amit y zott 2008

Strategic Management JournalStrat. Mgmt. J., 29: 1–26 (2008)

Published online 22 August 2007 in Wiley InterScience (www.interscience.wiley.com) DOI: 10.1002/smj.642

Received 17 November 2004; Final revision received 2 July 2007

THE FIT BETWEEN PRODUCT MARKET STRATEGYAND BUSINESS MODEL: IMPLICATIONS FOR FIRMPERFORMANCE

CHRISTOPH ZOTT1* and RAPHAEL AMIT2

1 Insead, Fontainebleau, France2 The Wharton School, University of Pennsylvania, Philadelphia, Pennsylvania, U.S.A.

We examine the fit between a firm’s product market strategy and its business model. We developa formal model in order to analyze the contingent effects of product market strategy and businessmodel choices on firm performance. We investigate a unique, manually collected dataset, and findthat novelty-centered business models—coupled with product market strategies that emphasizedifferentiation, cost leadership, or early market entry—can enhance firm performance. Ourdata suggest that business model and product market strategy are complements, not substitutes.Copyright 2007 John Wiley & Sons, Ltd.

INTRODUCTION

A central objective of strategic management re-search has been to understand the contingenteffects of strategy on firm performance. Contin-gency theory suggests that there is no optimalstrategy for all organizations and posits that themost desirable choice of strategy variables altersaccording to certain factors, termed contingencyfactors (Donaldson, 1996). Accordingly, strategicmanagement scholars have examined a wide rangeof contingency factors, such as aspects of theenvironment, organization structure (Miller, 1988),technology (Dowling and McGee, 1994), and mar-keting choices (Claycomb, Germain, and Droege,2000), and explored how these and other factorsinteract with strategy variables to determine firmperformance.

One focus of that literature considers structuralforms as contingency factors. An important early

Keywords: product market strategy; business model; per-formance; contingency theory; competitive strategy∗ Correspondence to: Christoph Zott, INSEAD, Boulevard deConstance, 77305 Fontainebleau Cedex, France.E-mail: [email protected]

contribution was made by Chandler (1962), whoconsidered the contingency relationship between afirm’s corporate strategy and its internal admin-istrative structure (specifically, divisional versusfunctional form). While this particular pair ofstrategy/structure variables has been thoroughlyaddressed (e.g., see Amburgey and Dacin, 1994),the literature has otherwise paid surprisingly ‘lit-tle attention to extending the question of strat-egy/structure fit issues for other structural formsof organization’ (Yin and Zajac, 2004: 365). Inthis paper, we address this gap by introducing thefirm’s business model as a new contingency factorthat captures the structure of a firm’s boundary-spanning exchanges and asking: How do the firm’sbusiness model and its product market strategyinteract to impact firm performance?

We address this question by elaborating on thebusiness model, which is a relatively new, yetrich and potentially powerful concept in the strat-egy literature. The business model is a struc-tural template that describes the organization ofa focal firm’s transactions with all of its exter-nal constituents in factor and product markets.It has been brought to the forefront of strategic

Copyright 2007 John Wiley & Sons, Ltd.

2 C. Zott and R. Amit

management thinking, and has become a partic-ularly important new contingency factor throughrecent rapid advances in information and commu-nication technologies—in particular, Internet andbroadband technologies—that have facilitated newtypes of technology-mediated interactions betweeneconomic agents (Geoffrion and Krishnan, 2003).These developments have enabled firms to changefundamentally the ways they organize and transactboth within and across firm and industry bound-aries (Mendelson, 2000). They have also enabledan emerging approach to enterprise-level design,as Nadler and Tushman (1997) have asserted. Thatapproach spawns ‘new designs that extend beyondthe corporation’s traditional outer walls,’ and ithelps managers ‘recognize the untapped opportuni-ties for competitive advantage that lie within theirown organizations’ (Nadler and Tushman, 1997:120). Thus, the focus of organization design seemsto have shifted from the administrative structureof the firm to the structural organization of itsexchanges with external stakeholders. Echoing thisshift, researchers have observed that the locus ofvalue creation increasingly extends traditional firmboundaries (Dyer and Singh, 1998; Gulati, Nohria,and Zaheer, 2000; Normann, 2001), and they havetherefore called for a broader conceptualization oforganizational boundaries beyond the legally rele-vant demarcation of the firm from its environment(Santos and Eisenhardt, 2005). The business modelrepresents this kind of broader concept.

The study of business models is an importanttopic for strategic management research becausebusiness models affect firms’ possibilities for valuecreation and value capture (Amit and Zott, 2001).Since product market strategies are also chosen,or emerge, in order to increase value creationand capture by firms, researchers and managersneed to know how business models and productmarket strategy, independently as well as jointly,impact the performance prospects of firms. In orderto improve our understanding of the contingenteffects of business model and product market strat-egy, we first examine conceptually how a firm’sbusiness model is distinct from its product marketstrategy. We then investigate analytically how var-ious product market strategies and business modelchoices interact to affect firm performance. Byexamining a unique manually collected dataset onbusiness strategy and business models, we estab-lish empirically that a firm’s product market strat-egy and its business model are distinct constructs

that affect firm performance. Specifically, we findthat novelty-centered business models, coupledeither with a differentiation or cost leadership strat-egy, enhance firm performance measured as themarket value of a firm’s equity. In addition, weascertain that a novelty-centered business modeltogether with early entry into a market have a pos-itive effect on performance.

This study attempts to make several contribu-tions to the strategy literature. First, it extendsthe scholarly perspective of structure as an impor-tant contingency factor from a concern with theadministrative structure of the firm to a focus onthe pattern of transactions the focal firm enableswith external stakeholders. Second, we argue the-oretically, and show empirically, that the busi-ness model is a valid and distinct construct fromreceived notions of a firm’s product market strat-egy. This, we believe, is a particularly importantpoint, because the business model, as a source ofvalue, can help explain why some firms outperformothers; it provides a rationale for value creation andappropriation that is distinct from (but may interactwith) the firm’s product market strategy. Third, weanalyze the contingent effects of business modeland strategy on firm performance, and we testthem empirically. We show that novel businessmodels can augment the competitive advantagerealized through superior product market strate-gies. That is, both product market strategy andstructure, as embodied by the business model, canenhance the firm’s competitive advantage indepen-dently as well as jointly, and therefore complementeach other.

The next sections present our theory and model,and explain the data and methods used to test themodel. We then present our results, and concludewith a discussion of our findings and implicationsfor future research.

THEORY

The contingency relationship of strategyand structure

Contingency theory seeks to understand the behav-ior of a firm by analyzing separately its constituentparts, making disaggregated one-to-one compar-isons of variables and their links with performance(Meyer, Tsui, and Hinings, 1993). A prominentconcern among contingency theorists has beento explore variables related to the strategy and

Copyright 2007 John Wiley & Sons, Ltd. Strat. Mgmt. J., 29: 1–26 (2008)DOI: 10.1002/smj

The Fit Between Product Market Strategy and Business Model 3

structure of firms (e.g., Doty, Glick, and Huber,1993; Galbraith, 1977; Miles and Snow, 1978;Mintzberg, 1979), and to examine their contin-gent effects on firm performance. For example,in his study of large American corporations andtheir approaches toward product market diversifi-cation, Alfred Chandler (1962) observed that majorincreases in volume, geographic dispersion, andvertical and horizontal integration of firms werefollowed by changes in their administrative activ-ity, which eventually led to the emergence of theM-form of organization. That line of reasoning,however, provoked the counterargument that ‘strat-egy follows structure’ (e.g., Bower, 1970), whichwas predicated on the logic that managerial cog-nition and skills mediate between structure andstrategy. The ensuing debate on the contingentrelationship between strategy, structure, and firmperformance flourished in the 1970s and 1980s.It has subsequently been revived through a closerempirical examination of dynamics and causal-ity (Amburgey and Dacin, 1994) and calls for anextension of the analysis to various forms of strat-egy and structure that had not previously beenconsidered (Nadler and Tushman, 1997; Yin andZajac, 2004).

In this paper, we seek to enrich the debateon the strategy/structure fit by shifting the focusfrom corporate to product market strategy, andby focusing on a structural construct that cap-tures the firm’s transactions with external parties,namely, the firm’s business model. We view prod-uct market strategy as the way in which a firmchooses to position itself against competitors inits addressable market spaces. We concentrate onsome salient aspects of a firm’s product marketstrategy, namely those decisions that affect themain drivers of customer demand: price, quality,and timing (Besanko, Dranove, and Shanley, 1996;Porter, 1985). A firm can leverage these drivers bymaking two fundamental strategic decisions: whattype of product market positioning approach toadopt, i.e., cost leadership and/or product/servicedifferentiation (Porter, 1985); and when to enterthe market (Lieberman and Montgomery, 1988).The answers to these questions are central to ourunderstanding of how firms that operate in com-petitive product markets create and appropriatevalue.1

1 There are many more possible facets of a firm’s productmarket strategy that could be explored in future research; for

The business model: a new structural concept

Technological progress has brought new opportu-nities for the creation of organizational arrange-ments (business models) among firms, partners,and customers (Geoffrion and Krishnan, 2003;Mendelson, 2000; Normann, 2001). The businessmodel is a structural template of how a focal firmtransacts with customers, partners, and vendors;that is, how it chooses to connect with factor andproduct markets. It refers to the overall gestaltof these possibly interlinked boundary-spanningtransactions. Consider the case ofPriceline.com Inc., a provider of an electronic pric-ing system known as a demand collection sys-tem (Hann and Terwiesch, 2003). Transactions areenabled through a reverse market auction mecha-nism for which the company has secured a businessmethod patent, indicating that the business modelis fairly innovative. Priceline allows customers toname the price at which they wish to transact, andthe company attempts to find a provider of theproduct or service within a specified price range.Priceline’s novel business model enables buyersto save money on a wide range of products andservices by trading flexibility about the choiceof brands, product features, timing, convenience,and/or sellers in return for prices that are lowerthan those charged through traditional retail chan-nels. At the same time, Priceline enables sellersto generate incremental revenue by disposing ofexcess inventory or capacity at prices lower thanthey offer through other channels, while protectingtheir brand.

