dynamic capital i ties

Strategic Management JournalStrat. Mgmt. J.,21: 1105–1121 (2000)

DYNAMIC CAPABILITIES: WHAT ARE THEY?

KATHLEEN M. EISENHARDT* and JEFFREY A. MARTINDepartment of Management Science and Engineering, Stanford University,Stanford, California, U.S.A.

This paper focuses on dynamic capabilities and, more generally, the resource-based view ofthe firm. We argue that dynamic capabilities are a set of specific and identifiable processessuch as product development, strategic decision making, and alliancing. They are neither vaguenor tautological. Although dynamic capabilities are idiosyncratic in their details and pathdependent in their emergence, they have significant commonalities across firms (popularlytermed ‘best practice’). This suggests that they are more homogeneous, fungible, equifinal, andsubstitutable than is usually assumed. In moderately dynamic markets, dynamic capabilitiesresemble the traditional conception of routines. They are detailed, analytic, stable processeswith predictable outcomes. In contrast, in high-velocity markets, they are simple, highlyexperiential and fragile processes with unpredictable outcomes. Finally, well-known learningmechanisms guide the evolution of dynamic capabilities. In moderately dynamic markets, theevolutionary emphasis is on variation. In high-velocity markets, it is on selection. At the levelof RBV, we conclude that traditional RBV misidentifies the locus of long-term competitiveadvantage in dynamic markets, overemphasizes the strategic logic of leverage, and reaches aboundary condition in high-velocity markets.Copyright 2000 John Wiley & Sons, Ltd.

The resource-based view of the firm (RBV) is aninfluential theoretical framework for understand-ing how competitive advantage within firms isachieved and how that advantage might be sus-tained over time (Barney, 1991; Nelson, 1991;Penrose, 1959; Peteraf, 1993; Prahalad andHamel, 1990; Schumpeter, 1934; Teece, Pisano,and Shuen, 1997; Wernerfelt, 1984). This per-spective focuses on the internal organization offirms, and so is a complement to the traditionalemphasis of strategy on industry structure andstrategic positioning within that structure as

Key words: dynamic capabilities; competitive advan-tage; resource-based view; dynamic markets; resources;high-velocity markets; organization theory; organi-zational change*Correspondence to: Kathleen M. Eisenhardt, Department ofManagement Science and Engineering, Stanford University,309 Terman, Stanford, CA 94305, U.S.A.

Copyright 2000 John Wiley & Sons, Ltd.

the determinants of competitive advantage(Henderson and Cockburn, 1994; Porter, 1979).In particular, RBV assumes that firms can beconceptualized as bundles of resources, that thoseresources are heterogeneously distributed acrossfirms, and that resource differences persist overtime (Amit and Schoemaker, 1993; Mahoney andPandian, 1992; Penrose, 1959; Wernerfelt, 1984).Based on these assumptions, researchers havetheorized that when firms have resources thatare valuable, rare, inimitable, and nonsubstitutable(i.e., so-called VRIN attributes), they canachieve sustainable competitive advantage byimplementing fresh value-creating strategies thatcannot be easily duplicated by competing firms(Barney, 1991; Conner and Prahalad, 1996; Nel-son, 1991; Peteraf, 1993; Wernerfelt, 1984, 1995).Finally, when these resources and their relatedactivity systems have complementarities, theirpotential to create sustained competitive advan-

1106 K. M. Eisenhardt and J. A. Martin

tage is enhanced (Collis and Montgomery, 1995,1998; Milgrom, Qian, and Roberts, 1991; Mil-grom and Roberts, 1990; Porter, 1996).

Recently, scholars have extended RBV todynamic markets (Teeceet al., 1997). The ration-ale is that RBV has not adequately explainedhow and why certain firms have competitiveadvantage in situations of rapid and unpredictablechange. In these markets, where the competitivelandscape is shifting, the dynamic capabilities bywhich firm managers ‘integrate, build, and recon-figure internal and external competencies to addressrapidly changing environments’ (Teeceet al., 1997:516) become the source of sustained competitiveadvantage. The manipulation of knowledgeresources, in particular, is especially critical insuch markets (Grant, 1996; Kogut, 1996).

Despite the significance of RBV, the perspec-tive has not gone unchallenged. It has been calledconceptually vague and tautological, with inatten-tion to the mechanisms by which resources actu-ally contribute to competitive advantage (e.g.,Mosakowski and McKelvey, 1997; Priem andButler, 2000; Williamson, 1999). It has also beencriticized for lack of empirical grounding (e.g.,Williamson, 1999; Priem and Butler, 2000). And,particularly relevant here, sustained competitiveadvantage has been seen as unlikely in dynamicmarkets (e.g., D’Aveni, 1994).

The purpose of this paper is to extend ourunderstanding of dynamic capabilities and in sodoing enhance RBV. Since dynamic capabilitiesare processes embedded in firms, we assume anorganizational and empirical lens, rather than aneconomic and formal modeling one (Barney, 1991;Peteraf, 1993). We examine the nature of dynamiccapabilities, how those capabilities are influencedby market dynamism, and their evolution over time.

We have several observations. First, dynamiccapabilities consist of specific strategic andorganizational processes like product develop-ment, alliancing, and strategic decision makingthat create value for firms within dynamic marketsby manipulating resources into new value-creatingstrategies. Dynamic capabilities are neither vaguenor tautologically defined abstractions. Second,these capabilities, which often have extensiveempirical research streams associated with them,exhibit commonalities across effective firms orwhat can be termed ‘best practice.’ Therefore,dynamic capabilities have greater equifinality,homogeneity, and substitutability across firms

Copyright 2000 John Wiley & Sons, Ltd. Strat. Mgmt. J.,21: 1105–1121 (2000)

than traditional RBV thinking implies. Third,effective patterns of dynamic capabilities varywith market dynamism. When markets are moder-ately dynamic such that change occurs in thecontext of stable industry structure, dynamiccapabilities resemble the traditional conception ofroutines (e.g., Cyert and March, 1963; Nelsonand Winter, 1982). That is, they are complicated,detailed, analytic processes that rely extensively onexisting knowledge and linear execution to producepredictable outcomes. In contrast, in high-velocitymarkets where industry structure is blurring,dynamic capabilities take on a different character.They are simple, experiential, unstable processesthat rely on quickly created new knowledge anditerative execution to produce adaptive, but unpre-dictable outcomes. Finally, well-known learningmechanisms guide the evolution of dynamic capa-bilities and underlie path dependence.

Overall, our work attempts to contribute to RBVby explicating the nature of dynamic capabilitiesin a way that is realistic, empirically valid, andnon-tautological. Our work also attempts to clarifyRBV’s logic of dynamic capabilities, resources,and competitive advantage. We argue that, sincethe functionality of dynamic capabilities can beduplicated across firms, their value for competitiveadvantage lies in the resource configurations thatthey create, not in the capabilities themselves.Dynamic capabilities are necessary, but not suf-ficient, conditions for competitive advantage. Wealso argue that dynamic capabilities can be usedto enhance existing resource configurations in thepursuit of long-term competitive advantage (RBV’slogic of leverage). They are, however, also veryfrequently used to build new resource configur-ations in the pursuit of temporary advantages(logic of opportunity). Most significant, we suggesta boundary condition. RBV breaks down in high-velocity markets, where the strategic challenge ismaintaining competitive advantage when the dur-ation of that advantage is inherently unpredictable,where time is an essential aspect of strategy, andthe dynamic capabilities that drive competitiveadvantage are themselves unstable processes thatare challenging to sustain.

DYNAMIC CAPABILITIES

Resources are at the heart of the resource-basedview (RBV). They are those specific physical

Dynamic Capabilities 1107

(e.g., specialized equipment, geographic location),human (e.g., expertise in chemistry), and organi-zational (e.g., superior sales force) assets that canbe used to implement value-creating strategies(Barney, 1986; Wernerfelt, 1984, 1995). Theyinclude the local abilities or ‘competencies’ thatare fundamental to the competitive advantage ofa firm such as skills in molecular biology forbiotech firms or in advertising for consumer prod-ucts firms. As such, resources form the basis ofunique value-creating strategies and their relatedactivity systems that address specific markets andcustomers in distinctive ways, and so lead tocompetitive advantage (e.g., configurations, Collisand Montgomery, 1995, 1998; Porter, 1996; corecompetencies, Prahalad and Hamel, 1990; leanproduction, Womack, Jones, and Roos, 1991).

