Foreword by C. K. Prahalad

During 1993 Microsoft’s market value became greater than that of IBM. Such an event would have been unthinkable just a few years before, given IBM’s total dominance of the computer industry. Yet, at the end of the 1970s, when they chose a small company–Microsoft–to be their partner in developing the operating system for their PCs, Big Blue’s senior executives little imagined that their decision was, in fact, paving the way for such a dramatic reversal of fortunes.

Allied to IBM, Microsoft was able to create and impose the international standard for personal computing and, on this basis, to become the world’s undisputed software leader. Its control of PC operating systems in fact turned out to be the company’s main asset for dominating the market for applications software.

These two examples demonstrate how alliances, formed initially to solve an immediate problem, can have major strategic consequences in the long term. It is therefore not enough, when engaging in cooperative strategies, simply to review the explicit objectives, as they are stated at the outset, pursued through an alliance; managers must in fact try to anticipate all the possible, and sometimes unwanted, consequences of the partnerships they set up and make sure that these consequences are not detrimental to the long-term interests of the company.

Anticipating the outcomes of alliances is of vital importance, notably when they involve the transfer of technology and know-how between the partner companies. This concern led General Electric, which had teamed up with a French company, SNECMA, to produce the CFM-56 aircraft engine, to seal the most sophisticated parts of the engine in order to prevent its partner from gaining access to them. But it is not enough to try to limit one’s partner’s access to sensitive know-how; companies must also take best advantage of the alliance to extend their own competencies. The case of the NUMMI joint venture, set up by General Motors and Toyota to produce small cars in the USA, is highly significant in this respect. Toyota successfully managed the careers of the Japanese executives who had worked for NUMMI; these executives were later given the responsibility of setting up and managing Toyota’s plants in the USA. Thus they were able to make full use of their experience with NUMMI, notably in managing American employees, negotiating with local unions and dealing with North American suppliers and subcontractors. In contrast, the General Motors executives who had worked for NUMMI were frequently given positions where they were unable to give the company the benefit of their understanding of Japanese management and production methods.

These problems in the management of alliances derive from the fundamentally ambivalent nature of this type of relationship. The allied companies are constantly obliged to trade off their own interests against those of the partnership. For instance, the partner firms in the Airbus consortium were surprised to find out that CASA, the Spanish partner in the venture, had agreed to act as a subcontractor in McDonnell-Douglas’s MD-83 program, an aircraft that was to compete directly with their own Airbus A320. A few years later, the press revealed that another Airbus partner, the German company DASA, was negotiating with Boeing for a share in a project to develop and manufacture a 600- to 800-seat jumbo jet, while Airbus was working on a competing project.

If managers are still looking for methods and models to help them manage their alliances, public authorities are also unclear about what attitude to adopt when confronted with these new strategies. Thus, while the Regional Development Department of the European Commission was paying subsidies to Ford and Volkswagen to encourage them to set up a joint production plant in Portugal to manufacture minivans, the Competition Department of this same Commission was scrutinizing the potentially anti-competitive–and consequently illicit–aspects of this partnership.

More fundamentally, when a company enters into multiple alliances, its very contours become increasingly blurred. For example, the senior managers of Philips’ household appliances business, associated in a joint venture with Whirlpool, found themselves faced with a thorny problem of loyalty: should they protect the interests of their former company (Philips) or work primarily for the benefit of the business, irrespective of its owner? The problem was all the more difficult to resolve as Philips had announced its ultimate intention of withdrawing from the venture in favor of Whirlpool. Similarly, executives of Matra-Hachette found it difficult to accept the fact that the Espace minivan–which their company had designed and manufactured, but which was sold under the Renault brand–was considered by the general public to be a Renault product.

When carried to their extreme, collaborative strategies can empty the partner firms of their industrial substance, by transferring most of their business to joint ventures. Thus Aerospatiale, a large European aerospace company, gradually divested most of its manufacturing activities in favor of multiple alliances forged in all its businesses: the Airbus and ATR consortia for aircraft; the Eurocopter and Euromissile joint ventures set up with DASA for helicopters and missiles respectively; Ariane for space launchers, etc. As a result, in the mid-1990s, about 80% of Aerospatiale’s sales were derived from collaborative projects undertaken in partnership with other companies. In certain cases, the firms that had initially formed the partnership ultimately become mere shareholders and all but disappear in the shadow of their joint venture, which takes over all operational activities. Who still recalls that Shell or Unilever were both started as joint ventures set up by Dutch and British firms?

