Organizational Technology

Transcription

1 Organizational Technology Michael L. Tushman and Wendy K. Smith Adapted from Chapter 17 of The Blackwell Companion to Organizations, Joel Baum (ed), Blackwell Publishers Ltd., UK 2002 pp Technological Change, Ambidextrous Organizations and Organizational Evolution Technical change is one of the core drivers of organizational fates (Tushman and Nelson, 1990; Nelson, 1995). While technological change accentuates organizational failure rates, there is substantial heterogeneity in organizational life chances (Barnett and Carroll, 1995). Some firms thrive during eras of ferment, other firms proactively destabilize their product class with technological discontinuities, even as most firms are swept away during Schumpeterian gales of creative destruction (eg Grove, 1996; Morone, 1993; Sorensen and Stuart, 2000; Carroll and Teo, 1998). The stream of research on organizational technology is interested in how organizations shape and are, in turn, shaped by technological change. This literature sheds substantial light on how organizational architectures, capabilities and senior teams affect both a firm s ability to shape technological change and to effectively compete when technologies change.

2 Technology and resource rich firms often fail to sustain their competitiveness at technology transitions. Consider SSIH, the Swiss watch consortium, Goodyear Tire, Polaroid, and Oticon, the Danish hearing aid firm. These organizations dominated their respective worldwide markets, SSIH and Goodyear through the 1970s and Polaroid and Oticon through the early 1990s. Each developed new technologies that had the capabilities to re-create their markets (e.g., quartz movements, radial tires, digital imaging, and in-the-ear [ITE] volume and tone control). But although SSIH, Goodyear, Polaroid and Oticon had the technology and the resources to innovate, it was smaller, more aggressive firms that initiated new technology in these four industries. SSIH, Goodyear, Polaroid and Oticon prospered until new industry standards what we will call dominant designs rapidly destroyed their market positions (Glassmier, 1991, Sull, 1999, Tripsas and Gavetti, 2000). Similar liabilities of success have been found in disk drives (Christensen and Bower, 1996), business equipment (Rosenbloom, 2000), photolithography (Henderson and Clark, 1990), typesetting (Tripsas, 1999), among others (Tushman and O Reilly, 1997; Miller, 1994). In the watch, tire, photography and hearing aid markets, it was not new technology that led to the demise of the Swiss, Americans or the Danes; indeed, SSIH, Goodyear, Polaroid and Oticon were technology leaders. Nor was the rapid loss in market share due to lack of financial resources or to governmental regulations. Rather, the rapid demise of SSIH and Goodyear and the losses at Polaroid and Oticon were rooted in organizational complacency and inertia. These pathologies of sustained success stunted their ability to renew themselves. This success syndrome is particularly 2

3 paradoxical in that each of these firms had the competencies, resources and technologies to proactively drive innovation streams. Innovation streams are patterns of innovations; some that build on and extend prior products (e.g., mechanical watches, bias ply tires, and behind-the-ear [BTE] hearing aids), while others destroy those very products that account for a firm s historical success (eg. analog to digital imaging). Innovation streams focus theoretical and empirical attention away from isolated innovations and toward patterns of innovation over time. This paradoxical pattern in which winners, with all their competencies and assets, become losers is found across industries and countries (see Hamel and Prahalad, 1994; Utterback, 1994; Christensen, 1998). It seems that building core competencies and managing through continuous improvement are not sufficient for sustained competitive advantage. Worse, building on core competencies (e.g., for the Swiss, precision mechanics) and engaging in continuous incremental improvement actually traps the organization in its past and leads to catastrophic failure as technologies and markets shift. Core competencies often turn into rigidities (Leonard-Barton, 1992; Benner and Tushman, 2000). Those firms caught by historically anchored inertia are unable to build, extend, or destroy their existing competencies in order to develop innovations that would create new markets (as Starkey did with ITE hearing aids) or rewrite the competitive rules in existing markets (as Seiko and Michlein did with quartz watches and radial tires). These inertial firms get selected out of their competitive arenas by basic ecological dynamics (Levinthal, 1997; Sorensen and Stuart, 2000; Carroll and Hannan, 2000). 3

4 But liabilities of success are not deterministic; core competencies need not become core rigidities. Some organizations are capable of proactively shifting bases of competition through streams of innovation (Brown and Eisenhardt, 1998; Tushman and O Reilly, 1997). These firms are able to develop incremental innovation as well as innovations that alter industry standards, substitute for existing products, and/or reconfigure products to fundamentally different markets. For example, in the watch industry, Seiko not only was able to compete in mechanical watches, but was also willing to experiment with quartz and tuning fork movements. Based on these technological options, Seiko managers made the decision to substitute quartz movements for their existing mechanical movements. In retrospect, the switch to the quartz movement led to fundamentally different competitive rules in the watch industry. Similarly, Starkey (a U.S. hearing aid company) was able to move beyond BTE hearing aids to ITE hearing aids by simply reconfiguring existing hearing aid components. This seemingly minor architectural innovation led to a new industry standard and to different industry rules anchored on sound quality and fashion. Dynamic capabilities are rooted in driving streams of innovation. Firms that survive technological transitions compete through patterns of innovation over time: incremental, competence-enhancing innovation (e.g., thinner mechanical watches); architectural innovation (e.g., Starkey s ITE hearing aid); taking existing technologies to new customer segments, and fundamentally new, often competence-destroying, innovation (e.g., Seiko s quartz movement). By driving streams of innovation, senior teams increase the probability that their firm will be able to shape industry standards, 4

