a2 †; a2  x  a3 ; 497
ˆ 0; x > a3 :
TFN example. Participants in the research survey used human reasoning to
respond to the questions and to rate the success/failure of the BPR efforts they
were involved in. Since the underlying data collected are approximate
reasoning and fuzzy in nature, TFNs are appropriate to estimate the degree of
fuzziness in the success/failure of past BPR efforts.
The only algebraic operation on TFNs that will be applied in this study is
the addition of TFNs. Addition of TFNs is proven to yield a TFN: define two
TFNs A and B by the triplets as A = (a1, a2, a3) and B = (b1, b2, b3). Thus the
addition of A and B will be (Kaufmann and Gupta, 1988):
A ‡†B ˆ a1 ; a2 ; a3 † ‡† b1 ; b2 ; b3 †
ˆ a1 ‡ b1 ; a2 ‡ b2 ; a3 ‡ b3 †; aTFN;

where (+) indicates a fuzzy addition.

Questionnaire development
As mentioned above the four CSFs (egalitarian leadership, collaborative
working environment, top management commitment, and change in

Figure 1.
TFN - example
BPMJ management systems) and one CFF (employee resistance) are used to
8,5 construct the questionnaire. These factors are further broken down into
smaller significant elements of three to five per factor. Each of these elements
is represented in a question in the questionnaire. These significant elements
are derived from the definitions of the critical factors in Lee's work (Lee,
1995).
498 This survey was intended to collect information from organizations that
have been through BPR efforts. The main purpose of this survey is not to
assess the success or failure of the BPR effort, but to collect information
reflective of the general organizational characteristics at the time of the
reengineering. The questionnaires were distributed to the leader of or
personnel involved in the BPR project at each organization. The survey
involved minimal personal exposure and the information obtained was held
in strict confidence. No reference was made to specific individuals, names of
organization, or types of organization in any report. The participation in this
study was completely voluntary, and the participants were free to withdraw
from the study at any time without penalty. In fact, a non-disclosure
agreement accompanied the survey for privacy protection. The survey
questionnaire is shown in Figure 2.
The questionnaire (Figure 2) was used to collect data from organizations
identified through various sources to have undergone BPR efforts.
Questionnaires were sent to 25 selected organizations in two phases. These
organizations were given a period of one month to reply to the survey. Two
weeks into the period follow up through e-mail and regular postal mails were
done to remind participating organizations of the survey.

Data analysis
Data collection was started as soon as the first survey reply was received. This
process took place in approximately two and a half months, and the information
collected was compiled into a library of ``organizational characteristics'' for
analysis. While compiling this library, preliminary application of the proposed
methodology was initiated with available information. For each organization, an
``organizational profile'' was created. This profile contained organizational
characteristics related to the CSFs and CFF.
To calculate the point for each critical factor two considerations were taken
into account: the score for each significant element of the factor, and the weight
for each significant element. A score of ``1'' (strongly agree) would yield a full
mark of 1 towards that element; a score of ``2'' (agree) would yield a mark of 0.5
towards that element; a score of ``3'' (neutral) would not yield any mark towards
the element; a score of ``4'' (disagree) would yield a negative mark of 0.5 towards
the element; and a score of ``5'' (strongly disagree) would yield a full negative
mark of ±1 towards the element. Each significant element would carry equal
weight towards each critical factor. For example, for a critical factor with four
significant elements of scores 2, 3, 1, 4 each, the point for that critical factor
would be:
Business process
reengineering
efforts

499

Figure 2.
Questionnaire
BPMJ
8,5

500

Figure 2.
 CF x† ˆ 0:25 0:5 ‡ 0 ‡ 10:5† Business process
ˆ 0:25: reengineering
efforts
After calculating all the factor points for an organization, the total CF points of
that organization would be the summation of these points. However, the point
for critical failure factor would carry a negative effect towards the total points.
Thus, the calculation for the total points for organization xyz would be: 501
X
CFxyz ˆ CSF 1† ‡ CSF 2† ‡ CSF 3† ‡ CSF 4† CFF:
After acquiring all the total CF points from the targeted organizations, the next
phase of the data analysis was initiated.
First, the BPR scores were scaled up by the factor of the number of
participating organizations, N. This was necessary to ensure the integrity of
the calculation concerning the new TFN that would be created by adding the
triplets from the survey, which were the past BPR effort ratings. In order for
the BPR scores to be meaningful, they have to be on the same basis as the new
TFN. Scaling the BPR scores is like adding N number of identical BPR triplets
so that the scores would be on the same plane as the new TFN. Second, the total
CF points (X-value) for these organizations and the scaled up BPR scores (Y-
value) was fitted onto a simple linear regression model:
Y ˆ mX ‡ b:
The underlying assumption was that the correlation between the total points
and the degree of BPR success is linear: an organization with high total points
has high degree of success. The regression model was used to find the score in
the TFNijk (membership value) of a new organization ``ijk'' by using the total
points calculated based on the same questions about the critical factors for
organization ``ijk''.
The TFNijk were applied after the regression model was generated. The
ratings provided by the organizations representing the ``total failure'' (a1),
``average'' (a2), and ``great success'' (a3), of a project were the triplets for the
TFN. By adding these triplets, a new TFN with new triplets was created:
X
TFNmodel ˆ TFN triplets†:

