Alianças Estratégicas na Implementação de Estratégias de Tecnologia - Yves Doz e Gary Hamel

37 The Use of Alliances in Implementing Technology Strategies YVES DOZ GARY HAMEL INTRODUCTION Strategic alliances are in vogue Many new strategic alliances are being created every year most of them with a substantial technological content or even a specific RD focus Doz 1992 While the phenomenon is not new for example alliances existed at the beginning of the 20th Century in such industries as nonferrous metals and chemicals the pace of alliance formation has accelerated since the late 1980s This paper first analyzes the reasons for the recently observed surge in the number of strategic alliances The intensification and sectoral spread of global competition is one central driving force calling for acceleratedquicker permutations of ever more complex skill sets Few firms have all the inhouse capabilities required to compete globally market access management and operational disciplines critical technologies and creative product and market ideas Most need to complement internal capabilities The first section of this paper analyzes what additional capabilities companies seek as they are confronted by growing often technologybased global competition and why alliances may give them access to or internalization of these capabilities 1 NO FIRM IS AN ISLAND In most industries companies are now racing towards what they see as the future high ground of industry leadership While a few global competitive battles featuring daily in the media such as those for high definition television notebook computers AIDS treatment civilian aircraft or motor cars capture a disproportionate share of the attention of researchers and policy makers alike many more such global races are going on for instance in the areas of new materials voice recognition computing electric vehicles communication networks membrane separation plant protection and growth regulation1 Some of the contestants are global companies that seek to expand their leadership in their traditional industry while others are smaller companies hoping to exploit significant innovations Yet other companies seeking an escape from declining industries are rushing into new fields in which they MANAGING LINKAGES are comparatively small and inexperienced and where their existing understanding skills and other intangible resources may be only tangentially useful Further new fields often result from hybrid technologies bringing several disciplines together eg bioelectronics or from creative ways of combining product innovation and process capabilities or from the integration of diverse technologies into complex systems In both cases the range of required technologies broadens To win the race to extend industry horizons and extract the benefits most companies increasingly find they are in need of an array of complementary assets Teece 1987 Complementary Capabilities Even large wellestablished firms which have the strongest heritage to defend see technology as the main determinant of competitive success in their core business As a result they may increasingly need access to new technology developed by other firms Only with new technologies can they leverage the market access capabilities they built eg distribution infrastructure and global brands capitalize on the management disciplines they cultivate such as quality management valueengineering or project management skills and continue to exploit and strengthen their reputation and brand image Market access skills and management disciplines are of little value unless they can be applied to a constant stream of new products and technologies Existing companies need innovative products to exploit existing distribution networks and to capitalize on the management disciplines they have cultivated innovation thus becomes the key to protecting the value and future longevity of past efforts Historically skills have been unevenly distributed between firms and between regions of the world Hamel 1990 European companies may create sciencebased inventions but often suffer from a poor understanding of markets and thus fail to bring those inventions to new markets Doz 1994 Hamel and Prahalad 1991 American companies may succeed at fundamental market innovation but fail to buttress initial successes with fastpaced improvements in cost and quality Japanese companies design production and product development processes for speed and efficiency De Meyer 1991 By pursuing manufacturing flexibility and agility they are able to engage in trialanderror product development and introduction Nevens Summe and Uttal 1990 Hamel and Prahalad 1991 Japanese companies have also often developed a capability to blend various technologies into collective knowhow that results in deeply rooted and widely held competencies for example in joint product and process development or in commercialization skills Prahalad and Hamel 1990 Fruin forthcoming Nonaka and Takeuchi 1995 Yet Japanese companies sometimes have to rely on others for market access or fundamental innovations in sciences While significant productivity gaps still result from historically different approaches and priorities toward management disciplines Womack et al 1990 approaches to these disciplines now tend to converge within an industry and the productivity differences between the survivors are narrowing It is therefore unlikely that differences in management disciplinesno matter how wide a gap they may create at some point in time between competitorswill result in sustainable longterm competitiveness differences between survivors1 While constant improvements in management disciplines can exhaust some weaker follower companies and lead to their demise survivors are increasingly evenly matched in costs quality and cycle time Further these improvements may only generate diminishing returns beyond 6 sigma quality for example what is the competitive advantage to be gained from further improvement Similarly as European American and Japanese companies rush to build marketing and manufacturing networks across the triads markets market access asymmetries may disappear and competitors with more or less matching worldwide market access capabilities emerge They will cover the world in approximately equal terms In summary increased global competition has forced companies which historically developed different capabilities and skills into a process of rapid convergence In this process they are increasingly trying both to close skill gaps visàvis their competitors and to develop new competencies and skills that their competitors will find difficult to imitate and match The Race to the Future This convergence process takes place along several dimensions global market access infrastructure and skills management disciplines that improve the integrity and the value of the products and competencies and technologies that allow product innovation and market creation Yet the speed at which a company can achieve convergence varies with more rapid progress in some dimensions Increasing familiarity with world markets has made competition for market access less dependent on skills and more on resources Management disciplines become increasingly wellcodified and thus more easily transferable although their successful implementation may still benefit from a mentor Competencies are much more difficult to codify and thus to transact Nevertheless competitive victory will belong to those who build new competencies the fastest The pace at which a company can move toward convergence thus depends on the type of new capabilities it requires Market access may require greater or fewer inputs of efforts and local knowledge depending on which market is to be penetrated The structure of distribution the type of products sold and possible protection policies all play a role Developing channels alone to penetrate a complex market in which protectionism may arise is a difficult slow and risky process Alliances with distributors can accelerate market penetration The early reliance of Japanese companies on US mass merchandisers in the consumer electronics industry gave the Japanese manufacturers the scale needed for automating productionand achieving very high qualitywithout having to invest in market development Hamel and Prahalad 1985 Similarly Japanese companies now use alliances with local companies to learn about markets for more sophisticated products such as computer systems and telecom networks Indeed historically market access alliances have generally been the most widespread form of strategic alliances and joint ventures eg Franko 1973 Harrigan 1986 Better understanding and longer experience of global markets may reduce the need for market access alliances and global competition increases the pressure for manufacturers to directly control their distribution It is thus not surprising to observe a decline of market access alliances relative to other alliances Doz 1992 Why Technology Alliances Through patents licensing rights and technology transfer agreements the market for technology may be relatively efficient The market for competencies may be much less efficient Developing a complex partly tacit competence is a problem of example as much as of cash in the absence of a mentor with whom to copractice the competence as in an apprenticeship it may be impossible to develop the competence autonomously At best independent development may be a long timeincompressible process Dierickx and Cool 1994 Alliances may allow one company to intercept the skills of another and give it the opportunity to close skill gaps much faster than internal development would allow For example General Motors used its joint venture with Toyota to learn lean manufacturing processes and new labor relation practices Similarly Thomson the French consumer electronics group relied on its alliance with JVC to learn to mass produce the micromechanic subsystems key to successful video cassette recorder production Hamel and Doz 1990 Many of the most critical current technologies are increasingly embedded ie firm and contextspecific It is not the abstract theoretical understanding of technologies but their effective daytoday practice that matters Take VLSI memories understanding their technology is not conceptually difficult but designing actual memories and running a highyield memory factory is harder The hurdle is practice not theory Embedded technologies based on systemic skills and tacit collective process knowhow are less easily imitable than the increasingly wellcodified management disciplines such as total quality management and concurrent engineering and harder to match than market access skills Prahalad and Hamel 1990 Dierickx and Cool 1994 Itami and Roehl 1987 Related to the embedded nature of technology we observe another difficulty the growing interdependence between previously standalone technologies and the need to incorporate a wider range of technologies into complex systems and integrated solutions Cantwell 1989 As the range of technologies to be integrated widens the required combination of technologies is increasingly difficult for single firms to master Managers also believe that the integration capabilities to bring these technologies together will establish a new hierarchy within industries System integrators who are close to top management issues and endusers will define solutions prime contractors will lead in their development and subcontractors will follow in their implementation The current convergence of information technology services software development and strategy consulting or the concerns of aerospace companies for system and mission integration skills bear witness to this trend at least in some industries The embedded and systemic characteristics of new technologies which makes them more akin to competencies have