5 take advantage of new markets for existing technology, and proactively introduce substitute products that, as they cannibalize existing products, create new markets and competitive rules (Teece, 1996; Burgelman and Grove, 1996; Hurst, 1995; Brown and Eisenhardt, 1998). These dynamic capabilities are rooted in a firm s ability to be ambidextrous-- to both learn and incrementally build on its past even as it simultaneously creates technological options from which senior teams make strategic bets (Duncan, 1976; Tushman and O Reilly, 1997). Because of powerful inertial processes accentuated in those most successful firms, these strategic bets must be coupled with discontinuous organizational changes (Sastry, 1997; Romanelli and Tushman, 1994; Gavetti and Levinthal, 2000). We discuss the topics of technology cycles, innovation streams, ambidextrous organizations, senior teams, and discontinuous organizational change in turn. Technology Cycles and Dominant Designs Technology cycles are composed of technological discontinuities (for example, quartz and tuning fork movements in watches) that trigger periods of technological and competitive ferment. During eras of ferment, rival technologies compete with each other and with the existing technological regime. These turbulent innovation periods close with the emergence of an industry standard or dominant design (Utterback, 1994; Anderson and Tushman, 1990). For example, in early radio transmission, continuouswave transmission was a technological discontinuity that threatened to replace spark- 5

6 gap transmission. Continuous-wave transmission initiated competition not only between this new innovation and spark-gap transmission but among three variants of the innovation: alternating-wave, arc, and vacuum tube transmission. This period of technological ferment led to vacuum tube transmission as the dominant design in radio transmission (Aitken, 1985; Rosenkopf and Tushman, 1994). The emergence of a dominant design ushers in a period of incremental as well as architectural technological change, a period that is broken at some point by the next substitute product. The subsequent technological discontinuity then triggers the next wave of technological variation, selection, and retention (see Figure 1). Technology cycles are seen most directly in nonassembled or simple products (e.g., glass, chemicals, skis, tennis racquets). For example, in crop fungicides, Ciba- Geigy s Tilt (propiconazol) was a new chemical entity that challenged Bayer s and BASF s products. Tilt triggered competition between chemical entities as well as between a vast number of propiconazol formulations. Ciba eventually created its EC 250 version, which became the industry standard in crop fungicides. Ciba s Crop Protection Division then initiated several product substitutes (including genetically engineered seeds) to cannibalize and replace propiconazol. These fundamentally new crop protection products initiated the next technology cycle in the crop protection market (Rosenkopf and Tushman, 1994). In more complex assembled products (e.g., computers or watches) and systems (e.g., radio or voice mail), technology cycles apply at the subsystem level. David Landes (1983) and Thomas Hughes (1983) provide rich historically anchored detail on the 6

7 Variation Technological substitution 1 Technological discontinuity 2 Competence enhancing innovations Competence destroying innovations Era of of incremental change 2 Elaborate dominant design Architectural innovation Market-based innovation Era of of ferment 2 Substitution Design competition Community driven technical change Variation Technological discontinuity 1 Era of of incremental change 1 Era of of ferment 1 Selection Dominant design 2 Selection Dominant design 1 Figure 1 Technology cycles over time (adapted from Rosenkopf and Tushman, 1994) watch and electric power industries from their respective births. These comprehensive histories illustrate the interplay between technology, organizations and communities. Watches, for example, are assembled products made up of at least four subsystems: energy source, oscillation device, transmission, and display. Each of these subsystems has its own technology cycle. In watch oscillation, the pin-lever escapement became the dominant design in the late nineteenth century. Escapements became better and better through incremental changes in the same fundamental design until the late 1960s. Between 1968 and 1972, escapements were threatened by both tuning fork and quartz oscillation. This period of technological competition among escapements, tuning fork oscillation, and quartz movement ended with the emergence of quartz oscillation as the dominant design in the subsystem. As with mechanical escapements, the emergence of 7

8 quartz movements as the dominant design led, in turn, to numerous incremental improvements in the quartz movement and sharp decreases in innovation in tuning fork and escapement oscillation devices. Not all subsystems are equivalent. Rather, products are composed of hierarchically ordered subsystems that are coupled together by linking mechanisms that are as crucial to the product s performance as are the subsystems themselves (Clark, 1985; Henderson and Clark, 1990; Baldwin and Clark, 2000). Those more core subsystems are either tightly coupled to other subsystems or are strategic bottlenecks (Hughes, 1983; Schilling, 2000; Tushman and Murmann, 1998). In contrast, peripheral subsystems are weakly connected to other subsystems. Shifts in core subsystems have cascading effects on other more peripheral subsystems (Ulrich and Eppinger, 1995). Further, subsystems shift in relative strategic importance as the industry evolves. In watches, oscillation was the key strategic battlefield through the early 1970s; then, once the quartz movement became the dominant design, the locus of strategic innovation shifted to the face, energy, and transmission subsystems. Similar dynamics of subsystem and linkage technology cycles have been documented in a variety of industries (Hughes, 1983, Van de Ven and and Garud, 1994; Baum, Korn and Kotha, 1995). The technological discontinuities that initiate technology cycles are relatively rare, unpredictable events triggered by scientific advance (e.g., battery technology for watches) or through a recombining of existing technology (e.g., Sony s Walkman or continuous aim gunfire) [Morison, 1966; Sanderson and Uzumeri, 1995]. Technological 8

