Strategy Research Worth Your Attention

Here is an article listing the names of 20 retail firms that have been more profitable than their competitors over the last 5 years (plus, there is an interactive chart showing average analyst recommendations for the stocks of the top six through Oct., 2011). When I tell you that, across a broad spectrum of measures, Target persistently outperforms its direct competitors, what explanations leap to mind?

Whether you are an active strategy scholar or someone whose only exposure to strategy was a first-year course in an MBA program somewhere, my guess is your train-of-thought goes something like: the fact that Target’s performance superiority is repeated year after year makes “sheer  luck” an unlikely explanation; therefore, Target must have superior capabilities; moreover, these capabilities must exhibit certain features (e.g., inimitability), otherwise competition would quickly erase its relative advantage. Then, you would be off and running, digging deeper into such possibilities as: superior information technology, better understanding of customers, economies of scale, excellent brand equity, more flexible something-or-other for “competing on the edge of chaos,” etc.

Sadly, it turns out, you may have jumped the gun when you ruled out the “sheer luck” hypothesis. Denrell, Fang and Zhao have a new paper forthcoming in SMJ entitled, “Inferring Superior Capabilities from Sustained Superior Performance.” In it, they explore the problem of inferring superior skills from data on relative firm performance. The paper is an instance of what I have come to consider classic Denrell: take a central central piece of conventional wisdom in management scholarship (the roots of which are probably the informal, intuitive conjectures in some famous paper from the 1980s), analyze the issue with a stochastic model that is at once simple and general, and demonstrate — yet again — that human intuition with respect to statistical inference is really, really bad. This is a salient trait in even the smartest among us. In this paper, the analysis shows why persistent superior performance may not imply superior capabilities. As a bonus, they go on to apply Bayesian methods to a large data set to infer the luck versus ability components of observed sustained advantages.

I always think of papers like this when I hear someone jumping on the “dude, your ivory tower social science is totally irrelevant to real-world business practitioners like me” bandwagon. (In fairness, this bandwagon is presently a fashionable place for fellow academics to be as well.) The moment we accept that humans are inherently bad at logical and probabilistic processing is the moment we flag the importance of learning rigorously derived general principles. Knowing how to decide whether the sustained ass-kicking your are receiving from your competitor is due to luck or ability is, actually, an important skill in real-world business management.

4 Comments on “Strategy Research Worth Your Attention”

  1. teppo says:

    Nice post. Haven’t read the paper, yet. There’s some similar intuition in Alchian, JPE 1950 (specifically, the Borel thought experiment about Parisians flipping coins).

  2. stevepostrel says:

    BTW, you also have to worry that the ass-kicking is actually caused by accounting fraud (WorldCom had the old AT&T managers going crazy trying to replicate their performance).

    I’ve looked at this paper and Denrell’s earlier work on which it’s based. The idea is that if there is path dependence in performance, an early success can lead to repeated future successes even if a firm’s “current” capabilities are no better or even worse than rivals.

    My problem is with the interpretation of “current” and “capability.” For example, if my past (possibly lucky) gain in market share caused me to go down the experience curve, and today I have lower costs because I’m farther down the curve, Denrell wants to say that my current capabilities are no better than rivals’. We are on the same experience curve, so if “capability” refers to the entire functional relationship between experience and unit cost, then Denrell is right. But if “capability” refers to the actual efficiency of current production practices, then Denrell is wrong–the firm that got the past lucky increase in production experience today uses different methods and produces at lower unit cost. Strategy in general is plagued by semantic confusion about these “along the curve” and “shift the curve” explanations for firm heterogeneity.

    This is also another good example of how formal methods are meaningless without reference to the non-formal interpretive link between the phenomena we care about and the formal variables in a model. Formal analysis plays the crucial role of eliminating inconsistencies and laying bare the relationships between different factors, but it can’t prove the relevance or appropriateness of a model to the concepts we are actually interested in.

  3. teppo says:

    Don’t know if this is in the papers – but there’s sort of a Matthew effect (cf. Merton, rich get richer) when it comes to performance – you might perform well in an initial round (perhaps simply due to luck), which may provide access to additional resources, and so on.

  4. stevepostrel says:

    There’s a host of postulated positive-feedback loops in the industrial organization literature. Much of the 1980s and 1990s theory work involved spinning out scenarios where firms would get “undeserved” persistent advantages: Network externalities, differentiation advantages of pioneering brands, experience curves, and so on. In each case, the semantic question of what we mean by a “superior current capability” is raised.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s