The business model can then be defined as‘the structure, content, and governance of trans-actions’ between the focal firm and its exchangepartners (Amit and Zott, 2001 : 511).2 It representsa conceptualization of the pattern of transactionallinks between the firm and its exchange partners.3

example, the firm’s product offering, served customer segments,or geographic markets.2 There are other definitions of the term business model, forexample, those that define it as the way a firm generates revenues(for an overview, see Ghaziani and Ventresca, 2005). For thepurpose of this article, however, we rely on the definitionproposed by Amit and Zott (2001), and on the distinctionbetween business and revenue models: a revenue model refers tothe specific modes in which a business model enables revenuegeneration.3 The business model construct is distinct from the value netstrategic analysis framework developed by Brandenburger andNalebuff (1996). The players in the value net, such as competi-tors and certain complementors, may or may not be part of the



Business models can be characterized by theirdesign themes, which capture the common threadsthat orchestrate and connect the focal firm’s trans-actions with external parties. The design themesdescribe the holistic gestalt of a firm’s businessmodel, and they facilitate its conceptualization andmeasurement. They are not mutually exclusive:several design themes may be present in any givenbusiness model. In this paper we focus on noveltyand efficiency as design themes (Zott and Amit,2007), because they are the corresponding themes(on the business model level) to product differen-tiation and cost leadership (on the product marketstrategy level), and thus are the appropriate con-tingency factors to consider. This choice of designthemes suits our theoretical purpose of exploringthe fit between business model and product marketstrategy.

Novelty-centered business models refer to newways of conducting economic exchanges amongvarious participants. The conceptualization andadoption of new ways of conducting transac-tions can be achieved, for example, by connectingpreviously unconnected parties, by linking trans-action participants in new ways, or by design-ing new transaction mechanisms (like Priceline).Efficiency-centered business models refer to themeasures firms may take to achieve transactionefficiency (i.e., reduce transaction costs for allparticipants); they do not refer to the outcome(efficiency) itself. The essence of an efficiency-centered business model is the reduction of trans-action costs (Williamson, 1975). This reductioncan derive from the attenuation of uncertainty,complexity, or information asymmetry, as wellas from reduced coordination costs and transac-tion risk. An example of an efficiency-centereddesign element would be the order-tracking fea-ture in Amazon’s business model, which is aimedat enhancing transaction transparency, and thus atincreasing efficiency. As a whole, Amazon’s busi-ness model is both efficiency- as well as novelty-centered, which illustrates the point made abovethat the design themes are not orthogonal.

The business model can be a source of com-petitive advantage that is distinct from the firm’sproduct market position (Christensen, 2001). Firmsthat address the same customer need, and pur-sue similar product market strategies, can do so

business model because some of them may not transact with thefocal firm.

with very different business models. Consider, forexample, the market for navigation software fordevices such as personal digital assistants, hand-held computers, or smart phones. Some firms inthat space offer non-wireless solutions directly tothe end-user in a one-shot transaction, while others,like the French company Webraska, offer wire-less navigation solutions that can be sold throughwireless carriers, and that require a very distinctset of ongoing exchanges between the firm, endusers, and the wireless carriers (Zott and Bancerek,2004). A firm with a distinct business model thatcreates more value than that of its rivals holds apotential advantage. All other things being equal,it has the possibility to capture more value for itsshareholders. Consequently, a business model mayaffect a firm’s performance outcomes, as does afirm’s product market strategy, and therefore itscontingent effects on product market strategy needto be considered. In Table 1 we contrast businessmodel and product market strategy.

Table 1 shows that product market strategy dif-fers from the business model mainly through itsfocus on the positioning of the firm vis-a-vis itsrivals, whereas the business model is a struc-tural construct that centers on the pattern of thefirm’s economic exchanges with external partiesin its addressable factor and product markets. Toillustrate this distinction, consider four firms (alltaken from our sample, on which we elaboratein the methods section below) that have adopteddifferent business model and product market strat-egy configurations: Priceline, NetBank, Didax, andMultex. Earlier we described how Priceline haschosen a product market strategy of cost lead-ership, and a business model centered on nov-elty. This choice can be contrasted with that ofanother firm—NetBank, an online bank—whichhas combined its cost leadership strategy with amore efficiency- than novelty-centered businessmodel. In terms of cost leadership strategy, Net-Bank clearly aims at providing cost-effective bank-ing services. It does not incur the cost of sup-porting a branch system, keeps its overhead low,and partners with outside providers of specializedservices and technologies who enjoy economiesof scale. Moreover, NetBank’s business model isefficiency-centered. NetBank enables fast trans-action processing, reduces customer search andinformation costs by providing rates and fees com-parisons, and provides lenders with information



Table 1. Business model and product market strategy

Business model Product market strategy

Definition A structural template of how a focal firmtransacts with customers, partners, andvendors. It captures the pattern of thefirm’s boundary spanning connectionswith factor and product markets

Pattern of managerial actions that explains howa firm achieves and maintains competitiveadvantage through positioning in productmarkets

Main questionsaddressed

How to connect with factor and productmarkets

What positioning to adopt against rivals

Which parties to bring together toexploit a business opportunity, andhow to link them to the focal firm toenable transactions (i.e., whatexchange mechanisms to adopt?)

What kind of generic strategy to adopt (i.e.,cost leadership and/or differentiation)?

What information or goods to exchangeamong the parties, and what resourcesand capabilities to deploy to enablethe exchanges?

When to enter the market?

How to control the transactions between What products to sell?∗

the parties, and what incentives to What customers to serve?∗

adopt for the parties? Which geographic markets to address?∗

Unit of analysis Focal firm and its exchange partners Firm

Focus Externally oriented: focus on firm’sexchanges with others

Internally/externally oriented: focus on firm’sactivities and actions in light of competition

∗ Elements of product market strategy marked with an asterisk are not addressed in this paper. They could be addressed in futurework.

about account registrations so that they can tai-lor their offerings better to customer preferences.At the time of our measurement, NetBank’s busi-ness model was less novelty-centered. Althoughthe firm had been a pioneer with its online bankingbusiness model, the model had already been copiedby competitors like Wingspanbank.com, and Net-Bank had not managed to sustain the innovativeedge of its business model. So, while Pricelineand NetBank both pursue cost leadership strate-gies, they have emphasized different design themesfor their respective business models.

Now consider Didax and Multex, firms that havecoupled a product differentiation strategy with dif-ferent choices of business models. Didax (lateracquired by Salem Communications and operat-ing as Crosswalk.com) was an online portal forChristians and Christian-related institutions thatoffered its clients products and services such asconsulting and IT management. The firm aimedat product differentiation by constantly develop-ing and marketing new services. For example, itintroduced real-time forums across new topicalareas, such as Christian life. At the same time,its business model was novelty-centered, because

it brought together a new range of participants(individuals, businesses, churches, nonprofit orga-nizations) who had novel incentives to participateand do business on a community portal, namelythose based on Christian values and the desire tosupport charity.

Didax can be compared and contrasted withMultex, a firm that combined a product differenti-ation strategy with an efficiency-centered businessmodel. Multex offered and distributed financialinformation and research via Web platforms toover 25,000 companies around the world (Reutersacquired the company in 2003). Multex clearlypursued a product differentiation approach: thefirm constantly introduced new service features,such as easy-to-read report formats, and it devel-oped proprietary technology to enhance its ser-vices (e.g., software to distribute research reportsquickly to specific authorized users). Its busi-ness model, however, was that of an efficiency-centered financial information integrator and dis-tributor. The transactions the firm offered to itsclients enabled fast access to complex information.The transactions were simple, mass-customized,and enabled the firm to reach a vast pool of



geographically dispersed clients—indicators of anefficiency-centered business model.

The conceptual distinction between the businessmodel and the product market strategy of firms,illustrated by these examples, leads us to proposea corollary, which we state explicitly because theconceptual arguments that support the conjecturethat business models and product market strategiesdiffer are relatively new (e.g., see Magretta, 2002),may not be widely known or accepted, and untilnow have not been empirically established.

Corollary: Business models (as, for example,measured by design themes) are distinct fromproduct market strategies (as, for example, mea-sured by generic strategies).

Having established the distinction between thebusiness model and product market strategy con-structs, what is the fit between them, and what arethe implications of it? Contingency theory impliesthat organizational effectiveness (measured, forexample, in terms of firm performance) is afunction of the fit between contingency factors.According to Galbraith (1977: 6) fit, or ‘coher-ence,’ ‘is the primary determinant of success.’ Forexample, it can be argued that alignment betweena firm’s administrative structure and its diversifica-tion strategy has positive implications on firm per-formance (Chandler, 1962). Recent research intothe relationship between strategy and structure hasconfirmed a moderating effect of these constructson firm performance (Mintzberg, 1990; Siggelkowand Levinthal, 2003). This research has highlightedthe usefulness of examining interactions betweensalient dimensions of strategy and structure on firmperformance. It has also established that alignmentbetween these factors could be expected to resultin higher performance.

The fit between product market strategy andbusiness model

To evaluate the implications of business modeland product market strategy on firm perfor-mance, we consider two main business modeldesign themes—novelty-centered and efficiency-centered—along with three product marketstrategy choices: cost leadership, differentiation(Porter, 1985), and the timing of entry into a mar-ket (Lieberman and Montgomery, 1988). As with

business model design themes, these product mar-ket strategy choices are not mutually exclusive,nor are they exhaustive. For example, a firm’smanagers could choose to pursue simultaneously astrategy of product differentiation, cost leadership,and early market entry.

Which business model fits best with the firm’schoice of product market strategy? Or, to put itanother way, what constitutes a good fit betweenthese constructs? The literature generally considerscoherent configurations of design elements thatmanifest themselves as peaks in the performancelandscape as good fit (Siggelkow, 2001). Twodesign elements (A and B) fit well if there arecomplementarities between them; that is, if themarginal benefit of A increases with the level of B,and if the levels of A and B are adjusted optimallyto achieve a local performance optimum (Milgromand Roberts, 1995).

We have developed a formal model that allowsus to investigate which combinations of businessmodel design themes and product market strate-gies fit well. This model helps us to theorize aboutthese relationships in a more structured and rig-orous way than would be possible through verbaltheorizing. It is useful because there has been littleprior theorizing on business models on which wecould draw in the development of our theory. Ourobjective, however, is not to derive a fully spec-ified model and closed-form analytical solutions.We intend to provide theoretical guidance for ourempirical analysis.