9 discontinuities rupture existing incremental innovation patterns and spawn periods of technological ferment that are confusing, uncertain, and costly to customers, suppliers, vendors, and regulatory agencies. Absent governmental regulation, a single dominant design emerges from periods of variation, or eras of technological ferment (see Noble, 1984, Cusumano et al, 1992; Anderson and Tushman, 1990). During windows of opportunity at the emergence of dominant designs, competing firms must switch to the new standard or risk getting locked out of the market (Christensen, et al, 1998; Tegarden et al, 1999). How do dominant designs emerge? Except for the most simple nonassembled products, the closing on a dominant design is not technologically driven because no technology can dominate on all possible dimensions of merit. Nor does the closing on a dominant design take place through the invisible hand of the market (Noble, 1984; Pinch et al, 1987). Rather, it occurs through social, political, and organizational competition among the alternative technological variants (Hughes, 1983; Baum, Korn, and Kotha, 1995; Yoffee and Cusumano, 1999; Tushman and Rosenkopf, 1992). Dominant designs emerge out of the struggle between alternative technological trajectories initiated and pushed by competitors, alliance groups, and governmental regulators each with their own political, social, and economic agendas. This social construction of technology has been thoroughly documented in a range of industries (see Bijker et al, 1987). In an unusually rich process oriented case study, Van de Ven et al, (1999) provide detail on the complex regulatory, organizational, and physician dynamics in the evolution of industry standards in the cochlear implant industry. This 9

10 case not only describes the evolution of standards at the subsystem level, it also describes how one firm, 3M, worked outside its boundaries to shape industry standard.s Similarly, the emergence of the VHS over Beta as a dominant design in the videocassette recorder industry illustrates the process of technology cycles, and the social and political influences on this process. Cusumano, et. al. (1992) used historical data on technological capabilities, mass market demands, and organizational strategies for six key companies developing videocassette recorders between 1975 and They demonstrated that though initially Beta was more technologically advanced (tapes held greater amount of information with a higher resolution, etc.) and initially captured more of the market, JVC was able to beat Sony through proactive alliances with strong producers and distributors. Dominant designs are watershed events in a technology cycle. Before a dominant design emerges, technological progress is driven by competition between alternative technologies. After a dominant design emerges, subsequent technological change is driven by the logic of the selected technology itself (see Figure 1). The closing on a dominant design shifts innovation from major product variation to major process innovation and, in turn, to incremental innovation to building on, extending, and continuously improving the selected variant. These periods of incremental innovation lead to profound advances in the now standard product (Hollander, 1965; Myers and Marquis, 1969; Abernathy, 1978). In contrast, the consequences of betting on the wrong design are devastating particularly if that design is a core subsystem (e.g., IBM s 10

11 losing control of the microprocessor and operating system in PCs to Intel and Microsoft). Technology cycles apply both for product subsystems and for linking technologies, and they apply across product classes the only difference between hightech (e.g., disk drive) and low-tech (e.g., concrete) industries is the length of time between the emergence of a dominant design and the subsequent discontinuity. Technology cycles highlight the points at which senior teams have substantial impact on firm and product class evolution versus where they only have minor impacts (Christensen et al, 1998; Hambrick and Finkelstein, 1996). During eras of ferment actions by senior teams affect both the nature of technical change as well as organizational fates. During eras of incremental change, however, managerial influence on technical progress in the existing trajectory is sharply limited. Senior teams can, however, destabilize their product class by initiating different types of innovation-- by driving innovation streams. Innovation Types and Innovation Streams For complex products the locus of innovation occurs both within subsystems and with those technologies that link subsystems together (Baldwin and Clark, 2000; Schilling, 2000). Decomposing products into components and linking mechanisms and clarifying target markets/customers helps untangle incremental, architectural, discontinuous, and market innovation types (Henderson and Clark, 1990; Christensen, 1998; Tushman and Murmann, 1998) [see Figure 2]. 11

12 Incremental innovations These are those innovations that push the existing technological trajectory for existing subsystem and linking mechanisms. Such innovations are associated with significant product improvement and enhanced customer satisfaction over time (Myers and Marquis, 1969; Hollander, 1965). New 8 Disk Drives Small Copiers Small Motorcycles Swatch (SSIH) Digital Imaging (Polaroid) Markets Existing 12 Disk Drives Mechanical Movements (SSIH) BTEar Hearing Aids Ciba / Tilt Bias Ply Tires Analog Imaging ITEar Hearing Aids Photolithography Typesetting Quartz Movement Ciba/ Seed Digital Imaging Radial Tires Incremental Architectural Discontinuous Innovation Type Figure 2 Innovation Streams (adapted from Tushman and O Reilly, 1997) Architectural innovations These involve shifts in subsystems and/or linking mechanisms (Henderson and Clark, 1990). These relatively simple innovation types are often initially targeted to new markets. Henderson and Clark (1990) identified architectural and modular innovation and explored their impacts on firms in the photolithography industry. Archival data and extensive interviews with the senior management teams of incumbents and new entrants indicated that the contact aligner and its successor the proximity aligner were 12

