CONSTELLATION STRATEGY: MANAGING ALLIANCE GROUPS

Becoming part of an alliance constellation, like the airline industry’s Star Alliance, can help a company compete and win. And while a constellation is, in some ways similar to a bi-lateral alliance, it also presents particular challenges. This author discusses these challenges, and some tactics that managers can use to make a constellation – and their own company’s strategy – succeed.

A few years ago, Cable & Wireless ran a series of glossy advertisements that proclaimed, “The corporation is dead. Long live the federation.” It was a catchy slogan that seemed to reflect a compelling vision for a global alliance constellation: The C&W federation would unite 50-plus firms from around the world into a new super-corporation. But a year later, the C&W federation had imploded in public view, as the partners found few benefits in deep multilateral co-operation.

Despite this failure, companies can compete-and win-by using a constellation of allied firms. (I define a constellation as a set of firms, linked through alliances, that compete in a specific business domain.) Airbus competed with great success against Boeing. The Star Alliance with its dozen or so members altered the nature of competition within the airline industry. Coca-Cola and Pepsi both managed complex constellations of bottlers and distributors around the world. And both Visa and MasterCard were successful organizations that governed hundreds of member firms.

In other industries too, alliance constellations have proliferated. Why? As is the case for simpler bilateral alliances, constellations can be used to pursue any number of strategic goals, including these:

  • Linking Markets: Companies sometimes form constellations to connect local markets and, in the process, provide customers with broader geographic coverage.
  • Combining Skills: Companies also form constellations to assemble a diverse basket of skills, sometimes to launch a totally new business.
  • Building Momentum: Constellations are also used to create market momentum-that is, to persuade customers, suppliers or competitors to adopt a new technology or business protocol.
  • Reducing Costs: On occasion, companies form constellations to reduce costs.
  • Sharing Risk: Companies also assemble multiple partners to share large investments or risks.

Given a set of goals and the basic logic of a constellation, how do you make it work? For starters, it is useful to have an intellectual construct-a framework for understanding constellations. While constellations share many traits with bilateral alliances, managing them also presents important new issues and accentuates others. In Mastering Alliance Strategy, my co-authors and I discuss a number of these issues, some of which I discuss in this article.

While many design elements must be considered when forming a constellation, our work suggests that firms must pay careful attention to four: group size, membership mix, internal rivalry and governance.

Group Size

The total size of a constellation, that is, the number of member firms, is a key design criterion. For some of the goals mentioned above, the more the merrier; for example, more partners help build momentum and may share risk or reach markets better than fewer partners. But for other goals, such as combining skills or even reaching scale economies, the number of partners is less important than the type and size of each partner. And, in all cases, the more partners are involved, the more separate interests will be represented, and so the more challenging will be the governance of the constellation.

Cable & Wireless’s strategy in designing its global telecommunications constellation seemed to be to assemble as many partners from as many countries as possible. Yet this didn’t matter much to C&W’s global customers, except for, perhaps, its stake in Hong Kong Telecom. AT&T, too, pursued many partners in its early constellation, called Worldpartners. Even so, in the battle for global telecommunication services, Concert, British Telecom’s joint venture first with MCI and later with AT&T, became a market leader with just two partners. BT understood that in the battle to provide voice, video and data services to multinational corporations, the competitive advantage of a group of  allied firms hinged on the size and reputation of the lead firms, as well as the nimbleness of the alliance itself, not on the sheer number of partners.

Several lessons flow from this and other cases. First, managers must realize that the ideal size and scope of a constellation depends critically on its goal and competitive context-different business strategies require different collections of players. Second, they should not increase the size of a constellation without good reason, as in itself, large size imposes management costs. Third, in setting the size and scope of a prospective constellation, managers should focus on the elements that are important for competitive success. Sometimes, success will indeed stem from having many sponsors, but more often than not it stems from having large, leading and capable partners, even if not many of them. And fourth, in developing the constellation, consider introducing partners in phases-often an early partnership will attract others, but just as often an overgrown constellation will need trimming later.

Membership Mix

At times, advantage for a group comes from its ability to assemble a diverse set of capabilities. What matters, then, is not the size or number of the pieces but, rather, that the right pieces are assembled snugly together. A good membership mix in a constellation is one that combines multiple skills and capabilities. Doing this requires, again, a keen sense of the success factors in a competitive domain.

Membership mix has been important in constellations in the market for personal digital assistants (PDAs) from the very start of this industry segment. A PDA combines the processing capabilities of a simple computer, the communications capability of a cellular phone, and the size, styling and durability of consumer electronics. To compete in this environment, early entrants thus knew that they had to “force” the convergence of elements from at least four industries-computer hardware, computer software, telecommunications and consumer electronics. Major companies in each of these industries entered the field, each arriving with their own particular strengths. IBM, Apple and Hewlett-Packard approached the business from their experience in computer hardware; Microsoft and Lotus, from computer software; AT&T, Motorola and BellSouth, from telecommunications; and Sharp, Casio, Tandy and Amstrad, from consumer electronics. Even so, they each assembled constellations, which gave them access to the technical capabilities they lacked.

