Competition and Business Strategy

The President and Fellows of Harvard College
Competition and Business Strategy in Historical Perspective

Author(s): Pankaj Ghemawat
Source: The Business History Review, Vol. 76, No. 1 (Spring, 2002), pp. 37-74
Published by: The President and Fellows of Harvard College
Stable URL: http://www.jstor.org/stable/4127751 .
Accessed: 16/01/2015 09:55
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp
.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact [email protected].
The President and Fellows of Harvard College is collaborating with JSTOR to digitize, preserve and extend
access to The Business History Review.
http://www.jstor.org
This content downloaded from 128.135.12.127 on Fri, 16 Jan 2015 09:55:51 AM

All use subject to JSTOR Terms and Conditions
Pankaj Ghemawat
Competition and Business Strategy in

Historical Perspective
A review of theories of competition and business strategy over
the last half-century reveals a fairly linear development of
early work by academics and consultants into efforts to
understand the determinants of industry profitability and
competitive position and, more recently, to add a time or his-
torical dimension to the analysis. The possible implications of
the emergence of a market for such ideas are also discussed.
"Strategy" is a term that can be traced back to the ancient Greeks, for
whom it meant a chief magistrate or a military commander in
chief. The use of the term in business, however, dates only to the twen-
tieth century, and its use in a self-consciously competitive context is
even more recent.
After providing some historical background, this essay focuses on
how the evolution of ideas about business strategy was influenced by
competitive thinking in the second half of the twentieth century. The
review aims not to be comprehensive but, instead, to focus on some key
topical issues in applying competitive thinking to business strategy.
Particular attention is paid to the role of three institutions --Harvard
Business School and two consulting firms, the Boston Consulting
Group and McKinsey & Company -in looking at the historical develop-
ment and diffusion of theories of business competition and strategy.
The essay concludes with some discussion of how the emergence of a
market for ideas in this broad domain is likely to affect future develop-
ments in this area.
PANKAJGHEMAWATis the Jaime and Josefina Chua Tiampo Professor of Business Ad-
ministration at HarvardBusiness School.
The author has drawn upon an earlier draft prepared by Dr. Peter Botticelli under his su-
pervision and has also benefited from helpful comments by Walter A. Friedman, Thomas K.
McCraw,and three referees.
Business History Review 76 (Spring 2002): 37-74. ? 2002 by The President
and Fellows of HarvardCollege.

PankajGhemawat / 38
Historical Background
Until the nineteenth century,the scope for applying(imperfectly)
competitive thinking to business situations appeared to be limited:
intense competitionhad emerged in many lines of business, but indi-
vidual firms apparentlyoften lacked the potential to have much of an
influenceon competitiveoutcomes.Instead,in most lines of business -
with the exception of a few commodities in which internationaltrade
had developed--firms had an incentive to remain small and to employ
as little fixed capital as possible. It was in this era that Adam Smith
penned his famous descriptionof marketforces as an "invisiblehand"
that was largelybeyondthe controlof individualfirms.
The scope for strategyas a way to control marketforces and shape
the competitive environmentstarted to become clearer in the second
half of the nineteenth century. In the United States, the building of
the railroads after 1850 led to the development of mass markets for
the first time. Along with improved access to capital and credit, mass
markets encouraged large-scale investment to exploit economies of
scale in production and economies of scope in distribution. In some
industries, Adam Smith's "invisible hand" was gradually tamed by
what the historian Alfred D. Chandler Jr. has termed the "visible
hand"of professionalmanagers.By the late nineteenth century,a new
type of firm began to emerge, first in the United States and then in
Europe:the vertically integrated, multidivisional (or "M-form")cor-
poration that made large investments in manufacturing and mar-
keting and in management hierarchies to coordinate those func-
tions. Overtime, the largest M-form companies managed to alter the
competitive environment within their industries and even across in-
dustrylines.'
The need for a formal approachto corporatestrategywas first ar-
ticulated by top executives of M-form corporations. Alfred Sloan
(chief executive of General Motors from 1923 to 1946) devised a strat-
egy that was explicitly based on the perceived strengths and weak-
nesses of its competitor, Ford.2In the 1930s, Chester Barnard,a top
executive with AT&T,argued that managers should pay especially
close attention to "strategicfactors,"which depend on "personalor
organizationalaction."3
'Alfred D. Chandler Jr., Strategy and Structure (Cambridge, Mass., 1963) and Scale and
Scope (Cambridge,Mass., 1990).
2See Alfred P. Sloan Jr., My Years with General Motors (New York, 1963).
3Chester I. Barnard, The Functions of the Executive (Cambridge, Mass., 1968; first pub-
lished 1938), 204-5.

Competition and Business Strategy / 39
The organizationalchallengesinvolvedin WorldWarII were a vital

stimulus to strategic thinking. The problem of allocating scarce re-
sources across the entire economy in wartimeled to many innovations
in managementscience. New operations-researchtechniques (e.g., lin-
ear programming)were devised, which paved the way for the use of
quantitativeanalysis in formal strategic planning. In 1944, John von
Neumann and Oskar Morgensternpublished their classic work, The
Theoryof Gamesand EconomicBehavior. This workessentiallysolved
the problemof zero-sumgames (most militaryones, from an aggregate
perspective)and framed the issues surroundingnon-zero-sum games
(most business ones). Also, the concept of "learningcurves"became an
increasinglyimportanttool for planning. The learning curve was first
discovered in the military aircraft industry in the 192os and 1930s,
whereit was noticedthat directlaborcosts tended to decreaseby a con-
stant percentageas the cumulativequantity of aircraftproduceddou-
bled. Learningeffectsfiguredprominentlyin wartimeproductionplan-
ning efforts.
WorldWar II also encouragedthe mindset of using formal strate-
gic thinking to guide managementdecisions. Thus, Peter Druckerar-
gued that "managementis not just passive, adaptivebehavior;it means
taking action to make the desired results come to pass."He noted that
economic theory had long treated markets as impersonal forces, be-
yond the controlof individualentrepreneursand organizations.But, in
the age of M-form corporations,managing "impliesresponsibilityfor
attemptingto shape the economicenvironment,for planning,initiating
and carryingthroughchanges in that economic environment,for con-
stantlypushingback the limitations of economic circumstanceson the
enterprise'sfreedom of action."4This insight became the rationalefor
business strategy-that, by consciouslyusing formal planning, a com-
pany could exert some positive controlover marketforces.
However,these insights on the nature of strategylargelylay fallow
for the decade after WorldWar II because wartime destructionled to
excess demand, which limited competition as firms rushed to expand
capacity. Given the enormousjob of rebuilding Europe and much of
Asia, it was not until the late 195os and 196os that manylarge multina-
tional corporationswere forcedto considerglobal competitionas a fac-
tor in planning.In addition,the wartimedisruptionof foreign multina-
tionals enabled U.S. companies to profit from the postwar boom
without effectivecompetitorsin many industries.
A more direct bridge to the developmentof strategic concepts for
business applicationswas providedby interservicecompetition in the
4 Peter Drucker, The Practice of Management (New York, 1954), 11.

Pankaj Ghemawat / 40
U.S. militaryafterWorldWarII. In this period,Americanmilitarylead-

ers found themselves debatingthe arrangementsthat would best pro-
tect legitimate competitionbetween military services while maintain-
ing the needed integration of strategic and tactical planning. Many
arguedthat the Army,Navy,Marines,and Air Forcewould be more ef-
ficient if they were unified into a single organization.As the debate
raged, Philip Selznick,a sociologist, noted that the Navy Department
"emergedas the defenderof subtle institutionalvalues and tried many
times to formulate the distinctive characteristicsof the various ser-
vices." In essence, the "Navyspokesmen attempted to distinguish be-
tween the Armyas a 'manpower'organizationand the Navy as a finely
adjustedsystem of technical,engineeringskills--a 'machine-centered'
organization.Facedwith what it perceivedas a mortalthreat,the Navy
became highly self-conscious about its distinctive competence."5The
concept of "distinctivecompetence"had great resonance for strategic
management,as we will see next.
Academic Underpinnings
The Second IndustrialRevolutionwitnessed the foundingof many
elite business schools in the United States,beginningwith the Wharton
School in 1881. HarvardBusiness School, founded in 1908, was one of
the first to promote the idea that managersshould be trained to think
strategicallyand not just to act as functionaladministrators.Beginning
in 1912, Harvardoffered a required second-year course in "business
policy,"which was designedto integratethe knowledgegained in func-
tional areas like accounting, operations, and finance, thereby giving
students a broaderperspectiveon the strategicproblemsfaced by cor-
porate executives. A course description from 1917 claimed that "an
analysis of any business problem shows not only its relation to other
problems in the same group, but also the intimate connection of
groups. Few problems in business are purely intra-departmental."It
was also stipulatedthat the policies of each departmentmust maintain
a "balancein accordwith the underlyingpolicies of the business as a
whole."'
In the early 1950s, two professors of business policy at Harvard,
GeorgeAlbertSmith Jr. and C. RolandChristensen,taught students to
question whether a firm's strategy matched its competitive environ-
ment. In readingcases, studentswere instructedto ask:do a company's
5 Philip Selznick, Leadership in Administration (Evanston, Ill., 1957), 49-50.

6 Official Register of Harvard University, 29 Mar. 1917, 42-3.

policies "fittogether into a programthat effectivelymeets the require-

ments of the competitivesituation"?7Studentswere told to addressthis
problemby asking:"Howis the whole industrydoing?Is it growingand
expanding?Or is it static; or declining?"Then, having "sizedup" the
competitiveenvironment,the student was to ask: "Onwhat basis must
any one companycompete with the others in this particularindustry?
At what kinds of things does it have to be especiallycompetent,in order
to compete?"8
In the late 1950s, anotherHarvardbusiness policy professor,Ken-
neth Andrews,built on this thinkingby arguingthat "everybusiness or-
ganization, every subunit of organization,and even every individual
[oughtto] have a clearlydefinedset of purposesor goals which keeps it
moving in a deliberatelychosen direction and prevents its drifting in
undesireddirections"(emphasisadded).As shown in the case of Alfred
Sloanat GeneralMotors,"theprimaryfunctionof the generalmanager,
over time, is supervisionof the continuous process of determiningthe
natureof the enterpriseand setting, revisingand attemptingto achieve
its goals."9 The motivationfor these conclusionswas suppliedby an in-
dustrynote and companycases that Andrewspreparedon Swiss watch-
makers,which uncoveredsignificantdifferencesin performanceasso-
ciated with their respectivestrategiesfor competingin that industry.'o
This format of combining industry notes with company cases, which
had been initiatedat HarvardBusiness Schoolby a professorof manu-
facturing,John MacLean,became the norm in Harvard'sbusiness pol-
icy course. In practice,an industrynote was often followedby multiple
cases on one or several companies with the objective, inter alia, of
economizingon students'preparationtime.'
By the 196os, classroomdiscussions in the business policy course
focused on matching a company's"strengths"and "weaknesses"--its
distinctive competence--with the "opportunities"and "threats"(or
risks)it facedin the marketplace.This framework,which came to be re-
ferredto by the acronymSWOT,was a majorstep forwardin bringing
explicitlycompetitivethinking to bear on questions of strategy. Ken-
neth Andrewsput these elements together in a way that became par-
ticularly well known. (See Figure 1.) In 1963, a business policy confer-
7George Albert Smith Jr. and C. Roland Christensen, Suggestions to Instructors on Pol-
icy Formulation (Chicago, 1951), 3-4.
8George Albert Smith Jr., Policy Formulation and Administration (Chicago, 1951), 14.
9Kenneth R. Andrews, The Concept of Corporate Strategy (Homewood, Ill., 1971), 23.
'oSee Part I of Edmund P. Learned, C. Roland Christensen, and Kenneth Andrews, Prob-
lems of General Management (Homewood, Ill., 1961).
1 Interview with Kenneth Andrews, 2 Apr. 1997.