13 exactly the same in their component subsystems, but altered the way that these subsystems interacted with one another. Over four such architectural innovations, incumbents treated the innovation as if it were an incremental innovation. Established organizations either failed to recognize the need to restructure their organization, to seek new markets, or to alter their production processes. In each case, incumbents failed to adapt to seemingly minor technical change. Similarly, Starkey s move into the fashion hearing aid market, Honda s early move to smaller motorcycles, the migration of disk drive technology from mainframes to personal computers, and Canon s smaller copiers are examples of architectural innovations that transformed their product classes. While these innovations were technologically simple, in each case incumbents fumbled their future (Kearns and Nadler, 1992; Henderson, 1995; Christensen, 1998). Discontinuous innovation This involves discontinuous technological change in a core subsystem. Such technical shifts trigger cascading changes in other less core subsystems and linking mechanisms (Tushman and Murmann, 1998). For example, in the photography industry, digital imaging was a competence destroying change in the camera s core subsystem. Tripsas and Gavetti (2000) document via interviews and industry data how Polaroid, a leading incumbent, was able to generate the technological know-how, but was unable to go to market with digital cameras due to profound organizational inertia and stunted senior team cognitive models. Sull et al (1997) and Sull s (1999) historical analysis of the tire industry demonstrates how difficult it is for incumbents to initiate discontinuous innovation even when they have all the requisite technical competencies. 13

14 Where incremental, architectural, and discontinuous innovations are defined by their technological impact on subsystems and/or linking mechanisms, market based innovations are those innovations that are targeted to new markets or customer segments. These often technically simple innovations are frequently missed by incumbents at their considerable peril. Christensen (1998) conducted a census of all disk drive produces from He found that in the transitions from 14 to 3.5 disk drives, in every case incumbents were displaced by new players. Each of these architectural innovations were initially most useful to newer, less demanding customers from the incumbents perspective. In every case, incumbents ceded these new markets to new players and in every case, these new players proceeded to subsequently move up market. Christensen (1998) observes that these technologically simple innovations are disruptive to an incumbent s existing organizational architectures (note that they are not technologically disruptive). Building on products as composed of subsystems and linking mechanisms, Sanderson and Uzumeri (1995) developed the notion of product platforms and families. A platform is a set of core subsystems; a product family is a set of products built from the same platform. These product families share traits, architecture, components, and interface standards. For example, once Sony closed on the WM-20 platform for their Walkman, they were then able to generate more than 30 incremental versions within the same family. Over a ten-year period, Sony was able to develop four Walkman product families and more than 160 incremental versions of those four families. Devoting sustained attention to technological discontinuities at the subsystem level (e.g., the flat 14

15 motor and the miniature battery), closing on a few standard platforms, and generating incremental product proliferation helped Sony control industry standards and outperform their Japanese, American, and European competitors in this product class (Sanderson and Uzumeri, 1995). Thus lying behind S-shaped product life cycle curves are fundamentally different innovation dynamics (Klepper, 1996). Eras of ferment are associated with discontinuous product variants. Dominant designs are associated with fundamental process innovation. After dominant designs emerge, the subsequent eras of incremental change are associated with product modularization and are fertile periods for incremental, architectural, discontinuous and market based innovation. Given the nature of technology cycles, then, the roots of sustained competitive advantage may lie in a firm s ability to proactively initiate multiple innovation types-- to initiate streams of innovation. Yet the external push for innovation streams runs counter to internal inertial forces. Incumbents, even when armed with technological capabilities, are held hostage to their successful pasts. It seems to be difficult for incumbents to develop the diverse competencies and organizational capabilities to shape and take advantage of dominant designs, to shape architectural innovation, to move to less demanding markets, or to introduce substitute products before the competition. Ambidextrous Organizations A firm s dynamic capabilities are rooted in its ability to drive innovation streams - to simultaneously create incremental, architectural, market and discontinuous 15

16 innovations (Tushman and O Reilly, 1997; Teece, 1996). Ambidextrous organizations are complex organizational forms composed of multiple internally inconsistent architectures that are collectively capable of operating simultaneously for short term efficiency as well as long term innovation (Duncan, 1976; Weick, 1979; Bradach, 1997). Such heterogeneous organizational forms build in the experimentation, improvisation, and luck associated with small organizations, along with the efficiency, consistency, and reliability associated with larger organizations (Eisenhardt and Tabrizi, 1995; Imai et al, 1985). Organizational architectures for incremental innovation are fundamentally different from those for all other innovation types. Continuous incremental improvement in products and processes and high-volume throughput are associated with organizations with relatively formalized roles and linking mechanisms, centralized procedures and processes, efficiency-oriented cultures and highly engineered work processes (Eisenhardt and Tabrizi, 1995; Burns and Stalker, 1961; Nadler and Tushman, 1997). Efficiency-oriented units drive continuous improvement, exploitation, and the elimination of variability and have relatively short time horizons (Levitt and March,1988; Levinthal, 1997). Such units are often relatively large and old, with highly ingrained, taken-for-granted assumptions and knowledge systems (Milliken and Lant, 1991) [see Figure 3]. In contrast to incremental innovation, discontinuous innovation emerges from entrepreneurial organizational architectures. Entrepreneurial units/organizations are relatively small; they have loose, decentralized product structures, experimental 16