Dynamic capabilities are the antecedent organi-zational and strategic routines by which managersalter their resource base—acquire and shedresources, integrate them together, and recombinethem—to generate new value-creating strategies(Grant, 1996; Pisano, 1994). As such, they arethe drivers behind the creation, evolution, andrecombination of other resources into new sourcesof competitive advantage (Henderson andCockburn, 1994; Teeceet al., 1997). Similar toTeece and colleagues (1997), we define dynamiccapabilities as:

The firm’s processes that use resources—speci-fically the processes to integrate, reconfigure, gainand release resources—to match and even createmarket change. Dynamic capabilities thus are theorganizational and strategic routines by whichfirms achieve new resource configurations as mar-kets emerge, collide, split, evolve, and die.

This definition of dynamic capabilities is simi-lar to the definitions given by other authors. Forexample, Kogut and Zander (1992) use the term‘combinative capabilities’ to describe organi-zational processes by which firms synthesize andacquire knowledge resources, and generate newapplications from those resources. Henderson andCockburn (1994) similarly use the term ‘architec-tural competence’ while Amit and Schoemaker(1993) use ‘capabilities.’

Dynamic capabilities as identifiable, specificprocesses

Dynamic capabilities are often described in vague


terms such as ‘routines to learn routines’ thathave been criticized as being tautological, end-lessly recursive, and nonoperational (e.g.,Mosakowski and McKelvey, 1997; Priem andButler, 2000; Williamson, 1999). Yet, dynamiccapabilities actually consist of identifiable andspecific routines that often have been the subjectof extensive empirical research in their own rightoutside of RBV.

Some dynamic capabilities integrate resources.For example, product development routines bywhich managers combine their varied skills andfunctional backgrounds to create revenue-producing products and services (e.g., Clark andFujimoto, 1991; Dougherty, 1992; Helfat andRaubitschek, 2000) are such a dynamic capability.Toyota has, for example, used its superior productdevelopment skills to achieve competitive advan-tage in the automotive industry (Clark and Fuji-moto, 1991). Similarly, strategic decision makingis a dynamic capability in which managers pooltheir various business, functional, and personalexpertise to make the choices that shape themajor strategic moves of the firm (e.g., Eisen-hardt, 1989; Fredrickson, 1984; Judge andMiller, 1991).

Other dynamic capabilities focus on recon-figuration of resources within firms. Transferprocesses including routines for replication andbrokering (e.g., Hansen, 1999; Hargadon and Sut-ton, 1997; Szulanski, 1996) are used by managersto copy, transfer, and recombine resources,especially knowledge-based ones, within the firm.For example, at the premier product design firm,IDEO, managers routinely create new productsby knowledge brokering from a variety of pre-vious design projects in many industries and frommany clients (Hargadon and Sutton, 1997).Resource allocation routines are used to distributescarce resources such as capital and manufactur-ing assets from central points within the hierarchy(e.g., Burgelman, 1994). At a more strategic level,coevolving involves the routines by which man-agers reconnect webs of collaborations amongvarious parts of the firm to generate new andsynergistic resource combinations among busi-nesses (e.g., Eisenhardt and Galunic, 2000). Dis-ney, for example, has historically excelled atcoevolving to create shifting synergies that drivesuperior performance (Wetlaufer, 2000). Patchingis a strategic process that centers on routinesto realign the match-up of businesses (i.e., add,


combine, and split) and their related resources tochanging market opportunities (Eisenhardt andBrown, 1999). Dell’s constant segmentation ofoperating businesses to match shifting customerdemands is an example of a superior patchingprocess (Magretta, 1998).

Still other dynamic capabilities are related tothe gain and release of resources. These includeknowledge creation routines whereby managersand others build new thinking within the firm, aparticularly crucial dynamic capability in indus-tries like pharmaceuticals, optical disks, and oilwhere cutting-edge knowledge is essential foreffective strategy and performance (e.g., Helfat,1997; Henderson and Cockburn, 1994; Rosenkopfand Nerkar, 1999). They also include allianceand acquisition routines that bring new resourcesinto the firm from external sources (e.g., Capron,Dussauge, and Mitchell, 1998; Gulati, 1999; Laneand Lubatkin, 1998; Powell, Koput, and Smith-Doerr, 1996; Ranft and Zeithaml, 1998; Zollo andSingh, 1998). Cisco Systems has, for example, avery effective acquisition process by which man-agers have assembled a changing array of prod-ucts and engineering know-how that drivesuperior performance. Similarly, biotech firmswith strong alliancing processes for accessingoutside knowledge achieve superior performance(Powell et al., 1996). Finally, although oftenneglected, exit routines that jettison resource com-binations that no longer provide competitiveadvantage are also critical dynamic capabilitiesas markets undergo change (Sull, 1999a, 1999b).

The identification of particular processes asdynamic capabilities has several implications. Forone, it opens up RBV thinking to a large, substan-tive body of empirical research that has oftenbeen neglected within the paradigm. This researchon capabilities such as product development andalliance formation sheds light not only on thesespecific processes, but also on the generalizednature of dynamic capabilities. So, contrary to thecriticism that dynamic capabilities lack empiricalgrounding (Williamson, 1999), dynamic capabili-ties as specific processes often have extensiveempirical research bases and management appli-cability.

More significant, the identification of specificroutines in terms of their relationship to alteringthe resource base addresses the tautology whicharises when the value of dynamic capabilities isdefined in terms of their effects on performance


(e.g., Priem and Butler, 2000; Williamson, 1999).That is, when the VRIN resources that drivecompetitive advantage are identified by observingsuperior performance and then attributing thatperformance to whatever unique resources thefirm appears to possess, the theory becomes tauto-logical. In contrast, by defining dynamic capabili-ties in terms of their functional relationship toresource manipulation, their value is defined inde-pendent of firm performance. This enables empiri-cal falsification.

Commonalities in key features, idiosyncrasyin details

Dynamic capabilities are often characterized asunique and idiosyncratic processes that emergefrom path-dependent histories of individual firms(Teeceet al., 1997). Yet, while dynamic capabili-ties are certainly idiosyncratic in their details, theequally striking observation is that specificdynamic capabilities also exhibit common featuresthat are associated with effective processes acrossfirms. These commonalities arise because thereare more and less effective ways of dealing withthe specific organizational, interpersonal, andtechnical challenges that must be addressed by agiven capability. In other words, just as there arebetter and worse ways to hit a golf ball or ski amogul field, there are more and less effectiveways to execute particular dynamic capabilitiessuch as alliancing, strategic decision making, andknowledge brokering. In popular parlance, thereis ‘best practice.’

Take, for example, the product developmentprocess, an important dynamic capability that hasbeen extensively researched (see Brown andEisenhardt, 1995, for a review). Effective productdevelopment routines typically involve the partici-pation of cross-functional teams that bringtogether different sources of expertise. Thesesources of expertise are essential for superiorproducts because each addresses a unique aspectof product quality or related production. Forexample, Imai, Ikujiro, and Takeuchi (1985) stud-ied seven product development efforts in fiveJapanese companies operating in several indus-tries. The products included the Fuji-Xerox FX-3500 copier, the City box-car by Honda, andthe Canon Sureshot camera. Performance wasmeasured in terms of the speed and flexibility ofdevelopment. The findings indicated that cross-


functional teams were essential for superior per-formance. The use of these teams enhanced therange of information that was available, and easedthe coordination and overlap of manufacturing,marketing, and design tasks during the course ofthe process.

Effective product development processes alsoinvolve routines that ensure that concrete andjoint experiences among team members, such asworking together to fix specific problems or par-ticipating in brainstorming sessions occur. Suchexperiences enhance innovation by breaking downthe thought worlds that arise because people withdifferent expertise not only know different things,but know those things differently. Concreteexperiences with others on the development teamcreate a common experience base and languagethat facilitates communication among functionallydistinct people. Dougherty (1992), for example,studied 18 product development projects in fivewell-established U.S. firms including Kodak andCampbell Soup. She found that common customervisits and feedback were essential for an effectiveproduct development process. Simply having liai-sons between groups was not enough to ensureeffective communication.

Effective product development processes alsohave extensive external communication that isoften facilitated by strong or ‘heavyweight’ teamleaders. For example, Ancona and Caldwell(1992) found that successful product developmentprocesses were characterized by extensive com-munication links outside of the group, particularlywhen those links were used by project teamleaders to buffer the group from outside influ-ences and to garner resources. Clark and Fujimoto(1991) similarly found that heavyweight leaderswho engaged in significant external communi-cation and vision setting led more productiveproduct development projects.

Commonalities that are related to more effec-tive routines exist for other dynamic capabilitiesas well. For example, successful acquisition proc-esses are characterized by preacquisition routinesthat assess cultural similarity and consistency ofvision (e.g., Larrson and Finkelstein, 1999) andpostacquisition routines that pay particular atten-tion to the speed of integration (Graebner, 2000)and the strategic redeployment of assets acrossthe two firms (Capronet al., 1998; Graebner,1999, 2000). Similarly, effective routines forcoevolving in order to capture synergies among


resources located in different parts of the organi-zation typically have common features. Theseinclude routines to ensure that business headsdevelop social bonds with one another, and sur-prisingly that the business heads are rewarded forindividual, not collective success (Christensen,1997; Eisenhardt and Galunic, 2000).