These examples–which could be extended almost indefinitely considering how widespread alliances have become in recent years–illustrate the variety and saliency of the problems raised by this new type of industrial organization. Thus, when they consider engaging their company in alliances, senior managers invariably have to face the following dilemmas:

• -Is an alliance really preferable to independent growth?

• -What firms will turn out to be suitable partners?

• -How should the partnership be organized?

• -Is a simple agreement sufficient or should a joint subsidiary be set up?

• -Will all the partners benefit equally from the alliance? Might there not be winners and losers?

• -How long will the alliance last? Should termination be planned from the very beginning?

• -Is it possible to cooperate loyally while still preserving the vital interests of the company?

• -How can the company effectively protect its technology and know-how?

• -Do alliances hinder or eliminate competition? Are certain forms of collaboration illicit?

It is impossible to provide a single answer, valid for all alliances, to each of these questions. Not all alliances are the same. A simple collaborative agreement between an original equipment manufacturer (OEM) and its subcontractors, a joint venture with a local partner to expand into a developing country, and a partnership between major global competitors for the joint development and production of a new car or aircraft correspond to extremely different circumstances. Strategies must therefore be tailored to each main type of alliance. This book presents an analytical framework making it possible to provide answers to the questions listed above, taking into account the specific characteristics of each individual alliance.

The book is organized on the basis of this framework, presented in Figure I.1, which enables readers to find their way in the text and to choose the themes and chapters corresponding more particularly to their specific interests and to the particular alliances with which they may be confronted.

Chapter 1 is aimed at defining precisely what we understand by ‘alliance’ and at describing its specific characteristics. This is made necessary by the fact that alliances are too often confused with more conventional interfirm strategies, such as mergers and acquisitions. Indeed, it is the very specific nature of alliances that makes them so difficult to manage.

In Chapter 2, we analyze the reasons for the recent burgeoning of alliances in virtually all industries and regions of the world. This will enable us to identify the strategic motives which, over the past few years, have driven ever-increasing numbers of companies to engage in collaborative strategies.

Readers interested in a more theoretical analysis of alliances should turn to Chapter 3 for a review of the approaches developed on this subject.

Chapter 4 deals with the heart of our problem. In distinguishing alliances between rival companies from partnerships between non-competing firms, it presents a typology of alliances that should enable readers to identify the specific form of cooperation with which they are confronted. They will then be in a position to turn directly to the chapter or chapters corresponding to the situations of greatest concern to them. Each of the following six chapters describes one of the main types of alliance, providing a detailed analysis and presenting guidelines for suitable action.

Chapters 5, 6 and 7 analyze partnerships between non-competing companies.

Chapter 5 focuses on joint ventures designed for international expansion, frequently set up by multinational corporations that decide to team up with a local partner to start operations in a new market, mainly in developing countries.

Chapter 6 is devoted to vertical partnerships between suppliers and customers, operating at different stages in the production process.

Chapter 7 deals with cross-industry agreements. These relationships are formed to exploit technological or commercial synergies, or to help one of the partner firms diversify into the other’s business.

Chapters 8, 9 and 10 are devoted to strategic alliances between competitors.

Chapter 8 analyzes shared-supply alliances in which the partners develop and manufacture common components, which each then incorporates into its own specific products.

Chapter 9 studies quasi-concentration alliances in which the allied companies develop and produce together a single common product.

Chapter 10 deals with complementary alliances. These partnerships bring together rival companies into an alliance where one firm will market a product initially designed by the other.

Finally, an original study on the evolutions and outcomes of strategic alliances is presented in Chapter 11. It reveals that the various types of alliances described in this book follow different patterns of evolution and lead to quite dissimilar outcomes. Therefore, correctly identifying the type of alliance a firm is about to enter is critically important. Indeed, it makes it possible to anticipate, before the alliance is actually formed, what its evolution, its outcome and its consequences are likely to be.

We suggest that readers wanting a rapid and synthetic view of the subject start with Chapters 1, 4 and 11. On this basis, and depending on their particular interests, readers looking for a more detailed presentation of a particular type of alliance may then choose among the other chapters as outlined in the preceding paragraphs.