significant implications for alliances The more embedded a technology the closer the cooperation needed to share and transfer it As a result we are likely to witness an increasing number of alliances driven by the motives to learn new skills and more attempts at more intimate closer alliances In sum alliances may provide vehicles for faster less costly and more permanent skill borrowing and internalization the ability to use and improve the new skill independently from the partner than other approaches to developing new competencies particularly when these skills are embedded and largely tacit Hamel 1990 In that sense alliances may be most useful at the cutting edge of the learning agenda to access and internalize technologies and knowhows that are embedded largely tacit uncodified and thus difficult to access via contractual approaches that do not involve a close collaboration between the partners This is the case of many new technologies which are themselves a growing source of competitive advantage when compared to market access and management disciplines Cospecialization or Internalization Embeddedness beyond a point can become a deterrent to internalization The very tacitness and embeddedness of a technology make it difficult to learn even in the copractice apprenticeship that characterizes many technology alliances The required learning effort may exceed the resources and the capabilities of the firm The risks involved may also be seen as too high because of uncertain imitability The outcome of the learning process may not be easily anticipated Access to the benefits provided by the use of the partners skills rather than internalization may thus be chosen It may not be necessary to learn the partners skills when access is achieved by proxy In fact internalization objectives may interfere with effective complementation and interweaving of skills Visible internalization attempts are likely to raise the partners suspicion which may erode goodwill and willingness to contribute the skills required for successful collaboration The relative balance in priorities between technology and skill complementation on the one hand and technology and skill internalization on the other is likely to depend on various factors First it depends partly on the scope of the learning opportunity Whether the skills to be exercised in the alliance extend to other activities is the key issue here For instance when Thomson Consumer Electronics started its alliance for VCR manufacturing with JVC back in 1983 it was clear to Thomsons management that the comanufacturing of VCRs provided an opportunity to learn a set of micromechanical precision manufacturing skills that would be useful not just for the joint production of VCRs or even for other VCRs made by Thomson but for a whole range of new products where precision manufacturing was also important such as CD video disks and players This provided a strong motivation for Thomson to learn JVCs skills Conversely Corning in an optical fiber alliance with Siemens did not see much value in trying to internalize Siemens telecommunication networking skills as they are not pertinent to Cornings other businesses Second the will to learn skills may depend on the perception by the partners of each others positions It was clear to Thomson Consumer Electronics for example that while it cooperated with JVC in Europe it also competed against JVC at the level of distribution brand image and market share in other regions and to some extent in Europe too Indeed the partner who took most of the JVs productionie competed most effectively in the end marketalso gained more influence on the JV as its economics became more dependent on the partner with higher sales Conversely both General Electrics and SNECMAs managers attributed a good part of the success of their very successful jet engine allianceCFMto the fact that they did not compete against each other in any segment of the civilian jet engine markets Mutual interest was also reinforced by competing against common rivals Pratt Whitney and Rolls Royce Third the nature of the skills themselves may facilitate or hinder cooperation The more distinct and different the more difficult the learning and the more cospecializedin the sense of jointly creating value that would not be available to the partners without these combined skillsthe more the continued goodwill of the partners is essential hence internalization is both more difficult and less attractive Similarly the more skills are embedded in other businesses which are not themselves part of the alliance the more difficult will be attempts to internalize If Siemens competence in transmission systems is more a result of its overall range of telecom equipment and software skills than the result of the cooperation with Corning Corning will be quite satisfied to access this cluster of competencies knowing full well that trying to encroach into them or substitute for them would be senseless Siemens may reach a similar conclusion if it observes that Cornings excellence in optical fibers is rooted in a cluster of competencies that are nurtured by many other Corning businesses While what is exchanged through the alliance is limited what is created in the alliance is important Conversely one can consider NUMMI the famous joint venture between General Motors and Toyota What was created in the NUMMI alliance is relatively minor a modest number of small cars assembled in a joint plant In GM and even Toyota in the United States the production of that plant is dwarfed by each partners own production Yet what is exchanged through the joint venture is critical a firsthand experience of copracticing Toyotas manufacturing methods for GM a direct experience of managing US unionized labor for Toyota Alliances that are fora for exchange rather than crucibles of new value creation are likely to be intrinsically more difficult for the partners to manage successfully Single Relationships or Networks Some companies alliance network is but the sum of individual discrete alliances with little coordination and cumulativeness between alliances Although the company may end up with many different alliances it may have entered each on its own merits as if other alliances did not exist and also manage each of them as a discrete operation While this is a viable approach where external constraints provide particularly strong incentives for cooperationas in the European aerospace industrythe discrete treatment of each partnership may also constitute a missed opportunity for leverage between alliances Corning for example is very clear about the nodal nature of its position at the center of a network of alliances each capitalizing on a cluster of core technologies in the inorganic chemistry area leveraged via focused alliances into a number of application domains Technologies and competencies likely to be applicable to several alliancesacross both geographies and productsare consciously cultivated Other companies use a virtual network of alliances over time as a ladder to learn from one alliance to the next and use the learning from each alliance to bargain for a stronger position in subsequent alliances While each alliance still needs to be managed effectively for its own sake on its own terms the fact that each belongs to a network opens new avenues for more ambitious technology strategies that build on the whole network and guide its further development in a strategically coordinated manner To engage in a network of alliances without managing them in a coordinated fashion results in missed strategic opportunities In sum we can categorize the management issues faced in technology alliances along two dimensions 1 Whether the main logic of the alliance is to create new value within the alliance proper mostly through technology and skill complementation and cospecialization or to use the alliance as a conduit to exchange skills and thus to transfer value between the partners mostly through skill learning and internalization 2 Whether the alliance should be seen and managed as a discrete relationship a single alliance as if other alliances did not exist or whether the management challenge is more accurately captured in terms of complementarity between alliances and evolution of a dynamic network of alliances over time These categories are summarized in Figure 371 below While many alliances obviously straddle several categories we will first consider single alliances that fit clearly in one category and net Individual Network of Alliance Alliance Capability Complementation GESNECMA Corning Glass Capability transfer Thomson JVC Fujitsu Fig 371 Technology alliance management categories work with a clear purpose We will then analyze more complex and more ambiguous alliances which straddle multiple categories The alliance between General Electric and SNECMA in the jet engine business mentioned earlier provides an interesting example of successful capability complementation whereas the Thomson JVC alliance provides an example of a companyThomsontrying to use the alliance to appropriate and internalize some of its partners capabilities Corning Glass provides an example of a network of alliances conceived and managed as a series of opportunities to leverage a common cluster of competencies across many alliances The history of Fujitsus relationships with its partners complemented by some other examples will show the value of a nodal position in a skill acquisition network 2 MANAGING VARIOUS TYPES OF TECHNOLOGY ALLIANCES Complementary Capabilities in an Individual Alliance The General ElectricSNECMA alliance which serves as an example of successful capability complementation in an individual alliance had many reasons to succeed Some were extrinsic to the alliance management itself others the result of the alliance design and management process In the early 1970s SNECMA had envisioned the need for a 10 ton thrust engine for the civilian market but lacked the experience and credibility to sell to airlines and maintain engines on a worldwide basis It turned to GE which had independently identified the same market segment had a military engine that could be adapted but lacked the resources needed The collaboration started in 1973 following the removal of a Department of Defense bar on the export of engine cores by GE to France Fortunately it was possible to mate GEs core with the SNECMAdeveloped systems mainly the main bypass fan without divulging GEspecific information other than the size and functional specifica tions of the coreallowing GE to get US government approval for the collaboration The task interdependence between the partners was kept very low with black box cores shipped to France but very little joint work between the two companies The engine was aimed at a market segment the socalled 10 ton engine referring to the engine unit thrust which everyone expected to be important but failed to materialize Although prototypes of the new engine coded CFM56 ran well as early as 1976 no sales were made The program was on hold to the deep embarrassment of the partners It was nearly cancelled by the French government and GEs corporate management in 1979 before the first sales were finally achieved to reengine older civilian transport planes and military tankers in the United States In retrospect managers of both partners saw a hidden benefit to this delay It had given them the time to proceed with the project and learn to work together in an unhurried fashion and it had tested their commitment to the project In sum despite the delay in the market development the alliance offered great value creation potential to both companies It allowed them to enter the civilian markets and to become challengers to Pratt Whitney and Rolls Roycefrom which they ultimately took market leadership in the 1980sin a way they