The existence of common features amongeffective dynamic capabilities does not, however,imply that any particular dynamic capability isexactly alike across firms. Take, for example,knowledge creation processes, a crucial dynamiccapability especially within high-technology firms.A common feature across successful knowledgecreation processes is explicit linkage between thefocal firm and knowledge sources outside thefirm. In the pioneering research of Allen andcolleagues (e.g., Allen, 1977; Allen, Piepmeier,and Cooney, 1971; Katz and Tushman, 1981),these linkages were a small number of ‘gatekeep-ers’ within the firm. These individuals maintainedactive communication with scientists at otherfirms, government laboratories, and universities.Similarly, Henderson and Cockburn (1994) foundthat external linkages were crucial to effectiveknowledge creation processes in their extensivestudy of the pharmaceutical industry. These link-ages, however, took the form of propublicationincentives by which scientists were rewarded formaintaining external links to the wider scientificcommunity through the use of publication inscientific journals as a promotion criterion. Simi-larly, Powell et al. (1996) found that knowledgecreation processes that included external linkagesin the form of significant alliance relationshipsled to superior R&D performance within biotechfirms. So, while external linkages are necessaryfor effective knowledge creation, those linkagescan take varied forms including informal personalrelationships, relationships driven by promotioncriterion, and formal alliances.

Commonalities across firms for effective speci-fic dynamic capabilities have several implications.First, they imply equifinality. That is, managersof firms that develop an effective dynamic capa-bility such as patching, knowledge creation, oralliancing processes very probably begin thedevelopment of that capability from differentstarting points, and take unique paths. Yet, sincethey end up with capabilities that are similar interms of key attributes, there are multiple paths(equifinality) to the same dynamic capabilities.


A recent study by Cockburn, Henderson, andStern (2000) illustrates this phenomenon. Theseauthors studied the emergence of propublicationincentives (as noted above, a common featureof effective knowledge creation processes in thepharmaceutical industry). They found that man-agers began at different starting points and trav-eled different paths before adoption of theseincentives. By happenstance, managers at somefirms were already rewarding scientists for theirpublications at the start of the study. Someadopted the practice sooner than others becausecutting-edge research was more relevant in theirparticular areas of therapeutic emphasis, orbecause they were located near major researchuniversities where firms were more influenced bythe norms of academic institutions. Still othersadopted the practice when senior leadershipchanged. Firms began with different initial con-ditions and propensities for adoption, and fol-lowed different adoption paths. But eventuallymanagers at most firms adopted propublicationincentives for their scientists.

Second, commonalities in key features of effec-tive dynamic capabilities imply that these routinesare more substitutable and fungible across differ-ent contexts than current theory suggests. In thecase of substitutability, as our example of knowl-edge creation processes suggests, effectivedynamic capabilities can differ in form and detailsas long as the important commonalities arepresent. In the case of fungibility, commonalitiesimply the efficacy of particular dynamic capabili-ties across a range of industries.

Third, commonalities imply that dynamic capa-bilities per se are not likely to be sources ofsustained competitive advantage. The thinking isas follows. According to the logic of RBV, sus-tained competitive advantage occurs when capa-bilities are not only valuable and rare, but alsoinimitable, immobile, and nonsubstitutable.Dynamic capabilities are typically valuable. Theymay be rare or at least not possessed by allcompetitors equally, as is apparent in much ofthe empirical research. Sustainability, however,breaks down for the latter conditions. Equifinalityrenders inimitability and immobility irrelevant tosustained advantage. That is, firms can gain thesame capabilities from many paths, and inde-pendent of other firms. So, whether they canimitate other firms or move resources is notparticularly relevant because managers of firms


can discover them on their own. Dynamic capa-bilities are substitutable because they need tohave key features in common to be effective, butthey can actually be different in terms of manydetails. This suggests that dynamic capabilitiesper se can be a source of competitive, but notsustainable, advantage.

Finally, commonalities suggest that the scaleof ‘idiosyncratic firm effects’ in the empiricalliterature (Brush, Bromiley, and Hendrickx, 1999;McGahan and Porter, 1997; Roquebert, Phillips,and Westfall, 1996; Schmalensee, 1985; Werner-felt and Montgomery, 1988) is probably over-stated. Simply using dummy variables for firmsleads to underspecified models that cannot capturekey organizational attributes of dynamic capabili-ties as drivers of performance. Table 1 contrastsour view with previous ones.

Market dynamism: moderately dynamic tohigh-velocity markets

The pattern of effective dynamic capabilitiesdepends upon market dynamism. In particular,dynamic capabilities vary in their reliance onexisting knowledge. Moderately dynamic marketsare ones in which change occurs frequently, butalong roughly predictable and linear paths. Theyhave relatively stable industry structures such thatmarket boundaries are clear and the players (e.g.,competitors, customers, complementers) are wellknown. In these markets, effective dynamic capa-bilities rely heavily on existing knowledge. Man-agers analyze situations in the context of theirexisting tacit knowledge and rules of thumb, andthen plan and organize their activities in a rela-tively ordered fashion (Burns and Stalker, 1966).They can develop efficient processes that arepredictable and relatively stable with linear steps,beginning with analysis and ending withimplementation (Helfat, 1997).

For example, Pisano (1994) studied the devel-opment of new manufacturing processes in asample of 23 process development projects inchemical- and biological-based pharmaceuticalcompanies. In the moderately dynamic chemicalindustry where there is deep theoretical and prac-tical knowledge, the routines for developing newmanufacturing processes were more effectivewhen they involved a structured and analyticprocess. Termed by the author ‘learning beforedoing’, managers relied on analyzing the situation


Table 1. Contrasting conceptions of dynamic capabilities

Traditional view of dynamic Reconceptualization of dynamiccapabilities capabilities

Definition Routines to learn routines Specific organizational and strategicprocesses (e.g., product innovation,strategic decision making, alliancing) bywhich managers alter their resource base

Heterogeneity Idiosyncratic (i.e., firm specific) Commonalities (i.e., best practice) withsome idiosyncratic details

Pattern Detailed, analytic routines Depending on market dynamism, rangingfrom detailed, analytic routines tosimple, experiential, ones

Outcome Predictable Depending on market dynamism,predictable or unpredictable

Competitive Advantage Sustained competitive advantage from Competitive advantage from valuable,VRIN dynamic capabilities somewhat rare, equifinal, substitutable,

and fungible dynamic capabilitiesEvolution Unique path Unique path shaped by learning

mechanisms such as practice,codification, mistakes, and pacing

to come up with an appropriate manufacturingprocess, and then implementing that processwithin the factory.

Similarly, Fredrickson (1984) examined stra-tegic decision making in the paint industry, aslowly evolving industry. He found that moreeffective decision making processes were linear.These effective processes were characterized bya sequence of problem solving steps that beganwith comprehensive collection of data, followedby development of alternatives, extensive analysisof those alternatives, and choice.

In some situations, existing tacit knowledge isfurther codified into detailed routines that pre-cisely specify steps and subdivide activitiesamong different individuals. Such routines deepenthe memory of firms for the routine (Argote,1999) and enhance the predictability of the proc-ess (Nelson and Winter, 1982). A good exampleis Eisenhardt and Tabrizi’s (1995) study of 72product development projects in the computerindustry. In the moderately dynamic mainframesector, more effective product development proc-esses were characterized by a linear progressionthrough progress gates, from specification throughprototype to design, test and finally manufacturingramp-up. Tasks within the development processwere distributed among suppliers and focal firms,which permitted the overlap of different processsteps without requiring extensive communicationduring the process.


In contrast, when markets are very dynamic orwhat is termed ‘high velocity’ (e.g., Eisenhardt,1989), change becomes nonlinear and less pre-dictable. High-velocity markets are ones in whichmarket boundaries are blurred, successful businessmodels are unclear, and market players (i.e., buy-ers, suppliers, competitors, complementers) areambiguous and shifting. The overall industrystructure is unclear. Uncertainty cannot be mod-eled as probabilities because it is not possible tospecify a priori the possible future states. In thesemarkets, dynamic capabilities necessarily relymuch less on existing knowledge and much moreon rapidly creating situation-specific new knowl-edge. Existing knowledge can even be a disadvan-tage if managers overgeneralize from past situ-ations (Argote, 1999).