Environmental Distinctive
Conditions Competence
and Trends
Capabilities:
Economic Financial
Technical Managerial
Physical Functional
Political Organizational
Social Reputation
History
Community
Nation
World
Corporate
Resources
Opportunities
and Risks As extendingor
Considerationof constraining
Identification all combinations opportunity
ssessment of Risk Identificationof

strengths and
weaknesses
Programsfor
increasing
capability
Evaluationto determine
best match of
ad resources
opportunit
Choice of Products
and Markets
Economic Strategy
Figure 1. Andrews's Strategy Framework.(Source: Kenneth Andrews, The Concept of Corpo-

rate Strategy, rev. ed. [Homewood, Ill., 1980], 69.)
ence was held at Harvardthat helped diffuse the SWOTconcept in

academiaand in managementpractice.Attendancewas heavy, and yet
the popularityof SWOT--whichwas still used by many firms, includ-
ing Wal-Mart,in the 199os- did not bring closure to the problem of
actually defining a firm's distinctive competence. To solve this prob-
lem, strategistshad to decide which aspects of the firmwere "enduring
and unchangingover relativelylong periods of time" and which were
"necessarilymore responsive to changes in the marketplaceand the
pressures of other environmentalforces."This distinction was crucial
because "thestrategic decision is concernedwith the long-term devel-

opment of the enterprise"(emphasis added).'2When strategychoices

were analyzedfrom a long-range perspective,the idea of "distinctive
competence"took on addedimportancebecause of the risksinvolvedin
most long-run investments.Thus, if the opportunitiesa firm was pur-
suing appeared"to outrun [its] present distinctive competence,"then
the strategisthad to consider a firm's "willingnessto gamble that the
lattercan be built up to the requiredlevel."'3
The debateover a firm's"willingnessto gamble"its distinctivecom-
petence in pursuit of opportunitycontinued in the 196os, fueled by a
boomingstockmarketand corporatestrategiesthat were heavilygeared
towardgrowthand diversification.In a classic 1960 article,"Marketing
Myopia,"TheodoreLevittwas sharplycritical of firms that seemed to
focus too much on deliveringa product,presumablybased on its dis-
tinctive competence, rather than consciously serving the customer.
Levittthus arguedthat when companiesfail, "itusually means that the
productfails to adaptto the constantlychangingpatternsof consumer
needs and tastes, to new and modified marketing institutions and
practices,or to productdevelopmentsin complementaryindustries."'4
However,anotherleadingstrategist,IgorAnsoff,arguedthat Levitt
was asking companies to take unnecessary risks by investing in new
productsthat might not fit the firm'sdistinctivecompetence.Ansoff ar-
gued that a company should first ask whether a new product had a
"commonthread"with its existing products. He defined the common
thread as a firm's "mission"or its commitment to exploit an existing
need in the marketas a whole.'5Ansoff noted that "sometimesthe cus-
tomer is erroneouslyidentifiedas the common thread of a firm'sbusi-
ness. In reality,a given type of customerwill frequentlyhave a range of
unrelatedproductmissions or needs."'6Thus, for a firm to maintainits
strategic focus, Ansoff suggested certain categories for defining the
common thread in its business/corporatestrategy.(See Figurea2.)An-
soff and others also focused on translating the logic of the SWOT
frameworkinto a series of concrete questions that needed to be an-
sweredin the developmentof strategies."7
In the 196os, diversificationand technologicalchanges increased
the complexityof the strategicsituations that many companies faced,
and intensifiedtheir need for more sophisticatedmeasures that could
12
Andrews, The Concept of Corporate Strategy, 29.
13Ibid., too.
14Theodore Levitt, "MarketingMyopia,"Harvard Business Review (July/Aug. 1960): 52.
15Igor Ansoff, Corporate Strategy (New York, 1965), o16-9.
16Ibid., 105-8.
'~7Michael E. Porter, "IndustrialOrganization and the Evolution of Concepts for Strategic
Planning," in T. H. Naylor, ed., Corporate Strategy (New York, 1982), 184.

Present Product New Product
Present Mission MarketPenetration Product Development
New Mission MarketDevelopment Diversification
Figure 2. Ansoffs Product/Mission Matrix as adapted by Henry Mintzberg. (Source: Henry

Mintzberg, "GenericStrategies,"in Advances in Strategic Management, vol. 5 [Greenwich,
Conn., 1988], 2. For the original, see Igor Ansoff, Corporate Strategy [New York, 1965], 128.)
be used to evaluate and compare many different types of businesses.

Since business policy groups at Harvard and elsewhere remained
stronglywedded to the idea that strategiescould only be analyzedon a
case-by-casebasis in orderto accountfor the unique characteristicsof
every business, corporationsturned elsewhere to satisfy their craving
for standardizedapproachesto strategymaking.'8A study by the Stan-
ford ResearchInstitute indicatedthat a majorityof large U.S. compa-
nies had set up formalplanningdepartmentsby 1963.19Some of these
internal efforts were quite elaborate. General Electric (GE) is a bell-
wetherexample:it used Harvardfacultyextensivelyin its executiveed-
ucation programs,but it also independently developed an elaborate,
computer-based"ProfitabilityOptimizationModel"(PROM)in the first
half of the 196os that appearedto explain a significantfraction of the
variation in the return on investment afforded by its various busi-
nesses.20 Over time, like many other companies, GE also sought the
help of private consulting firms. While consultants made important
contributionsin many areas, such as planning, forecasting, logistics,
and long-rangeresearchand development(R&D),the followingsection
tracestheir earlyimpacton mainstreamstrategicthinking.
The Rise of Strategy Consultants

The 196os and early 1970s witnessed the rise of a number of strategy
consulting practices. In particular, the Boston Consulting Group
'8Adam M. Brandenburger,Michael E. Porter, and Nicolaj Siggelkow, "Competitionand

Strategy: The Emergence of a Field," paper presented at McArthur Symposium, Harvard
Business School, 9 Oct. 1996, 3-4.
19 Stanford Research Institute, Planning in Business (Menlo Park, 1963).
20 Sidney E. Schoeffler, Robert D. Buzzell, and Donald F. Heany, "Impact of Strategic
Planning on Profit Performance,"Harvard Business Review (Mar./Apr. 1974): 139.

(BCG),founded in 1963, had a major impact on the field by applying

quantitativeresearchto problems of business and corporatestrategy.
BCG'sfounder, BruceHenderson,believed that a consultant'sjob was
to find "meaningfulquantitativerelationships"between a companyand
its chosen markets.21In his words,"goodstrategymust be based prima-
rily on logic, not. .. on experiencederived from intuition."22Indeed,
Hendersonwas utterlyconvincedthat economictheorywould someday
lead to a set of universalrules for strategy.As he explained,"[I]nmost
firms strategytends to be intuitiveand based upon traditionalpatterns
of behaviorwhich have been successful in the past. ... [However,]in
growthindustriesor in a changingenvironment,this kind of strategyis
rarelyadequate.The acceleratingrate of changeis producinga business
world in which customarymanagerialhabits and organizationare in-
creasinglyinadequate."23
In orderto help executivesmake effectivestrategicdecisions, BCG
drew on the existing knowledgebase in academia:one of its first em-
ployees, SeymourTilles, was formerlya lecturerin Harvard'sbusiness
policy course. However,it also struckoff in a new directionthat Bruce
Hendersonis said to have describedas "thebusiness of selling power-
ful oversimplifications."24 In fact, BCGcame to be known as a "strat-
egy boutique"because its business was largelybased, directlyor indi-
rectly, on a single concept: the experience curve (discussed below).
The value of using a single concept came from the fact that "in nearly
all problem solving there is a universe of alternativechoices, most of
which must be discarded without more than cursory attention."
Hence, some "frameof referenceis needed to screen the ... relevance
of data, methodology,and implicit value judgments"involved in any
strategydecision. Giventhat decision makingis necessarilya complex
process,the most useful "frameof referenceis the concept. Conceptual
thinking is the skeleton or the frameworkon which all other choices
are sorted out."25
BCGand the ExperienceCurve. BCGfirst developedits version of
the learning curve-what it labeled the "experiencecurve"-in 1965-
66. Accordingto BruceHenderson,"itwas developedto try to explain
price and competitive behavior in the extremely fast growing seg-
ments"of industries for clients like Texas Instrumentsand Black and
21Interviewwith Seymour Tilles, 24 Oct. 1996. Tilles credits Henderson for recognizing
the competitiveness of Japanese industry at a time, in the late 196os, when few Americans be-
lieved that Japan or any other countrycould compete successfully against American industry.
22
Bruce Henderson, The Logic of Business Strategy (Cambridge,Mass., 1984), 1o.
23BruceD. Henderson, Henderson on Corporate Strategy (Cambridge,Mass., 1979), 6-7.
24Interview with Seymour Tilles, 24 Oct. 1996.
25Henderson, Henderson on Corporate Strategy, 41.

Decker.26As BCGconsultantsstudied these industries, they naturally

asked why "one competitoroutperformsanother (assuming compara-
ble managementskills and resources)?Are there basic rules for suc-
cess? There, indeed, appearto be rules for success, and they relate to
the impact of accumulatedexperience on competitors'costs, industry
prices and the interrelationbetweenthe two."27
The firm's standard claim for the experience curve was that for
each cumulativedoubling of experience, total costs would decline by
roughly 20 to 30 percent due to economies of scale, organizational
learning,and technologicalinnovation.The strategicimplicationof the
experiencecurve, accordingto BCG,was that for a given product seg-
ment, "theproducer... who has made the most units should have the
lowest costs and the highest profits."28sBruce Hendersonclaimed that
with the experience curve "the stability of competitive relationships
should be predictable,the value of marketshare change should be cal-
culable,[and] the effectsof growthrate should [also] be calculable."29
From the Experience Curve to Portfolio Analysis. By the early
1970s, the experiencecurvehad led to another"powerfuloversimplifi-
cation"by BCG:the "Growth-Share Matrix,"which was the first use of
what came to be knownas "portfolioanalysis."(See Figure3.) The idea
was that after experiencecurves were drawn for each of a diversified
company'sbusiness units, their relative potential as areas for invest-
ment couldbe comparedby plottingthem on the grid.
BCG'sbasic strategy recommendationwas to maintain a balance
between "cashcows" (i.e., mature businesses) and "stars,"while allo-
cating some resourcesto feed "questionmarks,"which were potential
stars. "Dogs"were to be sold off. In more sophisticatedlanguage,a BCG
vice presidentexplainedthat "sincethe producerwith the largeststable
marketshare eventuallyhas the lowest costs and greatestprofits,it be-
comes vital to have a dominant market share in as many products as
possible. However, market share in slowly growing products can be
gained only by reducingthe share of competitorswho are likelyto fight
back."If a product market is growing rapidly, "a company can gain
share by securing most of the growth. Thus, while competitorsgrow,
26 Bruce Henderson explained that, unlike earlier versions of the "learningcurve," BCG's
experience curve "encompasses all costs (including capital, administrative, research and
marketing) and traces them through technological displacement and product evolution. It is
also based on cash flow rates, not accounting allocation." Bruce D. Henderson, preface to
Boston Consulting Group, Perspectives on Experience (Boston, 1972; first published 1968).
27 Boston Consulting Group,Perspectives on Experience, 7.
28 Patrick Conley, "Experience Curves as a Planning Tool," in Boston Consulting Group
pamphlet (1970): 15.
29 Bruce Henderson, preface, Boston Consulting Group,Perspectives on Experience.