17 cultures, loose work processes, strong entrepreneurial and technical competencies, and relatively young and heterogeneous human resource profiles (McGrath and MacMillan, 2000). These units generate the experiments/options from which the organization s senior team can learn about the future (Levitt and March, 1988; McGrath, 1999; Leonard-Barton, 1995). These units build new experience bases, knowledge systems, and networks to break from the larger organization s history. They generate variants from which the senior team can make bets on possible dominant designs, new customer segments, and technological discontinuities (Burgelman, 1994; Nonaka, 1988). In contrast to the larger, more mature, efficiency-oriented units, these small entrepreneurial units are inefficient and rarely profitable and have no established histories (see Figure 3). These entrepreneurial units may be created internally or acquired externally through acquisition, contract research, joint ventures or alliances (Teece, 1996; Roberts and Berry, 1985; Silverman, 1999). Architectural innovations take existing technologies and link these technologies in novel ways; they are built not on new technological breakthroughs, but rather on integrating competencies from both the efficiency oriented as well as the entrepreneurial subunits. While technologically simple, architectural innovations require fundamentally different linking structures, incentives, competencies and cultures from the existing architecture (see Iansiti and Clark, 1994). Because of the difficulties of building linkage capabilities in the context of incremental innovation, architectural innovations are often not initiated by incumbent industry leaders (Brown and Eisenhardt, 1995; Henderson, 1995). Similarly, taking existing products to new 17

18 markets must deal with the resistance of a firm s current customers and existing resource allocation processes. These sources of inertia often hold successful incumbents hostage to their pasts (Christensen, 1998; Rosenbloom and Christensen, 1994). Sales Ambidextrous Senior Team drives multiple strategies and innovation stream within business unit; anchored by single vision Tasks Mgt. Team Individuals Culture Exploration Ciba Crop Protection/ genetically engineered seeds; Seiko/quartz; Polaroid/ Digital Imaging Tasks Mgt. Team Mgt. Team Organization arrangements Exploitation Ciba Crop/Tilt; Seiko/mechanical; Polaroid/ Analog Imaging Individuals Culture Organization arrangements Strategic reorientations Time Figure 3 Ambidextrous Organizations: Exploitation and Exploration Within Business Units While incremental, architectural, and discontinuous innovations require fundamentally different organizational architectures, to drive streams of innovation these contrasting architectures must reside in within a single business unit. Senior management s challenge is to build into a single organization multiple internally consistent organizational architectures that are themselves inconsistent from each other (see Figure 3). In ambidextrous organizations, single organizations host multiple cultures, structures, processes, management teams, and human resource capabilities in order to be incrementally innovative while at the same time creating products that may 18

19 make the existing product line obsolete (Bradach, 1997; Duncan, 1976; Tushman and O Reilly, 1997). Ambidextrous organizational forms build in organizational capabilities to simultaneously explore and exploit, to decrease variance as well as to increase variance at the same time. March (1991) suggests that organizational learning is accentuated when there is both the exploitation and exploration simultaneously. Exploitation of knowledge extends existing knowledge resulting in predictable, positive returns (at least in the short term.) In contrast, exploration is inherently experimental and often inconsistent with previous knowledge. Exploration accentuates variation which in the short term is risky. Using a computer simulation that includes the internal strategic pressure and external competitive pressure, March demonstrates the survival value of balancing these inconsistent learning modes. Similarly, Burgelman (1991) identified contrasting learning modes in his field study of innovation at Intel. Induced strategic processes involve the continuation of current organizational strategy. These tended to be incremental changes, such as the continued development of memory chips. In contrast, autonomous strategic processes, which created the possibility of new products often to different customers, were typically outside the scope of the extant strategic initiative. Intel s senior team was less enthusiastic about funding these ventures because the uncertainty for market success was higher. Burgelman s analyses document how innovation streams require fundamentally different learning modes operating in parallel. 19

20 Ambidextrous organizations build in cultural, structural, and demographic contradictions. These internal contradictions are necessary if the organization is to be able to produce streams of innovations. Yet these contradictions create instability and conflict between the different organization units between those historically profitable, large, efficient, older, cash-generating units and the young, entrepreneurial, experimental, cash-absorbing units. Because the power, resources, and traditions of organizations are usually anchored in the older, more traditional units, these units usually work to ignore, sabotage, or otherwise trample entrepreneurial units (Cooper and Smith, 1992; Leonard-Barton, 1992; Miller, 1994; Morone, 1993; Hamel and Prahalad, 1994). Independent of the firm s boundary, ambidextrous organizations require subunits that are highly differentiated, weakly linked internally, but tightly integrated through the senior team (Tushman and O Reilly, 1997). The certainty of today s incremental advance often works to destroy the potential of tomorrow s architectural, market, or discontinuous advance. For example, Tripsas and Gavetti (2000) s inductive analysis of Polaroid s experience with digital cameras documents the senior team difficulties in building these complex organizational forms. As the market leader in the instant photography industry, Polaroid developed incremental innovations of their hallmark SX-70 instant camera (eg. improving the photo quality subsystem), while at the same time seeking new markets with digital imaging cameras. Such a discontinuous innovation had the potential to disrupt their exiting film/analog franchise. The incremental improvements in the instant camera occurred within an existing organizational structure, while the development of the 20