could not have achieved on their own given limited skills and credibility particularly after sales service SNECMA and limited resources GE The partners were also highly strategically compatible in fact sharing a similar objective Further in the case of GESNECMA the partners were also organizationally compatible Both companies shared similar technical values and their top management knew and respected each other The fact that the jet engine world is relatively small and that both the head of GEs jet engine business Mr Neuman and the CEO of SNECMA Mr Ravaud were not conventional executives steeped in corporate hierarchies but also deeply committed to the jet engine business facilitated the cooperation Each partner undertook significant efforts to adjust to the other A joint venture was created as a small thirty or so person group with its headquarters in Cincinnati where GEs aircraft engine business is centered but headed by a Frenchman and more or less equally staffed by secondees from both partners as a way to facilitate the relationship with the more geographically distant French partner SNECMA sent its managers to participate in GEs internal management development programs while GE executives stationed in France made efforts to blend with the culture and learn the language of SNECMA SNECMA also reorganized some aspects of its operations so their structure would match that of GE establishing direct counterparts in each organization Following the initial development period and in particular once the engine entered service and the lifting in the early 1980s of certain US Department of Defense restrictions on technology sharing the two partners worked closely on improving the basic engine and developing new versions for new aircraft programs this time and started to learn from each otherparticularly on manufacturing process materials technologies and quality management The fact that the cooperation had been established on the basis of workshare limiting the haggling over the respective values of their contributions to one momentous but early negotiation process made further collaboration and exchange of technical data easier Although both companies had strong proud cultures they were also receptive and their managers were willing to learn from each other Further the high stakes of the alliance and the joint crossing of the desert in the late 1970s cemented stronger bonds between them The workshare split following the initial negotiation also facilitated cooperation by designing out of the relationship many of the more contentious issues that might have marred cooperation The two organizations were thus able to learn to work together and to learn from each other unhampered by organizational discrepancies or by recurring conflicts in the collaboration process This situation can be contrasted with that of a less successful alliance the ATTOlivetti attempt to work together in the mid1980s The two organizations were hardly compatible Olivetti was a nimble organization used to competing in a very tough IBMcompatible computer market and to making quick sometimes intuitive decisions in that fastchanging competitive environment ATT was emerging from a long period as a telephone monopoly in which it had set its own pace and created its own markets Its decisionmaking processes were slow and explicit relying on strategic planning procedures and a wide basis of consensus or at least coalition among product line geographic and functional executives In contrast Olivetti was led by a very small team grouped around Carlo de Benedetti Further the ATTOlivetti partnership started not with a joint venture and a broad onceandforall agreement but as a series of periodically renewed supply contracts between the partners The renegotiation of these contracts threw the partners back into a conflictual mode every six months or so while the deep differences in their organizations and decisionmaking styles and speeds made a joint approach to problemsor even a mutual understanding of issuesextremely difficult Predictably the alliance never developed into a true collaboration and fell in disarray when irreversible commitments were called for to a new joint product line in minicomputers or when economic circumstances became more difficult with the fall in the value of the dollar for personal computers and typewriters As the relationship between SNECMA and GE strengthened the initial cospecialization decreased As the engine went into service and the development priorities shifted to improving its reliability and durability the clearcut division of tasks between the partners that shaped the early phases of the alliance could not be maintained Improvements and reactions to problems faced by engines in use needed closer and less predictable patterns of collaboration and information exchange Improvements interactively affected all subsystems irrespective of which partners had designed and supplied them This led to closer collaboration between them and decreased specialization By 1991 the skill bases had converged very significantly leading even to a role reversal in the development of some new products SNECMA for example is developing the high pressure core of a new generation version of the CFM engine while GE develops the low pressure fan Although one can thus argue that skill internalization took place quite extensively over time leading to a possible weakening of the alliance the partners were kept together by other forces As mentioned earlier they have carefully avoided competition in the civilian marketsby being partners in all their programs albeit not necessarily through the CFM joint ventureand limited competition in the military markets Further the extremely high costs of new product development as well as the limited market size for new engines now strongly encourage further cooperation almost irrespective of technology and skill complementarity In other words the level of mutual benefits versus the dissolution of the alliance as well as that of trust and commitment between the partners is now such that the alliance is likely to continue and further deepen in the future In planning for or assessing individual skillbased alliances it is important as illustrated by the analysis of the GESNECMA collaborationin contrast to the ATTOlivetti attemptsto focus on a few critical issues 1 How will the alliance create value In particular what is the amount of cospecialization between the technologies and skills of the partners offering the potential of value creation not otherwise available to the partners It is important also to consider the risk of each partner encroaching on the technology and competence domain of the other The more limited that risk the more genuinely cooperative rather than potentially competitive the alliance Lastly it is useful to compare the alliance as a means to combine cospecialized skills to alternatives such as acquisitions mere suppliercustomer agreements licensing and technology transfer agreements and internal development 2 What is the strategic compatibility between the partners Even when the alliance creates value it is possible for the partners to disagree on priorities and strategies and thus fail to reach a lasting agreement External factors totally independent from the alliance may also destabilize it It is thus essential to assess the potential compatibility between partners priorities and the resilience of the alliance to external factors 3 How compatible are the partners organizations and cultures As we discussed SNECMA and GE were compatible and SNECMA took practical measures to make itself even more compatible whereas ATT and Olivetti were essentially incompatible but recognized this only slowly Exposing the partner organizations to each other led to convergence in the GESNECMA casesas the starting positions were already close enough to allow the management of both companies to quickly reorganize areas of required improvement ATTs and Olivettis starting positions were so remote that even identifying areas and approaches for improvement proved difficult 4 How to make the process of collaboration convergent When the two organizations and cultures are deeply different it is important to pay explicit attention to the design of the collaboration process This might mean matching the sequence of mutual commitments demanded from each partner with the level of understanding and trust achieved together Farreaching collaborative agreements may not be easily obtained too early in an alliance until partners are close enough Hence rather than focus on the upfront resolution of all differences it may be more practical to design an accelerated convergence process to foster mutual understanding between the partners 5 How effective is the design of the alliance Here again the contrast between CFM and ATTOlivetti is instructive The structure of the CFM alliance effectively designed out many of the potentially contentious areas between the partners such as transfer pricing With these out of the way collaboration could proceed unimpeded Conversely the design of the ATTOlivetti alliance as a series of contracts led to divisive discussions being reentered periodically 6 How balanced are contributions and benefits over time In principle alliances where benefits accrue early to one partner and late to the other are perfectly feasible and often very useful In practice however they are rife with tension unless the level of mutual trust is very high and the later benefits are not substantially more uncertain than the earlier ones 7 How strong are the expectations of future benefits Successful cospecialization alliances have a tendency to grow and continue over time creating a constantly extended shadow of the future which itself contributes to current trust and cooperation The more the time horizon of the collaboration can be extended the more likely the collaborative behavior in the shorter term When partners attend carefully to the above issues alliances based on technology and skill complementation and cospecialization are the most robust alliances since the benefits they provide their partners remain dependent on the continuation of the alliance One can observe for example that Corning has relied extensively on such alliances in its many businesses and that most of these alliances have proved very resilient over time Corning uses a set of alliance selection and management principles that consistently addresses the issues raised above In addition however Corning faces a set of distinct issues that stem from its reliance on not just a few individual alliances but on a whole network Managing a Network of Complementary Capability Alliances Corning Glass is unique among major corporations in deriving the majority of its turnover from joint ventures and alliances as a matter of its own choice From the outset Corning had been a technologydriven specialty glass company It established an RD lab as early as 1908 Early on Cornings management had entered a series of joint ventures to combine its technologies with partner companies offering complementary technologies skills and capabilities for market access and understanding In 1937 an alliance with PPG gave it access to the building industry in 1938 an alliance with Owens Illinois provided an entry into glass fibers and in 1943 Dow Corning was created to develop applications for the silicon products invented by Corning Over time building on these early experiences Corning developed a strategic approach to technology alliances It sought partners providing a high degree of complementarity and cospecializationmainly by bringing application end product and user market knowledge and skillswhile minimizing the threat of encroachment posed by these partners For example in a more recent alliance with Siemens to develop the optical communication fiber