Obligatory navel-gazing preamble
My backlog for Cranberry Blog has gotten a little bit bloated, I’ll admit. I currently have a few dozen posts that I’ve started and failed to finish. I think this is partly because I keep wanting to add more.
In the spirit of actually publishing things, I thought I’d try an experiment: posting shorter articles in series. This article will be my first in my first series. We’ll see how it goes.
Introducing the thesis of the series
I continue to struggle with several pretty ubiquitous precepts of Silicon Valley startup culture. They generally revolve around an ambiguous aversion to management, to MBAs — and by extension to corporate strategy.
In many ways, I do sympathize with the instinct.
When I began my stint in strategy consulting — after working in the startup ecosystem myself — I frequently felt the urge to roll my eyes. Generic thought leadership targeted stodgy Fortune 100 executives felt …a little icky, a little silly.
I remember on one of my first days in the job, I made an infamous three-icon slide designed to communicate “the Cloud is an emerging trend that will disrupt your industry.” To me, it felt bizarre to talk about the cloud, especially because the cloud didn’t strike me as anything close to an emerging development.
[As an aside … I rapidly realized how wrong I was. There is an enormous gap in software adoption between us in the tech industry and everyone else. It’s sometimes hard to appreciate until you have the rare privilege to experience the contrast.]
In any event, as sympathetic as I’ve been to this anti-MBA sentiment, I have an abiding concern about one particularly flavor of the sentiment, that startups don’t need strategy — that they only need execution.
This series will build a case in defense of corporate strategy for early-stage companies.
The strategy-execution framework
There’s one framework — or at least a prevailing taxonomy — with which I expect most readers have some familiarity. We frequently partition the universe of activities into two categories: strategy and execution.
Most commonly, we use ‘strategy’ to describe activities like planning or deciding. We talk about strategy as some broad, general, and ethereal thing. It’s hypothetical. It’s a group of people with inflated titles in a conference room. It’s a bundle of mandates passed down the hierarchy to successive layers of subordinates. It’s a CEO using elliptical language in an investor meeting. We’re going to be the world’s best platform to the future of work.
Then we use ‘execution’ to describe pretty much everything else, all the doing activities, all the details that don’t make it to the pitch deck. Execution is a persistently rapid customer support team. Execution is a highly-paid developer navigating the vicissitudes of CSS. Execution is the sales operations analyst who quietly cleans up edge cases in data that no CRO would ever notice.
It’s a pretty linear path to an outcome. First you strategize, then you executive, then you get some kind of an outcome, good or bad.
We frequently get into debates over the relative importance of these things, particularly in the context of key hires. Do you want the VP who lives for daily process? Or do you want the VP who can set a vision?
Here’s a pretty solid example of how people talk within the strategy-execution framework. (Shreyas is a pretty worthwhile follow in general.)
Later in the series, I will argue that this way of thinking about strategy is really very inadequate. I’ll argue that remarkably few people have experience with corporate strategy and that treating planning as strategy just conjures an illusion of coherent company direction.
But this is a separate topic.
For now, let’s focus on the problems with execution.
Problems with talking about execution
Thinking about execution in this framework has major problems.
Execution subsumes an overwhelming variety of different activities, to the extent that it’s no more specific a term than doing stuff.
We see this vague usage constantly in professional sports interviews. Almost invariably, some pundit, player, or coach will assert the importance of execution. The team just has to go out and execute. It’s never been clear to me what that means — we all just accept it as a milquetoast nothing statement.
Russell Westbrook plays with the cliche in a pretty funny way here — it becomes pretty obvious by the second or third answer just how unhelpful it is to talk about execution.
But somewhat more importantly as a problem, execution lies downstream of strategy by definition. It’s simultaneously more proximate to the outcome than strategy and a dependent outcome of strategy. Moreover, we design strategy based on our reasonable expectations of downstream execution patterns.
As a result, we can freely choose to attribute any given outcome to execution. At least speaking empirically, all of the following are equally unfalsifiable:
The strategy was good, but the execution flopped.
The strategy was good, but the execution was better.
The strategy was wrong, but the execution compensated.
The strategy didn’t matter; it was all in the execution.
Even on the rare occasion we observe something that resembles a counterfactual strategy, we’re left unable to compare different execution patterns — because the strategies dictate the execution.
Obviously it’s reasonable to take an opinion in any given scenario. I do this somewhat often. But I do think we all need to exercise some epistemic modesty when comparing strategy and execution.
The execution-outcome tautology
One particular flavor of the strategy doesn’t matter position uniquely bugs me.
The reasoning generally starts with the following seemingly safe statements:
Statement 1: Execution is either good or bad
Statement 2: Outcomes are either good or bad
Statement 3: It is possible for strategy and execution to differ in quality
Statement 4: We know execution was good if we see good outcomes
Statement 5: We know execution was bad if we see bad outcomes
These premises logically compel the following conclusions.
You get good outcomes if and only if you have good execution
You get bad outcomes if and only if you have bad execution
Execution quality determines outcome quality
Strategy doesn’t necessarily determine outcome quality
Let’s approach this with some very simple algebra.
Suppose x represents the quality of execution, and it takes some binary form, either 0 or 1, where 1 represents good execution. This is just rephrasing the premise of Statement 1.
Suppose y represents the quality of outcomes, which identically takes some binary form, either 0 or 1, where 1 represents a good outcome. As before, this is just rephrasing the premise of Statement 2.
We know from Statement 4 that y = 1 implies x = 1, and from Statement 1, we know this identically means y = 1 implies x != 0.
We know from Statement 5 that y = 0 implies x = 0, and from Statement 2, we know this identically means y = 0 implies x != 1.
We can get a table like the below, which shows us that x = y.
We’ve set ourselves up to establish (pretty trivially) an equivalence between execution and outcomes. If we take the statement So the problem has to rest with our premises.
Statement 1 is kind of dubious on its face. There aren’t many things for which quality is binary, but this isn’t important; Statement 2 has the same similarly unimportant problems as Statement 1.
Statement 3 seems pretty true. Even if we suppose strategy and execution are correlated in quality, it’s pretty hard to assert that a difference is impossible.
Statements 4 and 5 are pretty clearly the problem. If we define the quality of execution in terms of the resultant outcomes, we pretty quickly wrap ourselves up in circular logic. Unfortunately, we are frequently in the habit of doing this.
Is it logically possible for the outcome to be good and the execution poor?
Strangely often, people will assert that it’s not possible. I’d argue it is — there are lots of examples of bad, expensive products that have taken off. But the problem with execution is it kind of describes everything. And it’s pretty easy to find something you can simultaneously describe as good and describe as execution.
It is logically possible for the outcome to be poor and the execution good?
I’d argue this happens all the time. But I meet people frequently who will argue this isn’t possible. But again, because execution describes essentially everything, it becomes trivially easy to rattle off a list of potential improvements to execution.
Closing thoughts
In any event, I find this whole strategy-execution framework completely bizarre and unhelpful. I do, however, encounter a puzzling number of people who repeat execution is everything, often using the kind of circular logic I’ve described above. I just have a really hard time taking that seriously.
I’m going to wrap here and continue in a successive post. I hope to have landed a few points, but let me know if anything here is strange, confusing, or incorrect.
In the next post (or multiples), I want to pose a defense of corporate strategy for startups and then put together a corporate strategy development framework.