Silos and the art of Empirical Theology

In reponse to my attempt to reconstruct the definition of a silo in a value-neutral way, Torp brings up an interesting empirical question about the relative proportions of healthy and unhealthy silos in the “wild,” and how you would add some empirical color to the discussion. It is reasonable to wonder whether any healthy silos actually exist, and ask how you might detect their existence and measure their “health.” I am going to argue here for an answer based on an analogy between macroeconomics and microeconomics that I hope you find surprising.

The Setup: A Micro-Macro Recursion Analogy

To respond to this thought, let me set up my micro-macro analogy, since that is the basis for my answer. In classical economics, the firm is supposed to be an organization that manifests an idealized form of collaboration within, while the larger embedding economy, which for this article I’ll take as synonymous with “market,” is supposed to manifest idealized competition among economic actors (which could be firms and/or individuals). The imperfection of real markets is readily recognized, but the underlying conceptual model of markets — what amounts to Darwinian competition — has remained solid. Economists from Hayek to Keynes to Friedman essentially work within the same competing-actors/invisible hand framework due to Adam Smith.

The firm though, is a different story. The ideal firm implicit in treatments in managerial economics (which I am just starting to explore beyond Economics 201 level, so don’t skewer me on this point) evokes a vision of a coordinated, collaboration machine dedicated to a single goal. Unlike the real market though, the real firm is not just a noisy, limited-information, bounded-rationality approximation of the the ideal thing. The inside of a firm (and this is easiest to see in the case of large firms) is so far from the notion of the ideal firm that the model should be discarded at a conceptual level. Why?

• Firms do not (and in general, should not) have a single unambiguous goal. At any given time, there are competing pressures from different goals: Wall Street quarterly result pressures, customer pressures, pressures from “value” investors to innovate, and pressures arising from the multiple possible interpretations of these ambiguous top-level drivers (defining “customer” as “current customer” or “current and potential future customer” leads to very different results).
• This shows up in the fact there has never been consensus on a single metric for even a single aspect of a company’s performance, let along a single metric for the entire performance. Performance on Wall Street, I am told, has variously been measured using a variety of metrics ranging from earnings per share (EPS) to a beast called EBITDA ( Earnings Before Interest, Taxes, Depreciation and Amortization).
• Enough stories abound about companies not collaborating internally that we can dismiss the possibility of a firm coming together in idealized collaborative execution even if it miraculously did achieve consensus on the goal.
• Tales of redemption and success that look like they are about collaboration towards the greater good (such as the great anecdotes in that classic on change management, The Wisdom of Teams) need to be taken with a grain of salt. These are tales of new, healthy silos dethroning old and unhealthy silos, not tales of all silos being banished on the path to the Grand Collaborative Firm.

This dynamic of dissent and lack of collaboration I will argue, is a good thing, and we shouldn’t be looking to fix it in a naive way. A highly collaborative organization marching to the tune of a single doctrine is not viable in the long term (though periods of such dictatorial discipline are probably good for short periods). The reason is that the market in which the firm operates is a complex place that sends a rich, noisy mixed signal into the innards of every firm within it. Different parts of the firm will process different data sets with different analytical methods. Marketing will highlight demographic trends, while engineering will be screaming about China, and research about missing the nanotechnology window of opportunity.

So what’s a good model for this tumult of a hundred contending voices? There’s no need to invent a new one: the same model that is used for the market at large is also a good model for the inside of a firm. Of course, we need to set the parameters differently, perhaps move closer to the nonzero sum end of the game theoretic spectrum (note that one of the “flaws” of the model of the market is that it does not fit an ideal zero-sum model). The key is that performance and growth of the company as a whole arise from internal competition among silos. This is hardly a revolutionary conclusion: the boundary between the firm and the market is a porous one, so we should expect the same sorts of models to work in both cases, albeit with different parametric characteristics (if you like the biological metaphors, think of cells emerging from a primordial chemical soup — the little bubbles of soup encased in semi-permeable membranes are still soup, even if they are stewing less vigorously).

So that allows me to propose some answers to Torp’s questions:

Who’s The “Maker”?

In the first part of his comment, Torp wonders (emphasis mine):

… if you took about 100 companies, and asked each one to recognize its top 10 silos, and then had them rate the silos as functional or dysfunctional… what the stats would be like…most functionally siloed organizations invariably have large amounts of 1 and 7 [key attributes of dysfunction in my definition], which get in the way of making them truly collaborative, and “company-first” thinking groups.

Let me get a minor point out of the way: whether you asked the question via a democratic poll or a CEO interview (or something in between) you are probably going to get an overestimate of the number and toxicity of the silos present in the organization. This is because normally functioning healthy silos per my micro-macro analogy should create a somewhat adversarial atmosphere of competition, and you should expect accounting and engineering to paint each other as “dysfunctional” silos just as Microsoft and Apple might call each other names. Bad times will only make this name-calling worse, since the economic jungle selects for survival, not truth-telling. But let us assume that our interviewer is smart enough to frame the questions right and analyze them right to get rid of this “healthy competition” bias.

The problematic part of Torp’s question is the “making them” part. Who is this maker? The CEO’s office (s/he and her/his close advisers) are as much a silo working in the internal economy to influence others as any other actor (the same is true of the market, since the government is just a particularly advantaged set of economic actors). So how do you “make” silos collaborative, “company-first” groups?

You, whether you are an individual employee or the CEO, shouldn’t even try to exercise influence with this particular intention in mind. Company-first slogans are fine for morale boosting and tapping into team spirit occasionally, and for corporate brand management, but in general, pure collaboration is a bad thing and should be resisted (call it “groupthink” and see if you agree with me now). Similarly “company-first” implies a consensual definition of “company-first” and whip-cracking around a single definition is only good for short periods of emergency recovery, but not a good practice in general. Every silo, healthy and unhealthy, will define “company first” differently. Sales will say: “company first = customer first,” engineering will say “company first = parts cost first” and research will say “company first = let’s get into nanotechnology”). In all but the most unhealthy periods, analogous to wartime or a national emergency at the level of nations, this dissent is necessary. This landscape of local definitions of the “greater good” is both inevitable and necessary. There is no such thing as a systems thinker (that’s a whole different post for later).

Aside: though the micro-macro analogy breaks down on several specific points, one bit survives. Just as money is the transactional medium for the market, mutual favors form the currency of the internal economy of a firm. A careful barter of favors through trust networks, built up one conversation at a time, is what keeps the firm humming.

And the same sort of force as in the Adam Smith model of the ideal economy directs this chaos towards growth: the invisible hand of (however imperfect) internal Darwinian competition.

Now this is probably not an appealing message to CEOs who need to believe in their greater ability to influence things in order to punch in to work every day. But a belief in an economic model of the internal behavior of a firm is not the same as fatalism. The same sorts of guided-evolution influence mechanisms work; you just need to frame your intervention in the same ways the federal government frames its approach to influencing the nation. You need to argue for your chosen level of intervention the way Keynesians argue with Friedmannians.

Measuring Silos

To get to Torp’s other question:

Any idea on how we can add some empirical color to this discussion? how does this really work in practice? How many shining examples are out there of human organizations that have formed “high-grade, high-value” siloes, without falling victim to the cons listed?

To that, here’s a quick suggestion based on my analogy: a sensible approach to measuring silos must borrow from the best in econometrics. Satisfaction surveys and the like are not likely to work. Measure the corporation like an economy. Map social networks and favor barter trades as best you can, without disrupting their functioning.