business Corning provided the fiber development and manufacturing skills Yet it relied on Siemens expertise in communication networks to develop the market and to help guide Cornings technology development efforts Corning was not about to develop expertise in communication systems while Siemens was not interested in nor able to contribute much to the development and manufacturing of the fiber itself The two firms were complementary and cospecialized and their overlap was limited to fiber connectors and splicers that both made Although one could argue that mere supply contracts rather than joint ventures might suffice Corning preferred equity joint ventures and treated these as strategic partnerships rather than mere investments Cornings top management felt that the early uncertainties about the new markets it sought to develop and the initial information asymmetries between partnersCorning knowing component material and process technology the partner knowing system technologies and market and customer characteristicswere best overcome via equity alliances pooling its interests and that of its partner in a single entity Corning was also increasingly concerned not just with application development and market access but also with combining its technologies with that of its partners and with developing markets rapidly ahead of competition Hence the market access speed provided by partners already engaging the appropriate market linkages was essential to Corning In each of its many alliances Corning applied these principles of complementarity cospecialization and rapid application development while minimizing overlaps and risks of encroachment between the partners and Corning In addition Corning had developed an expertise in managing partnerships and a commitment to their success among its top management which allowed Corning to be much more successful than less experienced companies in the management of alliances Yet Cornings identity as a network of alliances raised several difficult issues First alliance partners had a tendency to pull Corning in too many directions Serving a variety of strategic alliances ran the risk of technology dispersion spreading limited RD resources and capabilities too thinly Cornings implicit strategic concept of being a cluster of core competencies leveraged via a series of alliances into discrete business domains made this dispersion pressure all the more worrisome Corning was an outstanding partnerable to contribute a lot to individual alliancesonly insofar as it achieved synergies across alliances If each alliance had an independent technology base economies of scale in core technologies and economies of scope across alliances would be lost making Corning a much less attractive partner In nearly all its alliances Cornings value to the partnerand hence Cornings main source of influence and bargaining power in the alliancewas rooted in its technology and competence leadership Synergies were essential to maintain that leadership and its affordability In a situation in which Corning was often allied with partners of larger size or greater resources notably Siemens Asahi Glass affordability remained essential but difficult This had led Corning to divest from some of its more mature partnerships and to partner some of its core activities TV tubes provide an example of this approach As their technology stabilized and as margins declined this became a less attractive business for Corning In Europe Corning divested from its joint venture with Thomson Videocolor while in the United States it sold part of the equity of the TV Glass business to Asahi Establishing a domestic alliance in one of its historically core businesses was a difficult decision for Corning but one it justified by the need to trans fer resources to more promising higher technology faster growing opportunities such as porous glass for flat screens used in professional goods and portable computers Selectivity in new alliances was also difficult to achieve For example Corning had entered one alliance with CibaGeigy in 1985 in the medical diagnostics business It subsequently discovered that Ciba was willing to invest significantly across a wide front in the medical diagnostic business and to use the joint venture as the vehicle to carry out that effort while Cornings interests were much more narrowly focused Corning decided not to commit the scale of resources and the diversity of efforts required to match Cibas ambition and the partners split after a few years Corning emphasizing its own development in the diagnostic laboratory business The ability to maintain strategic control in a network of alliances with strong partners where the core technologies and core competencies synergies across alliances are weak is highly questionable For a company such as Corning or for other companies heavily involved in a network of alliances this may be the central strategic dilemma We can thus summarize the critical issues that a network of technologyleveraging alliances brings in addition to the issues present in single alliances we discussed in the previous sections in the following way 1 How broad is the scope of application of existing core competencies and technologies across the alliance network Obviously the broader the application scope relative to the resource commitment required to maintain the technology or the competence the better the hub or nodal position in the network Related applications provide economies of scale and scope by giving the opportunity to practice and exercise competencies and technologies common to several alliances Conversely as we argued earlier an alliance network with no synergies across the competencies and technologies required in each alliance would be of limited value12 2 Can core technology and core competence leadership be maintained by the nodal partner As we argued Corning remains valuable to its partners only insofar as it maintains leadership in the crucial core competence areas 3 Are the various partners compatible By and large Cornings partners have been in different downstream industries their offerings are not substitutes of each other As a result there are few interdependencies if any between Cornings various partners Where Corning entered multiple partnerships in different parts of the world in the same businesssay TV tubesproblems emerge with industry globalization in particular when the partner is a downstream endproduct supplier The partnership with Asahi or that with Thomson may well have cost Corning other potential customers such as Sony or encouraged some downstream competitors to remain or become vertically integrated upstreamsuch as Philips producing its own TV tube glass envelopes Having partners that are separate from each other in product and market positions generally facilitates the role of the nodal partner 4 Are the partners equally able and willing to invest in business growth The nodal partner can maintain its leadership position only insofar as its technical and financial resources are not stretched too thin by multiple partners who can outspend or outsmart the nodal partner in its various application and market areas 5 Is the managerial capability of the nodal partner up to the task Singly each of the alliances in a network raise the partner compatibility and process issues discussed in the previous subsection The diversity of partners is both a source of difficulty and potentially of learning as the diversity calls for more generalized alliance management skills Originally Corning had treated each alliance separately and the gentlemanly but loose country club atmosphere of Corning in the 1960s had allowed the management of each alliance to flexibly adjust to the requirements of the task and of the partner organization In the 1980s the strategy of the company became tighter and Corning started to manage the various alliances as part of an overall strategic design Yet apparently by the late 1980s Corning had developed enough expertise in selecting and managing strategic alliances to be able to rein in its various alliances without serious conflicts with the partners except when conflicts originated in genuine differences of strategic priority as with CibaGeigy Transferring Capabilities via an Alliance Corning was careful to maximize stable complementation and cospecialization between its capabilities and that of its partners and to minimize the actual capability overlap and potential encroachment on its part and that of its partners Alliances can also provide a vehicle to reduce cospecialization and complementarity as a way to erode the influence of a partner in the alliance and possibly to regain competitive independence when the skills of the partner have been successfully absorbed Taking the JVCThomson example we have already shown why the motives to learn the partners skills may be important In this section we consider the how of the learning process building further from the same example13 In 1983 at the start of the alliance Thomsons objective was to learn micromechanical mass manufacturing skills from JVC The mechanical system of the VCR called a mechadeck is made of the extremely precise assembly of a series of rotating heads and drums including the main drum around which the tape from the cartridge rotates in a helical pattern where tolerances are extremely tight The precision assembly of the mechadeck and the quality of the design and the manufacturing of its components largely determine the quality of the VCR to the customer in particular its image sharpness and its reliability The initial design of the alliance originally including Thorn as a third partner planned that JVC would provide components process equipment and would license product designs to a joint venture located in Europe with plants mainly in Germany and Britain staffed by workers provided by Thomson and Thorn and comanaged by JVC and the European partners Products would be made with both the JVC brand and the brands of the European partners and sold separately in their respective channels Thomson and Thorn would also become allies of JVC in promoting the VHS standard and in dealing with regulatory authorities in the EEC In particular the joint venture called J2T was a way for JVC to overcome the import restrictions imposed by the French and by other European countries on VCRs from Japan By deciding products and processes by providing components and by setting productivity standards JVC exercised almost full control over the cost levels of J2T By deciding which models were made in Europe as against supplied directly from Japan and by setting its own pricing policies JVC also influenced end product prices and thus the margins of its European partners in Europe Initially following the nationalization of the company by the French Socialist government in 1981 Thomson had been barred from participating in the joint venture with JVC In 1983 it bargained its entry in the alliance allowing it broader market access in Europe against a commitment from JVC to increase the European content of the product which also involved setting up a component and subsystem plant in France outside of J2T to be owned and run by Thomson with JVCs technical assistance This provided more freedom to Thomson to develop its own skills in micromechanics To learn these skills from JVC Thomson engaged in several complementary approaches First Thomson engineers analyzed both the product and the process and broke down the capability to design and manufacture VCRs into several complementary steps Capability to assemble JVCsupplied subassemblies using JVCsupplied equipment and process controls Capability for autonomous improvements in assembly efficiency through improvements to JVCspecific assembly line layout and procedures Capability for specifying and bringing on stream European