Modular Production

- A New Concept
. . . which promises to piLt production executives
once again in top management, and to give consumers
wider ranges of choice among products.

By Martin K. Starr great variability — and, consequently, risk and

uncertainty — from this area of enterprise. On
In the course of its normal development, an the other hand, by succeeding in this fashion,
industrial society begins with an orientation that they have relegated to other areas, such as
is focused primarily on production. This is finance, research, and marketing, the real lever-
hardly surprising, since technological mastery age factors in top management decision making.
of a process lies at the root of incipient economic But now change is occurring once again.
progress. Witness, for example, the impact that Many enterprise managers are aware that some-
such process inventions as the printing press, thing new is creeping into the production sphere.
the spinning jenny, the mandrel lathe, and the For a \A'hile it seemed as if automation were its
open-hearth furnace lia\e had on society. Dur- name. At least the more journalistic spokesmen,
ing this stage, managers of the production func- sensing a dramatic change, felt obliged to play
tion find themselves actively participating in the on this theme. On careful examination, ho\^'-
topmost managerial aff'airs. ever, it becomes apparent that automation is
Then, for evident and fundamental reasons, only a tool ol' tliis new force — and even hardly
the situation can change. Production decisions, that in its present form.
though recognized as vital, cease to l)e critical.
They no longer need to be made at the higlicst
levels of the enterprise. Feasible production The New Capacity
alternatives appear much alike, offering little Many production managers are not prepared
choice. Production decisions become involved to understand what is going on. Some hardly
with increments of efficiency. even sense that they are in a truly different era.
This second stage has more or less character- The reasons for this are clear. There is no
ized the technologies and methodologies of pro- precedent to go by, and no means for an indi-
duction management for the past 25 years in vidual whose interest centers in one functional
the West. The production function during this area to recognize the existence of the trend.
period could be sensibly described as "reliable." The change that we are talking about can
Production decisions have become relatively pre- be briefly described as the consumer's demand
dictable. for maximum productive variety (or maximum
It is a tribute to the pioneers of production choice). To achieve this variety, what I call
management that they have been able to remove "modular" or "combinatorial" productive capaci-
131
132 HBR Nov.-Dec. 196s

"Modular production" is the name Martin K. Starr gives to a newly developing capacity to design
and manufacture parts which can be combined in the maximum number of ways. In this article he
deals with such aspects of modular production as the following:
• Its significance — The production manager and his function will once again assume crucial impor-
tance in top management planning and control.
• What is new — Modular production makes possible the manufacture of a far greater and much
more genuine variety of products.
• Development — The new concept will not come into being overnight, since it poses many practical
difficulties, but industry will surely see more and more of it.
• Implications for marketing — The new produetion capacity eomes at an opportune time, for con-
sumers are demanding more uniqueness and individuality in products.
— The Editors

