January 22, 2017

Competitive vs. Emergent Strategy

At a certain point yesterday, a follower on Twitter asked me some questions about strategic theory. I'll try to frame and answer them as best as possible, as they relate to a project I am currently working on. The best reference for this is Henry's Mintzberg book Strategy Safari. The original schools of the organizational sides of strategic theory – the design and planning schools – tend to assume a stable external and internal environment. They just differ on how strategies are designed and implemented. The primary distinction between design and planning is that design theorists tend to see the responsibility for strategy-making as vested in an chief executive or top brass that uses a simple and informal planning process. The planning school obviously favors a much more formalized and stage-driven planning process. Both of them tend to see strategy-making as the imposition of a top-down derived strategy on an passive internal and external environment.

Obviously the assumption of competition and opponents thwarts things. Thus comes along Michael Porter's competitive strategy (quote is from Mintzberg, Ch. 4):

In fact, the positioning school did not depart radically from the premises of the planning school, or even those of the design school, with one key exception. But even the subtle differences also served to reorient the literature. Most notable in this school has been one simple and revolutionary idea, for better and for worse. Both the planning and design schools put no limits on the strategies that were possible in any given situation. The positioning school, in contrast, argued that only a few key strategies— as positions in the economic marketplace— are desirable in any given industry: ones that can be defended against existing and future competitors. Ease of defense means that firms which occupy these positions enjoy higher profits than other firms in the industry. And that, in turn, provides a reservoir of resources with which to expand, and so to enlarge as well as consolidate position. Cumulating that logic across industries, the positioning school ended up with a limited number of basic strategies overall, or at least categories of strategies— for example, product differentiation and focused market scope. These were called generic.

In the context of military theory, this idea ended up being incorporated into the planning assumptions of US Cold War defense-industrial strategy. Today it is still used in thinking about post-Cold War Great Power threats:

Strategy has to do with how a state or other political actor arrays its resources in space and time in order to achieve its political objectives against a competitor. The key features of any strategy are rationality (the existence of political objectives and a plan to achieve them) and interaction with a competitor who seeks at the very least to achieve different objectives if not thwart our ability to achieve our aims. Competitive strategies are a particular family of strategy...

...[c]ompetitive strategies are generally pursued to achieve limited aims. That is, they are meant to change a competitor’s decision-making calculus and thus his strategic behavior. They do not seek the overthrow of an adversary. In this regard, the competitive strategy that the United States pursued against the Soviet Union succeeded beyond the wildest imaginations of even its most enthusiastic supporters.

I think that competitive strategy is very useful as an analytical tool for thinking about social conflict as well. It's based on the idea of long-term iterated interaction, and isn't necessarily aimed at gaining a strategic decision. The more and more I look at various kinds of conflicts, the more I wonder if the entire idea of decisive victory is illusory (except for a few notable exceptions). So the benefit of competitive strategy is that it focuses on strategic investments that are designed to manipulate an adversary's decision-making calculus and get them to do what we want while optimizing our own competitive advantages. In other words, if we assume that the end of conflict is unforeseeable then being the best competitor is the best substitute for not knowing how to end it.

However, there are still several problems with competitive strategy. It assumes too much structure on behalf of the strategist's organization and that of the competitor. Or the environment, for that matter. When all three are fragmented and in constant flux, its rigidity analytically may be counterproductive. The key assumptions of the competition school is that the form of the competition remains stable, the behavior of adversaries is overdetermined, and all of the competitors are able to evaluate and select clearly defined strategic alternatives. What if the competition is ultimately chaotic and unpredictable? What if the behavior of the opponent is not overdetermined? What if the ability to select and evaluate clear competing alternatives is lacking – either because the clear alternatives do not exist or they cannot be implemented in a fractious organization?

It also still borrows much of the design/planning schools' emphasis on top-down rationalized planning. Is there something better we can use that is more realistic and emphasizes more of merely discovering the right strategy through trial and error and experimentation? The answer is yes, of course. This is the description of "emergent strategy" from Ch. 7 of Mintzberg.

In work carried out at McGill University’s Faculty of Management, in which strategy was defined as pattern or consistency in action, deliberate strategy was distinguished from emergent strategy (as we noted in Chapter 1). Deliberate strategy focuses on control— making sure that managerial intentions are realized in action— while emergent strategy emphasizes learning— coming to understand through the taking of actions what those intentions should be in the first place. Only deliberate strategy has been recognized in the three prescriptive schools of strategic management, which, as noted, emphasize control almost to the exclusion of learning. In these schools, organizational attention is riveted on the realization of explicit intentions (meaning “implementation”), not on adapting those intentions to new understandings. The concept of emergent strategy, however, opens the door to strategic learning, because it acknowledges the organization’s capacity to experiment. A single action can be taken, feedback can be received, and the process can continue until the organization converges on the pattern that becomes its strategy. Put differently, to make use of Lindblom’s metaphor, organizations need not nibble haphazardly. Each nibble can influence the next, leading eventually to a rather well defined set of recipes, so that it all ends up in one great big feast!

Instead of control (implement the intent of the top dog), the emphasis is instead on the idea of strategy as creating an environment in which learning and adaptation can take place. This approach has flaws of its own. First, without any kind of discipline, structure, and direction it is very difficult for strategy to "emerge" out of tactical expedients. The argument of Shimon Naveh and others is precisely that the tactical autonomy of the German military represented a kind of degenerative form of self-indulgence (akin to an really, really long guitar solo in a shitty metal band) that could never really did much of value. Another obvious risk is strategic drift – never settling on anything because the appeal of the novel is too tempting to resist. Learning is difficult and risky and the result may simply be the emergence of a strategy no one wanted.

However, everything involves tradeoffs. Nothing is perfect. Trying to achieve centralized control over strategic options and implementing them via a top-down, rationalized process seemingly leads in practice to a combination of counterproductive rigidity and wild goal-shifting. The future is likely in some synthesis of competitive strategy and learning/emergent strategy. Enough assumption of underlying structure to guide the process of learning in a disciplined way. Enough freedom for learning to be able to be open to new ideas and new possibilities. The competitive perhaps might even be to steer the opponent's learning in a negative direction or trap them into limiting their own strategic options. Consider the problem of controlling learning in robots these days; its very easy for a cleaning robot to "learn" that it can succeed by disabling its own sensors (hoila! I can't see anything to clean ergo there is nothing to clean!) Yet the alternative is depending entirely on hand-coding to direct the robot, which renders it fragile to things its programmers did not anticipate. Assuming your goal is to make sure that the office is cleaned inefficiently, you have put your adversary on the horns of a dilemma. That is the Russian contribution to strategy, in the form of "reflexive control" theory.

Tags: concept