We build on the model developed by Branden-burger and Stuart (1996) for value creation in asimple static setting with one firm, one customer,and one supplier. We adapt this model to focuson the transactions that a business model enables,rather than on a particular product or service. LetP(t) be the price that a customer pays for a goodor service acquired in transaction t , or for the rightto participate in the transaction. The focal firm hasadopted a business model of type d, where d isa vector describing the extent to which the busi-ness model emphasizes the design themes noveltyand efficiency. It has also adopted a product mar-ket strategy s, where s is a vector describing theextent to which the firm emphasizes differentia-tion, cost leadership, and early entry timing. Forsimplicity, denote that firm as F d

s ≡ F . Denote thefocal firm’s suppliers and partners (other than cus-tomers) as i, where i is an index ranging from 1to I , the total number of suppliers and partners



in the business model. Let Ri(t) be the revenuesthat focal firm F gets from partner i in a particulartransaction t . Let Ci(t) denote the flow of revenuesfrom F to i, and let CF (t) be F ’s costs of provid-ing its own resources, such as financial capital orintellectual capital (e.g., patents, trademarks). Thenthe value appropriated by firm F in transaction t

can be expressed as:

VF(t) = P(t) + �iRi(t) − �iCi(t) − CF (t) (1)

The total value appropriated (TVA) by firm F isthe value appropriated in all types of transactionst that the business model enables, where t is anindex ranging from 1 to T , and T denotes thenumber of transaction types, n(t) is the averagenumber of transactions of type t conducted:

TVA = �t [VF(t) × n(t)] (2)

Inserting Equation 1 into Equation 2 yields:

TVA = �t

{[P(t) + �iRi(t)

−�iCi(t) − CF (t)] × n(t)

}(3)

Table 2 summarizes the variables of the model.TVA as a proxy for firm F ’s performance is con-tingent on F ’s business model, d, and its productmarket strategy, s. If d and s are choice vari-ables of the firm, their impact on each term onthe right-hand side of Equation 3 must be consid-ered in order to understand their collective impacton TVA. Following Siggelkow (2002), a usefulthought experiment for evaluating the fit betweena particular business model design theme and aparticular product market strategy is to considerwhether the marginal value of a business modeldesign theme would be affected (in particular,whether it would increase) if a firm were to putmore emphasis on the respective product market

Table 2. Summary of variables

Variable Meaning Affected by which business modeldesign theme/product market strategy

Expected sign ofeffect on variablea

t Index denoting the transaction type(t = 1, . . . , T )

T Total number of transaction types Business model novelty +P(t) Price that a customer pays for a good acquired Business model novelty +

in transaction t , or for the right to participate Differentiation strategy +in the transaction Cost leadership strategy −

Early market entry timing +F Focal firm

i Index denoting F ’s suppliers and partners(i = 1, . . . , I )

I Total number of F ’s suppliers and partners

Ri(i) Revenues that F gets from supplier/partner i ina particular transaction t

Ci(t) Costs for focal firm F of collaborating with Business model novelty −supplier/partner i in a particular transaction t Business model efficiency −

Cost leadership strategy −Early market entry timing −

CF (t) Focal firm’s costs for providing its own Business model efficiency −resources for transaction t Cost leadership strategy −

VF (t) Value appropriated by firm F in transaction t

n(t) Average number of transactions of type t Business model efficiency +Cost leadership strategy +Early market entry timing −

TVA Total value appropriated by firm F

a That is, expected sign of the effect of a marginal increase in emphasis on the respective business model design theme or productmarket strategy on the variable.



strategy (or vice versa). This thought experimentis consistent with the definition of fit as indicativeof complementarity (Milgrom and Roberts, 1995;Siggelkow, 2001). From this we go on to explorethe marginal effects of business model designthemes and product market strategies on TVA.

Novelty-centered business model and TVA

A novelty-centered business model refers to theconceptualization and adoption of new ways ofconducting economic exchanges among transac-tion participants. Novelty primarily aims at cre-ating new types of transactions, i.e., increasingT . It also strengthens the focal firm’s bargain-ing power vis-a-vis other business model stake-holders. The focal firm is the innovator, and itsbusiness model is the locus of innovation. Thehigher the degree of business model novelty, thehigher the switching costs for the focal firm’s cus-tomers, suppliers, and partners, as there may notbe readily available alternatives to doing businesswith the focal firm (Zott and Amit, 2007). Conse-quently, stronger emphasis on a novelty-centeredbusiness model will have a positive effect on P(t)

and will exert downward pressure on Ci(t) due tothe increased bargaining power of the focal firm.Hence, a marginal increase in a firm’s emphasis ona novelty-centered business model may affect TVAin Equation 3 through T (+), P (+), and Ci(−).4

Next, we examine the marginal effect on TVA ofchanging a particular product market strategy, andfollow that with an analysis of the impact of such achange on the marginal value of a novelty-centeredbusiness model.

First, consider product market differentiation.A stronger emphasis on differentiation will influ-ence customers’ willingness to pay positively, andtherefore make it easier for the focal firm tocharge higher prices to customers, P(t).5 Hence, amarginal increase in a firm’s emphasis on differ-entiation may affect TVA mainly through P(+);

4 The sign in parentheses gives the expected direction of changein the respective variable from a marginal increase in novelty-centered business model design: e.g., T (+) means that Tincreases in novelty-centered design.5 Product market differentiation may also engender higher costs,Ci(t) and CF (t), but the overall effect could still be performance-enhancing, as Porter (1985: 14) has observed: ‘A firm that canachieve and sustain differentiation will be an above averageperformer in its industry if its price premium exceeds the extracosts incurred in being unique.’ We thank a reviewer for pointingout this insight.

this is how product market differentiation indepen-dently affects firm performance.

In addition, we consider the interaction betweendifferentiation strategy and novelty-centered busi-ness model design. A focus on innovation in mul-tiple domains (business model, product marketstrategy) may be mutually reinforcing, for exam-ple, by harnessing the creative energy of managersand employees, and by delivering new productsand services to customers in new ways, therebyincreasing the number of transaction types (T ).Moreover, a firm that focuses all its activities andtransactions on innovation may become an evenmore skillful innovator over time (Zott, 2003). Atthe same time, as Hargadon and Douglas (2001)have suggested, stakeholder acceptance of a prod-uct innovation hinges on whether or not the inno-vation is presented and brought to market in famil-iar ways. Business model novelty, however, doesnot preclude familiarity. A novelty-centered busi-ness model could rely on familiar design elements,combined in new ways that mediate betweenthe innovative features of a differentiated prod-uct offering and the expectations, norms, and rulesof the institutional environment. The developmentof an online Christian community in the case ofDidax, for example, represented a novel approachto business model design that relied on designelements (e.g., the concept of a Christian commu-nity) with which the users were already familiar.These elements facilitated customers’ acceptanceof an innovative product and service offering. Onbalance, we could expect a positive joint effecton TVA: a marginal increase in the degree ofproduct market differentiation will strengthen themarginal performance benefit of business modelnovelty (and vice versa).

Second, consider cost leadership. A strongeremphasis on cost leadership implies lower pricescharged to customers, P(t), as well as lower inputand production costs, Ci(t) and CF (t) (Porter,1985). Furthermore, new segments with customershighly sensitive to price can be addressed, and cus-tomers within given segments will be motivated toincrease their number of repeat transactions, thusraising n(t).6 In other words, a marginal increasein a firm’s emphasis on cost leadership may affect

6 This argument is based on one of the following two assump-tions: (1) a marginally increased emphasis on cost leadershipdoes not affect the (perceived) quality of the product and hencedoes not lead to a lower willingness to pay (WtP) by customerswho, as a result, might abandon the firm; (2) if, however, the



TVA through P(−), Ci(−), CF (−), and n(+); thisis how cost leadership independently affects firmperformance.

Moreover, a greater emphasis on cost leader-ship can also, on balance, enhance the marginaleffect of a novelty-centered business model onTVA. A more pronounced cost leadership approachinteracts positively with the firm’s strengthenedbargaining power over its suppliers through anincreased novelty-centered business model as itputs additional downward pressure on Ci . Again,new customers will have distinct reasons to bedrawn to the firm and engage in transactions withit—business model novelty and low cost—whichenhances the positive impact of a novelty-centeredbusiness model on n. Could there also be apotential substitution effect? Different stakehold-ers within the organization might work at cross-purposes, some trying to lower cost, others focus-ing on creating a novel business model. Firms likePriceline and Dell show that this need not be thecase: firms can have novel business models thatsupport and enhance the effectiveness of their costleadership strategies. Organization members canimplement low-cost strategies within a given novelbusiness model; they do not necessarily representconflicting objectives. Customers, too, may wel-come low prices, and in return accept the newways of doing business that help achieve theselow prices (e.g., in Priceline’s case, the reverseauction to sell airline tickets; and in Dell’s case,phone ordering and the tailoring of computers tocustomers’ needs). Conversely, adopting these newways of conducting transactions is helped by lowercosts, which give customers an increased incen-tive to adapt their behavior. Therefore, overall, wecould expect a positive joint effect of cost leader-ship and novelty-centered business model on TVA.

Third, consider timing of market entry. Firmsthat enter markets earlier may enjoy considerableadvantages arising from the creation of customerswitching costs, brand awareness, and reputation,thus allowing these firms (at least initially) tocharge higher prices P(t) (Lieberman and Mont-gomery, 1988). Early market entrants can also gainby learning (Arrow, 1962), accumulating propri-etary knowledge (Dierickx and Cool, 1989), and

WtP of customers declines, with the risk that some customersmight be lost that way, a lower price, P(t), compensates forthis loss, so that on average n(t) (the number of transactions)still increases. We thank an anonymous reviewer for pointingthis out.

by preempting scarce resources, thus lowering thecosts of resources, Ci(t) (Lieberman and Mont-gomery, 1988). As the early entrant attempts toaddress—and perhaps even create—a new mar-ket, the number of transactions, n, is initially likelyto be limited (although it could become high lateron (eBay is a case in point). In other words, amarginal increase in a firm’s emphasis on earlymarket entry timing may affect TVA through P(+),Ci(−), and n(−); this is how the timing of marketentry independently affects firm performance.7

A greater emphasis on early market entry canalso, on balance, enhance the marginal effect ofa novelty-centered business model on TVA. Mov-ing into a market earlier allows the firm to capturethe rents from business model innovation, whichcan be considered entrepreneurial rents, i.e., rentsthat accrue between the introduction of an innova-tion and its diffusion (Rumelt, 1987). In particu-lar, the positive effect of a novelty-centered busi-ness model on P may be more pronounced, andmore sustainable if the firm enters a market early.Although there could be a concern that early mar-ket entry coupled with a novelty-centered businessmodel might represent a high hurdle for achievingstakeholder (in particular, customer) acceptance,novel business models can include familiar designelements. Hence, on balance we could expect apositive joint effect on TVA.

In summary, this analysis suggests that couplinga novelty-centered business model with a productmarket strategy of differentiation, cost leadership,or early market entry represents good fit.

