2I will not address all these questions here, but Why do firms enter into such looseinstead will focus on fundamental issues that agreements, and willingly endure theshape the answers to these managerial associated difficulties and risks? In simplequestions. My intent is to outline an approach terms: the alternatives are less attractive forthat can guide analysis and management. the given situation. One alternative to an alliance is an arm’s-length contract. In many situations, such contracts do not provideFrom Firm to Alliance to Constellation sufficient incentives for firms to collaborate deeply. Another alternative is a merger orWhat is an Alliance? To avoid confusion, acquisition. In many cases, such an approachsome definitions are in order. First, let’s define is infeasible, or too expensive or risky. As an“alliance.” There is finally some agreement arrangement short of merger but deeper thanabout this between the academic work on an arm’s-length contract, an alliance mayeconomics of the firm (Baker, Gibbons, and strike just the right balance.Murphy, 2001) and the managerial literature(Bamford, Gomes-Casseres, and Robinson, What is a Constellation? There is less2003). It is well understood, for example, that agreement on what constitutes aalliances can be used to fulfill any number of “constellation.” I define it here simply as a setcorporate goals, including gaining scale, of firms linked together through alliances andreducing costs, accessing new skills, products, that competes in a particular competitiveor markets, and sharing risk. The real question domain. The key here is that the alliance formis whether such a goal is best achieved with an (defined above) is used as the thread toalliance or another organizational approach. “stitch” together the members of the group –Here too, it is usually agreed that an alliance this creates a unit that is looser than if thecan take on a wide range of forms, from members were merged through completeclassic stand-alone equity joint ventures to ownership, yet tighter than if the members justnon-equity relationships, including enhanced had short-term, arm’s-length transactions withsupplier agreements, contractual research each other.collaborations, marketing affiliations, licenses,and multipartner consortia. This definition leaves open the actual motivation and structure of the constellation.What do these arrangements have in common? Again, any number of goals might be pursuedThree characteristics. First, all alliances are by the group and the group can be structuredagreements between two or more separate in many ways. For example, some typicalfirms that involve ongoing resource goals of constellations are:contributions from each to create joint value.Typical partner contributions include • Expanding market reach rapidly and withtechnology, staff, customers, brands, capital, low investmentand equipment. Second, all alliances are in • Exploiting multiple geographic or verticalsome sense an “incomplete contract”— a markets simultaneouslyphrase from the economics of law that refers to • Developing, making, and selling “systeman agreement in which the terms cannot be products” with various componentscompletely specified and agreed at the outset. • Spreading and gaining industry acceptanceAs a result of these first two conditions, all for a technical standardalliances share a third characteristic: jointdecision making to manage the business and • Hedging bets and creating options forshare the value. future actions
3• Promoting various products or services the single firm has more required capabilities under one umbrella brand in-house than do the members of the constellations. The underlying theme of myJust as with alliances, constellations too can framework is that the design of a constellationtake on a variety of forms. Examples of affects how it competes and that the positionpossible forms are these: of a firm among and within constellations influences the gains made by the firm. (For• A production joint-venture with several related concepts, see Nohria and Garcia-Pont, owners 1992; Jarillo, 1988; and Hagel, 1996;• An R&D consortium with multiple Lorenzoni and Ornati, 1988; and Normann and members Ramírez, 1993.)• A jointly-owned “shared utility” serving various clients Implications for Managers Most of the managerial literature assumes implicitly that• A co-marketing network or franchising the firm is the primary unit of competition. system with multiple local channels This assumption needs to be modified for a• A committee to set standards (official and world with multiple alliances and competing de facto) constellations. In businesses where collective• The alliance portfolio of a business unit or competition is important, managers need to of a company govern not only the activities within the strict boundaries of their firm, but also theirWhile the forms and purposes of constellations alliances and constellations outside thesemay vary, they share certain characteristics. boundaries. Even without further analysis ofStrategic thinkers today commonly see the collective competition, therefore, we canfirm as a bundle of resources or capabilities, define two clear implications for managers:administered to achieve competitive advantagefor the firm (the original statement in Penrose, • When competitive performance depends on1956; a recent synthesis is Collis and the firm’s alliances, managers need to payMontgomery, 1997). These capabilities might attention to two sets of actions: (1) Theconsist of physical assets, intellectual assets, initial design of alliances (i.e., settingbrands, technical know-how, and so on. I see a goals, choosing a partner, and crafting theconstellation as an alternative way to govern structure) and (2) the management ofsuch a bundle of capabilities (see also Powell, alliances after start-up (i.e., building1990). Just like the firm, the constellation relationship, adjusting plans, and makingattempts to govern these capabilities so that it joint decisions). Elsewhere, I and othersgains competitive advantage in the have discussed guidelines for doing this;marketplace. But, unlike the firm, the the rest of this paper does not discuss theseconstellation is governed by a system of actions at the level of individual alliances.alliances, not by full organizational integrationand full control through ownership. • When the firm uses constellations to compete, success will depend on a parallelIn other words, while I agree that the firm can set of actions, i.e. constellation design andbe seen as a resource bundle that competes in constellation management. These actionsthe market, I argue that not all resource will require management to think broadlybundles need to be organized as firms -- some about its business and its capabilities, andmay well be organized as constellations. As a often demand an outside-in perspectiveresult, one might see single firms competing that seeks to shape the competitiveness ofagainst constellations; this simply means that the whole value-chain surrounding the