High Share Low Share
"Star" "Question
Figure 3. BCG'sGrowth-ShareMatrix. (Source: Mark"
Adapted from George Stalk Jr. and Thomas M.
Hout, Competing Against Time [New York,
1990], 12.)
S "CashCow" "Dog"
the companycan grow even faster and emerge with a dominant share
when growtheventuallyslows."30
Strategic Business Units and PortfolioAnalysis. Numerousother
consultingfirms came up with their own matricesfor portfolioanalysis
at roughlythe same time as BCG.McKinsey& Company'seffort,for in-
stance,began in 1968 when Fred Borch,the CEOof GE, asked McKin-
sey to examine his company'scorporatestructure,which consisted of
two hundred profit centers and one hundred and forty-five depart-
ments arrangedaroundten groups.The boundariesfor these units had
been defined accordingto theories of financial control, which the Mc-
Kinseyconsultantsjudged to be inadequate.They arguedthat the firm
should be organizedon more strategiclines, with greater concern for
externalconditions than internal controls and a more future-oriented
approachthan was possible using measures of past financial perfor-
mance. The study recommendeda formal strategic planning system
that would divide the company into "naturalbusiness units," which
Borchlater renamed "strategicbusiness units," or SBUs. GE's execu-
tives followedthis advice,which took two years to put into effect.
However, in 1971, a GE corporateexecutive asked McKinseyfor
help in evaluatingthe strategic plans that were being written by the
company'smany SBUs. GE had already examined the possibility of
using the BCGgrowth-sharematrixto decide the fate of its SBUs, but
its top managementhad decidedthen that they could not set priorities
on the basis of just two performancemeasures.And so, after studying
the problemfor three months, a McKinseyteam producedwhat came
to be known as the GE/McKinseynine-block matrix. The nine-block
matrixused about a dozen measures to screen for industry attractive-
30Conley, "ExperienceCurves as a Planning Tool," 10-11.

Attractiveness
Industry
High Medium Low
Investment
High Growth and
and Selective Selectivity
Growth Growth
C e u Selective Harvestl
Medium Growth Selectivity Divest
0
ZC
Harvest/ Harvestl
Low Selectivity Divest Divest
Figure 4. Industry Attractiveness- Business Strength Matrix. (Source: Arnoldo C. Hax and
Nicolas S. Majluf, Strategic Management: An Integrative Perspective [Englewood Cliffs,
N.J., 19841, 156.)
ness, or profitability,and anotherdozen to screen for competitiveposi-

tion, althoughthe weights to be attachedto them were not specified.31
(See Figure4.)
Another,more quantitative,approachto portfolioplanningwas de-
veloped at roughlythe same time under the aegis of the "ProfitImpact
of MarketStrategies"(PIMS) program,which was the multicompany
successor to the PROMprogramthat GE had started a decade earlier.
By the mid-197os, PIMS contained data on six hundred and twenty
SBUs drawn from fifty-seven diversified corporations.32These data
were used, in the first instance, to explorethe determinantsof returns
on investment by regressing historical returns on variables such as
market share, product quality, investment intensity, marketing and
R&Dexpenditures,and several dozen others. The regressions estab-
lished what were supposed to be benchmarksfor the potential perfor-
mance of SBUs with particularcharacteristicsagainst which their ac-
tual performancemightbe compared.
31Interview with Mike Allen, 4 Apr. 1997.

32Sidney E. Schoeffler, Robert D. Buzzell, and Donald F. Heany, "Impact of Strategic
Planning on Profit Performance," Harvard Business Review (Mar./Apr. 1974): 139-40,
144-5.

In all these applications,segmenting diversifiedcorporationsinto

SBUsbecame an importantprecursorto analyses of economic perfor-
mance.33This forced "de-averaging" of cost and performancenumbers
that had previouslybeen calculatedat more aggregatedlevels. In addi-
tion, it was thoughtthat, with such approaches,"strategicthinkingwas
appropriatelypushed 'downthe line' to managerscloser to the particu-
lar industryand its competitiveconditions."34
In the 1970s, virtuallyeverymajorconsultingfirm used some type
of portfolio analysis to generate strategy recommendations.The con-
cept became especiallypopularafter the oil crisis of 1973 forced many
large corporationsto rethink,if not discard, their existing long-range
plans. A McKinseyconsultant noted that "the sudden quadruplingof
energycosts [due to the OPECembargo],followed by a recession and
rumors of impending capital crisis, [meant that] setting long-term
growthand diversificationobjectiveswas suddenlyan exercisein irrel-
evance."Now, strategicplanning meant "sortingout winners and los-
ers, setting priorities, and husbanding capital." In a climate where
"productand geographicmarketswere depressed and capitalwas pre-
sumed to be short,"35portfolioanalysisgave executives a readyexcuse
to get rid of poorly performingbusiness units while directing most
availablefunds to the "stars."Thus, a survey of the "Fortune500" in-
dustrialcompaniesconcludedthat, by 1979, 45 percentof them had in-
troducedportfolioplanningtechniquesto some extent.36
Emerging Problems. Somewhat ironically, the very macroeco-
nomic conditions that (initially) increased the popularityof portfolio
analysisalso began to raise questions about the experiencecurve. The
high inflationand excess capacityresultingfrom downturnsin demand
inducedby the "oilshocks"of 1973 and 1979 disruptedhistoricalexpe-
riencecurvesin manyindustries,suggestingthat BruceHendersonhad
oversold the concept when he circulateda pamphlet in 1974 entitled
"WhyCosts Go Down Forever."Anotherproblem with the experience
curvewas pinpointedin a classic 1974articleby WilliamAbernathyand
Kenneth Wayne, which argued that "the consequence of intensively
pursuinga cost-minimizationstrategy [e.g., one based on the experi-
ence curve] is a reducedabilityto make innovativechanges and to re-
33See Walter Kiechel III, "CorporateStrategists under Fire,"Fortune (27 Dec. 1982).
34Frederick W. Gluck and Stephen P. Kaufman, "Using the Strategic Planning Frame-
work,"in McKinsey internal document, "Readingsin Strategy"(1979), 3-4.
35J. Quincy Hunsicker, "Strategic Planning: A Chinese Dinner?" McKinsey staff paper
(Dec. 1978), 3.
36 Philippe Haspeslagh, "PortfolioPlanning: Uses and Limits,"Harvard Business Review
(Jan. /Feb. 1982): 59.

spond to those introducedby competitors."37Abernathyand Wayne

pointedto the case of HenryFord,whose obsession with loweringcosts
had left him vulnerableto AlfredSloan'sstrategyof productinnovation
in the car business. The concept of the experiencecurve was also criti-
cizedfor treatingcost reductionsas automaticratherthan somethingto
be managed,for assuming that most experiencecould be kept propri-
etary instead of spilling over to competitors, for mixing up different
sourcesof cost reductionwith very differentstrategicimplications(e.g.,
learning versus scale versus exogenous technical progress), and for
leading to stalemates as more than one competitorpursued the same
genericsuccess factor.38
In the late 1970s, portfolioanalysiscame under attackas well. One
problemwas that, in many cases, the strategicrecommendationsfor an
SBUwere very sensitiveto the specificportfolio-analytictechniqueem-
ployed.For instance, an academicstudy appliedfour differentportfolio
techniquesto a group of fifteen SBUs owned by the same Fortune500
corporation;it foundthat only one of the fifteen SBUsfell into the same
portion of each of the four matrices, and only five of the fifteen were
classified similarly in terms of three of the four matrices.39This was
only a slightly higher level of concordance than would have been
expected if the fifteen SBUs had been randomly classified four sepa-
ratetimes!
An even more serious problem with portfolio analysis was that
even if one could figure out the "right"technique to employ, the me-
chanicaldeterminationof resourceallocation patterns on the basis of
historicalperformancedata was inherentlyproblematic.Some consult-
ants acknowledgedas much. In 1979, Fred Gluck,the head of McKin-
sey's strategic management practice, ventured the opinion that "the
heavy dependence on 'packaged'techniques [has] frequentlyresulted
in nothing more than a tighteningup, or fine tuning, of currentinitia-
tives within the traditionallyconfiguredbusinesses."Evenworse, tech-
nique-based strategies "rarelybeat existing competition"and often
leave businesses "vulnerableto unexpectedthrustsfrom companiesnot
previouslyconsideredcompetitors."40Gluckand his colleaguessought
to loosen some of the constraintsimposed by mechanisticapproaches,
37William J. Abernathy and Kenneth Wayne, "Limits of the Learning Curve,"Harvard

Business Review (Sept./Oct. 1974): 111.
38Pankaj Ghemawat, "Building Strategy on the Experience Curve,"Harvard Business
Review (Mar. /Apr.): 1985.
39YoramWind, Vijay Mahajan, and Donald J. Swire, "AnEmpirical Comparison of Stan-
dardized Portfolio Models," Journal of Marketing 47 (Spring 1983): 89-99. The statistical
analysis of their results is based on an unpublished draft by PankajGhemawat.
40Gluck and Kaufman, "Usingthe Strategic Planning Framework,"5-6.

4. StrategicManagement:
Createthe Future
3. ExternallyOrientedPlanning:
ThinkStrategically r
Dynamic
Analysis
StaticAnalysis
2. Forecast-BasedPlanning:
Predictthe Future
1. FinancialPlanning: t
MeetAnnualBudget
Figure 5. Four Phases of Strategy. (Source: Adapted from Frederick W. Gluck, Stephen P.
Kaufman, and A. Steven Walleck, "TheEvolution of Strategic Management," McKinsey staff
paper [Oct. 1978], 4. Reproducedin modified form in Gluck, Kaufman,and Walleck, "Strategic
Managementfor CompetitiveAdvantage,"Harvard Business Review [July/Aug. 1980], 157.)
proposing that successful companies devise progressive strategies to

take them throughfour basic stages. Each stage requiresthese compa-
nies to grapplewith increasinglevels of dynamism,multidimensional-
ity, and uncertainty,and they thereforebecome less amenableto rou-
tine quantitativeanalysis.(See Figure5.)
The most stinging attackon the analyticaltechniques popularized
by strategyconsultantswas offeredby two Harvardprofessorsof pro-
duction, Robert Hayes and WilliamAbernathy,in 1980. They argued
that "these new principles [of management],despite their sophistica-
tion and widespreadusefulness,encouragea preferencefor (1) analytic
detachmentratherthan the insight that comes from 'hands on experi-
ence' and (2) short-termcost reductionratherthan long-termdevelop-
ment of technologicalcompetitiveness."41 Hayes and Abernathyin par-
ticularcriticizedportfolioanalysis as a tool that led managersto focus
on minimizingfinancialrisks ratherthan on investing in new opportu-
nities that requirea long-termcommitmentof resources.42They went
on to compare U.S. firms unfavorablywith Japanese and, especially,
Europeanones.
These and other criticismsgraduallydiminishedthe popularityof
portfolioanalysis.However,its rise and fall did have a lasting influence
on subsequent work on competition and business strategybecause it
highlightedthe need for more carefulanalysis of the two basic dimen-
sions of portfolio-analyticgrids: industry attractivenessand competi-
41 Robert H. Hayes and William J. Abernathy, "ManagingOur Way to Economic Decline,"

Harvard Business Review (July/Aug. 1980): 68.
42 Ibid., 71.