Effective dynamic capabilities in high-velocitymarkets are simple, not complicated as they arein moderately dynamic markets. Simple routineskeep managers focused on broadly important is-sues without locking them into specific behaviorsor the use of past experience that may be inappro-priate given the actions required in a particularsituation. Often these routines consist of a fewrules that specify boundary conditions on theactions of managers or indicate priorities,important in fast-moving markets where attentionis in short supply.

Eisenhardt and Sull (2000) discussed the useof simple routines in high-velocity markets. They


described, for example, how Yahoo’s very suc-cessful alliancing process is largely unstructured,consisting of a two-rule routine that sets theboundary conditions for managers wishing toforge alliances. The rules are: no exclusivealliance deals and the basic service provided bythe deal (e.g., online greeting cards, party plan-ning services, etc.) must be free. There is littleelse to the routine. These rules set the boundaryconditions within which Yahoo managers havewide latitude for making a variety of alliancingdeals.

Similarly, Burgelman’s (1994, 1996) study ofIntel’s resource allocation process illustrates asimple routine, in this case one that specifiespriorities. At a time of extreme volatility in whichAsian manufacturers disrupted world markets withsevere price cutting and accelerated technologicalimprovements, Intel managers followed a simpleproduction rule, ‘margin-per-wafer-start’ thatdetermined the resource allocation for manufac-turing capacity (Burgelman, 1996: 205). Accord-ingly, as margins for memory chips decreasedand margins for microprocessors increased, Intelbegan producing proportionally more microproc-essors. By following this simple prioritization,Intel managers flexibly allocated resources andultimately morphed into a microprocessor com-pany well before senior managers recognizedthe transition.

While dynamic capabilities are simple in high-velocity markets, they are not completely unstruc-tured or ‘organic’ (e.g., Burns and Stalker, 1966;Lawrence and Lorsch, 1967). Indeed, if therewere no structures, these processes would fly outof control and exhibit no coherence. Therefore,simple routines provide enough structure (i.e.,semistructure) so that people can focus theirattention amid a cacophony of information andpossibilities, help provide sense making about thesituation, and be confident enough to act in thesehighly uncertain situations where it is easy tobecome paralyzed by anxiety.

Brown and Eisenhardt’s (1997) study of multi-ple product development processes is an illus-tration. The authors found that firms with highlystructured processes such as extensive gating pro-cedures produced new products quickly, but thatthose products often were not well adapted tomarket conditions. But, firms without some sim-ple rules were equally ineffective. Developers atthese firms had difficulty delivering products on


time to hit market windows, and consistentlyreinvented technical solutions. In contrast, firmswith the most successful product developmentprocesses relied on limited routines for prioritysetting, a business vision that bounded possibleproducts, and adherence to deadlines, but littleelse in the way of routines.

In high-velocity markets, absence of detailed,formal routines is not indicative of extensive useof tacit knowledge or complex social routinesthat cannot be codified, although these may bepresent. Rather, dynamic capabilities strikinglyinvolve the creation of new, situation-specificknowledge. This occurs by engaging in experien-tial actions to learn quickly and thereby compen-sating for limited, relevant existing knowledge byrapidly creating new knowledge about the currentsituation. So, dynamic capabilities often use pro-totyping and early testing to gain new knowledgequickly. Such actions create rapid learningthrough small losses and immediate feedback(Argote, 1999; Sitkin, 1992). Dynamic capabili-ties in these markets proceed in at iterativefashion. As managers adjust to new informationand changing conditions, they engage in morerecycling through steps such as developing alter-natives and implementation that would be linearin less dynamic markets. Dynamic capabilitiesalso rely more on real-time information, cross-functional relationships and intensive communi-cation among those involved in the process andwith the external market. Real-time informationalerts people early on to the need to adjust theiractions since problems and opportunities are spot-ted more quickly than when individuals weremore distant from information. Real-time infor-mation also builds intuition about the marketplacesuch that managers can more quickly understandthe changing situation and adapt to it (Eisenhardt,1989). Finally, dynamic capabilities in these mar-kets are characterized by parallel considerationand often partial implementation (e.g.,prototyping) of multiple options. Such optionsprovide fallback positions, which are useful sincesituations can change rapidly. They also givemanagers a sense of confidence to act quickly.The emotional inability to cope with uncertaintyis a major factor that slows down managers inhigh-velocity markets (Eisenhardt, 1989).

Pisano’s (1994) study of the process to developnew manufacturing procedures in chemical- andbiological-based pharmaceutical firms mentioned


above, is consistent with this thinking. The authorfound that ‘learning-by-doing’ (as contrasted with‘learning-before-doing’ described above) wasadvantageous in the more rapidly changingbiotech industry. In this context, it was effectiveto engage in greater experimentation and proto-typing with early testing of processes. Similarly,studies of strategic decision processes (e.g.,Eisenhardt, 1989; Judge and Miller, 1991; Wallyand Baum, 1994) found that experiential actionslike creating multiple alternatives (in addition toactions such as using real-time information) wasrelated to more effective strategic decision mak-ing processes in high-velocity markets. Thesefindings contrasted significantly with the linear,analytic process that Fredrickson (1984) found inthe less dynamic paint industry. Finally, Eisen-hardt and Tabrizi (1995) found that more andearlier testing, and more prototypes were featuresof effective product development processes in thefast-paced work station and personal computingmarkets. These experiential processes contrastedwith the detailed, linear processes that were effec-tive in the less dynamic, mainframe sector. Takentogether, these studies support the view that effec-tive dynamic capabilities in high-velocity marketsare experiential with extensive and frequent useof prototyping, real-time information, experimen-tation, and multiple alternatives.

While dynamic capabilities in high-velocitymarkets consist mostly of simple rules and real-time knowledge creation, they may have detailedroutines to deal with aspects of the process whereprior knowledge and/or codification are partic-ularly useful. Very often, this more detailedscripting exists at the end of a process wheresuch scripting helps to ensure fast, coordinatedexecution of complex details.

For example, Terwiesch, Chea, and Bohn(1999) examined the process of developing manu-facturing processes in the disk drive industry.They found that, while most of the processinvolved prototyping a variety of manufacturingalternatives, once decided, implementation of thechosen approach occurred in a highly scriptedfashion. Adler’s (1999) study of developingmanufacturing processes in the automotive indus-try had similar results for the importance ofexperimentation followed by highly rationalizedimplementation of the chosen option. Brown andEisenhardt’s (1998) study of multiple productdevelopment also indicated that most of the proc-


ess was experiential except for a highly scriptedroll-off routine to move developers from the endof one project to the beginning of the next.

The effects of market dynamism on dynamiccapabilities have several implications. One is thatsustainability of the capabilities themselves varieswith the dynamism of the market. In moderatelydynamic markets, dynamic capabilities resemblethe traditional conception of routines (Cyert andMarch, 1963; Nelson and Winter, 1982; Zolloand Winter, 1999). That is, they are complicated,predictable, analytic processes that rely exten-sively on existing knowledge, linear executionand slow evolution over time. As managers con-tinue to gain experience with these routines, theygroove the processes more deeply such that theybecome easily sustained and even inertial. Codi-fication of the routines through the technologyor formal procedures enhances that sustainability(Argote, 1999). Therefore, the capabilitiesbecome robust.

In contrast, in high-velocity markets, dynamiccapabilities take on a different character. Theyare simple (not complicated), experiential (notanalytic), and iterative (not linear) processes.They rely on the creation of situation-specificknowledge that is applied in the context of simpleboundary and priority-setting rules. But sincethese routines are simple, there is little structurefor managers to grasp and so they become easyto forget (Argote, 1999). This tendency to forgetis exacerbated by the high turnover and rapidgrowth that often accompanies firms in high-velocity markets. In more technical terms, theseimprovisational processes are dissipative, meaningthat they require constant energy to stay on track(Prigogine and Stengers, 1984). They are in thecontinuously unstable state of slipping into eithertoo much or too little structure that is sometimestermed the ‘edge of chaos’ (Kauffman, 1995).What is challenging to manage then is the optimalamount of structure (Eisenhardt and Bhatia,2000). Therefore, dynamic capabilities themselvesbecome difficult to sustain in high-velocity mar-kets. In moderately dynamic markets, competitiveadvantage is destroyed from outside the firm. Inhigh-velocity markets, the threat to competitiveadvantage comes not only from outside the firm,but also more insidiously from inside the firmthrough the potential collapse of dynamic capa-bilities.

The following quotes from several managers


in the computing industry capture this instability.As one manager described, ‘We do everythingon the fly … I’ve done some things at IBM andother companies where there is a very structuredenvironment—these companies are failing andwe’re leading the way. I’m not comfortable withthe lack of structure, but I hesitate to mess withwhat is working.’ At the other extreme, anothermanager described, ‘It is real easy for the divisionto just sort of put its head down in blinders andjust go run forward and implement … We’ve gotto force ourselves to step back’ (Brown andEisenhardt, 1997: 28).