21 digital camera technology occurred in its own separate laboratory, with 90% of its employees hired new for this project. While the products and technology were available, Polaroid s senior team was unable to take the strategic actions to capitalize on their digital investments. As Polaroid s senior team was unable to change their organizational, strategic and cognitive models, Polaroid remained trapped in its historically rooted analog market. Given the inertial, defensive, and political dynamics that exist in successful organizations, senior management teams must not only separate, protect and legitimize entrepreneurial units, they must also make the strategic and organizational decisions to take advantage of this internal and/or external experimentation. While ambidextrous organizational forms can enhance dynamic organizational capabilities, they are difficult to implement. In industry after industry, either exploitation drives out exploration or today s customer trumps tomorrow s. To drive innovation streams, senior teams must build in and sustain internally inconsistent architectures even as they are the locus of strategic integration. Ambidextrous Senior Management Teams Strategic integration across innovation streams is anchored in a senior team s understanding of these innovation and organizational dynamics and their symbolic and substantive actions (Tripsas and Gavetti, 2000; Pfeffer, 1982). Internal structural heterogeneity provides the senior team with the ability to improvise in the present even 21

22 as it experiments for the future (Gavetti and Levinthal, 2000; Brown and Eisenhardt, 1998; Iansiti and Clark, 1994). If these diverse capabilities can be integrated, they permit the organization to innovate for both today and tomorrow; but without integration, the potential of an ambidextrous organization is lost. A clear, emotionally engaging vision provides a strategic anchor from which senior teams can balance the contrasting requirements of innovation streams (Hamel and Prahalad, 1994; Collins and Porras, 1994). Simple, direct competitive visions create a point of clarity within which an organization can simultaneously host incremental and discontinuous innovations (Nonaka, 1998; Hurst, 1995). Sustained, consistent commitment to a unit s vision, even as strategies and objectives change, reinforces that vision. Commitment is further reinforced by senior management continuity and by their consistent behaviors in support of the vision (Pfeffer and Sutton, 2000). Through such clarity and consistency of vision the senior team can support the internally contradictory organizational architectures associated with ambidextrous organizations and still be seen as consistent and credible. For example in Ciba s Crop Protection Division, the senior team s aspiration of keeping crops healthy permitted the division to simultaneously work on chemical and biological technical trajectories. The senior team s composition and demography and its ways of working together are also powerful tools for achieving integration in ambidextrous organizations (Hambrick, 1998). A study of the top management teams in 24 firms in the electronic industry suggested that homogeneity in age and tenure but heterogeneity in expertise in top management teams lead to more positive team dynamics and increased ability to 22

23 make adaptive changes (O Reilly et al, 1993). While there are benefits in senior management continuity, there are also benefits in creating highly heterogeneous, demographically young senior teams. Such teams have the benefit of consistency in vision from the top along with the ability to import new team members with different competencies and expertise (Smith et al, 1994; Virany et al, 1992). For example, as Microsoft grew, Bill Gates broadened his senior team with managers from outside Microsoft with marketing, organizational, and technical skills (Cusumano and Selby, 1995). Highly effective senior teams have diverse competencies to handle the contrasting innovation demands of ambidextrous organizations. In contrast, senior teams that are demographically old and homogeneous are typically dominated by historically anchored perspectives and simple cognitive models (Milliken and Lant, 1991; Miller, 1994). Inertial senior teams get stuck in routinized processes and are unable to deal with the contrasting demands imposed by innovation streams (Hambrick and D Aveny, 1988; Boeker, 1989; Louis and Sutton, 1989). Organizations can develop diverse senior team competencies by importing executives from outside the firm, by creating diverse career experiences internally, by building heterogeneous teams within the organization, and often by executive team turnover (Hambrick, 1994, 1998; Zajac and Westphal, 1996). Effective senior teams are also able to work together in a way that takes advantage of their internal differences. Effective senior teams have internal processes that enable them to handle greater information and decision alternatives and to deal 23

24 with diverse points of view and contrasting opinions (Edmondson, 1999; Eisenhardt, 1989). In a survey of 53 management teams, Smith, et al (1994) demonstrate that senior team processes such as formal communication, informal communication, and social integration have an impact upon successful team performance independent of senior team demography. Diverse, self-critical senior teams with effective group processes not only get their own work accomplished but model appropriate ways to deal with conflict and cross-cutting priorities in the larger organization. In contrast, senior teams that send mixed messages, that cannot resolve their own conflicts, and that do not collaborate internally create highly unstable, politically chaotic organizations (Williams and O Reilly, 1998; Pfeffer and Sutton, 2000). Finally, senior management teams can create formal roles, structures, processes, and rewards to facilitate strategic integration. Particularly for architectural innovation, the development roles and formal linking mechanisms encourage integration within the senior team and across diverse parts of the organization. Team-based rewards that measure and value diverse types of innovation and collaborative team behaviors motivate team members to work together (Nadler and Tushman, 1997; Iansiti and Clark, 1994; Brown and Eisenhardt, 1995). Innovation Streams and Discontinuous Organization Change Within ambidextrous organizations, entrepreneurial units provide learning-bydoing data, variation, and luck to drive possible new dominant designs, architectural 24