Efficiency-centered business model and TVA

An efficiency-centered business model aims atreducing transaction costs for all transaction par-ticipants. This explains the likely negative effectsof a marginal emphasis in such business modeldesign on CF (t) and Ci(t). By reducing transac-tion costs, an efficiency-centered business modelmay also lead to higher transaction volume, n(t);more new customers will be drawn to transact withthe focal firm, and existing customers may transactmore frequently as a result of the lowered transac-tion costs. Hence, a marginal increase in a firm’s

7 As our arguments suggest, the strength of these effects maydepend on the time when they are measured—right at the earlymarket entry, or sometime afterwards. Although our model doesnot contain time as a variable, it could be extended in thatdirection.



emphasis on an efficiency-centered business modelmay affect TVA through CF (−), Ci(−), and n(+).8

To evaluate the possible fit between an effic-iency-centered business model and a particularchoice of business strategy, we examine whetherthe marginal value of an efficiency-centered busi-ness model would increase if a firm were to putmore emphasis on a particular product marketstrategy.

First, consider the strategy of differentiation.As shown above, a marginal increase in a firm’semphasis on product market differentiation mayaffect TVA independently primarily through P(+).It is not clear whether and how product dif-ferentiation would affect the marginal effect ofan efficiency-centered business model on TVAthrough CF , C, and n. Hence, the joint effectof differentiation and an efficiency-centered busi-ness model on TVA is likely to be indetermi-nate.

Second, consider cost leadership. We have seenthat a marginal increase in a firm’s emphasison cost leadership may affect TVA independentlythrough P(−), Ci(−), CF (−), and n(+). A focuson low costs in multiple domains (business model,product market strategy) may be mutually reinforc-ing by focusing managers’ efforts on cost savingsacross transactions, products, and processes, andby delivering low-cost products and services tocustomers in low-cost ways, thereby reinforcingthe marginal effects of efficiency-centered designon TVA through CF and n. Moreover, a firm thatfocuses on cost reductions in all its activities andtransactions may become even more skillful atreducing costs over time, thus decreasing CF evenfurther. Hence, we might expect a positive jointeffect of cost leadership and an efficiency-centeredbusiness model on TVA.

Third, consider timing of market entry. A margi-nal increase in a firm’s emphasis on early mar-ket entry timing may affect TVA independentlythrough P(+), CF (−), and n(−). However, it isnot clear whether and how early market entrytiming, on balance, affects the marginal benefitof an efficiency-centered business model. On theone hand, one could argue that early adoptersare important for early market entrants, and thosemight not be that price-sensitive; efficiency might

8 Again, the sign in parentheses gives the expected direction ofchange in that variable from a marginal increase in the degreeof efficiency-centered business model design.

not be so important to them. On the other hand,efficiency-centered business model design refersto efficient transactions (e.g., lower search costsfor customers) and does not necessarily involvelower price tags for products or services. Early(product) adopters might enjoy and value increasedtransaction efficiency. Therefore the joint effectof early market entry and an efficiency-centeredbusiness model on TVA is likely to be indetermi-nate.

In summary, this analysis suggests that couplingan efficiency-centered business model with a prod-uct market strategy of cost leadership representsgood fit, whereas the fit with either a product mar-ket strategy of differentiation or with early marketentry cannot be clearly predicted.

DATA AND METHODS

Sample

We collected data on a sample of firms thathad gone public in Europe or in the UnitedStates between April 1996 and May 2000. Thissample selection strategy enabled us to create adataset of about 300 firms that conducted part oftheir business over the Internet (e.g., firms likeeTrade, Guess, and Priceline), and served as fer-tile ground to investigate Internet-enabled businessmodels. These firms were likely to experimentwith, and take advantage of, the possibilities thatadvanced information and communication tech-nologies offered for the design of their businessmodels. Consistent with the observation that suchbusiness models span firm and industry boundaries(Amit and Zott, 2001), we constructed a cross-industry sample.

We randomly sampled 170 firms on their busi-ness model characteristics and product marketstrategies. We considered companies that had re-cently gone public because at the time we con-ducted the study there were not many establishedfirms in the public domain that used the Inter-net to enable their business models. Also, rela-tively young firms have fewer lines of businessthan older, more established corporations, and theirbusiness models are easier to describe and mea-sure as they involve fewer transaction types andexchange partners. Data collection from initialpublic offering documents also ensured the avail-ability and consistency of the data on business



models and business strategies. This is an acknowl-edged method for studying firms’ strategies (e.g.,Dowling and McGee, 1994).

Data collection

The data collection proceeded in two stages. In thefirst stage, we built composite scales for businessmodel design themes, and we identified and mea-sured the relevant items on the basis of a contentanalysis of IPO prospectuses. In the second stage,we followed a similar procedure to build compos-ite scales for, and measure, relevant dimensions ofproduct market strategies.

To determine the scales for the business modeldesign themes, we relied on measurement scalesdeveloped by Zott and Amit (2007). To collect thedata, we hired 11 part- or full-time research assis-tants (primarily MBA students), and trained themas expert raters to analyze assigned sample com-panies. We thus built on the common techniqueof using expert panelists in management research(see, for example, MacCormack, Verganti, and Ian-siti, 2001). We carefully selected our raters andtrained them in data collection and data analysis.On average, it took a rater about 2 days to col-lect data on a given business model, to understandthe model, and to assess it. Data sources includedprimarily IPO prospectuses (Dowling and McGee,1994), but also annual reports, investment analysts’reports, and Web sites. The business model datawere collected from May 2000 to June 2001. Dur-ing that period, we were able to take one measure-ment of the design themes for each of the businessmodels in our sample. The lack of readily avail-able data on business models obliged us to drawon primary sources of data and construct a unique,manually collected dataset; it also prevented usfrom collecting time-series data, which are prefer-able in studies that can draw on secondary sourcesof data (Bowen and Wiersema, 1999).

We validated inter-rater reliability by assigninga randomly chosen business model to two differentexpert raters, and by conducting a pair-wise com-parison of responses, yielding a Cronbach alpha of0.81, and a Pearson correlation coefficient of 0.72.Raters were in broad agreement with each otherfor 82 percent of the individual items. We repeatedthe test periodically, and found that all indicatorsof reliability improved further. We also purifiedthe efficiency scale by dropping two items (see

Appendix), which resulted in an improved relia-bility of the measure.

With respect to product market strategy scales,we drew on the strategy and management literatureto establish measures of product market position-ing through differentiation, cost leadership, andtiming of market entry. We found that most of theempirical work on Porter’s (1985) generic strate-gies, for example, had been conducted on thebasis of surveys administered to managers. A fewresearchers (e.g., Dowling and McGee, 1994) haveused IPO prospectuses to measure these items.We then adapted these survey-based instrumentsin order to analyze the content of our primary datasource.

We iteratively selected items to measure prod-uct market strategy dimensions. We started with51 items, derived from the literature, that measuredvarious aspects of generic firm-level strategy. Afterpilot testing these items on our sample firms, werefined some items and dropped others, mainly onthe basis of data availability. As a result of this pro-cess, and following further scale purification basedon an exploratory factor analysis, we retained threeitems referring to differentiation and four itemsreferring to cost leadership. We also retained asingle-item measure for market-entry timing. Tworaters then used these measures to analyze indepen-dently all 170 firms in our sample. The businessstrategy data were collected during the fourth quar-ter of 2003, using the same sources that we hadconsulted earlier to measure the business modeldesign themes. Thus, our data reflect the productmarket strategies that sample firms had adoptedbetween 2000 and 2001.

Inter-rater reliability on the business strategymeasures was established by conducting a pair-wise comparison of responses for five randomlychosen firms, yielding a Cronbach alpha of 0.92,and a Pearson correlation coefficient of 0.91.Raters were in agreement with each other on 77percent of the individual items (on a five-pointscale). All initial differences were resolved throughdiscussions, so the final agreement was 100 per-cent.

Econometric modeling and estimationapproach

We conducted a confirmatory factor analysis (CFA)and a partial least squares (PLS) regression anal-ysis in order to establish the discriminant validity



of our business model and product market strat-egy constructs. We then proceeded to analyze thedata using multivariate regression techniques. Weconfirmed that conventional assumptions underly-ing ordinary least squares (OLS) regression anal-ysis held in our dataset. First, after performing alogarithmic transformation of our dependent vari-able, we found that the null hypothesis of nor-mality could not be rejected at the 5 percentlevel of significance using a Shapiro–Wilk test.Second, being concerned with heteroscedastic-ity in a cross-sectional study (see Bowen andWiersema, 1999), we used White’s general testto detect evidence of heteroscedasticity. We thencorrected the p-values and t-statistics of estimatesusing White’s variance–covariance matrix (White,1980).

As a third measure to verify the validity of ourmodel, we tested for multicollinearity among inde-pendent variables by calculating variance infla-tion factors (VIF) (Kleinbaum et al., 1998) inregression models that contained only first-orderterms before mean-centering our measures. TheVIF levels that we observed were smaller than2, hence much smaller than the critical thresh-old of 10, thus eliminating the concern aboutmulticollinearity among first-order terms in theregression analysis. Multicollinearity may, how-ever, arise due to the introduction of the inter-action term, in which case mean centering canbe applied to all first- and second-order variablesas a standard and valid procedure to attenuatemulticollinearity (Aiken and West, 1991). Interac-tion terms are entered as orthogonalized effects,and this approach yields interaction variables thatare uncorrelated with their component variables.The VIF levels that we observed in regressionmodels containing first- and second-order termsafter mean centering our first-order measures wereagain all smaller than 2. Our model specifica-tion, therefore, proved robust to multicollinear-ity.

Independent variables

Two latent variables characterize the design themesof a business model (novelty and efficiency), andanother three latent variables characterize the prod-uct market positioning of the firm (differentiation,cost leadership, and timing of entry). We used 13items for novelty, 11 items for efficiency, three

items for differentiation, four items for cost lead-ership, and one item for timing of market entry(see the Appendix for details on the scales). Giventhe difficulty of obtaining objective measures ofbusiness models and product market strategy, wedeemed the use of perceptual measures obtainedfrom expert raters appropriate (Dess and Robin-son, 1984). The strength of each of these itemswas measured using five-point Likert-type scales,which we coded into a standardized score. Aftercoding, we aggregated the item scores for eachcomposite scale into an overall score using equalweights (Mendelson, 2000). This process yieldeddistinct quantitative measures of business modeland product market strategy.

We validated the internal consistency and relia-bility of our measures using standardized Cronbachalpha coefficients, which were 0.71 for the busi-ness model novelty measure, 0.70 for the businessmodel efficiency measure, 0.66 for the differen-tiation strategy measure, and 0.76 for the costleadership strategy measure. Hence, our measuressufficiently satisfy Nunnally’s (1978) guidelines,which suggest 0.7 as a benchmark for internal con-sistency.