COMPLEMENTARY ALLIANCES

HOW TO WIN THE RACE TO LEARN

In many complementary alliances, however, one of the partners may have a strong interest in capturing valuable capabilities contributed by the other. When the appropriability of the capabilities in question make such an objective realistic, the alliance is likely to degenerate into a "race to learn". When Nestlé and General Mills formed the Cereal Partners Worldwide alliance (see Box 10.4), neither of them could ignore the fact that their respective contributions were based on capabilities that were valuable and attractive to the other partner. Nestlé had previously tried to develop its own cereal business and had failed miserably while General Mills lagged far behind Kellogg’s, its arch rival in the cereal business, in all markets outside North America. It could be expected that Nestlé would eventually want to control a business of its own in cereals, one of the fastest-growing segments of the packaged foods industry; conversely, General Mills could not accept to rely on a joint venture for all its international operations, especially as most of the growth in the cereal business was driven by demand emanating from foreign markets.

All of the ingredients for a race to learn to flare up within CPW were thus combined. In this race, however, General Mills was at a disadvantage, since learning about breakfast cereal manufacturing technology is a priori easier and faster than developing a worldwide brand and the ability to deal with local distributors in many different parts of the world. Aware of this, General Mills requested that the agreement with Nestlé include a number of safeguards. First, Nestlé had to commit not to acquire General Mills, through a hostile takeover bid or otherwise, for at least 10 years after the joint venture was formed. Second, in the event of the alliance being terminated prematurely, Nestlé agreed to not enter the market for cereal in North America for at least 10 years. Finally, most products introduced by CPW would carry trade marks owned by General Mills (such as Cheerios, Golden Grahams, etc.) which Nestlé could not use if the alliance were broken.

Cereal Products Worldwide—a General Mills–Nestlé Global Alliance

During the 1980s, the ready-to-eat cereal market experienced high growth, particularly in Europe, where habits were switching from more traditional breakfast foods to cereals, as well as in many other markets outside North America. The dominant player in this market was undisputedly Kellogg’s, with market shares of 40% to 50% in all major markets.

In the late 1980s, General Mills was Kellogg’s strongest competitor in the USA and had increased its market share from 21% to 27% between 1980 and 1989. This success was made possible by innovative products such as Honey Nut or Apple Cinnamon Cheerios. General Mills’s presence outside the US, however, was extremely limited, leaving Kellogg’s unchallenged worldwide.

Nestlé, the Vevey, (Switzerland) based world leader in packaged foods, tried to enter the cereal market as early as 1983. Having met with limited success, it realized that it needed to improve its production process and also that it lacked the necessary technologies.

General Mills and Nestlé made the first contact in 1989 and agreed within 30 days to create Cereal Partners Worldwide (CPW), a 50–50 joint venture headquartered in Lausanne (Switzerland). The initial contract was for 10 years. By 1998, General Mills and Nestlé were estimated to have invested close to $1 billion each in their joint venture. Their aim was quite simple: to compete with Kellogg’s everywhere in the world, except in North America where General Mills was already No. 2.

General Mills contributed a line of successful products with well-known brand names, as well as its manufacturing and technical expertise. Nestlé gave CPW access to its distribution and sales network throughout the world, and to its marketing skills in each country; Nestlé also granted the joint venture the right to use its two main brand names in cereal, Chocapic and Nesquik. Each of the partners appointed four to five executives to CPW’s board of directors.

The first markets targeted by CPW were those where Nestlé had already established a presence in cereal products (France, Spain and Portugal). In 1992, CPW acquired a significant share of the British market (15%) by purchasing the cereal division of Ranks Hovis McDougall, which owned such brands as Shreddies and Shredded Wheat. In 1995, ahead of schedule, CPW reported its first pre-tax profit. Since then, CPW has also entered Asia, Latin America and Eastern Europe. By 1998, it was present in over 60 countries worldwide and was the second largest cereal company outside North America.

The difference in size between the two partners (Nestlé’s annual sales were five times greater than those of General Mills) is speculated to have led them to include a "standstill" provision in the CPW agreement: neither partner would attempt to take control over the other during the alliance’s first 10 years and, if a takeover were planned, an official announcement would have to be made at least 10 years before the actual purchase could take place.

To control the transfer of valuable capabilities, partner firms engaged in complementary alliances can also choose organizational arrangements that limit or, on the contrary, favor learning in particular areas or fields of expertise. Indeed, the way in which tasks are carried out within the alliance has a strong impact on the transfer of know-how between the partners. For example, in alliances where manufacturing is pooled in a jointly operated factory, or where marketing and sales are entrusted to teams mixing managers from both partners, much more interpartner learning is likely to take place. More generally, all joint activities favor the transfer and acquisition of skills; by physically working together, teams from both partner firms will learn from one another. Organizing work within the alliance in an appropriate way can therefore allow one of the partners to acquire valuable capabilities from the other while limiting unwanted transfers in the other direction. More specifically, a partner should request joint management and joint operations in those areas where it wants to acquire capabilities, while arguing for a strict allocation of tasks between the partners in domains where it wants to protect proprietary skills.