process equipment to supplant that supplied or specified by JVC Capability for developing advanced product features independently from JVC Capability for separate manufacture and assembly of precision components Capability for design and manufacture of a VCR independent of JVC Capability for simultaneous and closely coupled advances in both product design and product manufacturing processes independent of J2T Design and manufacturing capability for next generation product eg handheld mini videocamera This breakdown was important insofar as it facilitated the journey to develop a micromechanical capability Each stage was carefully defined with a practical milestone In contrast we observed alliances where the will to learn was strong but where the partner failed to decompose the learning task into a series of steps When the initial distance between the partners was substantial this led to a permanent learning gap The will to learn was not operationalized into practical achievable and measurable steps and the learning agenda appeared so daunting that it in effect discouraged any serious attempt to learn Identifying the learning objectives precisely and disaggregating the resulting learning agenda into specific achievable steps however is not enough This only points the way toward a feasible journey To actually carry out the learning requires several other approaches First Thomson structured the alliance to allow these steps to be taken As local content had to be increased to meet the EECwide agreement on Japan Thomson and JVC planned a sequence of local content increases which clearly paralleled the learning agenda developed by Thomson and allowed successive steps to be taken The alliance was also physically structured to allow learning In the Berlin assembly plant for example Thomson colocated J2Ts assembly lines with some lines fully under Thomsons control These latter lines were first used to experiment processes and to develop experience in manufacturing and then to manufacture Thomsons own VCR types rather than the ones licensed from JVC Workers who were supplied by Thomson were rotated between the J2T lines and Thomsons own allowing for an efficient and immediate transfer of learning At the beginning JVC underestimated Thomsons ability to learn and was not too concerned with Thomsons learning efforts Later on market access and good relationships with the European governments became important enough to JVC to trade these implicitly for support to Thomsons learning JVC was not too concerned since its rate of skill renewal was comparable to or better than Thomsons rate of skill catchup For example Thomson was not able to independently develop and produce the miniaturized VCR or the smaller and cheaper video cameras using the new VHSC or Video8 formats14 Second Thomson worked hard to improve its receptivity to JVCs skills Beyond the colocation of activities it used other approaches For example it hired former watchmakers from the Jura mountains who had been put out of business by the Japanese watch industry lured by the dual prospect of exercising their skills and getting even Although VCRs are made to tighter tolerance than clocks and many watches the technologies and mindsets required to make both types of products are relatively close Thomson also engaged in Japanesestyle small group activities to articulate exchange and diffuse the learning occurring in the plants This took place both at the level of individual steps and more importantly at that of understanding the interactions between these steps As Thomsons capabilities increased the interdependency between product features process characteristics and manufacturability and quality for example became clearer and clearer Also key to improving receptivity were the constant reminders from top management that the most important goal was to learn not the economic and financial performance of the alliance The priority given to the learning agenda was clearly communicated and repeated over time to all involved of dollars to reconfigure its own plants patterned on Toyotas to no avail Adler and Cole 1993 Conversely in the example we have discussed extensively Thomson was fortunate not to have to unlearn much about micromechanics The technology area was essentially new to Thomson at least insofar as high volume high precision manufacturing was concerned 3 What is the nature of the capabilities we need to develop Understanding the nature of required capabilities is essential to designing the interface between the partners or even in deciding in favor of a partnership over other firms of collaboration The more the capabilities belong to the realm of competencies tacit knowhow in action embedded organizationally systemic in interaction and cultivated through learning by doing the more copractice is essential Indeed without the NUMMI experience GM might have permanently missed the essence of Toyotas management processes despite the rather abundant literature available on them Where the capabilities are more explicit and less like competencies copractice and apprenticeship become less important and the capabilities may be transferred by other means typically technology licensing perhaps supported by some specified technical assistance Even in the case of an alliance a less intimate form of cooperation may be sufficient 4 How should the learning process be operationalized once the capabilities to be developed have been ascertained We dwelled at some length on this question in our description of the J2T alliance The key points to recall are breaking down the learning agenda into a sequence of practical steps with each new step building on the previous ones maximizing managerial receptivity and establishing a learning schedule 5 How should we establish the boundaries between the partners to allow observation and copractice The partner who wants to learn and has the capability to do so is likely to try to establish porous boundaries and to leave the interface loosely defined In the early phases of its collaboration with ICL as a way to learn about users for example Fujitsu tried to convince ICLs management to let Fujitsu engineers accompany account managers on visits to their clients in the UK At the same time ICL was defending a much more arms length relationship in which it would provide technical specs for Fujitsu to develop products against these specs Conversely in the J2T case dedicating only part of the Berlin plant to J2T and rotating the workforce between sections of the plant provided a crude but effective mechanism for both copractice and independent replication and experimentation Similarly GM has used NUMMI as a learning and training ground for its production managers who were transferred back to GMs own operations after having been seconded to NUMMI for a while 6 How can we learn through practice Practice alone is no guarantee of learning unless the right mindset and skills are present To some extent in learning situations believing is seeing Crossan and Inkpen 1994 Grasping the more subtle competencies can only come to the trained eye and the trained hand Thomson saw its use of watchmakers engineers and technicians as a cornerstone in its learning process Too much distance between the learning starting point and the new required skills and the implicit mindset they call for makes learning impossible Such a gap diminishes the chance that the partnership will allow the lagging partner not just to copy the skills of the leading partner but also to independently develop these skills further 7 How compatible are skill learning and internalization with the continuation of the alliance To a large extent JVC was locked in by the time its management realized the extent of Thomsons learning Thomsons very incompetence at the beginning of the alliance served Thomson well it was easy to learn from an unsuspecting JVC who had discounted Thomsons skills As the extent of Thomsons learning became visible JVC was too dependent on the JVCgiven European protectionism and the political salience of the venture to pull out or even to trigger a visible crisis with Thomson JVC did tighten up Thomsons access to its plants in Japan labs had always been offlimits but relatively late in the learning process Similarly although obviously displeased Toyota was hard put to prevent GM from appointing to the Saturn project the managers who had been posted at NUMMI 8 How can we benefit from the learning investment Clearly there are several ways to benefit from the learning One already discussed at some length involves widening the scope of application of the newly learned technologies and competencies thus increasing the leverage from the investment Another is to use the results from the learning to renegotiate a more favorable alliance with the partner for instance by rebalancing financial flows in favor of the partner who has learned A look at the history of royalty payments between Xerox and FujiXerox is instructive in that regard one can observe a progressive rebalancing of the relationship over thirty years increasingly favorable to FujiXerox as it assumes a growing product development and sourcing platform role for the whole Xerox group Gomez Casseres 1991 A third benefit is to use the learning from one alliance to improve ones position in the next as Thomson did in its alliance with Toshiba subsequent to J2T This positioned each alliance in the broader context of an intertemporal learning network where the learning ladder extends not just between steps in an alliance but also in a more ambitious and longerterm process from alliance to alliance Integrating a Network of CompetenceAcquisition Alliances Alliance networks can play many different roles In our discussion of Corning we stressed one of these roles the competence leverage multiplier role Networks can also allow one to build a stronger competitive coalition This can help establish a standard JVCs desire to secure the leadership position of the VHS standard did play a role in its initial willingness to cooperate with Thomson and then other independent consumer electronics suppliers Thorn and Telefunken in 197778 Alternatively the network can help achieve a sufficient global presence to challenge the leader in an industry this was one of General Electrics mo vestment and Thomson was only able to play an independent role in the VCR business when very keen Korean competition had driven most of the profits out of that business Was it a phyrric victory Perhaps not Thomson could not afford not to be in the VCR business nor to be in it in a permanent dependence on Japanese competitors Yet the micromechanical technologies central to VCR quality have applications beyond VCRs in products and businesses that are of interest to Thomson We can summarize the key issues raised in individual alliances whose main purpose is skill learning and capability transfer around a few questions 1 How clear is the learning agenda A broad learning goal is of little help unless it is translated into a specific learning agenda with welldefined capabilities rooted in the mastery of practical tasks and preferably arranged in a logical sequence of capability development or a learning ladder Internalizing new skills requires the operationalization of the learning goal 2 How open are we to learn Both the J2T story and the above plea for clarity and practicality in the learning agenda may understate the difficulty of the issue because they ignore the unlearning barriers In some skill learning alliances focusing too early on a clear learning agenda may be counterproductive The initial mindset of the partner may make it oblivious to the more important learning and lead to a focus on learning which does not require a deep shift in mindset For example General Motors probably lost several years and many billions of dollars in