ties — that is, capacities to design and manu- modify its own position in an environment that
facture parts which can be combined in numer- is undergoing change at an accelerating rate.
ous ways — are required, as well as compatible The problems that must be faced by the next
managerial abilities. generation of production managers are numer-
As has been suggested, the drive toward pro- ous. Once again production managers are likely
ductive variety is being forced by external fac- to be assigned top-management positions. All
tors. It is made possible because of internal industries will not be similarly affected; their
ones. An overview is needed for the trend to be adaptations can be expected to follow different
perceived. We can state that: timetables. How to speed up or slow down the
rate of change will be a subject of interest; but
1. The force for this ehange comes from the in the long run the forces for change are irresist-
marketplace. ible, and the degree of change will be recognized
2. The means for change resides in tlie produc- as having been predetermined and not under
tion management area, speeifieally in production's the manager's control.
use of —
Accordingly, I shall undertake in this article
. . . methodology derived from the management
to describe modular production and to point out
seienees;
where the effects of the developing combina-
. . . teehnology derived from the physical sei-
torial methods are likely to be felt most strongly.
enees ;
I shall try to focus on the kind of managerial
. . . data processing ability obtained from elec-
training that will be required in order to pro-
tronic computers.
vide some assurance of success in meeting new
In other words, the consumer is demanding challenges as they are raised. The evolution of
ever greater variety from which to choose. And equipment, alterations in organizational struc-
new methodologies and technological achieve- ture, changes in managerial methods — these
ments have developed within the production are some of the other issues I shall examine.
area which permit the consumer to force this
issue (albeit with varying degrees of success).
Nature of Aietamorphosis
Irresistible Forces The center of change resides in the produc-
The speed with which such change will oc- tion management area, which has long been a
cur is influenced by competitive factors. If stepchild of industrial drama. However, pro-
there is general management resistance, new duction management cannot swing the change
forms can be expected to grow at the expense by itself. For those companies that will evolve
of the reluctant. In any case, various align- successfully by maintaining or increasing their
ments of industrial power will arise according total share of a growing market, a new form of
to the ability of each organization to assess and effort which will cut across many functional
 HBR REPRINT SERVICE
Best Selling Reprint Series
Over the last ten years, the editors have grouped articles on a single topic which have
appeared in HBR and offered them as "packages" designed to help executives keep up to date
on the newest tools and techniques of decision making.
Out of the three dozen series now available, ten were in especially heavy demand during the
last year.
Believing that readers might like to have these "best sellers" called to their attention, the
Editors list them below.
CAPITAL INVESTMENT—15 articles on subjects ranging from STATISTICAL DECISION—these popular series present special
the evolution of investments to mathematical models and techniques, analytical approaches, applications to different
administrative procedures $5.00 business functions, and uses of quantitative analysis in policy
PLANNING—organization, steps, requirements, and making.
special
problems treated in detail in 16 analyses $5.50 Part I: 16 articles . $4.50
Part I I : 14 articles . $4.50
HUMAN RELATIONS—17 popular articles including such clas- Part I I I : 7 articles . $2.00
sics as "The Foreman: Master and Victim of Double
INVENTORY POLICY—5 analyses in depth of lot size, safety
Talk" $6.00
stocks, scheduling, reordering methods, forecasting, and
MARKETING PLANNING & STRATEGY—valuable lessons from, other leading problems $2.00
corporate experience and industrial research on marketing ORGANIZATIONAL DEVELOPMENT—what makes for strong,
philosophy, goals, policies, sales tactics, pricing, forecasting, flexible, growing organizations?
promotion, vital trends, and related topics. Part I: 17 articles . $6.00
Part I: 15 articles . $5.00 Part I I : 6 articles . $2.00
Part I I : 15 articles . $5.00 CONTROL—human problems with budgets, internal auditing,
Part III: 15 articles . $5.00 and general control policies. 15 articles . . . $5.00
EXECUTIVE DEVELOPMENT—covers on-the-job coaching, in- PRODUCTION MANAGEMENT—12 articles describing important
company programs, university training, and other aspects. new concepts in supervising, controlling, and planning
16 articles $5.50 operations $4.00
A complete list of available series is printed over.

As a service to its readers, the HARVARD BUSINESS REVIEW makes iWIHUII" 'reprints of oil articles published in

reprint this and previous issues. Postage is paid on orders shipped via parcel post or third class mail. If shipment via air or
other special means is requested, charges will be billed.

price These prices apply to total number of reprints (of the same or assorted articles) ordered at one time.
Single reprint $ 1.00 First 100 $ 39.70
Two reprints 1.30 Additional lOO's up to 1,000, each 20.00
schedule Three reprints
Four to 99, each
1.50
0.40
First 1,000
Additional lOO's over 1,000, each
215.00
15.00

REPRINT and SERIES ORDER FORM Ship this order to:

PLEASE PRINT
QUANTITY ISSUE YEAR AUTHOR OR SERIES PR/CE DATE

NAME

TITLE

COMPANY

STREET

CITY

STATE ZIP CODE

PAYMENT ENCLOSED f l B I L L COMPANY

 SPECIAL READERS MAY BE INTERESTED to know that many important management prob-
lems have been covered by series of articles. Made up of reprints of articles dealing
REPRINT with different aspects of a subject, each series is selected by the Editors to give
SERIES rounded coverage and a diversity of viewpoints:

Advertising Management—how to plan, organize, and measure Management Information—what are the unique gains, traps, and
the results of advertising are among the questions analyzed in methods of harnessing data processing that management should
this selection of 11 articles $3.50 know about? 9 articles $3.00
Businessmen & Government—trends in government, industry Money & Credit—5 probing analyses of different aspects of
relations with Washington, executives' political activities, and the Report of the Commission on Money and Credit $2.00
other topics are discussed. 11 articles . . . . $3.50 Organizational Development—what makes for strong, flexible,
Capital Investment—15 articles on subjects ranging from the growing organizations?
evolution of investments to mathematical models and admin- Part I: 17 articles . $6.00
istrative procedures . $5.00 Part I I : 6 articles . $2.00
Case Method—14 cases on different management problems, di- Philosophy of Business—what is the proper role of the manager in
gested into readable form, and with commentaries by business management and of management in society? 13 articles $4.50
authorities $4.50 Planning—organization, steps, requirements, and special problems
Consumer Motivation—13 articles on implications of behavioral treated in detail in 16 analyses $5.50
science research for public relations, product strategy, and Procurement—covers two areas rich in possibilities for cost sav-
marketing thinking $4.00 ings and increased efficiency: materials handling and purchas-
Control—human problems with budgets, internal auditing, and ing. 11 articles $3.50
general control policies. 15 articles $5.00 Production Management—12 articles describing important new
Decentralized Management—14 articles emphasizing profit re- concepts in supervising, controlling, and planning opera-
sponsibility as well as advanced intracompany pricing tech- tions $4.00
niques $4.50 Religion & Business—9 articles including the group that won
Executive Compensation—specific figures and trends in compen- the 1955 Citation of the Layman's Movement for a Christian
sation; how to decide salary levels, bonuses, and stock options. World $3.00
14 articles . . . . $4.50 R & D Management—examines a number of key issues, including
Executive Development—covers on-the-job coaching, in-com- research planning, technical probability, and organization. 14
pany programs, university training, and other aspects. 16 articles $4.50
articles $5.50 Retailing Strategy—what do the significant new trends in re-
Finance—analyzes corporate earnings, stockholder relations, small tailing mean for management? 8 articles . . . $2.50
business financing, new techniques, and other matters. Selling & Sales Management—14 articles combining factual
Part I: 17 articles . $6.00 and case data from the field with proposals for executive
Part I I : 11 articles . $3.50 action $4.50
Human Relations—17 popular articles including such classics as Small Business—these 8 articles probe a range of vital topics,
"The Foreman: Master and Victim of Double Talk" $6.00 from ownership and control to marketing management $2.50
Inventory Policy—5 analyses in depth of lot size, safety stocks, Statistical Decision—these popular series present special tech-
scheduling, reordering methods, forecasting, and other leading niques, analytical approaches, applications to different business
problems $2.00 functions, and uses of quantitative analysis in policy making.
Labor Relations—probing analyses of wildcat strikes, white-collar Part I: 16 articles . $4.50
unionism, featherbedding, and other key topics. 10 articles $3.50 Part I I : 14 articles . $4-50
Leasing—one of the fiercest management controversies in recent Part I I I : 7 articles . $2-00
years is highlighted in this group of 7 articles and 7 letters $2.00 Stock Options—with unrest about stock options growing both m
Marketing Planning & Strategy—valuable lessons from corporate and out of Congress, these analyses are more timely than ever
experience and industrial research on marketing philosophy, —5 articles, plus letters $2.00
goals, policies, sales tactics, pricing, forecasting, promotion, World Business—informative, helpful counsel for firms engaged
vital trends, and related topics. in overseas operations, covering trends, problems, strategies,
Part I : 15 articles . $5.00 and opportunities abroad.
Part I I : 15 articles . $5.00 Part I : 13 articles . $4-50
Part I I I : 15 articles . $5.00 Part I I : 14 articles . $4-50
These series can be obtained by using th< order card below. Simply list name of series.

FIRST CLASS

PERMIT No. 24756

Boston, Mass.

BUSrNESS REPLY MAIL

No postage stamp necessary if mailed in the United States

POSTAGE WILL BE PAID BY

Reprint Service
HARVARD BUSINESS REVIEW
Soldiers Field
Boston, Mass. 02163
 Modular Production 135
organizational areas is required. We can call output diversity. "Management science," as it is
this new orientation a "synthesis" to distinguish called, has resolved such basic problems as schedul-
it from "analysis," the euphemism which epito- ing work, providing adequate inventories, and ex-
mizes the traditional production management ercising quahty control, thus permitting splendid
approach. managerial control over an enormous variety of
products in the product mix.
Production managers are heing called on to
Significantly, management science has been eas-
find (within their potential set of tools and con- ily and directly assimilated by only one functional
cepts) the operational and economic means for area of business — production management. This
introducing real diversity in production output, observation has been made frequently; yet its rel-
using a given configuration of plant and equip- evance has been overlooked. Management science
ment. Real diversity means far more than the and tbe production function are natural partners.
illusion of an "adequate" product mix. It signi- And as management science unfolds, the produc-
fies the capability to produce a sequence of items tion function responds.
(or small batch of items), each of which can (3) Because of consumers' attraction to diversity,
differ markedly from preceding and subsequent the ability to produce a real variety oflFers the kind
items (or batches). To achieve differentiation, of high leverage that attracts top management at-
there will be small batches of units with sep- tention. The change represents a fundamental de-
arate setup charges. And yet, because of the parture from prior conditions in the sense that
technology and methodology employed, the total previously one company's ability to mass-produce
charges, including those for setup, can be sig- identical items could be eopied by other firms. But
nificantly less than the total revenue obtained for modular production there are no patterns avail-
able with which to model and manage the produc-
in the marketplace. tion system. A great deal of innovation and cre-
ativity is called for, and with these will be found
Spotlight on Production the inevitable partners, risk and uncertainty, and
It is not difficult to see why production man- real vulnerability to astute eompedtive practices.
agement decisions must be at the core of this
activity. Three reasons stand out: Control of Variability
(1) The essence of the depicted capability is Our understanding of the production of real
technological. The notion of high-volume, low- variety can be improved by taking a brief his-
cost, automated mass production will eventually torical excursion. In the early 1900's, students
give way to adaj^tive automation capable of prodLic- of production management began to learn how
ing a seqLience of unique outputs at no sacrifice of to combine many different input elements in a
volume and at no significant increase in cost. Al- transformation system so as to produce a stream
ready, in fact, this trend is in evidence. Thus: of output units that were sufficiently similar to
• In the automobile industry, a large and still be interchangeable in every important detail. In
increasing number of ordering options for a new effect, to acbieve mass production it had been
car are readily available to the consumer. necessary to learn how to remove or control the
• A major petroleum company has designed inherent variabilities of the inputs and the pro-
a gasoline pump that permits the consumer to cess. This accomplishment can be illustrated as
mix his own blend. shown in EXHIBIT I.
• Self-fitting clothing, such as stretch socks, Outputs from such a production system could
has a kind of built-in diversity. be mixed together in a bin and withdrawn in
any order for assembly with other parts that had
• Increases are occurring in the available va- been similarly produced. This gigantic step for-
riety of soap colors, type faces for typewriters,
and shades, styles, and scents for lipsticks, nail- ward in concept revolutionized the production
polish, perfumes, sun lotions, and other cos- process. In retrospect, the step seems to have
metics. been inevitable. For instance, while Eli Whit-
ney was developing the notion of interchange-
• There is increased diversity in the size and
able parts in the United States, Leblanc was al-
type of TV sets now available.
so making the concept operational in France —
In all of these cases, it is growth in technological but neither man was aware of the other's work.
capability that has permitted such diversity to occur. (It may be noted as well that the growth of
(2) The new methodology of production man- modular production partakes of the same sense
agement is capable of providing controls for high of inevitability.)
136 HBR Nov.-Dec.
EXHIBIT I. CONTROL WITH MASS PRODUCTION