Industry
Attractiveness
Figure 6. Two Basic Dimensions of Strategy.
Competitive
Advantage
tive position. Although these two dimensions had been identified

earlier-in the GeneralSurveyOutlinedevelopedby McKinsey& Com-
pany for internal use in 1952, for example--portfolio analysis under-
scoredthis particularmethodof analyzingthe effects of competitionon
business performance.U.S. managers,in particular,proved avid con-
sumers of insights about competitionbecause the exposureof much of
U.S. industry to competitiveforces increased dramaticallyduring the
1960s and 1970s. One economist roughly calculated that heightened
import competition, antitrustactions, and deregulationincreased the
share of the U.S. economy that was subject to effective competition
from 56 percentin 1958 to 77 percentby 1980.43The next two sections
describeattempts to unbundlethese two basic dimensions of strategy.
(See Figure6.)
Unbundling Industry Attractiveness

Thus far, we have made little mention of economists'contributions
to thinkingaboutcompetitivestrategy.On the one hand, economicthe-
ory emphasizes the role of competitive forces in determining market
outcomes. However, on the other hand, economists have often over-
looked the importance of strategy because, since Adam Smith, they
have traditionallyfocused on the case of perfect competition:an ideal-
ized situationin which large numbersof equallyable competitorsdrive
an industry'saggregateeconomic profits (i.e., profits in excess of the
opportunitycost of the capitalemployed) down to zero. Under perfect
competition,individualcompetitorsare straitjacketed,in the sense of
having a choice between producing efficiently and pricing at cost or
shutting down.
Some economists did addressthe opposite case of perfectcompeti-
tion, namely pure monopoly,with Antoine Cournotprovidingthe first
43William G. Shepherd, "Causesof Increased Competition in the U.S. Economy, 1939-

1980," Review of Economics and Statistics (Nov. 1982): 619.

definitive analysis-as well as analysis of oligopoly under specific

assumptions-in 1838.44Work on monopoly yielded some useful in-
sights, such as the expectationof an inverserelationbetweenthe profit-
abilityof a monopolizedindustryand the price elasticityof the demand
it faced--an insight that has remained central in modern marketing.
Nevertheless,the assumptionof monopolyobviouslytook things to the
other, equallyunfortunate,extreme by ruling out all directlycompeti-
tive forcesin the behaviorof firms.
This state of affairs began to change at an applied level in the
1930s, as a number of economists, particularlythose associated with
the "Harvardschool,"beganto arguethat the structureof many indus-
tries might permit incumbentfirms to earn positive economic profits
over long periodsof time.45EdwardS. Mason arguedthat the structure
of an industry would determine the conduct of buyers and sellers--
their choices of criticaldecisionvariables- and, by implication,its per-
formancealong such dimensions as profitability,efficiency,and inno-
vativeness.46Joe Bain,also of the HarvardEconomicsDepartment,ad-
vanced the research program of uncovering the general relation
between industry structureand performancethrough empiricalwork
focused on a limited numberof structuralvariables-most notably,in
two studies publishedin the 1950s. The first study found that the prof-
itabilityof manufacturingindustriesin which the eight largest compet-
itors accountedfor more than 70 percentof sales was nearlytwice that
of industrieswith eight-firmconcentrationratios of less than 70 per-
cent.47The second study explainedhow, in certain industries, "estab-
lished sellers can persistently raise their prices above a competitive
level without attractingnew firms to enter the industry."48Bain identi-
fied three basic barriersto entry: (1) an absolute cost advantageby an
establishedfirm (an enforceablepatent, for instance); (2) a significant
degreeof productdifferentiation;and (3) economies of scale.
Bain'sinsightsled to the rapidgrowthof a new subfieldof econom-
ics, known as industrialorganization,or "IO"for short, that explored
the structuralreasons why some industrieswere more profitablethan
others. By the mid-197os, severalhundred empiricalstudies in IO had
44Antoine A. Cournot, Recherches sur les Principes Mathematiques de la Theorie des

Richesses (Paris, 1838), sects. 26, 27; and Jurg Niehans, A History of Economic Theory (Bal-
timore, 199o), 18o-2.
45Economists associated with the Chicago School generally doubted the empirical impor-
tance of this possibility-except as an artifact of regulatory distortions.
46Mason's seminal work was "Priceand Production Policies of Large-Scale Enterprise,"
American Economic Review (Mar. 1939): 61-4.
47Joe S. Bain, "Relationof Profit Rate to Industry Concentration:American Manufactur-
ing, 1936-1940," Quarterly Journal of Economics (Aug. 1951): 293-324.
48Joe S. Bain, Barriers to New Competition (Cambridge,Mass., 1956), 3 n.

Industry Return on Equity

Drugs 21.4
Printingand Publishing 15.5
Petroleumand Coal 13.1
MotorVehicles and Equipment 11.6
Textile MillProducts 9.3
Ironand Steel 3.9
Figure 7. Differences in the Profitabilityof Selected Industries, 1971-1990. (Source:Anita M.

McGahan, "Selected ProfitabilityData on U.S. Industries and Companies,"HarvardBusiness
School Publishing, No. 792-066 [1992].)
been carried out. While the relation between structuralvariables and

performanceturned out to be more complicatedthan had been sug-
gested earlier,49these studies reinforcedthe idea that some industries
are inherentlymuch more profitableor "attractive"than others, as in-
dicatedbelow. (See Figure7.)
Harvard Business School's Business Policy Group was aware of
these insights from acrossthe CharlesRiver:excerptsfrom Bain'sbook
on barriersto entry were even assigned as required readings for the
business policy course in the early 196os. But the immediateimpact of
IO on business strategywas limited. Although many problems can be
discernedin retrospect,two seem to have been particularlyimportant.
First,IO economistsfocusedon issues of publicpolicyratherthan busi-
ness policy: they concerned themselves with the minimizationrather
than the maximizationof "excess"profits.Second,the emphasisof Bain
and his successors on using a limited list of structuralvariablesto ex-
plain industryprofitabilityshortchangedthe richnessof modernindus-
trial competition("conduct"within the IO paradigm).
Both of these problems with applying classical IO to business-
strategic concerns about industry attractiveness were addressed by
MichaelPorter,a graduateof the Ph.D. programofferedjointly by Har-
vard'sBusiness School and its EconomicsDepartment.In 1974, Porter
prepareda "Noteon the StructuralAnalysis of Industries,"which pre-
sented his first attemptto turn IO on its head by focusing on the busi-
ness policy objectiveof profit maximization,ratherthan on the public
policy objectiveof minimizing"excess"profits.50In 1980, he released
his landmarkbook, CompetitiveStrategy, which owed much of its suc-
49See, for instance, Harvey J. Goldschmid, H. Michael Mann, and J. Fred Weston, eds.,
Industrial Concentration: TheNew Learning (Boston, 1974).
50Michael E. Porter, "Note on the Structural Analysis of Industries," Harvard Business
School Teaching Note, no. 376-054 (1983).

cess to Porter'selaborateframeworkfor the structuralanalysis of in-

dustry attractiveness.Figure 8 reproduces Porter's "five forces" ap-
proachto understandingthe attractivenessof an industryenvironment
for the "average"competitorwithin it. In developingthis approachto
strategy, Porter noted the trade-offs involved in using a "framework"
ratherthan a more formal statistical "model."In his words, a frame-
work "encompassesmany variablesand seeks to capture much of the
complexity of actual competition. Frameworksidentify the relevant
variablesand the questions that the user must answer in order to de-
velop conclusionstailoredto a particularindustryand company"(em-
phasis added).5'In academicterms, the drawbackof frameworkssuch
as the five forces is that they often rangebeyond the empiricalevidence
that is available.In practice,managersroutinelyhave to considermuch
longer lists of variables than are embedded in the relatively simple
quantitativemodels used by economists. In the case of the five forces, a
survey of empiricalliteraturein the late 198os-more than a decade
after Porter first developed his framework- revealed that only a few
pointswere stronglysupportedby the empiricalliteraturegeneratedby
the IOfield.52(Thesepoints appearin bold print in Figure8.) This does
not mean that the other points are in conflictwith IO research;rather,
they reflect the experience of strategy practitioners,including Porter
himself.
In managerialterms, one of the breakthroughsbuilt into Porter's
frameworkwas that it emphasized "extendedcompetition"for value
ratherthan just competitionbetween existing rivals. For this reason,
and because it was easy to put into effect, the five-forces framework
cameto be used widelyby managersand consultants.Subsequentyears
witnessed refinementsand extensions, such as the rearrangementand
incorporation of additional variables (e.g., import competition and
multimarketcontact) into the determinants of the intensity of five
forces.The biggest conceptualadvance,however,was one proposed in
the mid-199os by two strategists concerned with game theory, Adam
Brandenburgerand BarryNalebuff,who arguedthat the process of cre-
ating value in the marketplace involved "four types of players--
customers, suppliers,competitors,and complementors."53 By a firm's
"complementors,"they meant other firms from which customers buy
51 Michael E. Porter,"Towarda Dynamic Theory of Strategy,"in RichardP. Rumelt, Dan E.

Schendel, and David J. Teece, eds., Fundamental Issues in Strategy (Boston, 1994), 427-9.
s2 Richard Schmalensee, "Inter-IndustryStudies of Structure and Performance,"in Rich-
ard Schmalensee and R. D. Willig, eds., Handbook of Industrial Organization, vol. 2 (Am-
sterdam, 1989).
53AdamM. Brandenburgerand BarryJ. Nalebuff, Co-opetition (New York, 1996).

Suppliers
Sources of Bargaining Power:
Switchingcosts
of inputs
Differentiation
Supplierconcentration
Presence of substituteinputs
Importanceof volumeto suppliers
Impactof inputson cost or differentiation
Threatof forward/backward integration
Cost relativeto totalpurchases in industry
New Entrants IndustryCompetitors

Entry Barriers: Factors affecting Rivalry:
Substitutes
Economies of scale Industry groWth
Brand identity Concentration and balance Threatdetermined by:
Capital requirements Fixedcosts/value added
Proprietaryproductdifferences Intermittentovercapacity Relativeprice performance
Switchingcosts Productdifferences of substitutes
Access to distribution Brandidentity Switchingcosts
Proprietarylearningcurve Switchingcosts Buyerpropensityto substitute
Access to necessary inputs Informational complexity
Low-costproductdesign Diversityof competitors
Governmentpolicy Corporatestakes
Expectedretaliation Exitbarriers
Buyers
Bargaining Power of Buyers:
Buyer concentration
Buyervolume
Switchingcosts
Buyerinformation
Buyerprofits
Substituteproducts
Pull-through
Pricesensitivity
Price/totalpurchases
Productdifferences
Brandidentity
Abilityto backwardintegrate
Impacton quality/performance
Decision makers'incentives
Figure 8. Porter's Five-Forces Frameworkfor Industry Analysis.
complementary products and services, or to which suppliers sell com-

plementary resources. As Brandenburger and Nalebuff pointed out, the
practical importance of this group of players was evident in the amount
of attention being paid in business to the subject of strategic alliances
and partnerships. Their Value Net graphic depicted this more complete
description of the business landscape-emphasizing, in particular, the
equal roles played by competition and complementarity. (See Figure 9.)