Figure 3 shows an example of the new TFN when it is constructed.

The membership function of this TFN was divided into eight segments, each
representing a degree of success: from ``total failure'' to ``failure'' to ``moderate
failure'' to ``minimal failure'' to ``minimal success'' to ``moderate success'' to
``success'' to ``great success''. These segments were determined by the values of
the final TFNs and the -cut level. Alternatively, the interval of TFNs can be
defined at level as shown:
† †
A ˆ ‰a1 ; a3 Š
ˆ ‰ a2 a1 † ‡ a1 ; a3 a2 † ‡ a3 Š;
BPMJ
8,5

502

Figure 3.
Methodology TFN ±
example

where a1, a2 and a3 were the three triplets of the TFN, and was the variable
used to determine the parallel cut to the x-axis of the TFN (Kaufmann and
Gupta, 1988). In order to have eight segments for the TFN, the -cut value was
set at an increment of 0.25.
This new TFN served as a base for computing the likelihood of an
organization succeeding in a BPR effort by using the membership value within
the TFN, found by utilizing the regression model, as an indicator. The
membership value was plotted onto the new TFN and the segment that this
value fell onto would indicate the risk level of that new organization if it
decides to invest in a BPR effort.
In order to find the risk level ( ), the indicator has to be set to be equivalent
to the following equations:
i† if TFNijk > a2 ; then
TFNijk ˆ A IJK† and A IJK† ˆ a3 a2 † ‡ a3
Thus ˆ TFNijk a3 = a2 a3 †

i† if TFNijk >a2 ; then

TFNijk ˆ A IJK† and A IJK† ˆ a2 a1 † ‡ a1
Thus ˆ TFNijk a1 = a2 a1 †:
After finding the value, it can be interpreted into more user-friendly terms.
As discussed above, there would be eight segments where the risk would be
assessed. Table I shows a summary of the interpretations.

Alpha-value TFNijk < a2 TFNijk < a2

0.00 to 0.25 Total failure Minimal success

Table I. 0.25 to 0.50 Failure Moderate success
Alpha-value risk 0.50 to 0.75 Moderate failure Success
interpretation 0.75 to 1.00 Minimal failure Great success
Results and validation Business process
Data collection reengineering
To ensure the consistency of the survey results, potential survey participants efforts
were prescreened. The BPR projects at these organizations were studied in
terms of the magnitude of the projects (e.g. how many departments were
involved) and the length of the projects (e.g. two months or two years). All of
the research surveys were sent to organizations with projects that were either 503
large (e.g. at least two departments were involved) or lengthy (e.g. lasted at
least one year) or both.
Research surveys were sent to organizations across North America to collect
data needed to build the proposed methodology. A total of 25 surveys were sent
and seven were returned, thus making a return rate of 28 percent. The typical
reasons given for declines of participation in the research and non-returned
surveys are:
. Personnel involved in previous BPR effort have since left the organizations.
. Lack of resources (e.g. manpower and time).
. Company policy not to participate in survey.
. Difficulty remembering previous BPR effort and the organizational
characteristics.
Due to the nature of the survey participation, the returned surveys are
anonymous in order to protect the organizational information provided by the
organizations. Hence, the specific types of industry these organizations operate
in are unrevealed. However, the industry types of the twenty-five organizations
were very diverse, ranging from governmental agencies to retail businesses.
The summary breakdown of these organizations is presented in Table II.
The majority of these BPR efforts had outside assistance from consultants
throughout the whole period of the effort or at a certain time of the
reengineering changes. Most of these organizations carried out such effort in
order to cut costs and boost productivity in general. Others focused primarily
on particular areas for improvements such as supply/demand logistics,
procurement systems, technology-related domains (e.g. management

Types of industry Frequency Percent

Governmental agency 5 0.20

Technology related 4 0.16
Consumer products 4 0.16
Manufacturing 3 0.12
Telecommunications 2 0.08
Automotive 2 0.08
Retail business 2 0.08
Healthcare 1 0.04 Table II.
Utilities 1 0.04 Industry breakdown of
Insurance 1 0.04 target organizations
BPMJ information systems and Internet approach), human resources and customer
8,5 service. Table III summarizes the foci of the BPR efforts.