Mass production requires control

(and reduction) of the variabilities
that are inherent in any set of inputs
(e.g., materials, workmanship) and
in the manufacturing process. In this
way, a stream of identical units of
output can be obtained.

The effective realization of the relationship the managerial ability to control the production
shown in EXHIBIT I was the result in no small of diverse outputs. So the marketing function,
measure of the efforts of such men as Frederick aimed at diversity, began where mass production
W. Taylor, Henry L. Gantt, and Frank Gil- (indicated by the single output line in EX-
breth. Few managers, no matter what their HIBIT II) left off, and its success is amply illus-
functional area, are unaware of the objectives trated by the important status of marketing in
and achievements of these early pioneers. Some- management in recent years.
what less well known, but totally appropriate to In general, marketing activities directed to-
this discussion, is the fact that when it was re- ward creating variety have represented an in-
alized that some variability in the outputs must creasing proportion of the total costs of an item.
always exist, Walter Shewhart and his colleagues The consumer has not been reluctant to accept
developed the function of statistical quality con- these services.
trol, whereby the level of fundamental or inher-
ent variability could be described and brought Increasing Expectations
under management direction. Initially, the marketing operation was most
As a final step, Henry Ford helped to com- effective when the diversity of production was
plete production's conquest over variability by minimal. Each firm established some unique
providing an operational instance of coordinat- qualities for its brands. The consumer was able
ed assembly for mass production. The result was to contrast these attributes with those of other
that the multiple inputs could be transformed brands. Differentiation was based on many
in both space and time to afford a controlled subtleties, not the least of which were the name,
stream of output. the package, and the image associations that
With this background in mind, let us turn were created. One basic chemical mixture could
from developments in the factory to trends in be positioned in the marketplace in sever-
the marketplace. al totally different ways. Even a single brand
might appear to possess quite diverse qualities
to different demographic segments of the popu-
Demand for Variety lation if the advertising and promotion could be
We know that the demand for variety origi- properly controlled.
nates at the consumer level. But how much But such differentiation was not to be enough.
variety does the consumer want? How much is Pressure exerted by the consumer for greater
he willing to pay for it? variety began to appear in several different
In the past, operating under a production forms. For instance, the consumer began to
configuration of the type illustrated in EXHIBIT question the "truth" of apparent variety. As a
I, marketing management stepped into the result, marketing, advertising, and promotion
breach and accepted the challenge to supply operations have come under the critical surveil-
the consumer with apparent variety. Production lance of consumer groups and the government.
facilities could not generate as much real vari- It appears, however, that disillusionment about
ety as the market was able to absorb, for the brand marketing, to the extent that it exists, has
technological capability was lacking, and so was not been connected with the purpose of market-
 Modular Production 137
EXHIBIT II. PRODUCT VARIETY WITH MASS PRODUCTION

Marketing management
supplies the consumer
with apparent variety
even though the pro-
duction output is based
on the concepts of
mass production.