Dependent variables

A firm’s stock market value reflects the market’sexpectations of future cash flows to shareholders,and can be viewed as a measure of perceived firmperformance, as opposed to realized performance,which is typically embodied in historical measuresof firm profitability (e.g., ROI, ROA). Given thelevel of uncertainty often associated with the trueprospects of firms that had a recent initial publicoffering (IPO), perceived performance operational-ized as stock market value is a measure that is par-ticularly germane to such a setting (Stuart, Hoang,and Hybels, 1999). Measures of realized perfor-mance, such as ROI, ROA, or Tobin’s q, are lessappropriate for these firms, which often have nega-tive earnings, few tangible assets, and low (or evennegative) book values. For instance, 134 firms inour sample (i.e., 86% of the sample firms for whichwe had the relevant accounting data) had negativeearnings. Five firms (i.e., 3% of the sample firmsfor which we had the relevant data) even had anegative book value.9

9 There are, of course, limitations in using stock market valuationas a dependent variable. The nature of our sample and the



We took measurements of the dependent vari-able at various time periods: annual average 2000,and average during the fourth quarter (Q4) of 2000.These time periods correspond well to the mea-surement of the independent variables. Since mostfirms in our sample have relatively low levels ofdebt, the market value of a firm’s equity is a goodapproximation of the market value of the wholefirm. We measured the market value of equity ata given date as the number of shares outstandingmultiplied by the firm’s stock price, taken fromthe combined CRSP and Datastream databases. Wethen took the logarithm of the market value ofthe equity in order to comply with the normalityassumption of OLS.

Control variables

We included further factors that might influencethe market value of a firm’s equity as controlvariables in the analysis because their omissionmight confound the analysis. On the firm level,we included variables that controlled for the ageand size (i.e., the number of employees) of thefirm. We also controlled for additional dimensionsof a firm’s strategy, such as the mode of marketentry, and its product and market scope (see theAppendix for details on these variables). On theindustry level of analysis we controlled for thedegree of competition and estimated market size.Our raters measured the degree of competitionon a four-point Likert scale based on informationfound in annual reports, prospectuses, competitors’SEC documents and Web sites, benchmark studies,Hoover’s Database (which lists each focal firm’smain competitors), as well as investment analysts’reports. The data on market size were obtainedfrom Forrester and other research reports and fromthe U.S. Department of Commerce.10 We also

period in which we collected the data could prompt concernsabout bias, due to an irrational bubble in the stock market.However, while the rationality of the markets during 1999–2000remains an open question (e.g., Pastor and Veronesi, 2006,offer a rational explanation for investors’ behavior and provideempirical evidence against the bubble hypothesis), our paper isnot predicated on the efficiency of capital markets. It centers onthe differential performance implications of alternative businessmodel designs and product market strategies, and our estimationmethod exploits their differential valuation by capital markets.Even if the companies in our sample were systemmaticallyovervalued, this would not distort the qualitative results of ourregression analysis.10 Some firms in our sample are first movers into new markets,relying on novelty-centered business models, which rendered

controlled for quadratic interaction effects amongour main variables.

RESULTS

Descriptive statistics

Table 3 provides an overview of the data we usein this study. Our sample firms have an averageage of 7 years (median of 4.3 years) in 2000, anda median of 270 employees. We note the largevariance among sample firms as evidenced by themedian, minimum, and maximum values of thesevariables. Our sample firms draw from relativelybroad and highly competitive market segmentsand focus on a narrow array of products. Thereare few early entrants into the market among oursample firms. Our sample, thus, consists mostly ofemerging growth companies that address relativelyestablished markets.

Table 3 also lists the Pearson correlations amongthe variables used in the regression analysis. Thecorrelations between a novelty-centered businessmodel and a differentiation strategy (0.148), andbetween an efficiency-centered business model anda cost leadership strategy (−0.064) are low, whichsupports the argument that business model designthemes and product market strategies are distinct.We also note that, while some correlations amongexplanatory variables are significant and relativelyhigh (e.g., between age and entry mode: 0.488),they do not appear to pose a multicollinearityproblem as the VIF are low for all these vari-ables.

Confirmatory factor analysis and partial leastsquares regression

A basic premise of this study is that the busi-ness model is distinct from product market strategy(see the corollary in the subsection of this article

measurement of these controls somewhat challenging. We esti-mated the market size and the degree of competition in theseand other cases based on the information we could find in thevarious sources mentioned above. For example, in 2000 (thetime period to which our independent and dependent variablesrefer) eBay did not have a serious direct competitor. The firmcompeted mainly with local online auctioneers (such as Ricardoin Europe), and with offline auction houses or even local fleamarkets. So we rated the degree of competitive threat it faced aslow (0.25 on a scale from 0 to 1). We estimated the firm’s mar-ket size as U.S. $4.7 billion based on Forrester and IDC reportsdating from 2002, in which we found calculations on the marketsize for online auctioneering in 2000.



Tabl

e3.

Pear

son

corr

elat

ions

and

desc

ript

ive

stat

istic

s

Var

iabl

ena

me

(acr

onym

)N

ovel

tyE

ffici

-en

cyD

iffe

ren-

tiat

ion

Cos

tle

ader

-sh

ip

Ent

ryti

min

gln

(mar

ket

valu

eav

g.Q

420

00)

ln(m

arke

tva

lue

avg.

2000

)

Com

peti-

tion

ln(m

arke

tsi

ze)

Age of firm

ln(e

mpl

o-ye

es)

Ent

rym

ode

Prod

uct

scop

eM

arke

tsc

ope

Inde

pend

ent

vari

able

Nov

elty

1.00

0E

ffici

ency

0.19

3∗1.

000

Dif

fere

ntia

tion

0.14

8†0.

053

1.00

0C

ost

lead

ersh

ip−0

.013

−0.0

64−0

.061

1.00

0E

ntry

tim

ing

0.23

8∗∗0.

004

0.19

7∗0.

164∗

1.00

0

Dep

ende

ntva

riab

leln

(mar

ket

valu

eav

erag

eQ

420

00)

0.17

6∗0.

790.

115

0.00

80.

125

1.00

0

ln(m

arke

tva

lue

aver

age

2000

)0.

241∗∗

0.12

00.

279∗∗

−0.0

370.

170∗

0.92

9∗∗1.

000

Con

trol

vari

able

Com

petit

ion

−0.4

76∗∗

−0.1

98∗∗

−0.1

51∗

0.02

5−0

.148

†−0

.148

†−0

.189

∗1.

000

ln(m

arke

tsi

ze)

−0.0

65−0

.004

−0.2

60∗∗

0.09

4−0

.052

0.21

7∗∗0.

105

0.17

9∗1.

000

Age

offir

m−0

.135

†−0

.101

−0.2

95∗∗

0.07

2−0

.026

0.21

9∗∗0.

044

0.07

10.

191∗

1.00

0ln

(em

ploy

ees)

0.06

7−0

.027

−0.1

64∗

0.11

−0.0

160.

632∗∗

0.54

7∗∗0.

012

0.33

9∗∗0.

459∗∗

1.00

0E

ntry

mod

e0.

069

−0.0

370.

443∗∗

−0.0

680.

075

−0.1

63∗

0.01

40.

007

−0.2

22∗∗

−0.4

88∗∗

−0.3

01∗∗

1.00

0Pr

oduc

tsc

ope

−0.0

60−0

.016

0.00

90.

054

0.09

2−0

.093

−0.1

44†

0.07

3−0

.076

0.10

6−0

.140

†−0

.134

†1.

000

Mar

ket

scop

e0.

155∗

0.10

7−0

.131

†−0

.153

∗−0

.060

−0.0

12−0

.026

−0.1

36†

0.04

50.

031

−0.0

35−0

.145

†0.

100

1.00

0

Des

crip

tive

stat

isti

csM

ean

0.38

20.

742

3.59

82.

657

2.14

751

788

30.

624

2241

07.

011

453.

971

3.76

51.

871

Med

ian

0.37

20.

750

3.66

72.

500

177

183

0.63

954

004.

327

04

41

S.D

.0.

138

0.12

40.

796

1.02

81.

590

1491

2262

0.17

569

111

7.9

3749

1.27

51.

011

1.04

7M

in.

0.07

70.

386

1.66

71

12

50

120

0.4

171

11

Max

.0.

814

15

55

1230

416

651

0.97

274

4000

4631

000

55

5N

170

170

170

170

170

161

169

170

170

170

170

170

170

170

Not

eon

desc

ript

ive

stat

istic

s:(1

)T

hein

depe

nden

tva

riab

les

are

indi

ces

that

have

been

code

dso

that

low

valu

esre

pres

ent

alo

wem

phas

is,

and

high

valu

esre

pres

ent

ahi

ghem

phas

ison

the

resp

ectiv

ebu

sine

ssm

odel

desi

gnth

eme,

orpr

oduc

tm

arke

tst

rate

gy.

Hig

hva

lues

ofE

ntry

timin

gin

dica

teea

rly

mar

ket

entr

ytim

ing.

(2)

Mar

ket

valu

ean

dm

arke

tsi

zear

egi

ven

inU

.S.

$m

illio

ns,

with

out

taki

ngth

elo

gari

thm

.(3

)Fi

rmsi

zeis

give

nas

num

ber

ofem

ploy

ees,

with

out

taki

ngth

elo

gari

thm

.(4

)H

igh

valu

esof

Ent

rym

ode

indi

cate

high

relia

nce

onst

rate

gic

part

ners

hips

and/

orjo

int

vent

ures

inde

velo

ping

,pr

oduc

ing,

orm

arke

ting

prod

ucts

.(5

)H

igh

valu

esof

Prod

uct

scop

ein

dica

tea

high

lyfo

cuse

dpr

oduc

tof

feri

ng.

(6)

Hig

hva

lues

ofM

arke

tsc

ope

indi

cate

ave

ryfo

cuse

dm

arke

tap

proa

ch.

∗∗p

<0.

01,

∗ 0.0

1<

=p

<0.

05,

†0.

05<

=p

<0.

1



entitled ‘The business model: a new structural con-cept’). Since the business model is a relatively newconstruct for strategic management research, it isincumbent upon us to validate that claim empiri-cally through establishing the discriminant validityof our main constructs. To do so, we performedtwo sets of analyses: confirmatory factor analysis(CFA), and partial least squares regression (PLS).If the results from these analyses converge, thisprovides strong support for our corollary.