The lessons are clear. First, managers must examine carefully what elements are needed to mount a credible strategy in a given business. This means thinking outside of the constraints of your own firm. Don’t start by asking what you have that helps you compete, but instead what a hypothetical competitor would need to succeed. Only after that should you consider precisely which of these critical elements you have, which are only weakly present in your firm, and which you lack completely. The third step is then to identify which other firms might have the elements that, if added to yours, would create a potentially winning combination. And, as with partner selection in bilateral alliances, one must be careful in this last step to consider not only what the other firm might add, but also how easy or difficult it might be to work with the other firm, as discussed next.

Internal Rivalry

Key among the factors that determine how firms work together in an alliance is the degree to which they compete with each other. But it is increasingly rare to find partner pairs that are totally devoid of competitive friction. And, again, this problem increases exponentially as we add members to a group. So, most alliance constellations cannot help but have some degree of internal rivalry; the key to success lies in minimizing this rivalry by careful choice of partners and then managing the rivalry by careful design of the structure and governance of the group. Japanese keiretsu, for example, traditionally had a more or less formal “exclusion rule” that stipulated there should be no duplication of activities among members.

But is exclusion always best? No. Some internal duplication-and thus rivalry-may in fact encourage innovation, increase flexibility and protect against uncertainties through diversification. Because of this, the lead firms in a constellation may choose to involve multiple partners in specific areas, even if it leads to a certain amount of internal friction. The point is to do so consciously and to weigh the pros and cons of internal rivalry. While some internal rivalry can spark innovation and limit dependencies, it also usually introduces tension and slows down collective decision-making.

Sometimes, internal rivalry can be managed by design of the roles and responsibilities of member firms. Narrow-scope alliances often live side by side more effectively than do broad ones. This means also that firms developing a new constellation must look ahead and not quickly “sign away” broad rights to the first partners; doing so may well pre-empt the addition of new partners or create internal conflicts when new partners do join. Finally, members of a constellation often benefit from creating separate organizational structures that can transcend the individual interests of the members, as Star Alliance did.

Governance Structure

Creating a separate organization is but one form of governing a constellation. Whether or not this model is used, a key issue in constellation design is how group decisions will be made and how member strategies will be coordinated and aligned.

A governance structure does not necessarily mean a formal, separate organization or consortium office. With one strong firm at the centre of a constellation, governance is usually managed by that firm, as part of managing its alliance portfolio. Boeing has no separate organization for managing its vast supplier network, nor does IBM for its network of software developers.

However, a formal and separate organization is often useful when a constellation is large and when it has a high degree of internal competition. Such an organization gives partners a more or less neutral way to establish common goals and rules of behaviour. Broadly, three types of governance structures are common in an alliance.

Drivers of growth

Alliance constellations often appear and grow rapidly in an industry. Once key players in an industry begin to use alliances for one reason or another, their rivals often are not far behind. Take the growth of alliances in the airline industry, shown in Figure 1. In the 1990s, a few new alliances were formed every year, leading to early constellations that remained limited in membership. But a number of important players continued to compete more or less as single entities. By the late 1990s, the pace of alliance formation picked up to a frenzy; new constellations were formed and grew rapidly. After 2000, almost all large airlines had joined a constellation, and the largest groupings counted over 10 members.

Three types of factors can drive the spread of alliances and the growth of constellations. The first is change in the business environment, such as in technology, regulation or demand. The second is imitation, in which managers follow organizational fashions, including those thought of as “best practices” at the time. The third is competitive rivalry, in which companies react to each other’s strategic moves.

Changes in the Environment

The emergence of a new technology often favours the formation of alliances. Thus, when the oil crisis of the 1970s created new demand for energy-efficient small cars, all the U.S. carmakers headed to Asia to seek allies that could help them build and sell small cars. But, often, the firms in an industry do not face exactly the same environmental forces and certainly do not interpret these forces in identical ways. And there is always more than one possible response to a given environmental change, leading some firms to choose alliances and others not.

In addition, the nature of an environmental change is seldom clear-cut from the start; more often, change evolves from competing trends that only become clarified in time. Alliances are attractive tools for managing risk in such conditions, because they help firms experiment with new strategies. This is one reason for the rise and decline of alliances in many high-tech businesses.

Imitation and Fads

Another popular explanation of why alliances spread rapidly is management fashion. Certainly, alliance strategies are not immune to the hype common to other management trends.