ing or with the concept of apparent variety, but random affair but is based on some personal pre-
with the lack of a commensurate measure of real disposition. "Keeping up with the Joneses," if it
vai'iety. The consumer is simply asking for a has any meaning, must be interpreted as being ae-
more significant choice. eeptably different from tbe Joneses where the dif-
Pressure for variety also helps to explain the ference ean be explained in terms of a purely per-
sonal and individual philosophy.
severe contraction in product life that has been
experienced in recent years, and the unusually Production Meets the Need
high failure rate for many new products which
require acceptance by a large market segment Marketing managers have succeeded to some
for economic success, yet are not sufficiently extent in presenting the consumer with a choice
unique to warrant such acceptance. It is also so that he can exercise his individuality. In
significant that the demand for variety appears part, this has been achieved by stressing non-
to follow a cyclical pattern. The oscillations are functional factors of quality which can serve to
distinguish one product from another. But the
correlated with production capabilities and the
consumer's drive for uniqueness appears to be
various phases of technological maturity. The
insatiable. For instance, the novelty of Euro-
consumer's expectations appear to be entirely in
pean cars led to an unexpected degree of accept-
keeping with the state of the art:
ance by U.S. consumers. The capricious affair
e As technology advances, the first reaction of of the U.S. consumer with compact cars has
the marketplace is to adopt the prototype model been of tlie same type. And the great increase
of the new technology as widely as possible. Only in consumer spending for such services as trav-
minimum brand differentiation and almost no vari- el and entertainment can also be accounted for
ety within a single brand are found. The issue of as another way to further one's individuality.
primary imiJortance to the consumer is the owner-
ship of the car, radio, or other product. What makes all this so significant is the tim-
ing. The drive toward highly personalized pos-
e Then, as soon as the prototype models are sessions conies at a time when the production
broadly distributed among consumers, a shift be- manager begins to find himself in a position to
gins to take place. Various market segments ap- deliver this real diversity without violating rea-
pear, and tbe aceumulation process proceeds in sonable economic bounds. The degree of per-
terms of particular brand names and special model sonalization that is required is more than that
numbers. delivered by monogrammed luggage, shirts, ties,
e Finally, the marketplace begins to diseard the and automobiles. It is more than what "key
notion of "keeping up with tbe Joneses" and em- clubs," unique shapes and colors of telephones,
phasizes instead the uniqueness of the individual's or antique household furnishings can provide.
possessions.^ This trend is based on the fact that It is embodied in the growing success of mail-
tbe quality of difference bas value for the con- order variety, in the concept of "do it yourself"
sumer. Acceptable difference is not, of course, a (with a great many options to do it "in your own
' See Ernest Dicliter, "Discovering the 'Inner Jones' " way"), and in other iforms which we shall men-
(Thinking Ahead), HBR May-June 1965, p. 6. tion presently.
138 HBR Nov.-Dec. 1965
EXHIBIT III. CONCEPT OF MODULAR PRODUCTION

COMBINATORIAL OUTPUTS
Management designs, develops, and pro-
duces those parts which can be combined O
in the maximum number of ways.

TRANSFORMATION Additional mari<eting

INPUTS variety can be added
ASSEMBLY (PROCESS) CONFIGURATIONS to the real variety
achieved by produc-
o tion. (See EXHIBIT II.)

Beginnings of the Trend modular capabilities. One accent is on design;