A second implication is that causal ambiguityof dynamic capabilities varies with market dyna-mism. In moderately dynamic markets, dynamiccapabilities are causally ambiguous because theyare complicated and difficult to observe (Simonin,1999). In contrast, in high-velocity markets,dynamic capabilities are causally ambiguousbecause they are simple. The extensive, experien-tial activity of effective dynamic capabilities inhigh-velocity markets obscures the fundamentalcommonalities that drive the effectiveness of thecapability. So, it is difficult to isolate causalityfrom the extensive, but unimportant idiosyncraticdetails. Sometimes even the managers themselvesdo not know why their dynamic capabilities aresuccessful. For example, the CEO of a majorbiotech firm told one of the authors, ‘We havethe best research process in the industry, but wedon’t know why.’ Further, many managers havea tendency to imitate more than is appropriate inthe mistaken belief that more detailed processesare better. Indeed, the counterintuitive insight isthat the complicated, highly adaptive movesrequired by high-velocity markets are driven bysimple rules. Table 2 links characteristics ofdynamic capabilities with market pace.

Evolution of dynamic capabilities

The literature characterizes dynamic capabilitiesas complicated routines that emerge from path-dependent processes (Nelson and Winter, 1982;Teece et al., 1997; Zollo and Winter, 1999).However, while path dependence appropriatelyemphasizes the encoding of inferences from theunique histories of firms into distinctive routines,path dependence is more accurately described interms of learning mechanisms that have beenidentified principally within the psychological


literature (Argote, 1999). These learning mecha-nisms guide the evolution of dynamic capabilities.

For example, repeated practice is an importantlearning mechanism for the development ofdynamic capabilities. Practice helps people tounderstand processes more fully and so developmore effective routines. The efficacy of suchexperience has been demonstrated in numerousempirical studies, including the vast literature onlearning curves in manufacturing (Argote, 1999).Similarly, Zollo and Singh’s (1998) research onbank acquisitions illustrates the role of repeatedpractice. The authors found that integration,relatedness and acquisition experience led toincreased performance. Specifically, repeatedpractice with homogeneous acquisitions (i.e.,those in the related markets) was positivelyassociated with the accumulation of tacit andexplicit knowledge about how to execute acqui-sitions and achieve superior acquisition perfor-mance.

While repeated practice per se can contributeto the evolution of dynamic capabilities, the codi-fication of that experience into technology andformal procedures makes that experience easierto apply and accelerates the building of routines(Argote, 1999; Zander and Kogut, 1995). Forexample, Kale, Dyer and Singh (1999), in a cross-industry study of alliances, found that concentrat-ing alliance experience in a dedicated alliancefunction was a more powerful predictor ofalliance success than experience alone. They sug-gest that a dedicated alliance function providesan important formalization mechanism throughwhich alliancing know-how (e.g., routines) canbe articulated, codified, shared and internalizedwithin the organization.

Mistakes also play a role in the evolution ofdynamic capabilities. Small losses, more thaneither successes or major failures, contribute toeffective learning (Sitkin, 1992). Success oftenfails to engage managers’ attention sufficiently sothat they learn from their experience. Major fail-ures raise defenses that block learning. In con-trast, small failures provide the greatest moti-vation to learn as such failures cause individualsto pay greater attention to the process, but donot create defensiveness that impedes learning.

The effects of mistakes were examined byHayward (2000) in his study of 241 acquisitionsin 120 U.S. firms in six market sectors. He foundthat a moderate number of small mistakes led to


Table 2. Dynamic capabilities and types of dynamic markets

Moderately dynamic markets High-velocity markets

Market definition Stable industry structure, defined Ambiguous industry structure, blurredboundaries, clear business models, boundaries, fluid business models,identifiable players, linear and ambiguous and shifting players, nonlinearpredictable change and unpredictable change

Pattern Detailed, analytic routines that rely Simple, experiential routines that rely onextensively on existing knowledge newly created knowledge specific to the

situationExecution Linear IterativeStable Yes NoOutcomes Predictable UnpredictableKey to effective evolution Frequent, nearby variation Carefully managed selection

superior acquisition skills. Similarly, Eisenhardtand Sull (2000) recounted how Yahoo managersdeveloped one of their rules for alliancing,described above, from a mistake. Yahoo managersformed an exclusive relationship with a majorcredit card firm. Shortly, they recognized thatthis alliance restricted flexibility, especially withregard to retailers, and terminated it at greatexpense. The ‘no exclusive deals’ rule emergedfrom this mistake. Similarly, in a study of long-term development of capabilities, Kim (1998)noted the importance of crises, both contrivedand real, for developing dynamic capabilities. Inhis investigation of the long-term building ofHyundai’s organizational competencies in theirautomotive business, Kim (1998) found that theinternal generation of a sense of failure (whichhe termed ‘constructed crisis’) was essential tothe motivation of the internal learning environ-ment. These crises created greater engagement inthe situation, and so increased learning withinHyundai.

The evolution of dynamic capabilities is alsoaffected by the pacing of experience. Experiencethat comes too fast can overwhelm managers,leading to an inability to transform experienceinto meaningful learning. Similarly, infrequentexperience can lead to forgetting what waslearned previously and so result in little knowl-edge accumulation as well (Argote, 1999). Forexample, in the study mentioned earlier, Hayward(1998) found that timing had an inverted ‘U’-shaped relationship with acquisition performance.Too many acquisitions done too frequentlyimpaired managers’ ability to absorb the lessonsof any particular acquisition. They needed timeto consolidate their learning. Yet, when there


were too few acquisitions spaced too far apart,managers did not have enough opportunities tohone their skill.

While basic learning mechanisms such as thosenoted above underlie the evolution of dynamiccapabilities, crucial aspects of that evolution alsodepend upon market dynamism. In moderatelydynamic markets, experience in closely related,but different situations, is particularly effective insharpening dynamic capabilities. Frequent, smallvariations help managers to deepen capabilitiesby elaborating them in current situations andextending them to related new ones. The resultis efficient, robust routines that keep pace withchanging markets and broaden opportunities forgrowth.

For example, Haleblian and Finkelstein (1999),in their study of 449 acquisitions, explored therelationships between acquisition experience andacquisition performance. Using the theoreticalframe of learning theory, the authors found thatmanagers with extensive experience were able todiscern similarities and differences between cur-rent and previous acquisitions, and so apply theiracquisition skills in a more discriminatory mannerthat was associated with superior performance. Incontrast, managers with moderate experience hadless nuanced acquisition capabilities. Similarly,Hayward (2000) found that moderate levels ofprior acquisition similarity were positively relatedto the development of acquisition capability.Managers appeared to create superior skill whenthey both reinforced their existing knowledge andyet also extended their experience into new typesof acquisitions.

In contrast, in high-velocity markets, the morecrucial aspect of evolution is selection, not vari-


ation. Variation happens readily in such markets.In contrast, selection is difficult because it ischallenging to figure out which experience shouldbe generalized from the extensive situation-specific knowledge that occurs. Which of themany experiences should be incorporated into theongoing routines for the capabilities and whichshould be forgotten? The temptation is to gen-eralize too quickly, and so to churn capabilitiestoo often on the basis of idiosyncratic events(Gersick, 1994; Sastry, 1999).

Finally, the order of implementation ofdynamic capabilities is consequential. That is,dynamic capabilities are often combinations ofsimpler capabilities and related routines, some ofwhich may be foundational to others and so mustbe learned first. Brown and Eisenhardt (1997)termed this property ‘sequenced steps.’ In theirstudy of multiple product development processesin six firms in the computer industry, theyobserved that multiple product developmentrequired the combination of three simplerdynamic capabilities: single product development,probing the future and linking routines from oneproduct development project to the next. Man-agers who built an effective dynamic capabilityto develop multiple products did so according to‘sequence steps’ that had to be executed in theproper order. Single product development skillsneeded to come first to provide the platform forfuture products, then skills related to probing thefuture for new product opportunities, and finallytime-pacing skills to create a product developmentrhythm connecting current products to futureones. Similarly, in his study of Hyundai, Kim(1998) found an appropriate sequencing of thelearning of capabilities from the simpler and morepredictable capabilities around manufacturingprocess creation to the more improvisationaldesign ones in the development of design rou-tines. Thus, effective implementation requiresknowing both the ingredients (i.e., key com-monalities of capabilities) and the recipe (i.e.,order of implementation).

DISCUSSION

The purpose of this paper is to explore dynamiccapabilities and more generally, RBV. In address-ing this agenda, we focused on the nature ofdynamic capabilities, the impact of market dyna-


mism, and their evolution. Our observations linkto several research areas.