its alliance with Toyota NUMMI by focusing initially on the plant layout material flows equipment and the like These were consistent with the mindset of GMs management but did not by themselves explain NUMMls remarkable performance Less visible to GM but more critical was the whole system of labor relations in the plant and the spirit a different system fostered It took several years for GM to discover that it had been focusing on only a small subset of Toyotas competencies Full realization only came after GM had spent too many billions mily in tives in initiating the relationship with SNECMA and a key interest of Fujitsu in its collaboration with Amdahl and ICL early on Contrary to other nodal partners we observed Fujitsu was able to coordinate its policies and to integrate its learning across alliances in a very effective way A central corporate alliance staff group headed by a senior executive was able to provide expertise and guidance to the various subunits of Fujitsu involved in various alliances and to encourage the sharing of learning between alliances Fujitsus nodal position in a wide range of alliances was actively exploited by that central group Beyond the interdependencies between coexisting alliances one can also consider an intertemporal network of alliances over time each contributing to the development of competencies on the part of the company For example Siemens had fallen behind in semiconductor technology when it concluded in the early 1980s that it could not be fully dependent on competitors or even on independent merchant suppliers such as INTEL for such critical components Playing a catchup game was relatively difficult though in an industry where the technology moves rapidly and where process skills are of the essence Siemens first entered an alliance with Toshiba which was also trying to catch up with industry leaders for similar reasons to Siemens Toshiba brought process manufacturing skills that Siemens lacked and a good experience of managing technology alliances Siemens brought other skills such as product development Strengthened by this initial alliance Siemens then allied with Philips first in the Mega project to develop 1 megabit memories While Siemens focused on DRAMs Philips focused on SRAMs thus minimizing the overlap in product and the potential competition while allowing a common investment in production skills and capabilities The project was launched in 1984 by Siemens and Philips joined in under the auspices of ESPRIT in 1985 By 1986 Siemens bought technology from Toshiba which had moved fast to develop its own 1 megabit memories The relative closeness of Philips and Siemens in their strategic interest in their technologies cultures and in geography facilitated a rapid start of the collaboration process The extensive use of parallel development was one other way to speed up the process By 1988 Siemens had a brand new facility in Regensburg successfully producing 1 megabit DRAM using a mix of Toshibas technology of Siemens own and of the joint results from the Mega investment In 1989 the Mega project was followed up by the Joint European Submicron Silicon Initiative JESSI JESSI was aimed to complement the Mega project as an investment into the overall European capabilities in semiconductors from material and manufacturing equipment providers to semiconductor users However the largest participants were still to be Siemens and Philips In 1990 Siemens entered an alliance with IBM to develop further the submission technology and build the next generation of memories The alliance was subsequently joined by Toshiba which in the meantime had captured the leadership position for 1 megabit DRAMs For Siemens one can summarize these ten years of effort as a way to reenter the race for advanced microelectronics Other companies had followed similar approaches For example following its demise at the end of World War II the German Aerospace industry engaged in a series of alliances starting in the early 1960s Playing an increasingly equal role to that of its partners it skillfully used the learning resulting from each alliance to gain strength in the subsequent ones Specific technologies appropriated in one alliance were used as bargaining chips to gain a better role in the next A network of alliances can also be used in certain junctures as a way of maintaining options in the face of uncertain technologies and markets The response of major pharmaceutical firms to the AIDS crisis for example fits that pattern Merck provides an interesting illustration It maintained a whole series of alliances to explore the various approaches to AIDS treatment vaccines better tests and to pursue various options in each approach In sum networks of technology alliances provide for a series of potential advantages which can be summarized around key issues 1 Can a network of alliances be used to make industry structures and terms of competition evolve toward ones own strength The most obvious example is provided by the VCR story where by quickly establishing a series of alliances to promote the VHS standard the JVCMatsushita group not only gained the advantage over Sony and scuppered Philips ambition to enter the VCR business on its own terms but also froze the technology for a number of years and shifted the competitive arena toward rapid production scale up mass manufacturing and distribution and cost reduction all areas in which Matsushita excels In general where standards play a role and where the competencies and skills of competitors differ widely the need for alliance networks to set the parameters of competition and to influence them in the interest of the focal company is significant 2 Can a nodal network position be used to cumulate and leverage competencies across alliances While Corning was essentially a competence developer building its own competencies to be applied in its various alliances Fujitsu as we discussed was partly a competence developer partly a competence exchanger especially insofar as it could use the same competencies across alliances By combining the learning from several technologically related alliances with different partners the nodal partner can normally build competencies faster and maintain or increase its leadership in the coalition 3 Can learning be cumulated over time between alliances and used as a source of growing bargaining position from alliance to alliance over time The experience of both Siemens and DASA the consolidation of the German aerospace industry suggests that this can be done given two conditions first the technological trajectory of the industry must be clear enough and second the strategic architecture of the company must be tight enough to inspire confidence in longterm competence development efforts Both semiconductors and aerospace are indeed relatively stable if technologically very demanding industries In more volatile environments learning cumulation may be less easily achievable 4 Can a network allow you to keep open a multiplicity of options at an affordable cost In highly uncertain technological areas like AIDS prevention and treatment an approach built on a network of relationships and constructing a type of bet focusing on individual technologies and research areas makes sense It acts both as an information network and as a portfolio of options In sum networks of alliances can provide many advantages but only provided the network is actively managed from the center In other words it must not be run merely like a collection of alliances Active management as we discussed briefly in the case of Fujitsu is not easy though Conclusion The main conclusions of this section are presented in summary form in Figure 372 3 TECHNOLOGY ALLIANCES AS LEARNING PROCESSES The preceding discussion of various types of alliances singly or in the context of an alliance net work may convey a sense of the management of alliances being a deterministic and analytical process where the appropriate questions can be raised and answers provided at the outset of the relationship This is rarely the case Alliances are often pursued as ways to explore new applications and new technologies or the match between the two By their very nature they confront uncertainty and information asymmetry indeed they are sometimes designed to overcome information asymmetry between the partners Further although a full explicit agreement is often seen as desirable seeking such an agreement early on can foreclose the possibility of collaboration given uncertainty information asymmetry and lack of familiarity that exists between the partners The need to learn in alliancesand not just in terms of skillsshould therefore not be underestimated And nor should the potential for dysfunctional learning be underestimated As an illustration of the need for learning and the difficulty of learning this section starts with the description of the alliance between Alza and CibaGeigy from which we draw various implications on the nature of technology alliances as learning processes CibaGeigy and Alza Alza was started in 1968 by Alex Zaffaroni a medical doctor and Syntex executive He had become convinced that the administration of drugs to patients could be considerably improved by better controlled release devices such as transdermal patches osmotic pills implantable slow release polymers or implantable minipumps Over the first ten years of its existence Alza gathered a firstrate group of scientists and technicians of drug delivery technologies and identified highly promising work Yet its products failed in the market they were more costly and harder to use than conventional products and addressed markets of limited size and by 1977 Alza was finding itself in an increasingly precarious financial position CibaGeigys pharmaceutical division which was short of really new products was keen to differentiate itself from competitors and wished to extend the life of some of its products or to use known chemical substances but ones whose use was difficult in conventional forms and dosages In 1976 CibaGeigy had carried out a preliminary study of new drug delivery methods and identified Alza as a source of potential technology When CibaGeigy found out that Alza was looking for funds and for marketing expertise it signaled interest The relationship between CibaGeigy and Alza started in 1977 when CibaGeigy reached a research agreement with Alza and also took a significant equity position in the company Alza was to develop for CibaGeigy applications of advanced drug delivery systems on which Alza had been working since its founding in 1968 in particular transdermal TTS and oral OROS slow release systems These were to be combined with active ingredients that were difficult to use in other forms eg nitroglycerine for angina pectoris or proprietary substances of CibaGeigy The patented life of these proprietary substances was to be extended by the use of the advanced delivery system The RD priorities were set at the beginning in the course of a joint research conference to be followed about twice yearly by a session of a joint research board to review progress and reassess priorities Priorities included both the development of new technologies eg implantable polymers and minipumps and the adaptation to CibaGeigys needs of technologies already largely developed by Alza Coordination was to be ensured by a series of complementary approaches scientific liaison desks were set up in all three locations including CibaGeigyUS project sponsors were appointed but only on a parttime basis on top of existing responsibilities in CibaGeigy a business liaison desk was established in CibaGeigy and internal coordination within CibaGeigy was to be achieved via an advisory team in Basle regrouping all key persons involved in the interface with Alza and in the work on the advanced delivery systems