a second is on the exercise of adequate mana-
The kind of timeliness that we observe here gerial controls.
is not unusual. There is, if anything, distinct There are numerous examples that can be
precedent for it — production has responded given of systems which have developed with a
to market demands before. Only the specifics tendency in this direction — jet engines, orig-
are different. inally designed for airplanes, being adapted for
use as integral components of power plants;
Elements of Change modular office wall panels and furniture per-
To appreciate the new trend, one can divide mitting a wide diversity of arrangements; and
the production system into two basic parts: the similar configuration adaptability of prefab-
ricated structures. But none of these examples
1. The transformation processes. can suffice to describe what would result from
2. Assembly operations. the total planning of a system of this kind; the
production manager can now deliver a much
Many different inputs (including materials, greater variety, at a reasonable cost, than his
skills, information, and power) enter the trans- present activities indicate.
formation process. They are combined in vari-
ous ways to produce a catalog of parts. It is the Retarding Factors
essence of the modular concept to design, de- At present, there are a number of difficulties
velop, and produce those parts which can be that stand in the way of the development of ad-
combined in the maximum number of ways. EX- vanced modular production:
HIBIT III is intended to present visually some of
(1) Many production managers are unaware of
the aspects of this statement. the market potential for highly diversified outputs.
This emphasis on maximizing the combina- Accordingly, they have not familiarized themselves
torial variety of assemblies from a given num- with the technological problems of combinatorial
ber of parts is the new element of interest to design whieh are involved.
production managers. It is what is meant by The reeent acquisition of new methodological
 Modular Production 139
abilities, however, is forcing production men to action until one's hand is forced — these atti-
examine implications for managerial control of tudes ^vill tend to be substituted for what is
richly diversified systems. And the enormous ma- ordinarily called entrepreneurial wisdom.
nipulative abilities of electronic computers may
reinforee the feeling that some new potential has Pressures for Change
entered tbe picture. In particular, recent develop-
ments in designing products by computer and tbe Meanwhile, however, there will he situations
use of visual input-output peripheral equipment where reluctance to move in the new direction
aids, sucb as those employed by tbe General Motors will seem absurd to businessmen involved. The
Corporation, have great significance for the achieve- totally new venture, the new or experimental
ment of sophisticated combinatorial variety in tbe autonomous branch, and industries that tradi-
design of modular parts. Tbese computer systems tionally are involved with substantial retooling
can store design components sucb as rectangles, — these are exemplary situations for the intro-
circles, force fields, and stress patterns. Indeed, a duction of modular capabilities. In time, such
great variety of component combinations can be instances may provide the momentum to force
tested by means of analytic and simulation pro- still other hands. It seems logical to suggest, as
cedures, eliminating tbe need to construct a proto-
type model for physical testing. well, that high-volume industries characterized
by short product life cycles will react early to
(2) Many marketing managers, though aware of the possibilities of modular production, and that
the market potential for highly diversified outputs, industries which characteristically are involved
are convinced that their traditional approach to tbe with expensive parts and components will not
consumer is sound. Witbout a complete revision overlook the potential benefits of modularity.
in perspective (which inertia prevents), tbe new Only in some ultimate sense can we expect
potential goals cannot be defined. And one eannot
measure how far short he falls from an imspecified that the overall industrial system of the present
goal. As a result, the demands of tbe marketplace will evolve to the point where it becomes truly
have not been translated into production terms. responsive to the selection process of the indi-
(Judging from all indications, researcb and de- vidual. What we can anticipate is tlie continued
velopment managers often suffer from tbe same fragmentation of the market to ever smaller
difficulty.) units that can be serviced with relatively more
unique outputs, with sov?e enterprising compa-
(3) From tbe viewj^oint of society, modular pro-
duction will bring with it significant eosts in tbe
nies capitalizing on the trend.
form of obsolescence. Obsolescence occurs not only
with respect to the present product line but al- Planning & Control
so with respect to present configurations of plant
and facilities. While it can be argued successfully At this point I shall depart from discussion
that changeover to modular production need not of general trends and problems in order to dem-
be total or swift, tbe fact remains that turning to onstrate in a more precise way what modular
tbe modular approach produces a great deal of production means to management. This demon-
unplanned obsolescence. In addition, tbe design
and engineering costs of entering into such produc- stration will allow us to draw some important
tion configurations can be exceptionally high. conclusions for those men entrusted with plan-
ning and control.
Management personnel must be trained to
cope \vith both the concepts (for planning) and Parts & Configurations
the realities (for operating). The background A better understanding of the nature of modu-
essential for this training Avill be the manage- lar capabilities for real productive variety can
ment sciences. In particular, the combinatorial be obtained by constructing a table of the kind
aspects of mathematical analysis will provide an sho^^'n in EXHIBIT IV. We see that, with N dif-
essential foundation — above and beyond which ferent kinds of parts, a total of M diffei-ent prod-
managerial intuitions can and will continue to uct configurations is derived. Some of these
be exercised. products require several units of a single part;
In view of the uncertainties, new risks, and for instance, PR4 requires two of PAo, and PRj
difficulties, it is predictable that many otherwise requires t\vo of PAj. The maximum possible
alert executives will tend to look the other way variety, i.e., maximum M, that can be obtained
when an issue pertaining to real productive va- with N different parts may be very large. Tbis
riety is raised. To wait and see, to postpone is especially so when we include both the differ-
 140 HBR Nov.-Dec. 1965
EXHIBIT IV. TABLE OF SPECIFICATIONS FOR THE of modular design might be, therefore, the num-
KIND AND NUMBER OF PARTS TO BE COMBINED ber of products (columns) that can be generated
FOR EACH OF A GIVEN NUMBER OF
DIFFERENT END PRODUCTS
from a given number of parts (rows).