Customers
Competitors------ Company------b Complementors
Suppliers
Figure 9. The Value Net. (Source:Adam M. Brandenburgerand BarryJ. Nalebuff, Co-opetition

[New York, 1996], 17.)
Otherstrategists,however,arguedthat some verylimiting assump-

tions were built into such frameworks.Thus, Kevin Coyne and Somu
Subramanyamof McKinseyargued that the Porter frameworkmade
three tacit but crucialassumptions:First,that an industryconsists of a
set of unrelatedbuyers,sellers, substitutes,and competitorsthat inter-
act at arm'slength. Second, that wealth will accrueto playersthat are
able to erect barriersagainstcompetitorsand potential entrants,or, in
other words, that the source of value is structuraladvantage.Third,
that uncertaintyis sufficientlylow that you can accuratelypredictpar-
ticipants'behaviorand choose a strategyaccordingly.54
Unbundling Competitive Position

The second basic dimension of business strategy highlighted by
Figure6 is competitiveposition. While differencesin the averageprof-
itabilityof industriescan be large, as indicatedin Figure7, differences
in profitabilitywithin industriescan be even larger.55Indeed, in some
cases firms in unattractiveindustriescan significantlyoutperformthe
averagesfor more profitableindustries, as indicated in Figure 10. In
addition,one might arguethat most businesses in most industryenvi-
ronments are better placed to try to alter their own competitiveposi-
tions, rather than the overall attractivenessof the industry in which
they operate.For both these reasons, competitiveposition has been of
greatinterestto business strategists.(See Figurelo.)
54Kevin P. Coyne and Somu Subramanyam, "BringingDiscipline to Strategy,"McKinsey

Quarterly 4 (1996): 16.
55See, for instance, Richard P. Rumelt, "How Much Does Industry Matter?"Strategic
Management Journal (March 1991): 167-85.

ROA(%)
20-
15-
OREGONSTEEL MILLS
WORTHINGTON
10-
NUCOR
5 ..
USX-US STEEL
INLANDSTEEL
0 .
* ARMCO
* BETHLEHEM
*LTV
-5- I I I I I SALES
0 1,000 2,000 3,000 4,000 5,000 6,000 ($ M)
Figure o10. Profitability within the Steel Industry, 1973-1992. (Source: David Collis and
Pankaj Ghemawat, "IndustryAnalysis: Understanding Industry Structure and Dynamics," in
Liam Fahey and Robert M. Randall, The Portable MBA in Strategy [New York, 19941, 174.)
Traditionalacademicresearchhas made a numberof contributions

to our understandingof positioning within industries, starting in the
1970s.The IO-basedliteratureon strategicgroups,initiatedat Harvard
by MichaelHunt'swork on broad-lineversus narrow-linestrategies in
the majorhome applianceindustry,suggestedthat competitorswithin
particularindustries could be grouped in terms of their competitive
strategies in ways that helped explain their interactions and relative
profitability.56A stream of work at Purdueexploredthe heterogeneity
of competitive positions, strategies, and performancein brewing and
other industries with a combinationof statisticalanalysis and qualita-
tive case studies. More recently,several academicpoints of view about
the sourcesof performancedifferenceswithinindustrieshave emerged-
views that are exploredmore fully in the next section. However,it does
seem accurate to say that the work that had the most impact on the
strategic thinking of business about competitive positions in the late
1970s and the 1980s was more pragmaticthan academicin its intent,
with consultantsonce againplayinga leading role.
56See Michael S. Hunt, "Competitionin the Major Home Appliance Industry,"DBA diss.,
Harvard University, 1972. A theoretical foundation for strategic groups was provided by
Richard E. Caves and Michael E. Porter, "From Entry Barriers to Mobility Barriers,"Quar-
terly Journal of Economics (Nov. 1977): 667-75.

CompetitiveCostAnalysis. Withthe rise of the experiencecurvein

the 1960s, most strategiststurned to some type of cost analysis as the
basis for assessing competitivepositions. The interest in competitive
cost analysissurvivedthe decliningpopularityof the experiencecurve
in the 1970s but was reshapedby it in two importantways. First, more
attention was paid to disaggregatingbusinesses into their component
activitiesor processes and to thinking about how costs in a particular
activitymight be shared across businesses. Second, strategistsgreatly
enrichedtheirmenu of cost driversto includemorethanjust experience.
The disaggregationof businesses into their component activities
seems to have been motivated,in part,by earlyattemptsto "fix"the ex-
periencecurveto deal with the rising real prices of many raw materials
in the 1970s.57The proposedfix involvedsplittingcosts into the costs of
purchasedmaterialsand "costadded"(value added minus profit mar-
gins) and redefiningthe experiencecurveas applyingonly to the latter.
The naturalnext step was to disaggregatea business's entire cost struc-
ture into activitieswhose costs might be expectedto behave in interest-
ingly different ways. As in the case of portfolio analysis, the idea of
splitting businesses into component activities diffused quicklyamong
consultantsand their clients in the 1970s. A templatefor activityanaly-
sis that became especiallyprominentis reproducedin Figure11.
Activityanalysis also suggested a way of getting aroundthe "free-
standing"conceptionof individualbusinesses built into the concept of
SBUs.One persistentproblemin splittingdiversifiedcorporationsinto
SBUs was that, with the exception of pure conglomerates,SBUs were
often related in ways that meant they shared elements of their cost
structurewith each other. Consultingfirms, particularlyBain and Stra-
tegic Planning Associates, both of whose founders had worked on a
BCGstudy of Texas Instrumentsthat was supposed to highlight the
problemof shared costs, began to emphasizethe developmentof what
came to be called "fieldmaps":matricesthat identified shared costs at
the level of individualactivitiesthat were linked across businesses, as
illustratedbelow.58
The second important development in competitive cost analysis
overthe late 1970s and early198os involvedenrichmentof the menu of
cost drivers considered by strategists. Scale effects, while officially
lumped into the experience curve, had long been looked at indepen-
dentlyin particularcases; even more specifictreatmentof the effects of
scale was now forcedby activityanalysisthat might indicate,for exam-
ple, that advertisingcosts were drivenby nationalscale, whereasdistri-
57This is based on my experience working at BCGin the late 1970s.

58Walter Kiechel III, "TheDecline of the Experience Curve,"Fortune (5 Oct. 1981).

Technology Manufacturing Distribution Marketing Service
Design Procurement Transport Retailing Parts

Development Assembly Inventory Advertising Labor
Figure 11. McKinsey's Business System. (Source: Adapted from Carter F. Bales, P. C. Chat-
terjee, Donald J. Gogel, and Anupam P. Puri, "Competitive Cost Analysis," McKinsey staff
paper [Jan. 1980], 6.)
bution costs were driven by local or regional scale. Field maps under-
scored the potential importance of economies (or diseconomies) of scope
across businesses rather than scale within a business. The effects of ca-
pacity utilization on costs were dramatized by macroeconomic down-
turns in the wake of the two oil shocks. The globalization of competition
in many industries highlighted the location of activities as a main driver
of competitors' cost positions, and so on. Thus, an influential mid-198os
discussion of cost analysis enumerated ten distinct cost drivers.59
Customer Analysis. Increased sophistication in analyzing relative
costs was accompanied by increased attention to customers in the pro-
cess of analyzing competitive position. Customers had never been en-
tirely invisible: even in the heyday of experience curve analysis, market
segmentation had been an essential strategic tool--although it was
sometimes used to gerrymander markets to "demonstrate" a positive
link between share and cost advantage rather than for any analytic pur-
pose. But, according to Walker Lewis, the founder of Strategic Planning
Associates, "To those who defended in classic experience-curve strat-
egy, about 80% of the businesses in the world were commodities."60o
This started to change in the 1970s.
Increased attention to customer analysis involved reconsideration
of the idea that attaining low costs and offering customers low prices
was always the best way to compete. More attention came to be paid to
differentiated ways of competing that might let a business command a
price premium by improving customers' performance or reducing their
(other) costs. While (product) differentiation had always occupied cen-
ter stage in marketing, the idea of looking at it in a cross-functional,
competitive context that also accounted for relative costs apparently
started to emerge in business strategy in the 1970s. Thus, a member of
Harvard's Business Policy group recalls using the distinction between
59Michael E. Porter, CompetitiveAdvantage (New York, 1985), ch. 3.

60Quoted in Kiechel, "TheDecline of the Experience Curve."

cost and differentiation,which was implicit in two of the three sources

of entry barriersidentified by Joe Bain in the 1950s (see above), to
organize classroom discussions in the early 197os.6' And McKinsey
reportedlystartedto applythe distinctionbetween cost and "value"to
client studies later in that decade.62 The first published accounts, in
Michael Porter'sbook CompetitiveStrategy and in a Harvard Busi-
ness Review articleby WilliamHall, appearedin 1980.63
Both Hall and Porterarguedthat successfulcompaniesusuallyhad
to choose to compete either on the basis of low costs or by differentiat-
ing productsthrough quality and performancecharacteristics.Porter
also identifieda focus option that cut across these two "genericstrate-
gies"andlinkedthese strategicoptionsto his workon industryanalysis:
In someindustries,thereareno opportunitiesfor focusor differen-
tiation-it's solelya cost game- andthis is truein a numberof bulk
commodities.In other industries, cost is relativelyunimportant
becauseof buyerandproductcharacteristics.64
Manyother strategistsagreedthat, exceptin such special cases, the

analysisof competitivepositionhad to coverboth relativecost and dif-
ferentiation.Therewas continuingdebate,however,aboutthe proposi-
tion, explicitlyput forthby Porter,that businesses "stuckin the middle"
should be expected to performless well than businesses that had tar-
geted lower cost or more differentiatedpositions. Others saw optimal
positioning as a choice from a continuum of trade-offs between cost
and differentiation,ratherthan as a choice between two mutuallyex-
clusive(and extreme)genericstrategies.
Porter'sbook, publishedin 1985, suggested analyzingcost and dif-
ferentiationvia the "valuechain,"a templatethat is reproducedin Fig-
ure 12. While Porter's value chain bore an obvious resemblance to
McKinsey'sbusiness system,his discussionof it emphasizedthe impor-
tance of regroupingfunctions into the activities actuallyperformedto
produce, market, deliver, and support products, thinking about links
between activities,and connectingthe value chain to the determinants
of competitiveposition in a specificway:
Competitive advantagecannotbe understoodby lookingat a firmas

a whole.It stems fromthe manydiscreteactivitiesa firmperforms
in designing,producing,marketing,delivering,and supportingits
61Interview with Hugo Uyterhoeven, 25 Apr. 1997.

62Interview with Fred Gluck, 18 Feb. 1997.
63Michael E. Porter, Competitive Strategy (New York, 1980), ch. 2; and William K. Hall,
"SurvivalStrategies in a Hostile Environment,"Harvard Business Review (Sept./Oct. 1980):
78 - 81.
64Porter, Competitive Strategy, 41-4.