and via a specialized RD group in CibaGeigyUS CibaGeigy and Alza expected to obtain within two to three years a first wave of applications both for TTS and OROS technologies and to research further products and systems Sixty percent of the joint RD budget was devoted to shortterm projects the balance to new technologies The alliance was thus initially focused on capability complementation Alza was providing delivery system technologies while CibaGeigy would provide substances and manufacturing registration and marketing expertise for the new products By 1979 some difficulties started to develop While projects made good progress it became clear that the OROS technologies needed more applicationspecific development and that progress on pharmacokinetics was required to guide this development In the absence of joint CibaGeigyAlza RD teams it was difficult to perform this interactive applicationspecific development between the system features and the substance characteristics The problem was compounded by the need to involve the manufacturing function which had been traditionally quite separate from research at CibaGeigy These difficulties made OROS a less immediate alternative to slow release technologies already developed within CibaGeigy TTS projects made good progress largely because they found maverick sponsors within CibaGeigy who developed strong personal commitments to the projects and to their counterparts at Alza Most CibaGeigy sponsors though had many other commitments no specific budget and no ability to modify or bypass the usual slow sequential development process whereby new drugs move from one test to another Further contacts were poor between CibaGeigyBasle and CibaGeigyUS Their relationship suffered from the rivalry between the subsidiary a struggle for independence and the consideration of the relationship with Alza in that context and the intent of divisional headquarters on achieving more international coordination within RD A tradition of consensual decision making did not facilitate the resolution of such tensions within CibaGeigy As a result Alza sometimes found itself involved in parallel relationships on the same topic with both CibaGeigyBasle and CibaGeigyUS in which the two CibaGeigy subunits did not communicate Faced with these delays and frustrations Alzas management reacted in several ways First the problems reinforced the view that CibaGeigy was perhaps not the ideal partner over the long run and that the cooperation was better seen as transitory with Alza regaining its independence after a few years Second in an attempt to get faster action on the part of CibaGeigy Alzas managers and scientists started to walk the corridors in CibaGeigy trying to get quick decisions This behavior violated the accepted norms of interaction in CibaGeigy and triggered adverse reactions on the part of CibaGeigys executives and personnel in particular as Alzas executives had a tendency to leverage their access to CibaGeigys top management This access was easier for them than for CibaGeigys own employees given the hierarchical nature of CibaGeigys processes TTS was favored by the relatively high informal status and willingness to bypass rules of its sponsors as compared to a more usual process for OROS Third Alza started to look for third party research projects as a way to generate revenues quickly and to make full use of its RD capability Alzas management was also concerned that unless it maintained a steady research effort upstream the value of the partnership with Alza to CibaGeigy and Alzas bargaining power in it would decline over time Within CibaGeigy those who had doubted the value of new drug delivery systems and of Alzas innovations saw slow progress as confirmation of their negative prior judgment CibaGeigys executives also started to realize that in the absence of third party contracts Alzas losses would mount leading to a cumulative negative cash flow of 57 million by 1986 Exercising its power of shareholder CibaGeigy imposed on Alza drastic savings which affected mostly the marketing and manufacturing functions but left RD relatively unscathed The Alza plant was sold and proceeds used to pay back loans Alzas management felt these cuts were inescapable if only to restore their partners confidence In 197980 CibaGeigy reorganized its development activities by therapeutic area teams to speed up product development as other pharmaceutical companies were doing too At the same time although Alza was authorized by CibaGeigy to look for third party contracts a slow and complicated vetting process by CibaGeigy deterred potential third party partners from entering an agreement with Alza Increasingly frustrated with Alzas intrusions in their decision making CibaGeigy executives resorted more and more to premeetings where they would reach decisions on a common CibaGeigy position to be presented to and if need be imposed on Alza Alzas managers obviously felt railroaded Openness in communication between the partners started to break down CibaGeigy was wary that information on development and manufacturing could leak to competitors while Alza started to suspect that CibaGeigy wanted to develop its own inhouse advanced drug delivery system capability including research As a result communication exchanges were reduced to their contractual minimum and the development of some applications further suffered The CibaGeigy researchers frustrated by the lack of open communication came to believe in 198081 that they had obtained from Alza about all there was to obtain and that neither further process knowhow nor major breakthroughs were likely to result from the partnership At the end of 1981 CibaGeigy decided to divest from Alza and phase out the research agreement over the following three years The cooperation agreement was revised to give Alza more generous royalties on the products it had developed and less exclusivity to CibaGeigy facilitating Alzas search for multiple third party agreements Although CibaGeigy remained a shareholder of Alza for a while and keeps cooperating with Alza on specified products the alliance was dissolved It was not seen by either party as a failure though Several TTSbased products introduced in the early 80s have become major successes for CibaGeigy while Alza once out of the alliance with CibaGeigy was able to quickly reach a number of third party agreements and to succeed as a contract researcher for a variety of health care companies The Nature of Learning in the AlzaCibaGeigy Alliance Contrary to the simpler alliances discussed in the earlier section it is difficult to pigeonhole the CibaGeigyAlza alliance Part of the difficulty stems from its evolution During the first two years 197779 the alliance was clearly seen as a capability complementation one However over time it became clear that for OROS at least capability complementation was actually rather difficult The task was just more difficult than expected and required more applicationspecific work than was provided for in the initial agreement By 1979 these difficulties were recognized but this sometimes led to more friction For example the more active involvement of Alzas managers in Basle led to adverse reactions on the part of CibaGeigys personnel Alza failed to achieve effective action by CibaGeigys rules and failed to change these rules or obtain a way to override them During the first two years of the alliance thus learning took place on several fronts The partners learned about the common task and found it more difficult than anticipated The partners also learned how difficult it was to work together successfully A set of differences of a somewhat similar nature to those observed in the relationship between ATT and Olivetti stood in the way of successful collaboration between Alza and CibaGeigy They also learned about their own goals and their partners and the fact that these goals were not always easily reconcilable For example the head of CibaGeigys pharmaceutical division had entered the alliance with both concern to not stifle entrepreneurship and innovation at Alza and a desire to use these as a role model for CibaGeigys own RD The first concern called for an arms length relationship in which Alza would be protected from CibaGeigy the latter called for more exposure and joint work Similarly Alzas management was keenly aware that its longterm value to CibaGeigy and potentially to third parties was critically dependent on new research But the agreements with CibaGeigy made Alzas future dependent on royalty payments stemming from the early introduction of already developed products As a result of these dual priorities the focus of Alzas technical skills and the balance between advanced research and product development support became difficult issues The partners learned not only about their own sets of goals and their potential internal contradictions but about each others CibaGeigy discovered that Alza had not given up its ambition to become a fully fledged autonomous pharmaceutical company while Alza discovered that CibaGeigy had not necessarily forsaken its own RD on advanced drug delivery systems This led to a heightened sense of possible encroachments Not all learning was constructive in that useful learning occasionally triggered dysfunctional behavior As discussed in the summary description of the AlzaCibaGeigy alliance Alzas understanding of the complexity of its partners decision making processes led to an active lobbying and corridor management attitude which aggravated a good part of CibaGeigys staff who were unused to such intrusive behavior This in turn led some within CibaGeigy to decommit personally from the relationship not the outcome sought by Alza Learning also led to modify the interface or the internal organization of the partners Although there were various reasons for CibaGeigys reorganizatiqn by therapeutic area development teams in 1979 the desire to facilitate the relationship with Alza and to speed up the development process did play a role Beyond 1979 the awareness of competition grew Alza was increasingly worried about its technology lead being eroded by CibaGeigys perceived slowness and by the rapid catchup of competitors especially that of Key Pharmaceuticals Learning in Technology Alliances This example illustrates the fact that technology alliances are evolutionary processes where the initial conditions are frequently questioned by intermediate learning outcomes of the very process of collaboration In managing technology alliances it is important to be attentive to at least five domains of learning Environment learning Results from a better appreciation of the environment of the partnership This better appreciation may just result from action getting feedback from the environment leading the partners to revise or sharpen their initial assumptions or from changes in the environment in the course of the partnership making some of the initial assumptions obsolete The internal environment of the partnership within each partner may also change as a function of other opportunities developing or closing or of internal organizational and political aspects such as the partnership gaining or losing powerful sponsors for reasons that have little to do with its evolution itself As partners may have more or less similar perceptions of the external environment and more or less compatible evolutions of their internal environment environmental learning may lead to convergence or divergence between them and to increasingly similar or different assumptions about the future environment Task learning Starting the cooperation work may also lead to learning about the task in particular when the task was broadly defined at the outset andor when it is complex or uncertain Depending on the novelty of the task of the partnership more or less task learning may be