Variety Variety of products Economic Model

of
parts PRi PH2 PR;1 PRi PRj PRM It is now possible to expand the notions that
PAi I 0 I I . . . 0 . . . 0 we have been developing into an economic
PA2 I
0 I 2 . . . 0 . . . 0
framework. As a convenient approximation, let
PAs 0 0 0 0 ... I . . . 0
PAi 0 0 T 0 . . . 0 . . . 0
us begin with the assumption that we can pro-
duce fixed but different amounts of each part
and that we have N such parts which can be
PA, 0 (3 ][ C5 ... 2 ... 0 assembled in various ways to yield a variety of
M products. If we know the costs of each part,
we can combine all of this information by means
PAN of a linear programming model to determine an
optimal product mix. The objective would be
NOTES: PAi denotes the part identified by the stock
number i. to maximize profits, subject to the constraints
PR) denotes the product variation listed in the imposed by the modular system of parts that has
finished goods catalog as j . (As shown been predesigned and by the productive capacity
above, the i^roduct j assembly requires one of the enterprise to produce these parts.
unit of part 3, two units of part i, and
one unit of part N. The sequence of as- Erom the fundamental theorem of linear pro-
sembly is not indieated.) gramming we know that the number of parts
PAN denotes the last part listed in our table. (as we have defined them) plus the number of
PRM denotes the last product variation listed any other resource constraints (such as storage
in our table; in this case, it is the product space, limited power, or scarce raw materials)
using only one unit of part N.
will be equal to the maximum number of prod-
ucts that can be included in the ultimate com-
ent possible combinations of the parts and vary- position of the product mix. But this rule pro-
ing numbers of them in combination. Thus: duces the somewhat absurd result that the prod-
If each part can appear zero, one, or two times uct mix will be only slightly larger in size than
(i.e., three options), then the maximum variety that the total number of parts that have been de-
can he derived with N parts will he 3^ — i (one is signed by the company. Such a result wovild
subtracted hecause there will he one configuration defeat the purpose and concept of modular pro-
in which none of the parts appears). Considering ductive capability.
only five parts (N = 5), the maximtmi variety (de-
fined in this way) will he 242.
What, then, can account for the fact that
With ten parts, the maximum variety rises to a
many more product configurations than are in-
somewhat astonishing 59,048. dicated by the far smaller number of parts
If, in addition, the ways and sequences with would, in fact, be included in the ultimate com-
which the parts can he comhined also add variety, position of the product mix? The answer must
then the numher of possihilities rises even faster. lie in the demand of the marketplace for unique-
ness. Once we have accepted the fact that the
Of course, many of the theoretical possibili- revenue curves for each product configuration
ties cannot exist and would have no appeal as a are not linear but fall off as the uniqueness of
product choice for any consumer. But the basic the product class diminishes, we have a satisfac-
idea of modular design is to have an inventory tory explanation for expanding the product mix.
of parts which can partake in many appealing EXHIBIT V compares a linear revenue curve
product configurations. Each order received by with one that decreases as the uniqueness of the
the company is translated into a unique assembly product class in the marketplace diminishes.
configuration. (The same sort of table as shown The linear form represents complete consumer
in EXHIBIT IV must have been designed long ago indifference to the amount and scope of product
for the first Erector set. It appears that indus- distribution. The nonlinear curve of diminish-
try is about to take seriously a fundamental ing returns, on the other hand, indicates de-
principle long embodied in many children's creasing consumer responsiveness as the extent
toys.) A useful measure of the effective degree of distribution and ownership of a particular
 Modular Production 141
product configuration inci'eases. (The decline in EXHIBIT V. LINEAR AND NONLINEAR
responsiveness may be hastened by repeated REVENUE CURVES
purchases over time.)
CUSTOMERS ARE INDIFFERENT TO
Optimal Product Mix EXTENT OF PRODUCT DISTRIBUTION