25 innovations, and/or product substitutions. Whether done internally, through alliances or acquisitions, the senior team can learn about alternative futures from these entrepreneurial units. In contrast, those more mature units drive sustained incremental innovation and short-term learning. The senior management team can then draw upon these diverse types of innovation and learning to make strategic decisions when to initiate a dominant design, when to move on new customers, what product variant to bet on, when to initiate an architectural innovation, and/or when to introduce a product substitute. The success of strategic choices of dominant designs, new customers, architectural innovation, and/or product substitutes are only known after the fact. Through building ambidextrous organizational capabilities, the senior management team maximizes the probability that they will have both the expertise and the luck from which to make industry-shaping decisions proactively rather than reactively. Ambidextrous organizations create options from which the senior team makes informed bets on the future (Burgelman and Grove, 1996; McGrath, 1999). While correct strategic bets can be known only in retrospect, managerial action within the firm and with collaborators, alliance partners, and governmental agencies can affect the ultimate closing on an industry standard or the success of a product substitute (Teece, 1987; Rosenkopf and Tushman, 1998; Cusumano and Yoffee,1998). At the closing of a dominant design, the strategic innovation requirements within the firm shifts from a focus on major product variation to major process 25

26 innovation and then to sustained incremental innovation (Abernathy and Clark, 1985; Utterback, 1994). At product substitution events, for architectural innovations, and/or for taking an existing product to new markets, strategic management shifts from incremental innovation to major product, market or architectural innovation. As strategic innovation requirements shift at these junctures, so too must the dominant organizational capabilities (Sorensen and Stuart, 2000). Those organizational architectures, those structures, roles, cultures, processes, and competencies, appropriate during eras of ferment are no longer appropriate during eras of incremental change. Similarly, those organizational architectures appropriate during eras of incremental change are no longer appropriate during eras of ferment (Foster, 1987; Benner and Tushman, 2000). Because structural and social inertia is so powerful, managers can attempt to rewrite their industry s rules only if they are willing to rewrite their organization s rules. At these strategic junctures, shifts in a firm s innovation stream can be executed only through discontinuous organizational change (Tushman and Romanelli, 1985; Greenwood and Hinnings, 1993; Miller, 1994; Rosenbloom, 2000). For example, IBM s 360 decision in mainframes was coupled with sweeping shifts in IBM s structure, controls, systems, and culture. In contrast, bold strategic moves or great technology uncoupled from organizational capabilities leads to underperformance (Tripsas and Gavetti, 2000; Virany et al, 1992). Similarly, Sony s superior Beta technology format was to unable to counter JVC s combination of an adequate VHS technology coupled with brilliant organizational capabilities and strategic alliances (Cusumano et al, 1992). 26

27 Innovation streams are, then, rooted as much in reconfigured organizational architectures as in technological prowess (Sull, 1999; Rosenbloom and Christensen, 1994; Rosenbloom, 2000). Innovation streams and the associated frame-breaking organizational changes are often initiated by transformed senior teams (Meyer et al, 1990; Ancona, 1990; Virany et al, 1992; Romanelli and Tushman, 1994). While reorientations are risky and often done incompetently (eg. Carroll and Teo, 1998; Henderson, 1993), persistence in the face of a changing innovation stream is even more so. Further, if strategic reorientations are not done proactively, they have to be done reactively as with Burroughs in mainframes, SSIH in watches, Polaroid in digital cameras, and Oticon in hearing aids. Reactive reorientations (turnarounds) are more risky than proactive reorientations because they must be implemented under crisis conditions and under considerable time pressure (Hambrick et al, 1998). Senior management s challenge in leading discontinuous change is fundamentally different from that in leading incremental change (Weick and Quinn, 1999). Senior teams must build the capabilities to manage both, if they are to effectively manage across an innovation stream (Nadler, et al, 1995; Barnett and Carroll, 1995). Those more effective discontinuous changes are initiated and directed by the senior team, are shaped by an integrated change agenda, and are rapidly implemented driven by the senior team s vision and consistent actions (Nadler, 1998). These challenges are often associated with shifts in the senior team and within middle management (Kanter et al, 1992; Pettigrew, 1985, 1987). 27