Our work suggests reframing the concept ofdynamic capabilities. Dynamic capabilities are nottautological, vague, and endlessly recursive assome have suggested (e.g., Priem and Butler,2000; Williamson, 1999). Rather, they consist ofmany well-known processes such as alliancing,product development, and strategic decision mak-ing that have been studied extensively in theirown right, apart from RBV. Their value for com-petitive advantage lies in their ability to alter theresource base: create, integrate, recombine, andrelease resources.

Dynamic capabilities also exhibit commonali-ties across firms that are associated with superioreffectiveness. So while the specifics of any givendynamic capability may be idiosyncratic to afirm (e.g., exact composition of a cross-functionalproduct development team) and path dependent inits emergence, ‘best practice’ exists for particulardynamic capabilities across firms. These com-monalities imply that dynamic capabilities areequifinal such that firms can develop these capa-bilities from many starting points and along dif-ferent paths. They are also more homogeneous,fungible, and substitutable than is usuallyassumed. Overall, these observations suggest amodified conception of dynamic capabilities.

Our work also suggests an expanded view ofroutines (Cyert and March, 1963; Nelson andWinter, 1982; Winter and Szulanski, 1999). Weargue that, in moderately dynamic markets, rou-tines in the form of dynamic capabilities areembedded in cumulative, existing knowledge.They involve analysis using existing knowledgeand rules of thumb, followed by implementation.When this existing knowledge is codified, theresulting routines are often detailed and specificwith predictable outcomes (Helfat, 1997; Nelsonand Winter, 1982). Therefore, in moderatelydynamic markets, dynamic capabilities exhibit theproperties suggested in the traditional researchwhere effective routines are efficient and robustprocesses (Cyert and March, 1963; Nelson andWinter, 1982).

In contrast, in high-velocity markets, dynamiccapabilities rely extensively on new knowledgecreated for specific situations. Routines are pur-posefully simple to allow for emergent adaptation,although not completely unstructured. Since newknowledge must be rapidly gained in each new


situation, experiential activities such as prototyp-ing, real-time information, multiple options, andexperimenting that generate immediate knowledgequickly replace analysis. In order to adapt tochanging information, routines are iterative andcognitively mindful, not linear and mindless.Although there may be pockets of detailed rou-tines where existing knowledge is relevant,dynamic capabilities are strikingly simple. There-fore, in high-velocity markets, effective routinesare adaptive to changing circumstances. The priceof that adaptability is unstable processes withunpredictable outcomes. Overall, this points to aricher conception of routines that goes beyondthe usual view of efficient and robust processes(Cyert and March, 1963; Nelson and Winter,1982) to include these more fragile, ‘semistructured’ones that are effective in high-velocity markets.

Our work also addresses the evolution ofdynamic capabilities. We observe that, while theevolution of dynamic capabilities occurs along aunique path for any given firm, that path isshaped by well-known learning mechanisms.Repeated practice, for example, accelerates theformation of dynamic capabilities (Argote, 1999).Small losses (Sitkin, 1992), crises (Kim, 1998),and paced experience (Hayward, 2000) can moti-vate more rapid evolution. In moderately dynamicmarkets, small and frequent variations throughrelated experience deepen capabilities (Haleblianand Finkelstein, 1999). In high-velocity marketswhere learning can be too rapid, selection ofwhat to keep from experience is more crucial(Gersick, 1994). Finally, the order of implemen-tation can be critical in dynamic capabilities thatare composed of several distinct capabilities(Brown and Eisenhardt, 1997). Taken together,these insights open the ‘black box’ of path depen-dence to reveal that the evolution of dynamiccapabilities is guided by well-known learningmechanisms.

Towards a new perspective on the resource-based view

Most significant, our work addresses the logicallinks among dynamic capabilities, resources, andcompetitive advantage, a problematic area withinRBV (Priem and Butler, 2000). We have threepoints. First, the argument that VRIN dynamiccapabilities are themselves the source of long-term competitive advantage in dynamic markets


misidentifies the source of that advantage. Asnoted earlier, effective dynamic capabilities havecommonalites across firms in terms of key fea-tures (popularly termed, ‘best practice’). There-fore, they violate the RBV assumption of persis-tent heterogeneity across firms. So, while firmswith more effective dynamic capabilities such assuperior product innovation and alliancing proc-esses are likely to have competitive advantageover firms with less effective capabilities,dynamic capabilities are not themselves sourcesof long-term competitive advantage.

So where does the potential for long-term com-petitive advantage lie? It lies in using dynamiccapabilities sooner, more astutely, or more fortu-itously than the competition to create resourceconfigurations that have that advantage. So, forexample, the acquisition capability of GE Capitalis well known, and competitors can readily copyit or independently develop it themselves. Butwhat is far more difficult to duplicate is theresource base of already acquired companies andthe related synergies among them that GE Capitalhas achieved and continues to build. This advan-tage is particularly enhanced when the relatedresource configurations are combinations oftightly woven, synergistic activities (Collis andMontgomery, 1995; Milgrom and Roberts, 1990;Porter, 1996; Prahalad and Hamel, 1990). There-fore, long-term competitive advantage lies in theresource configurations that managers build usingdynamic capabilities, not in the capabilities them-selves. Effective dynamic capabilities are neces-sary, but not sufficient, conditions for competi-tive advantage.

Second, RBV thinking overemphasizes the stra-tegic logic of leverage. While certainly someresource configurations do lead to long-term com-petitive advantage and some situations such asthose with significant scale economies or networkeffects favor the emergence of such advantages,long-term competitive advantage is infrequentlyachieved in dynamic markets. Rather, the realityis that competitive advantage is often short term.In these situations, it makes sense for managersto compete by creating a series of temporaryadvantages. Their strategic logic is opportunity(Lengnick-Hall and Wolff, 1999).

For example, D’Aveni (1994) described theCoke vs. Pepsi duopoly in which the competitorsleapfrogged one another for decades with tempo-rary advantages in new products, technical and


organizational innovations, and advertising.Neither firm could consistently gain the upperhand. Rather, each prospered by rapidly movinginto new sources of advantage. Similarly, Rob-erts’ (1999) study of the pharmaceutical industryindicated that persistent high performance wasdriven by temporary advantages in the form ofnew products. More successful firms appeared topossess a product development dynamic capa-bility that led to a superior product flow, butone that only rarely led to long-term positionaladvantage. In both of these situations, creating aseries of moves and counter-moves to out-maneuver the competition and build tem-porary advantage led to superior performance(D’Aveni, 1994).

Overall, dynamic capabilities are best concep-tualized as tools that manipulate resource con-figurations. Sometimes it is effective to use thesetools to enhance existing resource configurationsand to strengthen current position using RBV’spath-dependent strategic logic of leverage. Here,the goal is long-term competitive advantage. Morefrequently, in dynamic markets, it makes senseto use dynamic capabilities to build new resourceconfigurations and move into fresh competitivepositions using a path-breaking strategic logic ofchange (see also Karim and Mitchell, 2000).Here, the goal is a series of temporary competi-tive advantages. The broad point is that a blend ofstrategic logics makes sense in dynamic markets.

Finally, high-velocity markets are a boundarycondition for RBV, a much needed addition tothe theory (Lengnick-Hall and Wolff, 1999; Priemand Butler, 2000). In such markets, firm managersmust cope not only with the external challengeof competition, but also with the internal chal-lenge of potentially collapsing dynamic capabili-ties. As significant, RBV’s path-dependent stra-tegic logic of leverage not only lacks a logic ofchange that is crucial in dynamic markets, butalso underplays the difficulty of predicting thelength of current advantage and the sources offuture advantage. Intel is a terrific example.Although the firm dominated its market for over adecade, its managers operated as if its competitiveadvantage could end at any time. Indeed, theirslogan was ‘only the paranoid survive.’

Similarly, RBV’s assumption of the organi-zation as a bundle of resources breaks down inhigh-velocity markets. In these situations,resources are added, recombined, and dropped


with regularity (Galunic and Eisenhardt, 2000;Galunic and Rodan, 1998). Being tightly bundledis usually problematic. RBV’s emphasis on long-term competitive advantage is often unrealistic inhigh-velocity markets. Short-term, unpredictableadvantage is the norm. Growth is a more usefulperformance metric than profit. Finally, RBV mis-ses the strategic role of time. Understanding theflow of strategy from leveraging the past to prob-ing the future and the rhythm of when, where,and how often to change is central to strategyin high-velocity markets (Brown and Eisenhardt,1998). Overall, while RBV centers on leveragingbundled resources to achieve long-term competi-tive advantage, strategy in high-velocity marketsis about creating a series of unpredictable advan-tages through timing and loosely structuredorganization. The strategic logic is opportunityand the imperative is when, where, and how oftento change.