The saturation phenomenon is well known

in marketing. It is usually based on the notion
of an exhaustion of "prime" targets. I am sug-
gesting that the size of the potential market
diminishes at a faster rate than can be ac-
counted for by exhaustion alone, or, in other
words, that the acquisition of a new customer
will remove more than one potential custom-
er from the marketplace. The reason is that -a
every customer added increases the consumer's 8
Q.
disenchantment with what he conceives to be ^_
O
boring, commonplace, or ordinary.
When a nonlinear form of the kind shown
in EXHIBIT V is encountered, the effect with s
respect to a linear program is to increase the
number of constraints. As a result, the size of /
the optimal product mix increases. Under such Decreasing uniqueness of product ^
conditions, and even with relatively few parts, as number of units owned or available increases
a great number of product configurations will least as significant as those which followed from
be required in order to maximize revenue. The the general adoption of the concept of inter-
more demanding the consumer is for uniqueness changeable parts.
in his possessions, the greater the curvature of
the revenue lines. And in turn, as the curva-
ture of a revenue line increases, the indicated Meeting the Challenge
size of the optimal product mix grows larger.
The need and the ability to provide real diver-
sity will certainly ^'ary in many -ways. However,
Interchangeable Modules management can anticipate some of the impor-
In terms of EXHIBIT IV, the traditional con- tant factors that are likely to operate to the ad-
cept of interchangeable parts can be interpreted vantage or disadvantage of a company, and plan
as follows: All units that belong to any particu- accordingly.
lar parts classification can be treated as being
identical. It does not matter whether a given Changing Opportunities
unit of a certain family of parts was made at Some of the more likely factors of differenti-
the beginning or end of a production run. Any ation will be discernible in terms of the charac-
member of this family of parts can combine with teristics of industrial sectors, demographic seg-
equal ease with any member of another family mentation, competitive structures, organization-
of parts in the row. Note that this interchange- al flexibility, executive attitudes, and a variety
ability exists within the rows of the table and not of exogenous factors which are imposed by the
between the rows of the table. economy and international trade. Since starting
Interchangcahle part modules, on the other ventures will possess advantages over going in-
hand, are designed so as to be highly transfer- terests with established production systems, the
able between columns; i.e., an interchangeable latter may find themselves being pushed at un-
module enters with equal ease into many dif- expected times and in unanticipated ways.
ferent product configurations. In contrast, pro- Organizations seeking to diversify will en-
duction as it is conceived today tends to isolate counter new opportunities to hranch out. Cre-
the columns. ating real productive variety — that which
Once the concept of interchangeable modules comes with modular production — is equivalent
becomes widely accepted, it may well be respon- to building into the system preplanned poten-
sible for developments that are astounding or at tials for diversification. In a related sense, ac-
 142 HBR Nov.-Dec. 1965
tivities such as long-range planning and master ers. These peripheral equipments permit almost
planning necessarily will span new horizons and instantaneous assessment of a designer's work.
assume new dimensions. Complex design constructions and evaluations
The marketplace will cease to be treated as a that formerly required many man-weeks of anal-
statistical phenomenon that can be described by ysis will be achieved so rapidly that a new sense
gas-law models, analogs of random walks, quan- of creativity will develop.
tum mechanics, or other such models used for de-
riving average responses and occasionally for Efficiency in Distribution
obtaining measures of variability. Instead, there The distribution problem which arises as a
will be a growth of Interest in the behavior of result of increased variety is certain to be enor-
the individual. The marketplace will be most mous. Here, too, the role of the computer is like-
appropriately described as the sum of individual ly to be crucial. The rapid communication of
behaviors. individual consumer demands to the production
The concept of the highly segmented market control units, the translation of these demands
will be found to apply quite well to common- into specifications appropriate for the transfor-
market type operations, where consumer diver- mation and assembly processes, and the distri-
sity occurs because of the structural variants in bution of production outputs to the appropriate
the environment (e.g., broad cultural differences consumers — all of these steps appear to be
between consumer segments, and significant pro- feasible because of new insights and abilities
duction variations caused by price and wage and the computer.
differentials, previously established underwriter
codes, ancl separate systems of national taxation). Organizational Steps
It is apparent that we are dealing with a situ-
Innovation in Production ation that will involve fundamental changes in
The essential ingredients of change for the the enterprise. What is required is the ability
production manager are twofold. On the one to manage a new kind of productivity. An ap-
hand, a veritable revolution in design ability is propriate organizational structure would provide
required in order to move in the direction of greater responsiveness to the market. In such
interchangeable modules. The roots of such ef- an organization, production would be in touch
forts already exist in the well-developed concept with consumers, with the contact mediated by
of standard parts (e.g., screw threads, gauges, the marketing function. An appropriate organi-
and light bulbs) and in standard process opera- zation would also allow the production manager
tions (e.g., synthetic time standards and com- to respond with sensitive perception to develop-
puter-controlled equipment). On the other hand, ing technologies. To achieve such results, a
it will he recognized that adaptive automation much higher level of functional integration is
rather than a fixed-output type is the ideal to called for. It is also quite clear that production
be desired. However, the primary significance management should once again participate in
of adaptive automation will be for assembly op- top management decision making.
erations and only secondarily for the transfor- Simulation methods can be expected to play
mation processes. a crucial role in the development of an appropri-
The use of the computer seems to be the key. ate organization. They will provide necessary
It will l)e a major asset when it comes to coping information, as well as assurance concerning
with the production of real variety. From an- the effect of proposed changes in relationships
other point of view, it is also a force for change. and procedures. Also, when management simu-
Wlien employed in conjunction with appropri- lates the design of modular parts, their assem-
ate models from the management sciences, such bly into many unique products, the managerial
as sequencing models, programming methods, control problems that result, and the distribution
line-])alancing and queuing models, and heuris- problems created by a highly segmented market,
tic procedures of many kinds, a new level of it can examine the important characteristics ot
managerial control of diversity can be achieved. the total system. Thus, simulation provides a
Radically different design techniques can also rational means for analyzing an enormously
IJC expected. Modular design is a realistic objec- complex system whose diversity promises not less
tive in the light of visual-aid inputs to comput- but more satisfaction born of less conformity-
Harvard Business Review Notice of Use Restrictions, May 2009

Harvard Business Review and Harvard Business Publishing Newsletter content on EBSCOhost is licensed for
the private individual use of authorized EBSCOhost users. It is not intended for use as assigned course material
in academic institutions nor as corporate learning or training materials in businesses. Academic licensees may
not use this content in electronic reserves, electronic course packs, persistent linking from syllabi or by any
other means of incorporating the content into course resources. Business licensees may not host this content on
learning management systems or use persistent linking or other means to incorporate the content into learning
management systems. Harvard Business Publishing will be pleased to grant permission to make this content
available through such means. For rates and permission, contact [email protected].