Firm Infrastructure
Human Resource Management

Support
Activities
Technology Development
Procurement
Primary Inbound Outbound Marketing&

Activities Logistics Operations Logistics Sales Service
Figure 12. Porter's Value Chain. (Source: Michael E. Porter, Competitive Advantage [New
York, 19851, 37.)
product. Each of these activities can contribute to a firm's relative

cost position and create a basis for differentiation.... The value
chain disaggregates a firm into its strategically relevant activities in
order to understand the behavior of costs and the existing and
potential sources of differentiation.65
Putting customer analysis and cost analysis together was promoted

not only by disaggregating businesses into activities (or processes) but
also by splitting customers into segments based on cost-to-serve as well
as customer needs. Such "de-averaging" of customers was often said to
expose situations in which 20 percent of a business's customers ac-
counted for more than 80 percent, or even 100 percent, of its profits.66
It also suggested new customer segmentation criteria. Thus, Bain &
Company built a thriving "customer retention" practice, starting in the
late 198os, on the basis of the higher costs of capturing new customers
as opposed to retaining existing ones.
Competitive Dynamics and History

The development of business systems, value chains, and similar
templates naturally refocused attention on the problem of coordinating
across a large number of choices linked in cross section that was high-
lighted, in a cross-functional context, in the original description of Har-
vard Business School's course on business policy. However, such atten-
tion tended to crowd out consideration of longitudinal links between
choices, which was emphasized by Selznick's work on organizational
commitments and distinctive competences and evident in Andrews's
focus on the aspects of firm behavior that were "enduring and unchang-
ing over relatively long periods of time."
65Porter, CompetitiveAdvantage, 33, 37.
66Talk by Arnoldo Hax at MIT on 29 April 1997.

The need to return the time dimension to predominantlystatic

ideas about competitive position was neatly illustrated by the tech-
niques for "value-basedstrategicmanagement"that began to be pro-
moted by consultingfirmslike SPAand Marakon,among others, in the
1980s. The development and diffusion of value-based techniques,
which connected positioning measures to shareholder value using
spreadsheetmodels of discountedcash flows, was driven by increases
in capital market pressures in the 198os, particularlyin the United
States:mergerand acquisitionactivitysoared;hostile takeoversof even
very large companiesbecame far more common; many companies re-
structuredto avoid them; levels of leverage generally increased; and
there was creeping institutionalization of equity holdings.67 Early
value-basedwork focused on the spread between a company or divi-
sion's rate of returnand its cost of capital as the basis for "solving"the
old corporate strategy problem of resource allocation across busi-
nesses. It quickly became clear, however, that estimated valuations
were very sensitive to two other, more dynamic, drivers of value: the
length of the time horizonoverwhich positive spreads(competitivead-
vantage)could be sustainedon the assets in place, and the (profitable)
reinvestmentopportunitiesor growthoptions affordedby a strategy.68
At the same time, analyses of business performancestarted to under-
score the treacherousnessof assuming that current profitabilityand
growth could automaticallybe sustained. Thus, my analysis of 700
business units revealedthat nine-tenths of the profitabilitydifferential
between businesses that were initially above average and those that
were initially below average vanished over a ten-year period.69 (See
Figure 13.)
The unsustainabilityof most competitiveadvantageswas generally
thoughtto reflectthe "RedQueen"effect:the idea that as organizations
struggledto adaptto competitivepressures,they would become stron-
ger competitors,sending the overalllevel of competitionspiralingup-
ward and eliminatingmost, if not all, competitiveadvantages.70In the
67F. M. Scherer and David Ross, Industrial Market Structure and Economic Perfor-
mance (Boston, 199o), ch. 5.
68Benjamin C. Esty, "Note on Value
Drivers," Harvard Business School Teaching Note,
no. 297-082 (1997).
69Pankaj Ghemawat, "Sustainable Advantage," Harvard Business Review (Sept./Oct.
1986): 53-8, and Commitment (New York, 1991), ch. 5.
70 The first economic citation of the "Red Queen" effect is generally attributed to L. Van
Valen. See L. Van Valen, "A New Evolutionary Law,"Evolutionary Theory 1 (1973): 1-30.
The literary reference is to Lewis Carroll'sAlice's Adventures in Wonderland and Through
the Looking Glass (New York, 1981; first published 1865-71), in which the Red Queen tells
Alice: "here, you see, it takes all the running you can do, to keep in the same place. If you
want to get somewhere else, you must run at least twice as fast.. ." (p. 127).

40
30
0 20
10
1 2 3 4 5 6 7 8 9 10
Year
Figure 13. The Limits to Sustainability.
late 1980s and early 199os, both academicsand consultantsstartedto

wrestlewith the dynamicquestionof how businesses might create and
sustain competitive advantage in the presence of competitors who
could not all be countedon to remaininert all the time.
From an academicperspective,many of the consultants'recom-
mendations regardingdynamics amounted to no more, and no less,
than the injunctionto try to be smarterthan the competition(for exam-
ple, by focusingon customers'futureneeds while competitorsremained
focused on their current needs). The most thoughtfulexception that
had a trulydynamicorientationwas workby GeorgeStalkand othersat
BCGon time-based competition.In an article published in the Har-
vard Business Review in 1988, Stalk argued:"Todaythe leading edge
of competitionis the combinationof fast responseand increasingvari-
ety. Companieswithoutthese advantagesare slippinginto commodity-
like competition,where customers buy mainly on price."71Stalk ex-
panded on this argumentin a book coauthoredwith Thomas Hout in
1990, accordingto which time-basedcompetitors"[c]reatemore infor-
mationand shareit more spontaneously.Forthe informationtechnolo-
71GeorgeStalk Jr., "Time-The Next Source of Competitive Advantage,"Harvard Busi-

ness Review (July/Aug. 1988).

gist, informationis a fluid asset, a data stream. But to the managerof a

business ... information is fuzzy and takes many forms--knowing a
customer'sspecial needs, seeing where the marketis heading ..."72
Time-basedcompetitionquicklycame to account for a substantial
fractionof BCG'sbusiness. Eventually,however,its limitationsalso be-
came apparent. In 1993, George Stalk and Alan Webber wrote that
some Japanesecompanieshad become so dedicatedto shorteningtheir
product-developmentcyclesthat they had createda "strategictreadmill
on which companieswere caught, condemnedto run faster and faster
but always staying in the same place competitively."73In particular,
Japanese electronicsmanufacturershad reacheda remarkablelevel of
efficiency,but it was an "efficiencythat [did] not meet or create needs
for any customer."74
For some, like Stalkhimself, the lesson from this and similar epi-
sodes was that there were no sustainable advantages: "Strategycan
neverbe a constant. ... Strategyis and alwayshas been a moving tar-
get."7sHowever, others, primarily academics, continued to work in
the 1990s on explanations of differences in performancethat would
continue to be useful even after they were widely grasped.76This aca-
demic work exploits, in different ways, the idea that history matters,
that history affects both the opportunities available to competitors
and the effectiveness with which competitors can exploit them. Such
workcan be seen as an attempt to add a historical or time dimension,
involvingstickiness and rigidities,to the two basic dimensions of early
portfolio analyticgrids: industry attractivenessand competitive posi-
tion. The rest of this section briefly reviews four strands of academic
inquiry that embodied new approaches to thinking about the time
dimension.
Game Theory. Gametheory is the mathematicalstudy of interac-
tions between playerswhose payoffsdepend on each other'schoices. A
generaltheoryof zero-sumgames, in which one player'sgain is exactly
equalto other players'losses, was suppliedby John von Neumann and
OskarMorgensternin their pathbreakingbook The Theory of Games
and EconomicBehavior.77Thereis no generaltheory of non-zero-sum
games, which affordopportunitiesfor cooperationas well as competi-
72Stalk and Hout, Competing Against Time, 179.

73George Stalk Jr. and Alan M. Webber, "Japan'sDark Side of Time," Harvard Business
Review (July/Aug. 1993): 94.
74Ibid., 98-9.
75Ibid., 101- 2.
76 This test of stability is in the spirit of the game theorists, John von Neumann and Oskar
Morgenstern. See their Theory of Games and Economic Behavior (Princeton, 1944).
77Ibid.

tion, but researchin this area does supply a languageand a set of logi-
cal tools for analyzing the outcome that is likely-the equilibrium
point-given specific rules, payoff structures,and beliefs if players all
behave"rationally."78
Economiststrainedin IO startedto turn to game theory in the late
1970s as a way of studyingcompetitordynamics.Since the early 198os,
well over half of all the IO articlespublishedin the leading economics
journalshave been concernedwith some aspect of non-zero-sumgame
theory.79By the end of the 198os alone, competitionto invest in tangi-
ble and intangible assets, strategic control of information,horizontal
mergers, network competition and product standardization,contract-
ing, and numerous other settings in which interactiveeffects were apt
to be importanthad all been modeled using game theory.80 The effort
continues.
Game-theoryIO models tend, despite their diversity,to share an
emphasis "on the dynamics of strategic actions and in particularon
the role of commitment."81 The emphasis on commitmentor irrevers-
ibilitygrows out of game theory'sfocus on interactiveeffects. Fromthis
perspective,a strategicmove is one that "purposefullylimits your free-
dom of action. ... It changesother players'expectationsaboutyour fu-
ture responses, and you can turn this to your advantage.Othersknow
that when you have the freedom to act, you also have the freedom to
capitulate."82
The formalismof game theoryis accompaniedby severalsignificant
limitations:the sensitivityof the predictionsof game-theorymodels to
details, the limited numberof variablesconsidered in any one model,
and assumptions of rationalitythat are often heroic, to name just a
few.83Game theory's empiricalbase is also limited. The existing evi-
dence suggests, nonetheless, that it merits attention in analyses of in-
teractionsbetween small numbers of firms. While game theory often
formalizespreexistingintuitions,it can sometimes yield unanticipated,
and even counterintuitive,predictions.Thus, game-theorymodeling of
78There is also a branch of game theory that provides upper bounds on players' payoffs if
freewheeling interactions between them are allowed. See Brandenburger and Nalebuff's Co-
opetition for applications of this idea to business.
79PankajGhemawat, Games Businesses Play (Cambridge, Mass., 1997), 3.
soFor a late 198os survey of game-theory IO, consult Carl Shapiro, "TheTheory of Busi-
ness Strategy,"RAND Journal of Economics (Spring 1989): 125-37.
s8Ibid., 127.
82Avinash K. Dixit and BarryJ. Nalebuff, Thinking Strategically (New York, 1991), 120.
Their logic is based on Thomas C. Schelling's pioneering book, The Strategy of Conflict
(Cambridge,Mass., 1979; first published in 196o).
83 For a detailed critique, see Richard P. Rumelt, Dan Schendel, and David J. Teece, "Stra-
tegic Management and Economics," Strategic Management Journal (Winter 1991): 5-29.
For further discussion, see Ghemawat, Games Businesses Play, chap. i.