4 firm. The rest of this paper discusses an The next step is to stitch these capabilities approach to doing so. together effectively. Because collective competition is still a young art, we have incomplete evidence of how best to organize aCreating Value: Constellation Strategy constellation. But, the evidence from a numberand Structure of industries suggests that an effective constellation requires a unifying force of someAs constellations become important in a kind – leadership at the core (Lorenzoni andbusiness, the competitive advantage of a firm Baden-Fuller, 1995), shared businesscomes to depend more and more on factors strategies, or common motivations (sometimesoutside the firm, in addition to the usual a common enemy helps too!). (See cases offactors under the firm’s control. Put simply, Coca-Cola, Visa, and Colliers in Bamford,the firm’s performance comes to depend on Gomes-Casseres, and Robinson, 2003; and ofthe capabilities of its allies and on how the RISC and PDA constellations in Gomes-relationships among allies are managed. As Casseres, 1996). A corollary to this is thatnoted earlier, the constellation’s control over competition among members erodes thethe resources inside its boundaries is cohesion of the constellation (Hwang andanalogous but not identical to the control that a Burgers, 1997). Other organizational elementsfirm exercises over the resources inside the matter too, but the point should be clear: Afirm. Still, to understand how value is created constellation can only hope to gain advantageby constellations, we can ask the same kinds from member resources if it is able to combineof questions that strategists ask about firms: and govern these resources effectively.What are the resources available to theconstellation? And how does the constellation Implications for Managers The discussionmanage these resources? Answers to these so far allows us to draw some preliminaryquestions go a long way toward determining guidelines for mangers about constellationthe “group-based advantages” created by the strategy and structure:constellation. • Constellation designers will face a tradeoffConstellation membership is the main between (1) expanding the group in andeterminant of what resources are assembled effort to assemble more and greaterinside the constellation. Because of the mix aggregate capabilities, and (2) keeping theand number of members, the constellation may group simple in an effort to promotebe able to count on a greater or lesser scale and effective governance. The appropriatescope of operations, on a variety of technical balance between expansion andcapabilities, on market presence in certain governance depends on the competitivesegments, and so on. Which of these resources context – e.g. the structure and behavior ofis actually relevant to competitive success, of rivals and the need for integration – andcourse, varies from industry to industry – e.g. the dynamics of the emerging group, e.g.scale and scope may not be important the alignment of incentives among players.everywhere, technical variety may or may notbe valuable, and so on. But, given a • Successful management of constellationscompetitive domain, managers ought to be requires careful mapping of theable to define the set of capabilities needed for competitive landscape and considerationsuccess. This is the first step in designing a of various options for membership andviable constellation. structure. This is not an activity that currently is regularly done by strategists in many firms; it also requires careful
5 monitoring and analysis of alliances of the IBM. Intel’s and Microsoft’s resources, firm’s rivals. however, were protected by copyright and by the firms’ efforts to block imitation and stay ahead of clones. As a general matter, the betterClaiming Value: a scarce, valuable resource contributed by aBargaining within the Constellation partner is protected by formal legal means, the greater will be the ability of that partner toAlthough constellations are created to generate exact value from the constellation. Intel andgroup-based advantages, they must yield value Microsoft also benefited from competitionat the firm level in order to attract and retain among systems vendors, i.e., among theirmembers. The game of competition may have suppliers of complements. IBM had no suchchanged, but we still keep score the old way. luck. (For a discussion of theoreticalGiven that a constellation creates some value, approaches to this case, see Gomes-Casseres,what determines how much of that value an 2003.)individual member can claim? Implications for Managers Again, becauseTo see why this question is crucial to the art of constellation management is young Icompeting in constellations, consider the can only suggest preliminary guidelines:history of the personal computer. The IBM PCwas launched in 1981 by a constellation • For a firm to gain from participation in acreated by IBM, with Intel supplying the constellation, it must be able to claim somemicroprocessor and Microsoft the operating of the value created by the collective. Thissystem. As a group, this triad created the means that it needs to position itself withinmicrocomputer format that within a few years the constellation so as to try and controldrove both the Apple II and the previously key, scarce resources or otherwisedominant CPM operating system to the increase its bargaining power vis-à-visperiphery of the market. Later, this IBM