In addition to all these factors (appropriability, value and attractiveness of contributed skills, organization of tasks within the alliance), the ability to take advantage of complementary alliances to acquire new capabilities also depends on the "absorptive capacity" of the firm (Cohen and Levinthal, 1990). To tap a wide range of skills as quickly as possible, the firm must, in particular, have an internal organization and culture that facilitate collective learning (Hamel, 1991; Doz and Hamel, 1998), encourages horizontal communication, favors the circulation of information and values openness and receptiveness to the outside world.

The influence of organizational choices on interpartner learning is particularly obvious in the case of the NUMMI joint venture formed by General Motors and Toyota (see Box 10.5). The organization of the alliance was deliberately designed to facilitate learning and skill transfers between the partners in specific areas, particularly in manufacturing, purchasing and human resource management. Had the objective of the alliance merely been to assemble and sell a few thousand Toyotas rebadged as Chevrolets, GM could have bought these cars from Toyota, or Toyota could have granted a manufacturing license to GM. The acquisition of new skills clearly motivated the decision to set up a jointly operated plant.

NUMMI—a General Motors–Toyota Alliance

In 1984, General Motors and Toyota set up the NUMMI (New United Motor Manufacturing Inc.) joint venture. NUMMI was formed to manage one of the GM’s former plants, located in Fremont, California. This plant, which dated back to 1963, had been shut down in 1982 because of strained labor relations and poor productivity.

The NUMMI factory, although it was less automated than most other GM facilities and was staffed with rehired former Fremont plant workers, rapidly achieved higher productivity levels than other GM manufacturing facilities in the USA. Quality also improved substantially and the Chevrolet Nova assembled at Fremont turned out to be as reliable as its twin model—the Toyota Corolla—manufactured in Japan.

These achievements came with the manufacturing processes and management methods transferred to the joint venture by Toyota. Both partners explicitly intended to use NUMMI as a tool for learning: GM would witness the Japanese manufacturing methods first hand and could then try to transfer them to its own production units, while Toyota would learn how to manage a factory in an American environment. Learning, for both partners, was focused on organizational skills linked to manufacturing. Because of the bilateral nature of the learning process, the NUMMI alliance had all the ingredients to turn into a "race to learn"; the first ally to achieve its objective would no longer need to collaborate. Time was of the essence as the Department of Justice, because of anti-trust concerns voiced by Chrysler, had mandated that the alliance be terminated by 1992. This time limit was eventually lifted and NUMMI was still operating in 1998; by that time, however, rumors of the Japanese partner exiting the alliance and possibly taking over NUMMI suggested that Toyota’s learning objectives had been achieved.

As soon as the alliance was formed, GM had set up procedures to transfer new skills—or to "nummize", as GM jargon had it—its own manufacturing facilities. GM executives were assigned to the joint venture for three-year periods and worked with Japanese managers. A liaison office was set up at NUMMI in order to circulate all the information gathered (through videos, databases, documents, factory visits) within GM. However, while those managers that held lasting positions at NUMMI actually acquired new capabilities and were able to have an influence on GM’s operations, it appeared that second-hand or more superficial exposure to the NUMMI experience had very little impact because the managers involved, although interested, remained skeptical about implementing the NUMMI methods in other plants.

In learning from NUMMI, GM faced a difficult challenge: only a handful of managers could be properly "nummized" while the valuable skills acquired in this process were to be transferred to the 100 or so manufacturing facilities that GM operated in North America. Toyota, on the other hand, only needed to train a few dozen executives in how to deal with the local environment in North America, in particular with labor and suppliers. These managers would then be used to staff Toyota’s two North American facilities.

It thus appeared that, because its objectives were more limited, Toyota was more favorably positioned in its " race to learn" with GM.

Interpartner learning is a significant aspect in most complementary alliances. This often leads to earlier termination, as whichever partner has first captured the desired capabilities is likely to pull out of the alliance before learning takes place in the opposite direction. Research on alliances has also shown that complementary alliances are more likely to end with a takeover by one partner than are other kinds of alliances, which supports the idea that, when the race to learn is over, the winner can operate on its own while the loser is left with the alternative of either forming a new alliance or of pulling out altogether and selling its stake in the partnership (Dussauge, Garrette and Mitchell, 1998).