shrinkagein, and exit from, decliningindustriesyielded the prediction

that, otherthingsbeing equal,initial size shouldhurt survivability.This
surprisingpredictionturns out to enjoy some empiricalsupport!84
TheResource-BasedViewof the Firm. The idea of looking at com-
panies in terms of their resourceendowmentsis an old one, but it was
revived in the 198os in an article by BirgerWernerfelt.85Wernerfelt
noted: "Thetraditionalconcept of strategy [put forth by KennethAn-
drews in 1971]is phrasedin terms of the resourceposition (strengths
and weaknesses) of the firm, whereas most of our formal economic
tools operateon the productmarketside."86While Wernerfeltalso de-
scribed resourcesand productsas "twosides of the same coin,"other
adherentsto what has come to be calledthe resource-basedview (RBV)
of the firm argue that superior product market positions rest on the
ownershipof scarce,firm-specificresources.
Resource-basedtheorists also seek to distinguishtheir perspective
on sustained superior performance from that of IO economics by
stressing the intrinsic inimitabilityof scarce, valuable resources for a
varietyof reasons:the abilityto obtain a particularresourcemay be de-
pendent on unique, historical circumstances that competitorscannot
recreate;the link betweenthe resourcespossessed by a firm and its sus-
tained competitive advantagemay be causally ambiguous or poorly
understood;or the resourceresponsiblefor the advantagemay be so-
cially complex and therefore"beyondthe abilityof firms to systemati-
cally manage and influence"(e.g., corporate culture).87Game-theory
IO, in contrast,has tended to focus on less extremesituations in which
imitationof superiorresourcesmay be feasiblebut uneconomical(e.g.,
becauseof preemption).
Resource-basedtheoriststhereforehave traditionallytended to see
firms as stuck with a few key resources,which they must deploy across
productmarketsin ways that maximizetotal profits ratherthan profits
in individualmarkets.This insight animated C. K. Prahaladand Gary
Hamel'sinfluentialarticle,"TheCoreCompetenceof the Corporation,"
84For a discussion of the original models (by Ghemawat and Nalebuff) and the support-
ing empirical evidence, consult Ghemawat, Games Businesses Play, ch. 5.
85In the same year, Richard Rumelt also noted that the strategic firm "is characterizedby
a bundle of linked and idiosyncratic resources and resource conversion activities." See his
chapter, "Towardsa Strategic Theory of the Firm,"in R. B. Lamb, ed., Competitive Strategic
Management (Englewood Cliffs, N.J., 1984), 561.
86BirgerWernerfelt, "AResource-based View of the Firm,"Strategic Management Jour-
nal 5 (1984): 171.In addition to citing Andrews's 1971book, The Concept of Corporate Strat-
egy, Wernerfelt referred to the pioneering work of Edith Penrose, The Theory of the Growth
of the Firm (Oxford, 1959).
87Jay B. Barney, "Firm Resources and Sustained Competitive Advantage," Journal of
Management (March 1991): 107-11.

which attackedthe SBU system of managementfor focusing on prod-

ucts ratherthan on underlyingcore competencies in a way that argu-
ablybounded innovation,imprisonedresources,and led to a decline in
investment: "In the short run, a company's competitiveness derives
from the price/performanceattributesof currentproducts.. . . In the
long run, competitivenessderivesfrom the . .. core competenciesthat
spawnunanticipatednew products."88
To many resource-basedtheorists,the core competenciesthat Pra-
halad and Hamel celebrate are simply a neologism for the resources
that the RBVhas emphasizedall along. Whetherthe same can be said
about another, more distinct, line of researchon dynamic capabilities
that emergedin the 199os is an open question.
Dynamic Capabilities. In the 199os, a number of strategistshave
triedto extendthe resource-basedview by explaininghow firm-specific
capabilitiesto perform activities better than competitors can be built
and redeployed over long periods of time. The dynamic-capabilities
view of the firm differsfromthe RBVbecause capabilitiesare to be de-
veloped rather than taken as given, as described more fully in a pio-
neeringarticleby DavidTeece, GaryPisano, and Amy Shuen:
If controlover scarceresourcesis the sourceof economicprofits,
then it follows that issues such as skill acquisition and learning
becomefundamentalstrategicissues. It is this second dimension,
encompassingskill acquisition,learning,and capabilityaccumula-
tion that... [we]referto as "thedynamiccapabilitiesapproach." ...
Rentsareviewedas not onlyresultingfromuncertainty... but also
from directedactivitiesby firmswhich createdifferentiatedcapa-
bilities, and from managerialeffortsto strategicallydeploythese
assets in coordinatedways.89
Takingdynamiccapabilitiesalso impliesthat one of the most strate-

gic aspects of the firm is "theway things are done in the firm, or what
mightbe referredto as its 'routines,'or patternsof currentpracticeand
As a result,"researchin such areasas managementof R&D,
learning."o90
product and processdevelopment,manufacturing,and humanresources
tend to be quiterelevant[to strategy]."9'Researchin these areassupplies
some specific content to the idea that strategyexecutionis important.
88C. K. Prahalad and Gary Hamel, "TheCore Competence of the Corporation,"Harvard
Business Review (May/June 1990): 81.
89David J. Teece, Gary Pisano, and Amy Shuen, "Dynamic Capabilities and Strategic
Management,"mimeo (June 1992): 12-13.
90David Teece and Gary Pisano, "The
Dynamic Capabilities of Firms: An Introduction,"
Industrial and Corporate Change 3 (1994): 540-1. The idea of "routines"as a unit of analy-
sis was pioneered by Richard R. Nelson and Sidney G. Winter, An Evolutionary Theory of
Economic Change (Cambridge,Mass., 1982).
91Teece, Pisano, and Shuen, "DynamicCapabilities and Strategic Management," 2.

The process of capability development is thought to have several

interesting attributes. First, it is generally "path dependent." In other
words, "a firm's previous investments and its repertoire of routines (its
'history') constrains its future behavior ... because learning tends to be
local." Second, capability development also tends to be subject to long
time lags. And third, the "embeddedness" of capabilities in organizations
can convert them into rigidities or sources of inertia -particularly when
attempts are being made to create new, nontraditional capabilities.92
Commitment. A final, historically based approach to thinking
about the dynamics of competition that is intimately related to the
three discussed above focuses on commitment or irreversibility: the
constraints imposed by past choices on present ones.93 The managerial
logic of focusing on decisions that involve significant levels of commit-
ment has been articulated particularly well by a practicing manager:
A decision to build the Edsel or Mustang (or locate your new factory
in Orlando or Yakima) shouldn't be made hastily; nor without
plenty of inputs. ... [But there is] no point in taking three weeks to
make a decision that can be made in three seconds--and corrected
inexpensively later if wrong. The whole organization may be out of
business while you oscillate between baby-blue or buffalo-brown
coffee cups.94
Commitments to durable, firm-specific resources and capabilities

that cannot easily be bought or sold account for the persistence ob-
served in most strategies over time. Modern IO theory also flags such
commitments as being responsible for the sustained profit differences
among product market competitors: thought experiments as well as
formal models indicate that, in the absence of the frictions implied by
commitment, hit-and-run entry would lead to perfectly competitive
(zero-profit) outcomes even without large numbers of competitors.95 A
final attraction of commitment as a way of organizing thinking about
competitor dynamics is that it can be integrated with other modes of
strategic analysis described earlier in this note, as indicated in Figure
92 Dorothy Leonard-Barton,"CoreCapabilities and Core Rigidities: A Paradox in Manag-

ing New Product Development,"Strategic Management Journal (1992): 111-25.
93For a book-length discussion of commitments, see Pankaj Ghemawat, Commitment
(New York, 1991). For connections to the other modes of dynamic analysis discussed in this
section, see chs. 4 and 5 of PankajGhemawat, Strategy and the Business Landscape (Read-
ing, Mass., 1999).
94Robert Townsend, Up the Organization (New York, 1970).
95See, for instance, William J. Baumol, John C. Panzar,and Robert D. Willig, Contestable
Markets and the Theory of Industry Structure (New York, 1982) for an analysis of the eco-
nomic implications of zero commitment; and Richard E. Caves, "EconomicAnalysis and the
Quest for Competitive Advantage,"American Economic Review (May 1984): 127-32, for
comments on the implications for business strategy.

Capabilities - N Resource N- Product Market

Sets)
(Opportunity Commitments Activities
4 A
...... . . . . . . . . . . .
Figure 14. Commitment and Strategy (Source: Adapted from Pankaj Ghemawat, "Resources
and Strategy:An IO Perspective,"HarvardBusiness School working paper [1991], 20, Fig. 3).
14. The ideas behind the figureare very simple. Traditionalpositioning

concepts focus on optimizingthe fit between productmarketactivities
on the right-handside of the figure.The bold arrowsrunningfrom left
to right indicate that choices about which activities to perform, and
how to perform them, are constrained by capabilities and resources
that can be variedonly in the long run and that are responsiblefor sus-
tained profit differencesbetween competitors.The two fainter arrows
that feed backfrom rightto left capturethe ways in which the activities
the organizationperformsand the resourcecommitmentsit makes af-
fect its future opportunityset or capabilities. Finally, the bold arrow
that runs from capabilitiesto resource commitments serves as a re-
minder that the terms on which an organizationcan commit resources
depend, in part, on the capabilitiesit has built up.
Markets for Ideas at the Millennium96

A teleology was implicitin the discussion in the last three sections:
startingin the 1970s, strategistsfirst sought to probe the two basic di-
mensions of early portfolio-analyticgrids, industry attractivenessand
competitiveposition, and then to add a time or historicaldimension to
the analysis. Dynamicthinking along the lines discussed in the previ-
ous section and others (e.g., options thinking, systems dynamics, dis-
ruptive technologies and change management,to cite just four other
areas of enquiry)has absorbedthe bulk of academicstrategists'atten-
tion in the last fifteen-plusyears. But when one looks at the practice of
strategyin the late 199os, this simple narrativeis complicatedby an ap-
parent profusion of tools and ideas about strategy in particularand
management in general, many of which are quite ahistorical. Both
points are illustratedby indexes of the influence of business ideas such
as, for example, importance-weightedcitation counts calculated by
96 For a more extended discussion of the ideas in this postscript, see Pankaj Ghemawat,
"Competition among Management Paradigms: An Economic Analysis," Harvard Business

School Working Paper (2000).

RichardPascale, admittedlywith a significant subjective component,

that are reproducedin Figure15.97A complete enumeration,let alone
discussion,of contemporarytools and ideas is beyond the scope of this
essay, but a few broad points seem worth making about their recent
profusionand turnover.Giventhe forward-lookingnature of this dis-
cussion, it is inevitablymore conjecturalthan the retrospectivesin the
previoussections.
Some of the profusionof ideas about strategyand managementis
probablyto be celebrated.Thus, there are advantagesto being able to
choose from a large menu of ideas ratherthan from a small one, espe-
ciallyin complexenvironmentswhere "onesize doesn't fit all"(and es-
peciallywhen the fixed costs of idea development are low). Similarly,
the rapid turnoverof many ideas, which appearsto have increased in
recentyears, can be explainedin benign terms as well.98 Thus,some ar-
gue that the world is changingrapidly,maybe faster than ever before;
others,that the rapidpeakingfollowedby a decline in attentionto ideas
may indicate that they have been successfullyinternalizedratherthan
discredited;yet others,that at least some of the apparentturnoverrep-
resentsa rhetoricalspurto action,ratherthan real changein the under-
lying ideas themselves.99
It seems difficultto maintain,however,that all the patternsevident
in Figure15 conformto monotonicideals of progress.Consider,for ex-
ample,what happenedwith business-processreengineering,the single
most prominententryas of 1995. Reengineeringwas popularizedin the
early 199os by MichaelHammerand James Champyof the consulting
firm CSCIndex.100Hammeroriginallyexplainedthe idea in a 1990 ar-
ticle in the Harvard Business Review: "Ratherthan embedding out-
dated processes in silicon and software,we should obliteratethem and
start over.We should ... use the powerof moderninformationtechnol-
97For additional discussion of the methodology employed, consult Richard T. Pascale,

Managing on the Edge (New York, 1990), 18-20.
98For some evidence that management ideas have become shorter-lived, see Paula P.
Carson, Patricia A. Lanier, KerryD. Carson, and Brandi N. Guidry, "Clearinga Path through
the Management Fashion Jungle: Some PreliminaryTrailblazing,"Academy of Management
Journal (December 2000).
99Richard D'Aveni, among many others, asserts unprecedented levels of environmental
change in Hypercompetition: Managing the Dynamics of Strategic Maneuvering (New
York, 1994). William Lee and GarySkarke discuss apparently transient ideas that are perma-
nently valuable in "Value-AddedFads: From Passing Fancy to Eternal Truths,"Journal of
Management Consulting (1996): 10-15. Robert G. Eccles and Nitin Nohria emphasize the
rhetorical uses of changing the wrappers on a limited number of timeless truths about man-
agement in Beyond the Hype: Rediscovering the Essence of Management (Boston, 1992).
'oo See Michael Hammer and James Champy,Reengineering the Corporation (New York,
1993). See also John Micklethwait and Adrian Wooldridge, The Witch Doctors (New York,
1996). Micklethwait and Wooldridge devote a chapter to CSCIndex.