CONCLUSION

This paper explores dynamic capabilities and,more broadly, RBV. Based on the sometimesneglected insights of organizational theory andempirical research, we conclude with what wehope is a more realistic, theoretically valid, andempirically accurate view. Dynamic capabilitiesinclude well-known organizational and strategicprocesses like alliancing and product developmentwhose strategic value lies in their ability tomanipulate resources into value-creating strate-gies. Although idiosyncratic, they exhibit com-monalities or ‘best practice’ across firms. Theirbroad structural patterns vary with market dyna-mism, ranging from the robust, grooved routinesin moderately dynamic markets to fragile semi-structured ones in high-velocity ones. They evolvevia well-known learning mechanisms.

More broadly, we conclude that long-termcompetitive advantage lies in resource configu-rations, not dynamic capabilities. In moderatelydynamic markets, RBV is enhanced by blendingits usual path-dependent strategic logic of lever-age with a path-breaking strategic logic ofchange. Finally, RBV encounters a boundary con-dition in high-velocity markets where the durationof competitive advantage is inherently unpredict-able, time is central to strategy, and dynamiccapabilities are themselves unstable. Here, the


strategic imperative is not leverage, but change.

ACKNOWLEDGEMENTS

An earlier version of this paper was presented atthe September 1999 Tuck/Consortium on Com-petitiveness and Cooperation (CCC) conferenceon the Evolution of Firm Capabilities. Weappreciate the helpful comments of Anil Gupta,Connie Helfat, Cynthia Montgomery, FilipeSantos, and the consortium participants.

REFERENCES

Adler PS. 1999. Flexibility versus efficiency? A casestudy of model changeovers in the Toyota pro-duction system.Organization Science10(1): 43–68.

Allen TJ. 1977. Managing the Flow of Technology:Technology Transfer and the Dissemination of Tech-nological Information within the R&D Organization.MIT Press: Cambridge, MA.

Allen TJ, Piepmeier JM, Cooney S. 1971.TechnologyTransfer to Developing Countries: The InternationalTechnological Gatekeeper. Massachusetts Institute ofTechnology: Cambridge, MA.

Amit R, Schoemaker PJH. 1993. Strategic assets andorganizational rent.Strategic Management Journal14(1): 33–46.

Ancona DG, Caldwell DF. 1992. Bridging the bound-ary: External process and performance in organi-zational teams.Administrative Science Quarterly37(4): 634–665.

Argote L. 1999. Organizational Learning: Creating,Retaining, and Transferring Knowledge. KluwerAcademic: Boston, MA.

Barney JB. 1991. Firm resources and sustained competi-tive advantage.Journal of Management17(1): 99–120.

Barney JB. 1986. Organizational culture: can it be asource of sustained competitive advantage?Academyof Management Review11(3): 656–665.

Brown SL, Eisenhardt KM. 1995. Product development:past research, present findings and future directions.Academy of Management Review20(2): 343–378.

Brown SL, Eisenhardt KM. 1997. The art of continuouschange: linking complexity theory and time-pacedevolution in relentlessly shifting organizations.Administrative Science Quarterly42(1): 1–34.

Brown SL, Eisenhardt KM. 1998.Competing on theEdge: Strategy as Structured Chaos. Harvard Busi-ness School Press: Boston, MA.

Brush TH, Bromiley P, Hendrickx M. 1999. The rela-tive influence of industry and corporation on busi-ness segment performance: an alternative estimate.Strategic Management Journal20(6): 519–547.

Burgelman RA. 1994. Fading memories: a processtheory of strategic business exit in dynamic environ-


ments.Administrative Science Quarterly39(1): 24–56.

Burgelman RA. 1996. A process model of strategicbusiness exit.Strategic Management Journal, Sum-mer Special Issue17: 193–214.

Burns T, Stalker GM. 1966.The Management of Inno-vation. 2nd edn, Associated Book Publishers: Lon-don.

Capron L, Dussauge P, Mitchell W. 1998. Resourceredeployment following horizontal acquisitions inEurope and North America, 1988–1992.StrategicManagement Journal19(7): 631–661.

Christensen C. 1997.Managing Innovation at NYPRO,Inc. (A) (B). Harvard Business School Publishing:Boston, MA.

Clark KB, Fujimoto T. 1991.Product DevelopmentPerformance: Strategy, Organization, and Manage-ment in the World Auto Industry. Harvard BusinessSchool Press: Boston, MA.

Cockburn I, Henderson R, Stern S. 2000. Untanglingthe origins of competitive advantage.Strategic Man-agement Journal(this issue).

Collis DJ, Montgomery CA. 1995. Competing onresources.Harvard Business Review73(4): 118–128.

Collis DJ, Montgomery CA. 1998. Creating corporateadvantage.Harvard Business Review76(3): 70–83.

Conner KR, Prahalad CK. 1996. A resource-basedtheory of the firm: knowledge versus opportunism.Organization Science7(5): 477–501.

Cyert RM, March JG. 1963.A Behavioral Theory ofthe Firm. Prentice-Hall: Englewood Cliffs, NJ.

D’Aveni RA. 1994. Hypercompetition: Managing theDynamics of Strategic Maneuvering. Free Press:New York.

Dougherty D. 1992. Interpretive barriers to successfulproduct innovation in large firms.OrganizationScience3: 179–202.

Eisenhardt KM. 1989. Making fast strategic decisionsin high-velocity environments.Academy of Manage-ment Journal32(3): 543–576.

Eisenhardt K, Sull D. 2001. What is strategy in the neweconomy?Harvard Business Review. (forthcoming).

Eisenhardt KM, Bhatia MM. 2000. Organizational com-plexity and computation. InCompanion to Organiza-tions, Baum JAC (ed.). Blackwell: Oxford, UK.(forthcoming).

Eisenhardt KM, Brown SL. 1999. Patching: restitchingbusiness portfolios in dynamic markets.HarvardBusiness Review77(3): 72–82.

Eisenhardt KM, Galunic DC. 2000. Coevolving: at last,a way to make synergies work.Harvard BusinessReview78(1): 91–101.

Eisenhardt KM, Tabrizi BN. 1995. Accelerating adap-tive processes: product innovation in the global com-puter industry. Administrative Science Quarterly40(1): 84–110.

Fredrickson JW. 1984. The comprehensiveness of stra-tegic decision processes: extension, observations,future directions.Academy of Management Journal27(3): 445–467.

Galunic DC, Eisenhardt KM. 2000. Architectural inno-vation and modular corporate forms. Working paper,


1-41. INSEAD/Stanford University, Fontainebleauand Stanford, CA.

Galunic DC, Rodan S. 1998. Resource recombinationsin the firm: knowledge structures and the potentialfor Schumpeterian innovation.Strategic Manage-ment Journal19(12): 1193–1201.

Gersick CJG. 1994. Pacing strategic change: the caseof a new venture.Academy of Management Journal37(1): 9–45.

Graebner M. 1999. A review of recent research onmergers and acquisitions. Working paper, StanfordUniversity, Stanford, CA.

Graebner M. 2000. Acquisitions of entrepreneurialfirms. Working paper, 1-85. Stanford University,Stanford, CA.

Grant RM. 1996. Toward a knowledge-based theory ofthe firm. Strategic Management Journal, SummerSpecial Issue17: 109–122.

Gulati R. 1999. Network location and learning: theinfluence of network resources and firm capabilitieson alliance formation.Strategic Management Jour-nal 20(5): 397–420.

Haleblian J, Finkelstein S. 1999. The influence oforganizational acquisition experience on acquisitionperformance: a behavioral learning perspective.Administrative Science Quarterly44(1): 29–56.

Hansen MT. 1999. The search-transfer problem: therole of weak ties in sharing knowledge acrossorganization subunits.Administrative Science Quar-terly (March) 44: 82–111.

Hargadon A, Sutton RI. 1997. Technology brokeringand innovation in a product development firm.Administrative Science Quarterly42(4): 716–749.

Hayward MLA. 1998. Is learning loopy? Evidence ofwhen acquirers learn from their acquisition experi-ences. Working paper (LRP WP45/1998), LondonBusiness School, London.

Hayward MLA. 2000. Acquirer learning from acqui-sition experience: evidence from 1985–1995. Work-ing paper, London Business School, London.

Helfat CE. 1997. Know-how and asset complementarityand dynamic capability accumulation.StrategicManagement Journal18(5): 339–360.

Helfat CE, Raubitschek RS. 2000. Product sequencing:co-evolution of knowledge, capabilities and products.Strategic Management Journal21(10–11): 961–979.

Henderson R, Cockburn I. 1994. Measuring com-petence? Exploring firm effects in pharmaceuticalresearch. Strategic Management Journal, WinterSpecial Issue15: 63–84.