We first conducted the confirmatory factor ana-lytic method proposed by Gatignon et al. (2002).The method consists of selecting pairs of con-structs and then conducting CFA for each pair. Inapplying this method, we first ran a CFA for eachpair of factors in an unconstrained measurementmodel. In this first model, the correlation betweenthe factors was estimated. Take novelty and dif-ferentiation as the chosen pair of factors. Noveltytraits loaded onto the novelty factor, and the dif-ferentiation traits loaded onto the differentiationfactor. Table 4 depicts the results from this analy-sis in the rows where the correlation between thefactors is reported as freely estimated (i.e., not setequal to 1). For example, the estimated correlationbetween novelty and differentiation was 0.19.

We then ran a CFA on a measurement modelwith only one factor, where the correlation betweenthe constructs of interest was constrained to be1. If the unconstrained model, where the corre-lation is freely estimated, improves the fit sig-nificantly compared to the constrained model,the two constructs are distinct from each other,although they can still be significantly correlated(Gatignon et al., 2002; Gatignon, 2003). To illus-trate this, consider novelty and differentiation. The

results from the CFA demonstrate that novelty-centered business model and differentiation inproduct markets are distinct constructs, althoughthey are positively correlated. This is confirmedby a significantly (at the 0.01 level) improvedconfirmatory factor analytic model when the cor-relation is estimated, compared to a measure-ment model where the correlation is constrainedto 1 (chi-squared = 260 − 186 = 74, degrees offreedom = 104 − 103 = 1). As Table 4 shows, weobtain similar results for all other pairs involv-ing generic product market strategies and businessmodel design themes, which provides support forour corollary.11

In addition to CFA, the literature suggests PLSas another method for assessing discriminant valid-ity. Using PLS, one can determine whether a con-struct shares more variance with its measures thanit shares with other constructs in the model (Hul-land, 1999). This is achieved by: (1) calculatingthe square roots of the average variance extracted(AVE) values, which measure the average varianceshared between a construct and its measures; andby (2) calculating the correlations between differ-ent constructs. A matrix can then be constructedwhere the square root of AVE is in the diagonal,and the correlations between the constructs are inthe off-diagonal. This matrix is shown in Table 5.For adequate discriminant validity, the diagonalelements should be greater than the off-diagonalelements in the corresponding rows and columns

11 We note that the CFA can also be used to assess the convergentvalidity of the constructs (Gatignon et al., 2002; Gatignon,2003), which we have confirmed in a separate analysis that isavailable upon request.

Table 4. Pair-wise confirmatory factor analysis

Low-cost strategy Differentiation strategy

Correlation Chi-squared d.f. p-value Correlation Chi-squared d.f. p-value

Efficiency 0.05 143.19 89 0 0.22 131.99 76 01 377.07 90 0 1 214.32 77 0

Novelty 0.07 193.89 118 0 0.19 186.26 103 01 418.34 119 0 1 260.00 104 0

The number of degrees of freedom (d.f.) for the chi-square test is calculated as Nvar × (Nvar + 1)/2 − Npar, where Nvar is thenumber of variables (i.e., the items for each scale) and Npar is the number of parameters to be estimated. For each variable weestimate a theta–delta and a lambda–X, and—if not constrained to 1—we also estimate the correlation between the constructs.Lambda–X is the matrix of factor loadings and theta–delta is the column vector of measurement errors (Gatignon, 2003:164). For example, in the case of novelty and differentiation strategy, when the correlation between these constructs is fixedto one, Nvar = 13(items for novelty) + 3(items for differentiation) = 16, and Npar = 32, hence d.f. = Nvar × (Nvar + 1)/2 − Npar =16 × (16 + 1)/2 − 32 = 136 − 32 = 104.



Table 5. Partial least squares analysis

Differentiation Cost leadership Efficiency Novelty

Differentiation 0.577 −0.061 0.053 0.148Cost leadership −0.061 0.500 −0.064 −0.013Efficiency 0.053 −0.064 0.302 0.193Novelty 0.148 −0.013 0.193 0.277

The table depicts the square root of average variance extracted on the diagonal, and correlations onthe off-diagonal.

(Fornell and Larcker, 1981). This is the case here,and is further evidence in support of the discrimi-nant validity of our constructs.

Hierarchical OLS regressions

Tables 6 and 7 show the results of selectedhierarchical OLS regression runs. Table 6 (Model1) supports the prediction that coupling a novelty-centered business model with a differentiationproduct market strategy represents good fit; thesevariables jointly produce a significant positiveeffect on performance. Furthermore, Table 6(Model 2) supports the hypothesized good fitbetween novelty-centered business models andcost leadership strategy (albeit only at the10% level of significance), and we also findsupport (see Model 3) for a positive interactionbetween novelty-centered business models andearly market entry timing. Regarding the fitbetween efficiency-centered business models andproduct market strategies, however, our empiricalanalysis (as shown in Model 4) does not supportthe expected good fit between efficiency-centeredbusiness models and cost leadership strategy;it produced insignificant results. We performedadditional analyses, not shown in detail here, inwhich we did not find any statistically significantinteraction terms involving efficiency-centeredbusiness models and product market differentiationor early market timing, which suggests neithergood nor bad fit between these variables. Theseresults are consistent with our model.

We conducted further sensitivity analyses onthe significant effects by testing different mod-els and different dependent variables. Some ofthese are shown in Table 7, which reports resultsfor models that included the interaction betweena novelty-centered business model and a strategyof early market entry. For example, Table 7 panel(A) refers to regressions that used the logarithm

of market value averaged over the fourth quar-ter of 2000, and panel (B) refers to regressionsthat used the logarithm of market value aver-aged over the entire year 2000. A comparison ofpanels (A) and (B) reveals that the results fromour analyses were consistent across different mea-surements of the dependent variable, albeit atdifferent levels of statistical significance, whichconfirmed the complementary nature of the par-ticular business model–product market strategyinteraction. In particular, concerning the interac-tions between novelty-centered business modelsand various product market strategies, we foundthat our data produce a positive coefficient on therelevant interaction terms in all regressions thatwe ran. That coefficient is statistically significantat the 5 percent level in a majority (though not all)of the models.

To corroborate and examine the results fromthese models further, we performed post hoc anal-ysis using plotting techniques suggested by Aikenand West (1991). Consider, for example, the resultson the interaction between product market dif-ferentiation and novelty-centered business modeldesign reported in the top panel of Table 6, Model1. The plots of differentiation on performance fordifferent values of novelty (mean value, one stan-dard deviation below the mean, one standard devi-ation above the mean) revealed that for highervalues of novelty the slope of the plotted regres-sion line was larger, and positive (see Figure 1).The plots of novelty on performance for differentvalues of differentiation (mean value, one standarddeviation below the mean, one standard devia-tion above the mean) revealed similar qualitativeresults, as well as the additional insight that theobserved positive interaction effect between dif-ferentiation strategy and novelty-centered businessmodel design is powerful: it trumps the indepen-dent effect of novelty-centered design on perfor-mance (see Figure 2). The slope of the regression



Table 6. OLS regressions to test interactions between business model design themes and product market strategies(dependent variable: ln (market value avg. quarter 4, 2000))

RHS variables Model 1 Model 2 Model 3 Model 4 Model 5Estimate

(S.E.)Estimate

(S.E.)Estimate

(S.E.)Estimate

(S.E.)Estimate

(S.E.)

Constant −0.19 −0.37 −0.39 −0.47 −0.29Independent variablesNovelty 1.26† 1.45† 1.31† 1.67∗ 0.93

(0.93) (0.92) (0.98) (0.96)Efficiency 0.95 0.63 0.84 0.76 1.34†

(0.82)Differentiation 1.80∗∗∗ 1.92∗∗∗ 1.99∗∗∗ 2.02∗∗∗ 1.77∗∗∗

(0.50) (0.53) (0.53) (0.55) (0.52)Cost leadership −0.44 −0.35 −0.43 −0.42 −0.47Timing of entry 0.16 0.15 0.04 0.19 −0.01

Control variablesCompetition −0.48 −0.66 −0.25 −0.40 −0.12log (market size) 0.071 0.08 0.06 0.08 0.07Age −0.01 −0.01 −0.01 −0.01 −0.01log (employees) 0.85∗∗∗ 0.87∗∗∗ 0.85∗∗∗ 0.87∗∗∗ 0.80∗∗∗

(0.09) (0.09) (0.09) (0.09) (0.09)Mode of entry −0.36 −0.31 −0.42 −0.34 −0.44Product scope −0.02 0.02 −0.06 0 −0.04Market scope −0.12 −0.06 0.13 −0.04 0.03

InteractionsNovelty∗ Differentiation 11.07∗ 10.61∗

(5.05) (5.19)(Novelty∗ Differentiation)2 −158.92 −122.8Novelty∗ Cost leadership 3.91† 3.17

(2.80)(Novelty∗ Cost leadership)2 −13.81 −24.87Novelty∗ Timing of entry 4.68∗ 3.63†

(2.24) (2.49)(Novelty∗ Timing of entry)2 8.32 18.90Efficiency∗ Cost leadership 4.00 2.86(Efficiency × Cost leadership)2 −16.97 −9.09R2 0.52 0.51 0.52 0.51 0.55Adj. R2 0.47 0.46 0.48 0.46 0.48N 161 161 161 161 161F 11.18∗∗∗ 10.76∗∗∗ 11.39∗∗∗ 10.67∗∗∗ 8.40∗∗∗

∗∗∗ p < 0.001; ∗∗ p < 0.01; ∗ 0.01 ≤ p < 0.05; †0.05 ≤ p < 0.1.

line is negative for low values of differentiation,and becomes positive for high values of differenti-ation. In other words, the plots shown in Figures 1and 2 show a positive interaction effect, whichsuggests that a novelty-centered business modeland a differentiation strategy are complements, notsubstitutes.12

12 The slopes of the simple regression lines shown in Figures 1and 2 differ significantly from one another. Aiken and West(1991: 19ff.) demonstrate formally that the corresponding t-test is equivalent to testing the significance of the coefficientof the interaction term in the regression. Since we observed astatistically significant coefficient of the interaction term in the

Finally, we note that even when the interac-tion terms reported in Tables 6 and 7 were statis-tically significant, the coefficients on some of thecorresponding main variables were insignificant.This corroborates the importance of consideringinteractions between product market strategies andbusiness models, over and above their independenteffects on firm performance.

regression (see Table 6), the corresponding slopes are signifi-cantly different from each other in the plots provided in Figures 1and 2.



Table 7. Hierarchical OLS regressions to test robustness of novelty/timing of entry interaction

(A) Dependent variable ln (market value avg. quarter 4, 2000)

RHS variables Model 6Estimate (S.E.)

Model 7Estimate (S.E.)