!,• :, , l'i , ? , 1II•.•c,,•,:
'V\\ /j
195 1619
/190 99119
/(
? ~I
? ! ' I
? ..
. . . ,?
1950~ 196 190 ~ 98
Figure 15. Ebbs, Flows, and Residual Impact of Business Fads, 1950-1995. (Source: Adapted from RichardT. Pas
199o], 18-20.)

ogy to radicallyredesignourbusiness processesin orderto achievedra-

matic improvementsin their performance."1o• Hammerand Champy's
book, Reengineering the Corporation, which came out in 1993, sold
nearly two million in
copies. Surveys 1994 found that 78 percentof the
Fortune500 companies and 6o percent of a broader sample of 2,200
U.S. companieswere engagedin some form of reengineering,on aver-
age with several projects apiece.102 Consultingrevenues from reengi-
neering exploded to an estimated $2.5 billion by 1995.103 After 1995,
however,there was a bust: consultingrevenuesplummeted,by perhaps
two-thirdsover the next three years, as reengineeringcame to be seen
as a euphemismfor downsizingand as companiesapparentlyshifted to
placing more emphasis on growth (implying, incidentally,that there
had been some excesses in their previouseffortsto reengineer).
Muchof the worrythat the extent of profusionor turnoverof ideas
aboutmanagementmay be excessivefrom a social standpointis linked
to the observationthat this is one of the few areas of intellectual en-
quiryin which it actuallymakes sense to talk about marketsfor ideas.
Unlike, say, twenty-five or thirty years ago, truly large amounts of
money are at stake, and are activelycompeted for, in the development
of "blockbuster" ideas like reengineering--a process that increasingly
seems to fit with the end state describedby Schumpeteras the "routini-
zation of innovation."Market-basedtheoretical models indicate that,
on the supplyside, privateincentivesto invest in developingnew prod-
ucts are likely, in winner-take-allsettings, to exceed social gains.'04To
the extent that market-based,commercialconsiderationsincreasingly
influencethe developmentof new ideas about management,they are a
sourceof growingconcern.
Concernsabout supply-side salesmanship are exacerbatedby the
demand-sideinformationalimperfectionsof marketsfor ideas, as op-
posed to more conventionalproducts. Most fundamentally,the buyer
of an idea is unable to judge how much informationis worth until it is
disclosedto him, but the seller has a difficulttime repossessingthe in-
formationin case the buyer decides, following disclosure, not to pay
very much for it. Partialdisclosure may avoid the total breakdownof
market-basedexchangein such situationsbut still leaves a residualin-
o10Michael Hammer, "ReengineeringWork: Don't Automate, Obliterate,"Harvard Busi-

ness Review (July/Aug. 1990): 104.
'02 Micklethwait and Wooldridge, The Witch Doctors, 29.
103 See James O'Shea and Charles
Madigan, Dangerous Company: The Consulting Pow-
erhouses and the Businesses They Save and Ruin (New York, 1997).
'04 For a general discussion, see Robert H. Frank and Philip J. Cook, The Winner-Take-
All Society (New York, 1995); for formal modeling and a discussion specific to the manage-
ment idea business, see Ghemawat, "Competitionamong Management Paradigms."

formation asymmetry.105 Performancecontractingis sometimes pro-

posed as an antidote to otherwiseineradicableinformationalproblems
of this sort, but its efficacyand use in the context of managementideas
seem to be limited by noisy performancemeasurement. Instead, the
market-basedtransferof ideas to companiesappearsto be sustainedby
mechanisms such as reputationand observationallearning. Based on
microtheoreticalanalysis,these mechanismsmay lead to "cascades"of
ideas, in which companiesthat choose late optimallydecide to ignore
their own informationand emulate the choices made earlierby other
companies.106Such fadlike dynamics can also enhance the sales of
productswith broad,as opposedto niche, appeal.107 And then there are
contractingproblemswithin, ratherthan between, firms that point in
the same direction.In particular,models of principal-agentproblems
show that managers,in orderto preserveor gain reputationwhen mar-
kets are imperfectlyinformed,may prefereitherto "hidein the herd"so
as not to be accountableor to "ridethe herd"in order to prove qual-
ity.?os The possible link to situations in which managers must decide
which, if any, new ideas to adopt should be obvious. More broadly,
demand-sideconsiderationssuggest some reasons to worryabout pat-
terns in the diffusionof new ideas as well as the incentives to develop
them in the first place.
Whethersuch worries about the performanceof marketsfor ideas
actuallymaketheir effects felt in the real world of managementis, ulti-
mately, an empiricalmatter. Unfortunately,the informationalimper-
fections noted above- and others, such as the difficulty of counting
ideas- complicatesystematicempiricalanalysis of productvarietyand
turnover in managementideas. A shared basis for understandingthe
historical evolution of ideas, which I have attemptedto providein the
specific context of competitivethinkingabout business strategy,is but
a first step in unravelingsuch complications.
105 See, for example, James J. Anton and Dennis A. Yao, "The Sale of Ideas: Strategic Dis-
closure, Property Rights, and Incomplete Contracts," unpublished working paper, Fuqua
School of Business, Duke University (1998).
106
See Sushil Bikhchandani, David Hirshleifer, and Ivo Welch, "Learningfrom the Be-
havior of Others: Conformity, Fads and Informational Cascades,"Journal of Economic Per-
spectives (1998): 15-70.
107 See Daniel L. McFadden and Kenneth E. Train, "Consumers'Evaluation of New Prod-
ucts: Learningfrom Self and Others,"Journal of Political Economy (Aug. 1996): 683- 703.
108 These models derive some of their real-world
appeal from the use of relative perfor-
mance measures to evaluate managers. See Robert Gibbons and Kevin J. Murphy, "Relative
Performance Evaluation of Chief Executive Officers,"Industrial and Labor Relations Re-
view (Feb. 1990o): 3oS-51S.


Competition and Business Strategy

Uploaded by

Copyright:

Available Formats

Competition and Business Strategy

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Competition and Business Strategy

Uploaded by

Copyright:

Available Formats

What is the main topic of the article?

What is the main topic of the article?

How does the author structure their analysis of the topic?

How does the author structure their analysis of the topic?

The President and Fellows of Harvard College

Competition and Business Strategy in Historical Perspective

This content downloaded from 128.135.12.127 on Fri, 16 Jan 2015 09:55:51 AM

Competition and Business Strategy in

This content downloaded from 128.135.12.127 on Fri, 16 Jan 2015 09:55:51 AM

This content downloaded from 128.135.12.127 on Fri, 16 Jan 2015 09:55:51 AM

The organizationalchallengesinvolvedin WorldWarII were a vital

4 Peter Drucker, The Practice of Management (New York, 1954), 11.

This content downloaded from 128.135.12.127 on Fri, 16 Jan 2015 09:55:51 AM

U.S. militaryafterWorldWarII. In this period,Americanmilitarylead-

5 Philip Selznick, Leadership in Administration (Evanston, Ill., 1957), 49-50.

This content downloaded from 128.135.12.127 on Fri, 16 Jan 2015 09:55:51 AM

policies "fittogether into a programthat effectivelymeets the require-

This content downloaded from 128.135.12.127 on Fri, 16 Jan 2015 09:55:51 AM

ssessment of Risk Identificationof

Figure 1. Andrews's Strategy Framework.(Source: Kenneth Andrews, The Concept of Corpo-

ence was held at Harvardthat helped diffuse the SWOTconcept in

This content downloaded from 128.135.12.127 on Fri, 16 Jan 2015 09:55:51 AM

opment of the enterprise"(emphasis added).'2When strategychoices

This content downloaded from 128.135.12.127 on Fri, 16 Jan 2015 09:55:51 AM

Present Product New Product

Present Mission MarketPenetration Product Development

New Mission MarketDevelopment Diversification

Figure 2. Ansoffs Product/Mission Matrix as adapted by Henry Mintzberg. (Source: Henry

be used to evaluate and compare many different types of businesses.

The Rise of Strategy Consultants

'8Adam M. Brandenburger,Michael E. Porter, and Nicolaj Siggelkow, "Competitionand

This content downloaded from 128.135.12.127 on Fri, 16 Jan 2015 09:55:51 AM

(BCG),founded in 1963, had a major impact on the field by applying

This content downloaded from 128.135.12.127 on Fri, 16 Jan 2015 09:55:51 AM

Decker.26As BCGconsultantsstudied these industries, they naturally

This content downloaded from 128.135.12.127 on Fri, 16 Jan 2015 09:55:51 AM

High Share Low Share

30Conley, "ExperienceCurves as a Planning Tool," 10-11.

This content downloaded from 128.135.12.127 on Fri, 16 Jan 2015 09:55:51 AM

High Medium Low

ness, or profitability,and anotherdozen to screen for competitiveposi-

31Interview with Mike Allen, 4 Apr. 1997.

This content downloaded from 128.135.12.127 on Fri, 16 Jan 2015 09:55:51 AM

In all these applications,segmenting diversifiedcorporationsinto

This content downloaded from 128.135.12.127 on Fri, 16 Jan 2015 09:55:51 AM

spond to those introducedby competitors."37Abernathyand Wayne

37William J. Abernathy and Kenneth Wayne, "Limits of the Learning Curve,"Harvard

This content downloaded from 128.135.12.127 on Fri, 16 Jan 2015 09:55:51 AM

proposing that successful companies devise progressive strategies to

41 Robert H. Hayes and William J. Abernathy, "ManagingOur Way to Economic Decline,"

This content downloaded from 128.135.12.127 on Fri, 16 Jan 2015 09:55:51 AM

tive position. Although these two dimensions had been identified

Unbundling Industry Attractiveness

43William G. Shepherd, "Causesof Increased Competition in the U.S. Economy, 1939-

This content downloaded from 128.135.12.127 on Fri, 16 Jan 2015 09:55:51 AM

definitive analysis-as well as analysis of oligopoly under specific

44Antoine A. Cournot, Recherches sur les Principes Mathematiques de la Theorie des

This content downloaded from 128.135.12.127 on Fri, 16 Jan 2015 09:55:51 AM

Industry Return on Equity

Figure 7. Differences in the Profitabilityof Selected Industries, 1971-1990. (Source:Anita M.