Survey data analysis

As revealed in the survey results, each of the participating organizations has
different combinations of organizational characteristics. By using the survey
504 results and the methodology discussed above, the ``organizational profile'' for
each of these organizations is constructed. Figures 4-10 represent the
``organizational profile'' for the survey participants as shown as a summary
trend based on the critical factors discussed and the survey results.
Four of the seven organizations exhibited negative critical factor (CF) points
and the other three exhibited positive CF points, thus making this a well-
balanced organizational base. As discussed above, a regression model of a
straight line will be built using these organizational profiles, to find the X-
coefficient and the Y-intercept values. Figure 11 shows the relationship
between the CF points of these organizations and the scaled BPR scores.
The values for the X-coefficient and the Y-intercept are 7.475 and 38.328
respectively, thus making the straight line of:
Y ˆ 7:475X ‡ 38:328:
After building the regression model, the final TFN was constructed. In order to
find the triplets for the final TFN, addition of the triplets of the seven survey
participants was carried out.
TFNmodel ˆ  TFN triplets†
ˆ 4; 37; 67†:
Thus, a1(model) = 4, a2(model) = 37, and a3(model) = 67. These values were
used in validating the methodology as used by a potential organization that
wishes to estimate its BPR risk level. By using these final triplets, the final TFN
was then established. The final TFN is shown in Figure 12.
After constructing the final TFN, the methodology is set to be tested by the
validating research participant, the Missouri Lottery.

Methodology validation
To validate the integrity of the proposed technique, the validation was performed
by testing the methodology using the survey results from the target research

Area of improvement Frequency Percentage

General improvement 7 28
Supply/demand logistics 4 16
Procurement systems 4 16
Technology-related 4 16
Table III. Human resources 4 16
Focus of BPR efforts Customer service 2 8
 Business process
reengineering
efforts

505

Figure 4.
Organizational
profile ± A

Figure 5.
Organizational
profile ± B

Figure 6.
Organizational
profile ± C

organization, the Missouri Lottery. The Missouri Lottery organizational information

from the survey was fed into the methodology and the estimation of risk level in
terms of success/failure was then compared to the rating given in the survey.
The Missouri Lottery is a ``profit-driven'' non-profit organization that was
approved in 1984 by the State of Missouri. This is a governmental agency that
started its operations in 1986, and is housed under the Missouri Department of
Revenue. The organization is governed by a five-member commission, which is
BPMJ
8,5

506

Figure 7.
Organizational
profile ± D

Figure 8.
Organizational
profile ± E

Figure 9.
Organizational
profile ± F

appointed by the Missouri Governor and approved by the State. The Missouri
Lottery main office is located in Jefferson City, with three regional offices
located in Kansas City, Springfield, and St Louis. There are five divisions
within the organization: the Executive Office, the Finance and Administration
Division, the Sales and Marketing Division, the Security Division, and the
Communication Division. There are a total of 175 employees from all offices,
and the average yearly sales for the last five years are US$425 million. The
organization has since provided the State of Missouri with US$1.3 billion in
support including educational funding.
 Business process
reengineering
efforts

507
Figure 10.
Organizational
profile ± G

Figure 11.
Regression model

Figure 12.
Final TFN

BPR project at the Missouri Lottery. The research project started in January
1999 and ended in June 2000. The working group consisted of members from
the University of Missouri and the Missouri Lottery representative as a
group leader. The group started the project by examining the various
business processes at the Missouri Lottery. As a result, a total of 25
processes were identified and documented. From the analysis of the
BPMJ processes, the group started to address several areas for improvements via
8,5 reengineering.
A major reengineering effort was recommended, and the group took a multi-
phase modular approach suggested by the upper management of the Missouri
Lottery. This approach was initiated by the introduction of an empowered task
force by the BPR group to dramatically improve the daily operations that
508 involve Missouri Lottery retailers' special business needs. Then, the second
task force that focused on the promotions and events by the Missouri Lottery
was proposed to enhance communications, planning and evaluations of these
activities. The third module of the effort was to formulate another task force
that would serve as a singular channel where Missouri Lottery can better assist
its external customers.
Validation target ± Missouri Lottery. The Missouri Lottery representative of
the working group agreed to participate in the research survey, and the survey
result is used to validate the proposed methodology.
An identical research survey was presented to the representative to collect
Missouri Lottery organizational characteristics and the degree of success of the
BPR effort conducted. Later the data analysis process as outlined above was
carried out to compute the critical factor points and the triangular fuzzy
Number for the organization. From the analysis, the Missouri Lottery has
organizational-characteristics scores of (where the subscript ML indicates the
Missouri Lottery):
CSF 1† ˆ 0:625; CSF 2† ˆ 0:100; CSF 3† ˆ 0:333;
CSF 4† ˆ 0:200; and CFF ˆ 0:125; thus CFML ˆ 1:383:

Figure 13 represents the ``organizational profile'' for the Missouri Lottery.

By substituting the total critical factor points into the regression model, the
triangular fuzzy number for the Missouri Lottery is:
TFNML ˆ 7:475 1:383† ‡ 38:328
ˆ 27:990:

Figure 13.
Organizational profile ±
Missouri Lottery
Recall the triplets calculated above, a1 (model) = 4, a2 (model) = 37, and a3 Business process
(model) = 67. Observe that TFNML < a2 (model), so the calculation of the reengineering
value for the Missouri Lottery is: efforts
AML ˆ ‰TFNML a1 model†Š=‰a2 model† a1 model†Š
ˆ 27:990 4:000†= 37:000 4:000†:
ˆ 0:727: 509

The triplets of BPR success from the Missouri Lottery were a1 (ML) = 0, a2 (ML)
= 6, and a3 (ML) = 8, as shown in Figure 14.
By using simple interpolation, ML of 0.727 is equivalent to TFNML of 4.362.
Remember that the result from the survey indicated a score of 4 by the survey
participant. Thus the difference of TFNML and the BPR score from the survey is 0.362.
By dissecting TFNML into eight segments that depict the eight levels of
success/failure, the increment of 1.5 would indicate the beginning of the
adjacent level on the range of 0 to 6 on the Missouri Lottery Triplets (e.g. 0.0 to
1.5 indicates the level of ``total failure''). Thus, the BPR score from the survey
would be under the category of ``moderate failure'' since the score falls between
the range of 3.0 and 4.5. At the same time, the Missouri Lottery's BPR effort
with (ML of 0.727 or TFNML of 4.362 would fall under the same category.
Subsequently, it indicates that the estimated risk level using the model is
aligned with the participant's rating of the BPR effort.

Conclusion and discussion

Observations and discussion
Observations have been made throughout the course of the research. The low
return rate of the research survey (28 percent) has constituted a small sample size
for the basis of the proposed methodology. However, BPR projects are large in
nature, in terms of both the investments of resources and time. Thus, there are
few organizations that have gone through a complete BPR effort and were able to
participate in the research survey. In addition, there are limited numbers of
organizations that are willing to participate in such research studies.

Figure 14.
Missouri Lottery
Triplets
BPMJ However, research shows that the process of inducting theory using case
8,5 studies is especially appropriate in new topic areas even when the sample size
is fairly small (Eisenhardt, 1989). There are three strengths of using such an
approach:
(1) likelihood of establishing novel theory (Cameron and Quinn, 1988);
510 (2) developing methodology is potentially testable with constructs that can
be readily measured and hypotheses that can be proven false (Hannan
and Freeman, 1977); and
(3) resultant theory is likely to be empirically valid (Mintzberg, 1979).
In the future, as increasing number of organizations undertake such endeavors,
more organizations are likely to be willing to participate in similar research. In
addition, these organizations may become more deeply involved in such
research instead of merely supplying information for the studies.
Another noteworthy aspect of the research survey was the fact that it was a
retrospective approach versus a longitudinal study because the data collected
through the survey were historical information. This matter might have
resulted in a potential preconception by the survey participants between the
BPR success/failure and the factors presented in the survey.
To recall the validation of the proposed methodology, the Missouri Lottery's
BPR effort has ML of 0.727 or TFNML of 4.362, and the survey indicated a four.
The difference of 0.362 can be attributed to the fact that the participant in the
survey answered the survey with a discrete number, and the calculation of
TFNML is in non-integer number. In a sense, the calculation of TFNML actually
reflects the participant's rating on the BPR effort.