Imai K, Ikujiro N, Takeuchi H. 1985. Managing thenew product development process: how Japanesecompanies learn to unlearn. InThe Uneasy Alliance:Managing the Productivity–Technology Dilemma,Hayes RH, Clark K, Lorens J (eds.). Harvard Busi-ness School Press: Boston, MA; 337–375.

Judge WQ, Miller A. 1991. Antecedents and outcomesof decision speed in different environments.Acad-emy of Management Journal34(2): 449–464.

Kale P, Dyer JH, Singh H. 1999. Alliance capability,stock market response, and long term alliance suc-cess. Working paper, University of Michigan: AnnArbor, MI.


Karim SZ, Mitchell W. 2000. Path-dependent and path-breaking change: reconfiguring business resourcesfollowing acquisitions in the U.S. medical sector,1978–1995.Strategic Management Journal21(10–11): 1061–1081.

Katz R, Tushman ML. 1981. An investigation into themanagerial roles and career paths of gatekeepersand project supervisors in a major R&D facility.R&D Management11(3): 103–110.

Kauffman SA. 1995.At Home in the Universe: TheSearch for Laws of Self-Organization and Com-plexity. Oxford University Press: New York.

Kim L. 1998. Crisis construction and organizationallearning.Organization Science9(4): 506–521.

Kogut B. 1996. What firms do? Coordination, identity,and learning.Organization Science7(5): 502–518.

Kogut B, Zander U. 1992. Knowledge of the firm,combinative capabilities, and the replication of tech-nology. Organization Science3: 383–397.

Lane PJ, Lubatkin M. 1998. Relative absorptivecapacity and interorganizational learning.StrategicManagement Journal19(5): 461–477.

Larrson R, Finkelstein S. 1999. Integrating strategic,organizational, and human resource perspectives onmergers and acquisitions: a case survey of synergyrealization.Organization Science10(1): 1–26.

Lawrence PR, Lorsch JW. 1967.Organization andEnvironment; Managing Differentiation and Inte-gration. Division of Research Graduate School ofBusiness Administration Harvard University: Bos-ton, MA.

Lengnick-Hall CA, Wolff JA. 1999. Similarities andcontradictions in the core logic of three strategyresearch streams.Strategic Management Journal20(12): 1109–1132.

Magretta J. 1998. The power of virtual integration: aninterview with Dell Computer’s Michael Dell.Har-vard Business Review76(2): 72–84.

Mahoney JT, Pandian JR. 1992. The resource-basedview within the conversation of strategic man-agement. Strategic Management Journal13(5):363–380.

McGahan AM, Porter ME. 1997. How much doesindustry matter, really?Strategic Management Jour-nal, Summer Special Issue18: 15–30.

Milgrom P, Qian Y, Roberts J. 1991. Complementari-ties, momentum, and the evolution of modern manu-facturing.American Economic Review81(2): 84–88.

Milgrom P, Roberts J. 1990. The economics of modernmanufacturing: technology, strategy, and organi-zation. American Economic Review80(3): 511–528.

Mosakowski E, McKelvey B. 1997. Predicting rentgeneration in competence-based competition. InCompetence-Based Strategic Management, Heene A,Sanchez R (eds.). Chichester: Wiley; 65–85.

Nelson RR. 1991. Why do firms differ, and how doesit matter? Strategic Management Journal, WinterSpecial Issue12: 61–74.

Nelson R, Winter S. 1982.An Evolutionary Theory ofEconomic Change. Belknap Press: Cambridge, MA.

Penrose ET. 1959.The Theory of the Growth of theFirm. Wiley: New York.

Peteraf MA. 1993. The cornerstones of competitive


advantage.Strategic Management Journal14(3):179–191.

Pisano GP. 1994. Knowledge, integration, and the locusof learning: an empirical analysis of process devel-opment.Strategic Management Journal, Winter Spe-cial Issue15: 85–100.

Porter ME. 1979. How competitive forces shape strat-egy. Harvard Business Review57(2): 137–145.

Porter ME. 1996. What is strategy?Harvard BusinessReview74(6): 61–78.

Powell WW, Koput KW, Smith-Doerr L. 1996.Interorganizational collaboration and the locus ofinnovation. Administrative Science Quarterly41(1):116–145.

Prahalad CK, Hamel G. 1990. The core competence ofthe corporation.Harvard Business Review68(3):79–91.

Priem RL, Butler JE. 2000. Is the resource-based ‘view’a useful perspective for strategic managementresearch? Academy of Management Review(forthcoming).

Prigogine I, Stengers I. 1984.Order Out of Chaos:Man’s New Dialogue with Nature. Bantam Books:New York.

Ranft AL, Zeithaml CP. 1998. Preserving and transfer-ring knowledge-based resources during post-acquisition implementation: a study of high-techacquisitions. Working paper, College of Businessand Economics, West Virginia University, Morgan-town, WV.

Roberts PW. 1999. Product innovation, product-marketcompetition and persistent profitability in the U.S.pharmaceutical industry. Strategic ManagementJournal 20(7): 655–670.

Roquebert JA, Phillips RL, Westfall PA. 1996. Marketsvs. management.Strategic Management Journal17(8): 653–664.

Rosenkopf L, Nerkar A. 1999. Beyond local search:boundary-spanning, exploration and impact in theoptical disc industry. Working paper, The WhartonSchool, University of Pennsylvania, Philadelphia,PA.

Sastry MA. 1999. Managing strategic innovation andchange. Administrative Science Quarterly44(2):420–422.

Schmalensee R. 1985. Do markets differ much?Amer-ican Economic Review75(3): 341–351.

Schumpeter JA. 1934.The Theory of Economic Devel-opment. 7th edn (transl. Opie R) Harvard UniversityPress: Cambridge, MA.

Simonin BL. 1999. Ambiguity and the process ofknowledge transfer in strategic alliances.StrategicManagement Journal20(7): 595–623.

Sitkin SB. 1992. Learning through failure: the strategyof small losses, InResearch in Organizational


Behavior, Staw BM, Cummings LL (eds.). Vol. 14:JAI Press: Greenwich, CT; 231–266.

Sull DN. 1999a. The dynamics of standing still: Fire-stone tire & rubber and the radial revolution.Busi-ness History Review73 (Autumn): 430–464.

Sull DN. 1999b. Why good companies go bad.HarvardBusiness Review77(4): 42–52.

Szulanski G. 1996. Exploring internal stickiness:impediments to the transfer of best practice withinthe firm. Strategic Management Journal, WinterSpecial Issue17: 27–43.

Teece DJ, Pisano G, Shuen A. 1997. Dynamic capabili-ties and strategic management.Strategic Manage-ment Journal18(7): 509–533.

Terwiesch C, Chea KS, Bohn RE. 1999. An exploratorystudy of international product transfer and pro-duction ramp-up in the data storage industry. Report99-02, Information Storage Industry Center, Gradu-ate School of International Relations and PacificStudies, University of California at San Diego, LaJolla, CA.

Wally S, Baum JR. 1994. Personal and structural deter-minants of the pace of strategic decision making.Academy of Management Journal37(4): 932–956.

Wernerfelt B. 1984. A resource-based view of the firm.Strategic Management Journal5(2): 171–180.

Wernerfelt B. 1995. The resource-based view of thefirm: ten years after.Strategic Management Journal16(3): 171–174.

Wernerfelt B, Montgomery C. 1988. Tobin’s q and theimportance of focus in firm performance.AmericanEconomic Review78(1): 246–250.

Wetlaufer S. 2000. Common sense and conflict: aninterview with Disney’s Michael Eisner.HarvardBusiness Review78(1): 114–124.

Williamson OE. 1999. Strategy research: governanceand competence perspectives.Strategic ManagementJournal 20(12): 1087–1108.

Winter SG, Szulanski G. 1999. Replication as strategy.Working paper, University of Pennsylvania, Phila-delphia, PA.

Womack JP, Jones DT, Roos D. 1991.The Machinethat Changed the World: The Story of Lean Pro-duction. HarperCollins: New York.

Zander U, Kogut B. 1995. Knowledge and the speedof the transfer and imitation of organizational capa-bilities. Organization Science6(1): 76–92.

Zollo M, Singh H. 1998. The impact of knowledgecodification, experience trajectories and integrationstrategies on the performance of corporate acqui-sitions. Academy of Management Best Paper Pro-ceedings, San Diego, CA.

Zollo M, Winter S. 1999. From organizational routinesto dynamic capabilities. Working paper WP 99-07,University of Pennsylvania, Philadelphia, PA.

dynamic capital i ties

Documents

dedicated alliance function

strategic decision making

strategic business exit

strategic decision processes

term competitive advantage

drive competitive advantage

longterm competitive advantage

dynamic capabilities resemble