Constant 4.46∗∗∗ 4.41∗∗∗ −0.21 −0.39Novelty 1.62† 1.47 1.65∗ 1.31†

(1.20) (1.00) (0.98)Timing of entry 0.28 0.16 0.15 0.04Novelty ∗ Timing of entry 6.65∗∗ 5.41∗ 4.39∗ 4.68∗

(2.76) (2.86) (2.39) (2.24)(Novelty∗ Timing of entry)2 20.30 6.75 8.32Competition −0.53 −0.25log(market size) 0.02 0.06Age −0.02 −0.02log(employees) 0.86∗∗∗ 0.85∗∗∗

(0.09) (0.09)Differentiation 1.99∗∗∗

(0.53)Cost leadership −0.43Efficiency 0.84Mode of entry −0.42Product scope −0.06Market scope 0.13R2 0.07 0.08 0.48 0.52Adj. R2 0.06 0.06 0.45 0.48N 161 161 161 161F 4.33∗∗ 3.40∗ 17.36∗∗∗ 11.39∗∗∗

(B) Dependent variable ln (market value avg. 2000)

RHS variables Model 6Estimate (S.E.)




Constant 5.35∗∗∗ 5.31∗∗∗ 1.92∗∗ 1.08†Novelty 2.00∗ 1.87∗ 1.82∗ 1.35†

(1.05) (1.08) (0.87) (0.82)Timing of entry 0.34 0.25 0.24 0.09Novelty ∗ Timing of entry 5.82∗∗ 4.82∗ 3.55† 3.28†

(2.33) (2.58) (2.30) (1.99)(Novelty ∗ Timing of entry)2 16.34 11.22 14.10Competition −0.34 −0.05log(market size) −0.06 −0.01Age −0.04 −0.03log(employees) 0.78∗∗∗ 0.80∗∗∗

(0.08) (0.08)Differentiation 2.43∗∗∗

(0.40)Cost leadership −0.53Efficiency 1.08∗

(0.65)Mode of entry 0.06Product scope −0.21Market scope 0.21R2 0.11 0.11 0.46 0.56Adj. R2 0.09 0.09 0.43 0.52N 169 169 169 169F 6.86∗∗∗ 5.28∗∗∗ 16.88∗∗∗ 13.88∗∗∗

∗∗∗ p < 0.001; ∗∗ p < 0.01; ∗ 0.01 ≤ p < 0.05; †0.05 ≤ p < 0.1.



-4

-3

-2

-1

0

1

2

3

4

-1 0

Differentiation Strategy

Y=l

n m

arke

t va

lue

Nl Nm Nh

1

Figure 1. Plot of differentiation on performance fordifferent values of novelty (mean value, Nm, one standarddeviation below the mean, Nl, one standard deviationabove the mean, Nh). This figure is available in color

online at www.interscience.wiley.com/journal/smj

-15

-10

-5

0

5

10

15

-1 0

Novelty-centered Business Model Design

Y=

ln m

arke

t m

alue

Dl Dm Dh

1

Figure 2. Plot of novelty on performance for differentvalues of differentiation (mean value, Dm, one standarddeviation below the mean, Dl, one standard deviationabove the mean, Dh). This figure is available in color

online at www.interscience.wiley.com/journal/smj

DISCUSSION AND CONCLUSION

Our theoretical and empirical analysis reveals thata firm’s product market strategy and its businessmodel are distinct constructs that affect the firm’smarket value. We show the discriminant validity ofthe business model construct and, using hierarchi-cal OLS regression techniques, we find significanteffects of its interaction with product market strat-egy on the perceived performance of firms, as mea-sured by market capitalization. More specifically,we find positive interactions (at various levels ofstatistical significance) between novelty-centeredbusiness models and various product market strate-gies, such as cost leadership. This suggests thatalthough a firm may get ‘stuck in the middle’

between two product market strategies such as dif-ferentiation and cost leadership (Porter, 1985), andalso between two business model design themessuch as novelty and efficiency (Zott and Amit,2007), the same need not be true for a combinationof product market strategy (e.g., cost leadership)and business model design theme (e.g., novelty).Our empirical analyses suggest that these conceptsare complements rather than substitutes.

With respect to efficiency-centered businessmodels, however, our analysis did not provide sup-port for the expected positive interaction betweenan efficiency-centered business model and costleadership strategy. Our other empirical findingson efficiency-centered business models were con-sistent with the formal model: they did not revealany complementarities with a differentiation strat-egy or with the timing of entry, and indeed no clearpredictions can be made with respect to any suchrelationship.

We believe that our study makes several impor-tant contributions. First, we establish the contin-gent role of a firm’s business model in the determi-nation of its market value. In doing so, we extendthe scholarly inquiry into structure as a contin-gency factor. Whereas the traditional focus in thereceived literature has been on the firm’s inter-nal administrative structure, our analysis centeredon boundary-spanning transactions between a focalfirm and its ecosystem of partners, customers, andsuppliers. We show that adopting a broader viewof organizations—one that transcends traditionalfirm boundaries—can be valuable for understand-ing wealth creation and performance. By doing so,our study may inspire new research on the rela-tionship between strategy and structure, and on theboundaries of firms.

Second, we explore the fit between a focal firm’sbusiness-level product market strategy and thedesign themes of its business model. We elaborateon the notion of ‘good fit’ between these constructsby offering a formal model and by performing amarginal effects analysis within our model. Thisconstitutes a theoretical extension of the literatureon the fit between strategy and structure.

Third, by empirically testing the model our studyunderscores the complementary nature of the busi-ness model and product market strategy relation-ship. It thereby highlights the need to examine thefirm’s business model as a source of competitiveadvantage. We suggest that competitive advantage



can emerge from superior product market position-ing, as well as from the firm’s business model.That is, although the business model is a struc-tural construct, it is of strategic importance to thefirm. Indeed, the empirical results presented in thispaper show that both can enhance the firm’s per-formance, independently as well as jointly, whichsupports previously held conjectures (e.g., Chris-tensen, 2001). Our study points to the need toinvestigate competition among various businessmodels within an industry (Markides and Charis-tou, 2004) in addition to considering product mar-ket competition. Such rivalry on a business modellevel may have implications both for the wealthcreation potential of a given business model andfor value capture by the focal firm. In order tounderstand these phenomena better, we need toknow more about the strategic effects of businessmodels and how they influence the positioning offirms in their competitive environment.

Finally, our study raises the issue of timingof business model and product market strategydesign. They may be determined simultaneously.For example, when entrepreneurs and managersdefine and refine their business models, they mayconcurrently identify customer needs and mapthem against the products and services offeredby competitors (McGrath and MacMillan, 2000).However, it is also conceivable that product mar-ket strategy follows business model design, or viceversa. Little research has been conducted so far onhow business models evolve and in particular howthey coevolve with the product market strategy ofthe firm. In this study, we hope to have laid someof the foundations necessary to explore fruitfullythese new avenues for research.

ACKNOWLEDGEMENTS

We gratefully acknowledge the financial supportof the Wharton–INSEAD Alliance Center forGlobal Research & Development. Christoph Zottacknowledges support from the Rudolf and ValeriaMaag Fellowship in Entrepreneurship at INSEAD.Raffi Amit acknowledges financial support fromthe Wharton e-Business Initiative (a unit of theMack Center) and the Robert B. Goergen Chairin Entrepreneurship at the Wharton School. Wethank Iwona Bancerek, Amee Kamdar, and JennyKoelle for their valuable research assistance. Wealso thank Gueram Sargsyan for helpful advice.

We are grateful to two anonymous referees andthe editor for their constructive and helpful sug-gestions.

REFERENCES

Aiken LS, West SG. 1991. Multiple Regression: Testingand Interpreting Interactions . Sage: Thousand Oaks,CA.

Amburgey TL, Dacin T. 1994. As the left foot follows theright? The dynamics of strategic and structural change.Academy of Management Journal 37: 1427–1452.

Amit R, Zott C. 2001. Value creation in e-business.Strategic Management Journal , June–July SpecialIssue 22: 493–520.

Arrow K. 1962. Economic welfare and the allocationof research for invention. In The Rate and Directionof Inventive Activity: Economic and Social Factors ,Nelson R (ed.). Princeton University Press: Princeton,NJ; 609–625.

Besanko D, Dranove D, Shanley M. 1996. Economics ofStrategy . Wiley: New York.

Bowen HP, Wiersema MF. 1999. Matching methodto paradigm in strategy research: limitations ofcross-sectional analysis and some methodologicalalternatives. Strategic Management Journal 20(7):625–636.

Bower JL. 1970. Managing the Resource AllocationProcess: A Study of Corporate Planning andInvestment . Harvard Business School Press: Boston,MA.

Brandenburger AM, Nalebuff BJ. 1996. Co-opetition .Doubleday: New York.

Brandenburger AM, Stuart H. 1996. Value-based busi-ness strategy. Journal of Economics and ManagementStrategy 5: 5–25.

Chandler A. 1962. Strategy and Structure: Chapters in theHistory of American Industrial Enterprise. MIT Press:Cambridge, MA.

Christensen CM. 2001. The past and future of competitiveadvantage. MIT Sloan Management Review 42:105–109.

Claycomb C, Germain R, Droege C. 2000. The effects offormal strategic marketing planning on the industrialfirm’s configuraion, structure, exchange patterns, andperformance. Industrial Marketing Management 29:219–234.

Dess GG, Robinson RB Jr. 1984. Measuring organiza-tional performance in the absence of objective mea-sures: the case of the privately held firm and con-glomerate business unit. Strategic Management Jour-nal 5(3): 265–273.

Dierickx I, Cool K. 1989. Asset stock accumulation andsustainability of competitive advantage. ManagementScience 35: 1504–1511.

Donaldson L. 1996. The normal science of structuralcontingency theory. In Handbook of OrganizationStudies , Clegg SR, Hardy C, Nord WR (eds). Sage:London; 57–76.



Doty DH, Glick WH, Huber GP. 1993. Fit, equifinality,and organizational effectiveness: a test of twoconfigurational theories. Academy of ManagementJournal 36: 1196–1250.

Dowling MJ, McGee JE. 1994. Business and technologystrategies and new venture performance: a study of thetelecommunications equipment industry. ManagementScience 40: 1663–1677.

Dyer JH, Singh H. 1998. The relational view: cooperativestrategy and sources of interorganizational competitiveadvantage. Academy of Management Review 23:660–679.

Fornell C, Larcker DF. 1981. Evaluating structuralequation models with unobservable variables andmanagement error. Journal of Marketing Research 18:39–50.

Galbraith JR. 1977. Organization Design . Addison-Wesley: Reading, MA.

Gatignon H. 2003. Statistical Analysis of ManagementData . Kluwer: Boston, MA.