28 If implemented through incremental change methods, reorientations run the risk of being sabotaged by the politics, structures, and competencies of the status quo (Virany et al, 1992; Kearns and Nadler, 1992). For example, in Ciba s Crop Protection Division, the transition from fundamentally different fungicides to EC-50 (their bet on a dominant design) was executed through sweeping changes in the division and through a new fungicide team. In contrast, breakthrough innovation at Xerox in the late 1970s and early 1980s and Polaroid in the 1990s were not coupled with corresponding organizational shifts. The politics of stability held Xerox and Polaroid hostage to their pasts (Smith and Alexander, 1990; Tripsas and Gavetti, 2000). Managing innovation streams is about managing internal paradoxes in the context of innovation streams: managing efficiency and innovation, tactical and strategic, incremental and discontinuous, today and tomorrow (Gavetti and Levinthal, 2000; Brown and Eisenhardt, 1998; Quinn and Cameron, 1988). Managing innovation streams is about consistency and control as well as variability, learning by doing, and the cultivation of luck. It is the crucial role of the senior team to embrace these contradictions and take advantage of the tensions and synergies that emerge from juggling multiple competencies simultaneously (Van de Ven et al, 1988; Hurst, 1995; Tushman and O Reilly, 1997). It is the senior team s role to bind these paradoxical requirements together through their substantive and symbolic actions. Conclusion 28

29 Even if periods of incremental change do build organizational inertia, organizations can create and shape innovation streams. Through building ambidextrous organizational forms and creating options from which the senior team initiates proactive strategic change, organizations can manage the rhythm by which each expiring strength gives birth to its successor. Prior organizational competencies provide a platform so that the next phase of an organization s evolution does not start from ground zero; evolution involves, then, learning as well as unlearning (Weick, 1979). Organizations can, then, renew themselves through a series of proactive strategic reorientations anchored by a common vision. Like a dying vine, the prior period of incremental change provides the compost for its own seeds, its own variants, to thrive following a reorientation in the subsequent period of incremental change. In this review we have taken a strong point of view on organizational evolution. As dynamic capabilities are not rooted in technology cycles, organizational architectures, senior teams, or change dynamics alone, our understanding of the nature of dynamic capabilities must be rooted in deeply understanding these modules of evolution and how they interact with each other. Yet each module is open to much debate - the nature of technology cycles, the nature of organizational architecture, the role of senior teams, and the nature of organizational change are all contested domains. Yet this energy is well founded, for these issues are both professionally interesting as well as managerially crucial. There is significant opportunity, therefore, in digging deeply into those controversies in technology cycles, organizational architectures, senior 29

30 teams and change, and perhaps even more so in building integrative theory and research across these interdependent domains. Acknowledgments We thank Lori Rosenkopf, Mary Tripsas and Joel Baum for their reviews and critical comments. References Abernathy, W. The Productivity Dilemma. Baltimore: Johns Hopkins University Press, Abernathy, W., and K. Clark. Innovation: Mapping the Winds of Creative Destruction. Research Policy, 14, 1985, Aitken, H. The Continuous Wave. Princeton: Princeton University Press, Aldrich, H. Organizations Evolving: Sage Publications, London, Ancona, D. Top Management Teams: Preparing for the Revolution. In Applied Social Psychology and Organizatoinal Settings, edited by J. Carroll. Hillsdale, NJ: L. Erlbaum Assoc., 1990, Anderson, P., and M. Tushman. Technological Discontinuities and Dominant Designs: A Cyclical Model of Technological Change. Administrative Science Quarterly, 1990, 35, Baldwin, C. and K. Clark. Design Rules: The Power of Modularity. Cambridge, Mass: MIT Press, Barnett, W., and G. Carrol. Modeling Internal Organization Change. American Review of Sociology, 1995, 21: Baum, J., H. Korn, and S. Kotha. Dominant Designs and Population Dynamics in Telecommunications Services. Social Science Research, 24, 1995,

31 Baum, J., T. Calabrese and B.S. Silverman, "Don't Go it Alone: Alliance Networks and Startup Performance in Canadian Biotechnology," Strategic Management Journal, 2000, 21(3), pp Benner, M. and M. Tushman. Process Management and Organizational Adaptation: The Productivity Dilemma Revisited. Working Paper, Harvard Business School, Bijker, W., T. Hughes and T. Pinch. The Social Construction of Technological Systems. Cambridge, Mass.: MIT Press, Boeker, W. Strategic Change: Effects of Founding and History. Academy of Management Journal 32, 1989, Bradach, J. Using the plural form in the management of restaurant chains. Administrative Science Quarterly, 42, 1997, Brown, S., and K. Eisenhardt. Product Development: Past Research, Present Findings and Future Directions. Academy of Management Review, 20, 1995, Brown, S., and K. Eisenhardt. Competing on the edge: Strategy as structured chaos. Boston, Mass.: Harvard Business School Press, Burgelman, R. Intraorganizational Ecology of Strategy Making and Organizational Adaptation. Organization Science 2:3, 1991, Burgelman, R. Fading Memories: A Process Theory of Strategic Business Exit. Administrative Science Quarterly, 39, 1994, Burgelman, R., and A. Grove. Strategic Dissonance. California Management Review, 38, 1996, Burns, T and G. Stalker. The Management of Innovation, Tavistock, London, Burton, M.D. "The Company They Keep: Founders' Models for High Technology Firms" in Schoonhoven, K. and E. Romanelli (eds). The Entrepreneurship Dynamic: Palo Alto, Stanford Univ. Press, in press. Carroll, G., and M. Hannan. The Demography of Corporations & Industries. Princeton University Press, Carroll, G. and A. Teo "Creative Self Destruction Among Organizations: An Empirical Study of Technical Innovation and Organizational Failure in the American Automobile Industry, " Industrial and Corporate Change 5:2, 1996,