Imitation sometimes can be a reflection of how companies learn. Followers will imitate successful leaders when the leaders have demonstrated that a new strategy is useful. This kind of imitation is common when a strategy has been dubbed “best practice” by industry players. In the 1990s, as more and more leading firms formed alliances, others began to study their methods in the hope of discovering keys to success. All too often, the results of these best-practice studies were disappointing, as the tactics of one company failed to fit the circumstances of other companies. Yet, there is little doubt that the positive examples of successful alliance practitioners have helped educate others and encouraged them to form new alliances.

More cynical observers see management imitation as a much less rational process. In this view, following the trend offers protection in numbers. Management gurus and consulting firms may unwittingly promote this process by raising one strategy or another as a model. The Internet boom of the late 1990s was replete with this kind of imitation; venture capital investors pursued one fad after another, including the rampant use of alliances. Driven by investor expectations, many dotcom start-ups created large constellations of allies just to show how popular they were.

Competitive Reaction

Rivalry of various kinds often lies at the root of constellation dynamics. Economists have long noticed that close rivals in an industry often follow each other’s strategic moves, whether it is entering a new market, offering new products, or using a new organizational form such as a constellation.

Firms often follow each other’s strategic moves to avoid falling behind. The reasoning in alliance formation is as follows. Assume that constellations A and B are rivals, and that A expands its group by adding new members that bring new capabilities to the group. If B decides not to follow, and A’s new strategy generates advantages, then B falls behind. But if B follows by expanding its own group, then the two rivals are likely to stay close competitors, regardless of how the strategy turns out. Such a process of imitation drove the growth of constellations in airlines, where Star, One-world and Qualiflyer seemed to follow each other’s moves closely.

Sometimes, being first is the key to success. When an industry has high economies of scale, high customer switching costs or steep experience curves, competitors can be expected to try to establish a dominant market share early. Firms can use alliances to shorten the time needed to establish a lead position. Or, if they are latecomers, they can more rapidly erode the position of the lead firm. Alliances may thus quicken the race for first-mover advantages.

Because of such first-mover advantages, firms sometimes make pre-emptive strikes. Tying up early gives you the best choice of partners and pre-empts rivals from tying up with an attractive partner. Much more than imitation is at work here. A constellation may execute a pre-emptive alliance not because it is in a rush to match the capabilities of a rival, but because it wants to keep the rival from assembling this set of capabilities.

Limits to growth

Just as there are several drivers to the growth of constellations, a variety of limits may inhibit their expansion beyond a certain point. The airline industry did not encounter these limits before 2000, but there were indications that some of the larger constellations were beginning to face constraints on growth.

The simplest of these constraints is overcrowding of the field. The pool of eligible partners will diminish because of the boom in alliance formation itself. Or, including the still-available partners into an existing constellation may threaten to increase conflicts of interest among group members. This form of saturation in the industry has been called “strategic gridlock,” because it tends to limit the strategic flexibility that alliances typically offer a firm.

Other limits to the growth of constellations stem not from the external environment, but from internal constraints. Managers often cite scarcity of management capacity as a constraint on alliance formation. Negotiating each agreement requires great effort, and major alliances require the continual, direct and personal involvement of top management. These demands on management increase with the size of the group and the complexity of member interactions.

In addition, co-operation becomes increasingly difficult as the number of members in a constellation rises. Conflicts may also arise when one of the partners has another alliance that influences the goals of the first alliance. Whatever the root causes, such conflicts of interest will add to the coordination costs of the group and limit the degree to which it can be integrated to implement a common strategy.

Managing constellations

The ways in which constellations arise, grow and change over time are still not completely understood, but the best thinking and experience to date suggests the following guidelines for managers:

  • Evaluate design tradeoffs between group size, membership mix, internal rivalry and governance structure. Often, decisions in one dimension will create challenges in another. The time to think this through is in advance, not after your constellation has expanded in an ad hoc fashion.
  • Map your constellation and those of your rivals. All too often, key decision makers in a company do not have a clear picture of their constellation, or the various constellations in which they play. Here a picture is truly worth a thousand words: Create a diagram and update it regularly. Do that for your constellations and for those of your rivals; analyzing the alliance strategy of rivals is now a key element in competitive intelligence. Evaluate the strengths and weaknesses of your constellation compared with theirs.
  • Be clear about why you follow your rivals, or why you don’t. Imitation among constellations is common, as we saw. But this does not mean you should follow blindly. Rather, understand what drives imitation in your industry, who might be worth learning from, and when you should desist from following an example. One senior executive was fond of sending little notes to subordinates, attached to the latest news of a rival’s move: “Why are we not doing this too?” he would ask. His point was not to encourage his managers to copy, but rather to get them to evaluate and fashion a response to rivals’ actions.
  • Adjust constellation structure to respond to new conditions. As with the single alliance, a group of multiple partners can never stand still-it has to be flexible to pursue new opportunities and fix problems. In a constellation with many members, this can be difficult, because of the multitude of interests involved. This is one reason why leadership from one or a few firms is critical to good constellation design.