There’s a flavor of strategy that worries me. It looks something like this: successful company did action and was successful, therefore if I do action, I will be successful.
I associate four general problems with this thinking:
The apocrypha problem
The causality problem
The external validity problem
The dynamic market problem
The apocrypha problem
Here, I take issue with [successful company did action]. Surprisingly frequently, stories about companies are simply untrue. I think there are generally three kinds of untrue stories that get told: urban legends, contested narratives, and flawed recollections.
Urban legends are either invented by third parties or twisted in elaborate games of ‘telephone’. They tend to be literally untrue but helpful in demonstrating a generally true principle. For example, we might hear a story of a company that spent nothing on marketing, or how the early work at Apple transpired primarily in a garage. We can interpret these as stylized representations of general truth, that these companies ran lean operations. I strongly suspect that many specific, entertaining stories about Sam Walton are untrue, but that they nonetheless illustrate his character.
Contested narratives evolve mainly from parties serving their own interests. These are reasonably easy to identify. If multiple versions of the same story exist, it’s unlikely that all are equally true. It isn’t necessary for anyone involved to be lying — they may all share some incomplete context — but we all know that some people lie. I consider the Amazon Web Services founding story to be an example of a contested narrative.
Flawed recollections develop accidentally. We all regularly and innocuously forget details. We might forget the order of events, the tenor of a specific conversation, or the precise context behind some decision. We may fall victim to false memories. It’s simply not necessary for a person to bear ill intent for him to relay an untrue narrative. We all do this.
In any event, there is no reason to expect good outcomes from imitating false stories.
The causality problem
It’s very challenging to determine which actions effect any given outcome. I suspect most of my readers know this, but I imagine two major barriers to causal identification: noise and omitted variables.
By noise, I mean that desirable outcomes can result from random chance. My favorites come from the world of speculation; Reddit’s Wall Street Bets offers countless examples here.
I use omitted variables mostly in the spirit of its statistical meaning; it takes a lot of work to develop causal identification. Let’s consider a silly example. Suppose my car rolls down a hill. All the while, I feather the brakes. Absent the discovery of gravity, we might be tempted to conclude that brakes caused my movement down the hill. After all, the downward motion is always and everywhere coincident with my braking.
The mere fact that some actions precede or success is no signal of causality. At most, one can infer that such actions did not entirely preclude success. The actions in question may prove an entirely uncorrelated waste of time — or worse still, they’re like the brakes, merely overcome by gravity.
Many companies succeed despite poor choices. As such, copying some aspect of a predecessor’s playbook may oversample poor choices and miss the the factors that actually caused success.
The external validity problem
Even if we believe some series of actions caused a desirable outcome for some company A, there’s no particular reason to believe those actions would cause a similar outcome for some company B.
These companies likely differ in arbitrarily many ways; any one of them could be sufficient to invalidate A’s playbook as a strategy for B.
They may have different value propositions among their customers
They may have customers with differing demand elasticities
They may be subject to different competitive dynamics
They may have different economic models
They may have different organizational cultures
They may operate in different time periods
They may operate in different geographies
They may not share the same regulatory scheme
[Can come up with countless others]
Even companies that appear very similar on the surface may differ in some way that matters.
The dynamic market problem
I consider this a special case of the external validity problem; it matters enough to warrant its own discussion.
Any market comprises agents, each of which maximizes expected returns. In a world with perfect information, these agents adopt any available strategies.
Once some company A implements a playbook, we should expect all participants in the market (customers and competitors alike) to respond. Competitors might copy A’s strategy; more importantly, they might adopt some altogether new strategy that eradicates A’s competitive advantage.
It’s hard to say exactly what would happen in any given case, but it’s certainly not a given that the market will yield comparable results for a copycat.