required When partners start with information asymmetry complementary but not very differentiated skills and use the alliance to develop new opportunities the task is unlikely to be defined accurately or precisely at the outset calling for more precise definition through extensive task learning as the partnering process unfolds Process learning Similarly the very process of collaboration itself can be an object of learning Partners discover each other Recognition of their differences in structure processes action routines and the like may lead partners to learn to overcome or even to exploit these differences in the collaboration process Some of the noise and ambiguity which are inherent to collaboration between dissimilar organizations may be eliminated over time as the partners learn to re late to each other more clearly and to cooperate more effectively Skill learning Some alliances may also offer windows into the partners respective skills and opportunities to the partners to learn new skills from each other Hamel 1990 1991 Skill learning may allow the partners to cooperate more closely as their respective skills converge but may also decrease the value of future cooperation between them in particular if skill learning ranked high on the partners hidden agenda and make current cooperation rife with tension if skill learning is asymmetric between partners Goal learning The collaboration process itself may also reveal the partners goals to one another The behavior of each partner in the alliance may provide clues about their hidden agenda and true motives beyond or beside the explicitly shared goals of the collaboration Further and more interestingly the collaboration process itself may lead the partners to clarify revise or refocus their own goals New unseen benefits to the cooperation may become visible or new outcomes may now be expected given the partners interpretation of the evolution of the partnership The observation common to all five dimensions of action learning outlined above is that the determinants of partnership outcomes cannot be reduced to partnership starting conditions Initial conditions obviously matter in setting and construing the learning processes but the outcome of these learning processes is not fully determined by their starting position Learning results in the revision of the initial conditions of the alliance such as task definitions or interface structures and processes and may even lead partners to reform their own organizations and decision making processes as discussed in the GESNECMA case17 For obvious reasons commitments are also affected by the tacit or explicit reevaluations of expected outcomes that result from learning in the course of the alliance It is therefore useful to consider alliances not as oneoff agreements to be implemented but as iterations of interactive cycles of cooperation over time This entails ongoing learning in several domains punctuated by periodic but not necessarily fully explicit nor shared reassessments leading on the part of each partner to changes in commitment to the alliance task definition interface mechanisms and approaches and possibly in each partys organization and management processes CONCLUSION This chapter has argued that alliances have a key role to play in the implementation of technology strategies and more generally of corporate strategies This is especially the case when the skill bases of partners are different and where the technologies involved are more tacit and systemic rather like competencies and less easily manageable and learnable than explicit and discrete technologies Second we have argued that it is important to distinguish alliances whose purpose is to combine cospecialized technologies between the partners in which each retains a discrete set of technologies from alliances whose goal is explicitly or implicitly to learn and internalize the partners skills via the alliance They raise very different issues and the strategic and managerial priorities are quite different between the two types of alliances Thirdly we have argued that the management of technology alliances is exercised at least at two levels individual alliances and a network of alliances It is important to manage actively both levels and the issues are quite different It is also important to recognize that alliance networks have to be managed differentlylike individual alliancesdepending on their main purpose Finally we have argued that alliances by their very nature cannot be planned and decided once and for all at the outset Instead they should be considered as multiple cycles of learning processes resulting in periodic reassessment and readjustment to the relationship It is perhaps worth ending with a word of caution While alliances often provide an attractive option partly because they are a vehicle to match commitments to learning as uncertainty and information asymmetries are removed their cost should not be underestimated Their costs in terms of management time and energy organizational stress and possible constraints on strategic freedom far exceed the visible costs on the PL statements or balance sheets Inexperienced managers should be wary of entering alliances without a full appreciation of these less discernible costs that manifest themselves only over time as the relationship develops These conclusions are not meant to dissuade companies from entering into alliances simply to encourage them to manage those alliances more actively Notes 1 The current superiority of lean over mass manufacturing disciplines in the car industry may force the elimination of some of the weaker car makers in Europe and the United States and their replacement with Japanese transplants Over the longer run however several US and European car manufacturers will adjust and regain competitiveness against the Japanese at least in this particular dimension 2 Although CFM did not cover the high end of the thrust range and thus the larger jet engines nor necessarily new innovative projects such as propfan engines GE had always offered SNECMA the opportunity to codevelop any new products and to participate significantly in all its existing programs 3 In the early 1990s Toyota succeeded in excluding the United Auto Workers from NUMMI In addition Toyota gets public relation benefits from NUMMIs success which points to mismanagement rather than an intrinsically undertrained and restive workforce as the cause of the decline of the US car industrys competitiveness 4 Although this particular case illustrates strategic similarity and alignment between partners sharing a common ambition such commonality should not be taken as a precondition for success which is why we use the word compatibility rather than commonality Partners pursuing different but compatible objectives in an alliance may actually find agreement sometimes easier as the benefits they expect from the alliance are not in competition with each other Reaching a package deal solution in which each partner finds satisfaction visàvis different objectives may be easier 5 With the exception of the reengining of old Mirage planes their products are not interchangeable but military aircraft producers do compete with for example the Dassault Mirage 2000 planes using SNECMA engines and their US competitors the F16 in particular using GE engines 6 It is possible though that a change in status of one or the other of the partners would lead to a restructuring of the alliance For example SNECMAs privatization or a divestiture by GE of its aeroengine group could lead to imbalance in feasible financial commitment between the partners possibly leading to a restructuring of the current relationship 7 For example the fluctuations in the value of the dollar its deep fall of the late 1980s in particular did create tensions in the transatlantic alliances we researched Jet engines being sold in dollars SNECMA had to make tremendous productivity efforts and to engineer smart hedging deals to avoid a conflict with GE on pricing and market development policies 8 It is important to note here that GE had already relied on SNECMA as a risk sharing subcontractor for the CF6 engine to be installed on the Airbus A300 before it engaged into the CFM alliance The experience of the CF6 project had allowed the two companies to get to know each other in the context of a collaboration where mutual commitments were lesser than in the 5050 CFM alliance In contrast again ATT and Olivetti had to rush into trying to make momentous commitments such as a new joint computer product line from a situation in which they knew and trusted each other relatively little 9 For obvious reasons the more partners expect to continue cooperation over the long term the more cooperative and truthful they are likely to be in the short term Negotiation theorists call this phenomenon the shadow of the future One of the more salient consequences of this is that partnerships that have a set date of terminationor renewalare likely to grow increasingly contentious as they get closer to that date This calls for renegotiation and renewal dates well in advance of the expiration date 10 There are companies such as Aero sp atiale Deutsche Aerospace or British Aerospace where alliances and partnerships account for a higher share of their total sales However many of these firms alliances have been mandated by external forces such as military and civilian administrations in the various European states The management of most of these companies explores rather than welcomes at least some of the alliances in which they are forced to participate 11 Historically Corning had missed several opportunities by failing to transform its innovations into products fast enough from electric bulbs where it lost out to General Electric to semiconductors where its subsidiary Signetics had been an alsoran sold to Philips in 1975 and to ophthalmic glass where Corning lost out to flexible contact lenses 12 This is a close analogy to the argument in favor of related diversification over unrelated diversification and is not exclusive to alliances although the questions of why other companies see Corning as an attractive CHAPTER SEVEN EXECUTIVE LEADERSHIP AND MANAGING INNOVATION AND CHANGE We have explored in detail how successful firms surmount the problem of managing inconsistent requirements across cycles of technological change What glue binds an ambidextrous organization that is capable of leading and adapting to conflicting demands in turbulent environments Executive leaders hold such enterprises together and direct them to capture value from innovation They balance continuity and change guiding companies through long periods of gradual adaptation punctuated by dramatic transformations that take place in relatively brief intervals Tushman et al begin this section by describing this pattern of convergence and upheaval Their model provides concrete advice for managers who must decide how to manage both types of change and must know when each type is appropriate Nadler and Tushman discuss how top executives actually implement the difficult yet necessary process of fundamentally transforming an organizations architecture Kolesar provides a detailed case study of how one executive effected a seismic change in a traditional industrial giant Step by step he describes how the CEO instilled a total quality culture in a firm whose institutional resistance to such a metamorphosis was rooted in its proud heritage and the lessons learned from decades of dominating its industry Howard concludes this section by providing a concrete example that illustrates how a CEO leads a traumatic reorganization that fundamentally alters his companys identity Changing his firms definition of its position in the world not only alters its strategy but requires a shift in its deep structure 581