Gatignon H, Tushman M, Smith W, Anderson P. 2002.A structural approach to assessing innovation:construct development of innovation locus, type andcharacteristics. Management Science 48: 1103–1122.

Geoffrion AM, Krishnan R. 2003. E-business andmanagement science: mutual impacts (Part 1 of 2).Management Science 49: 1275–1286.

Ghaziani A, Ventresca MJ. 2005. Keywords and culturalchange: frame analysis of business model public talk1975–2000. Sociological Forum 20: 523–559.

Gulati R, Nohria N, Zaheer A. 2000. Strategic networks.Strategic Management Journal , March Special Issue21: 203–215.

Hann I, Terwiesch C. 2003. Measuring the frictional costsof online transactions: the case of a name-your-own-price channel. Management Science 48: 1563–1579.

Hargadon AB, Douglas Y. 2001. When innovations meetinstitutions: Edison and the design of the electric light.Administrative Science Quarterly 46: 476–501.

Hulland J. 1999. Use of partial least squares (PLS)in strategic management research: a review of fourrecent studies. Strategic Management Journal 20(2):195–204.

Kleinbaum DG, Kupper LL, Muller KE, Nizam A. 1998.Applied Regression Analysis and Other MultivariableMethods (3rd edn). Duxberry Press: Pacific Grove,CA.

Lieberman MB, Montgomery DB. 1988. First-moveradvantages. Strategic Management Journal , SummerSpecial Issue 9: 41–58.

MacCormack A, Verganti R, Iansiti M. 2001. Developingproducts on ‘Internet time’: the anatomy of aflexible development process. Management Science47: 133–150.

Magretta J. 2002. Why business models matter. HarvardBusiness Review 80: 86–92.

Markides C, Charistou CD. 2004. Competing with dualbusiness models: a contingency approach. Academy ofManagement Executive 18: 22–36.

McGrath R, MacMillan I. 2000. The EntrepreneurialMindset . Harvard Business School Press: Boston, MA.

Mendelson H. 2000. Organizational architecture andsuccess in the information technology industry.Management Science 46: 513–529.

Meyer AD, Tsui AS, Hinings CR. 1993. Guest co-editors’ introduction: configurational approaches toorganizational analysis. Academy of ManagementJournal 36: 1175–1195.

Miles RE, Snow CC. 1978. Organization Structure,Strategy, and Process . McGraw-Hill: New York.

Milgrom P, Roberts J. 1995. Complementarities andfit: strategy, structure, and organizational change inmanufacturing. Journal of Accounting and Economics19: 179–208.

Miller D. 1988. Relating Porter’s business strategies toenvironment and structure. Academy of ManagementJournal 31: 280–308.

Mintzberg HT. 1979. The Structuring of Organizations .Prentice-Hall: Englewood Cliffs, NJ.

Mintzberg HT. 1990. The design school: reconsideringthe basic premises of strategic management. StrategicManagement Journal 11(3): 171–195.

Nadler DA, Tushman ML. 1997. Competing by Design .Oxford University Press: Oxford, U.K.

Normann R. 2001. Reframing Business: When the MapChanges the Landscape. Wiley: Chichester, U.K.

Nunnally JC. 1978. Psychometric Theory . McGraw-Hill:New York.

Pastor L, Veronesi P. 2006. Was there a NASDAQ bubblein the late 1990s? Journal of Financial Economics 81:61–100.

Porter ME. 1985. Competitive Advantage: Creating andSustaining Superior Performance. Free Press: NewYork.

Rumelt R. 1987. Theory, strategy, and entrepreneurship.In The Competitive Challenge, Teece DJ (ed.).Ballinger Series on Innovation and OrganizationalChange. Longman: London; 137–158.

Santos FM, Eisenhardt KM. 2005. Organizational bound-aries and theories of organization. Organization Sci-ence 16: 491–508.

Siggelkow N. 2001. Change in the presence of fit: therise, the fall, and the renaissance of Liz Claiborne.Academy of Management Journal 44: 838–857.

Siggelkow N. 2002. Evolution toward fit. AdministrationScience Quarterly 47: 125–159.

Siggelkow N, Levinthal DA. 2003. Temporarily divide toconquer: centralized, decentralized, and reintegratedorganizational approaches to exploration and adapta-tion. Organization Science 14: 650–669.

Stuart TE, Hoang H, Hybels RC. 1999. Interorgani-zational endorsements and the performance ofentrepreneurial ventures. Administrative Science Quar-terly 44: 315–349.

White H. 1980. A heteroscedasticity-consistent covari-ance matrix estimator and a direct test for het-eroscedasticity. Econometrica 48: 817–838.

Williamson OE. 1975. Markets and Hierarchies: Analysisand Antitrust Implications . Free Press: New York.

Yin X, Zajac E. 2004. The strategy/governance structurefit relationship: theory and evidence in franchisingarrangements. Strategic Management Journal 25(4):365–383.



Zott C. 2003. Dynamic capabilities and the emergenceof intraindustry differential firm performance: insightsfrom a simulation study. Strategic ManagementJournal 24(2): 97–125.

Zott C, Amit R. 2007. Business model design and theperformance of entrepreneurial firms. OrganizationScience 18: 181–199.

Zott C, Bancerek I. 2004. Webraska (A) & (B). INSEADCase Studies G1368 (A, B). INSEAD: Fontainebleau,France.



Appendix: Scale Composition

Items composingefficiency-centeredbusiness model designtheme scale

Scale (code) Retainedin finalscale

Mean S.D. Min. Max.

Inventory costs for participants inthe business model are reduced

Strongly agree (1), Agree (0.75),Disagree (0.25), Stronglydisagree (0)

√0.79 0.25 0 1

Transactions are simple from theuser’s point of view


√0.80 0.23 0 1

The business model enables a lownumber of errors in theexecution of transactions


√0.68 0.25 0 1

Costs other than those alreadymentioned for participants inthe business model are reduced(e.g., marketing and sales,transaction processing,communication costs)


√0.51 0.34 0 1

The business model is scalable(i.e., can handle small as wellas large number of transactions)


√0.80 0.22 0 1

The business model enablesparticipants to make informeddecisions


√0.82 0.18 0.25 1

Transactions are transparent: flowsand use of information,services, goods can be verified


√0.78 0.24 0 1

As part of transactions,information is provided toparticipants to reduce theasymmetric degree ofknowledge among themregarding the quality and natureof the goods being exchanged


√0.71 0.27 0 1

As part of transactions,information is provided toparticipants about each other


√0.66 0.28 0 1

Access to a large range ofproducts, services andinformation, and otherparticipants is provided


0.85 0.23 0 1

The business model enablesdemand aggregation

Yes (1), No (0) 0.12 0.32 0 1

The business model enables fasttransactions


√0.81 0.25 0 1

The business model, overall, offershigh transaction efficiency


√0.79 0.22 0 1

Reliability α 0.70



(Continued )

Items composing novelty-centeredbusiness modeldesign theme scale

Scale (code) Retainedin finalscale

Mean S.D. Min. Max.

The business model offers newcombinations of products,services and information

Strongly agree (coded as 1), Agree(0.75), Disagree (0.25),Strongly disagree (0)

√0.42 0.30 0 1

The business model bringstogether new participants


√0.27 0.22 0 1

Incentives offered to participantsin transactions are novel


√0.32 0.25 0 1

The business model gives accessto an unprecedented variety andnumber of participants and/orgoods


√0.39 0.30 0 1

The business model linksparticipants to transactions innovel ways


√0.32 0.24 0 1

The richness (i.e., quality anddepth) of some of the linksbetween participants is novel


√0.36 0.27 0 1

Number of patents that the focalfirm has been awarded foraspects of its business model

0 (0), 1–2 (0.33), 3–4 (0.66), >4(1)

√0.054 0.19 0 1

Extent to which the businessmodel relies on trade secretsand/or copyrights

Radically (1), Substantially (0.66),A bit (0.33), Not at all (0)

√0.52 0.28 0 1

Does the focal firm claim to be apioneer with its business model?

Yes (1), No (0)√

0.4 0.49 0 1

The focal firm has continuouslyintroduced innovations in itsbusiness model


√0.6 0.33 0 1

There are competing businessmodels with the potential toleapfrog the firm’s businessmodel


√0.47 0.30 0 1

There are other important aspectsof the business model that makeit novel


√0.30 0.27 0 1

Overall, the company’s businessmodel is novel


√0.53 0.31 0 1

Reliability α 0.71



(Continued )

Items composingdifferentiation strategyscale

Scale Retainedin finalscale

Mean S.D. Min. Max.

Importance and use ofproduct–service-related patents

SCALE: 1 = not important at all,2 = slightly important,3 = moderately important,4 = important, 5 = veryimportant

√3.05 1.30 1 5

Importance of new productdevelopment, innovation andR&D activity


4.24 1.01 1 5

Emphasis on growth by acquiring,or merging withR&D/technology intensive firms


3.45 1.30 1 5

Branding and advertising as partof firm’s marketingstrategy/approach


√4.15 1.23 1 5

Differentiation strategy SCALE: 1 = do not use thisstrategy at all, 2 = strategy isnot important, 3 = use thisstrategy a bit, 4 = employ thisstrategy, 5 = very importantstrategy

√3.59 0.55 1 5

Reliability α 0.66

Items composing costleadership strategy scale


Mean S.D. Min. Max.

Offering products/services at lowprices/prices lower thancompetition


√3.18 1.55 1 5

Minimizing product-relatedexpenditures, in particularthrough process innovations


√2.79 1.57 1 5

Emphasizing economies of scaleand scope with products andservices


√1.82 1.37 1 5



(Continued )

Items composing costleadership strategy scale


Mean S.D. Min. Max.

Low-cost strategy SCALE: 1 = do not use thisstrategy at all, 2 = strategy isnot important, 3 = use thisstrategy a bit, 4 = employ thisstrategy, 5 = very importantstrategy

√2.84 1.02 1 5

Reliability α 0.76

Items for other strategyvariables


Mean S.D. Min. Max.

Timing of market entry (being thefirst to enter a market, and/orfirst to introduce products orservices in a market, orrealizing first mover advantagein another way)


√2.15 1.59 1 5

Mode of market entry (relying onstrategic partnerships, and jointventures in order to develop,produce, distribute, or marketproducts/services)


√3.97 1.27 1 5

Breadth of product offering(pursuing a narrow, focusedproduct scope)

SCALE 1: 1 = not important atall, 2 = slightly important,3 = moderately important,4 = important, 5 = veryimportant

√3.76 1.01 1 5

Breadth of targeted marketsegments (pursuing a narrow,focused market scope)


√1.87 1.05 1 5


amit y zott 2008

Documents

Transcript of amit y zott 2008