PC other members in the group.constellation slowly fell apart, but Microsoftand Intel went on to develop the powerful • This often raises a Catch-22 inWintel alliance. The main lesson here is that constellation growth: By sharing itsthe constellation created tremendous group- capabilities generously, a lead firm in abased advantages (it established the dominant constellation can attract strong partners,industry standard), but the firms within the and perhaps erode the power of rivalconstellation benefited to different degrees. constellations. But, this growth may wellIBM, it turned out, ended up with the least come at the cost of the firm’s ability toclaim on the joint value, even though it appropriate value from its constellation.initiated the constellation, held a central One way out of this dilemma is to begin theposition, and was much larger than its constellation with generous sharing, and topartners. try to increase the firm’s bargaining power over time.The key reason for this outcome lies in thenature of the resources each party contributedto the joint enterprise. In IBM’s case, its Conclusion: Managing beyond the Firmresources were marketing, manufacturing, andthe architecture of the product. To IBM’s This short paper has emphasized that modernsurprise, Compaq and a slew of IBM-clone competitive strategy depends increasingly onmakers were able to imitate the architecture managing resources that lie beyond theand then out-manufacture and out-market traditional boundaries of the firm. Managers in
6businesses that rely on cooperation withgovernments have long known this, as have The right time to address these issues is beforemanagers in entrepreneurial businesses, which alliances have spread too far in an industry.must continually leverage internal resources Alliances often spread in waves as one firmwith external ones. But the spread of alliances reacts to its rivals and before long the wholein a large variety of businesses means that industry is populated by constellations. Whenmany more firms need to manage beyond their this happens, “strategic gridlock” can precludeboundaries. new alliances and severely restrict the scope of constellation design. At the same time, firmsTo help managers in alliance-intensive that did not prepare their alliance strategy inindustries, I propose that they think more advance, will find that an ad-hoc group is nobroadly about the resource bundle with which match to a well-designed constellationthey compete. These bundles don’t lie (Bamford, Gomes-Casseres, and Robinson,exclusively within their firm, but are created 2003).by the alliance constellation around their firm.Because of this, it is critical for these A final piece of advice: If you think you don’tmanagers to identify, shape, and manage these have to compete in constellations, think again.alliance constellations. This is not a typical It may creep up on you without warning. And,managerial assignment, certainly not for line if upon reflection, an alliance constellationmanagers concerned with profit and loss on might be in your future, learn fast how tospecific operations within the firm. But, for manage beyond your firm. The approach inthose firm leaders thinking about strategy and this paper can serve as a starting point.about how the firm should be positioned in itsenvironment, the kind of analysis suggested *****here is critical.Benjamin Gomes-Casseres is Professor of International Business at the International Business School atBrandeis University. Previously, he was at the Harvard Business School (1985-1995). His research, teaching,and consulting have long focused on alliance strategy and management. He is co-author of Mastering AllianceStrategy: A Comprehensive Guide to Design, Management, and Organization (Jossey-Bass, 2003) and authorof The Alliance Revolution: The New Shape of Business Rivalry (Harvard University Press, 1996). His articleshave appeared in Harvard Business Review, Sloan Management Review, Journal of International BusinessStudies, Economic Behavior and Organization, Financial Times, and in other journals and books. He maintainsa website on alliance strategy at www.alliancestrategy.com and can be reached at firstname.lastname@example.org.
1ReferencesBaker, George, Robert Gibbons, and Kevin J. Murphy, “Relational Contracts and the Theory of theFirm,” Quarterly Journal of Economics, 2001.Bamford, James, Benjamin Gomes-Casseres, and Michael Robinson. Mastering Alliance Strategy: AComprehensive Guide to Design, Management, and Organization. San Francisco: Jossey-Bass/Wiley, 2003.Collis, David J. and Cynthia A. Montgomery. Corporate Strategy: Resources and the Scope of theFirm. Chicago: Irwin, 1997.Gomes-Casseres, Benjamin, “Competitive Advantage in Alliance Constellations,” StrategicOrganization, Vol 1 (3), August, 2003, pp. 327-335._______. The Alliance Revolution: The New Shape of Business Rivalry. Cambridge, MA: HarvardUniversity Press, 1996.________, “Group Versus Group: How Alliance Networks Compete,” Harvard Business Review,July-August, 1994, pp. 62-74.Hagel, John III, “Spider Versus Spider,” McKinsey Quarterly, 1996, Number 1.Hwang, Peter and Willem P. Burgers, “The Many Faces of Multi-Firm Alliances: Lessons forManagers,” California Management Review 39 (Spring 1997): 101-117.Jarillo, J. Carlos, “On Strategic Networks,” Strategic Management Journal, January-February, 1988,pp. 31-41.Lorenzoni, Gianni and Charles Baden-Fuller, “Creating a Strategic Center to Manage a Web ofAlliances,” California Management Review, 37 (Spring 1995): 146-163.Lorenzoni, Gianni and Oscar A. Ornati, “Constellations of Firms and New Ventures,” Journal ofBusiness Venturing, 3: 41-57 (1988).Nohria, Nitin and Carlos Garcia-Pont, “Global Strategic Linkages and Industry Structure,” StrategicManagement Journal, Summer, 1992, pp. 105-124.Normann, Richard and Rafael Ramírez, “From Value Chain to Value Constellation: DesigningInteractive Strategy,” Harvard Business Review (July–August 1993): 65–77.Penrose, Edith T. Theory of the Growth of the Firm, 3rd ed. New York: Oxford University Press,1995. First published in 1959. Second edition published in 1980.Powell, Walter, “Neither Market Nor Hierarchy: Network Forms of Organization,” Annual Review ofSociology, 24: 57-76.