Conclusion
Most research on BPR has focused on approaches to document business
processes, and on techniques to identify and select potential candidates for
reengineering. Limited research has been conducted in the area of assessing the
risk of such efforts, especially on quantitative risk estimation.
By using real world data collected from organizations that have experienced
radical changes through BPR efforts, a research model based on triangular
fuzzy numbers has been developed and validated. The resultant research model
was used to develop a tool that allows any organization considering BPR to
quantitatively estimate the potential risk level of those BPR efforts before
committing resources to BPR. Heretofore, no such quantitative BPR risk
assessment tool was available. Having such a quantitative BPR risk
assessment tool will improve management's a priori insights into the potential
outcomes of BPR. Continuing work is expected to show that such insights will
improve the overall success rate of BPR initiatives.

References
Bashein, B., Markus, L. and Riley, P. (1994), ``Precondition for BPR success'', Information Systems
Management, Vol. 11 No. 2, pp. 7-13.
Burgman, M., Ferson, S. and Akcakaya, H. (1993), ``Risk assessment in conservation biology'', Business process
Population and Community Biology Series, Chapman & Hall, New York, NY.
Cameron, K. and Quinn, R. (1988), ``Organizational paradox and transformation'', Paradox and
reengineering
Transformation, Cambridge, MA. efforts
Crowe, T. and Rolfes, J. (1998), ``Selecting BPR projects based on strategic objectives'', Business
Process Management Journal, Vol. 4 No. 2, pp. 114-36.
Eisenhardt, K. (1989), ``Building theories from case study research'', Academy of Management 511
Review, Vol. 14 No. 4, pp. 532-50.
Glasserman, P., Heidelberger, P. and Shahabuddin, P. (2000), ``Variance reduction techniques for
estimating value-at-risk'', Management Science, Vol. 46 No. 10, pp. 1349-64.
Goll, E.O. (1992), ``Let's debunk the myths and misconceptions about reengineering'', APICS The
Performance Advantage, December, pp. 29-32.
Hannan, M. and Freeman, J. (1977), ``The population ecology of organizations'', American Journal
of Sociology, Vol. 82, pp. 929-64.
Hoskins, C.G. (1973), ``Distinctions between risk and uncertainty'', Journal of Business Finance,
Vol. 5 No. 1, Spring, pp. 10-19.
Jorion, P. (1996a), ``Risk2 measuring the risk in value at risk'', Financial Analysts Journal, Vol. 52
No. 6, November/December, pp. 47-56.
Jorion, P. (1996b), Homepage, available at: www.gsm.uci.edu/~jorion/oc/case3.html (accessed
September 2000).
Kaufmann, A. and Gupta, M. (1988), Fuzzy Mathematical Models in Engineering and Management
Science, North-Holland, New York, NY.
Lee, J. (1995), ``An exploratory study of organizational/managerial factors influencing business
process reengineering implementation: an empirical study of critical success factors and
resistance management'', University of Nebraska, Lincoln, NE.
Mintzberg, H. (1979), The Structuring of Organizations: a Synthesis of the Research, Prentice-
Hall, Englewood Cliffs, NJ.
Nola, A. and Sessa, S. (1989), Fuzzy Relation Equations and Their Applications to Knowledge
Engineering, Kluwer Academic Publishers, Boston, MA.
Quinn, R. and Cameron, K. (1988), Paradox and Transformation: Toward a Theory of Change in
Organization and Management, Ballinger, Cambridge, MA.
Rowe, W.D. (1977), An Anatomy of Risk, Wiley, New York, NY.
Sharpe, W.F. (1994), ``The Sharpe ratio'', Journal of Portfolio Management, Vol. 21 No. 1, pp. 49-59.
Spaulding, D. (1997), Measuring Investment Performance: Calculating and Evaluating Investment
Risk and Return, McGraw-Hill, New York, NY.
Stanton, S., Hammer, M. and Power, B. (1992), ``From resistance to results: mastering the organizational
issues of reengineering'', Insights Quarterly: The Executive Journal of Business Reengineering,
Vol. 2 No. 2, pp. 6-16.
Stanton, T., Hammer, M. and Power, B. (1993), ``Reengineering getting everyone on board'', IT
Magazine, Vol. 25 No. 4, April, pp. 22-7.
Valimaki, J. and Tissari, T. (1997) ``Risk management focus in business reengineering initiatives'',
Proceeding of the IPMA Symposium on Project Management: Managing Risks in Projects,
E & FN Spon, London, pp. 233-42.
Zadeh, L. (1973), ``Outline of a new approach to the analysis of complex systems and decision
processes'', IEEE Transactions on Systems, Man, and Cybernetics, Vol. SMC-2, pp. 28-44.