Growth is a cascade of choices
You sit down with a number. Revenue needs to grow 30%. Market share needs to move. The board said "double down on embedded." Something concrete enough to feel like a starting point.
What happens next is, in theory, straightforward. You map the options. You weigh them against evidence. You choose. Then you cascade down — each choice opening a new set of options, a new round of weighing, a new choice. As Roger Martin puts it in "Playing to Win": strategy is an integrated set of choices. Not a plan. Not a list of initiatives. A cascade of bets, each one narrowing the space of the next.
Map → weigh → choose → repeat one level down.
The logic is clean. The problem is in the very first move.
The first cut
Before you can weigh anything, you have to decompose. Break the big number into smaller pieces you can reason about. And the very first way you break it — the axis you choose — determines almost everything that follows.
Richard Rumelt makes a related argument in "Good Strategy Bad Strategy": diagnosis is the most neglected step. Everyone jumps to the guiding policy and coherent actions. Nobody spends enough time understanding the problem. Martin gives you the cascade. Rumelt tells you that the step before the cascade — framing what you're actually solving — is where most strategies quietly fail.
My natural instinct is to reach for audience. Who are we trying to reach, what do they believe, what behavior would we need to change? I start building segments, mapping needs, looking for the underserved pocket where a shift in perception could unlock growth.
But I keep meeting smart people who reach for something else entirely. For a while I thought this was just preference — some people think in audiences, others think in channels. Styles.
It's not. The patterns I keep seeing in rooms tend to cluster around a handful of first instincts, and each one carries an implicit theory about how growth works:
Audience: the constraint is that the right people don't see us as relevant. Growth comes from understanding a group deeply enough to change their perception. Channel: the product has proven appeal, the constraint is distribution — we need to put it in front of more people, in more places. Product: people try what we have and it's not enough — growth requires building something better. Business model: people want what we offer but the economics don't work for them — growth means restructuring how value is captured, not what value is delivered. Geography: markets differ structurally, and what works in one region won't transplant without adaptation.
Five axes. Five different conversations, initiative sets, org chart implications. And five competing hypotheses that can each be wrong in their own way.
Not all of these are even the same kind of move. Some are choosing where to look — audience, geography. They scope the problem but don't tell you what to do once you're there. Others are choosing what to change — product, pricing, channels. They jump straight to a lever. Part of what makes the first cut so hard to examine is that these feel like the same type of decision, and they're not. A room where one person says "we need to win enterprise" and another says "we need to fix onboarding" isn't having a disagreement. They're talking past each other at different altitudes.
The audience-first person builds a beautiful positioning strategy for a segment that never encounters the product. The channel-first person optimizes distribution of a message nobody cares about. The product team builds what nobody asked for while the pricing was the problem all along. I've watched a team spend two quarters repositioning for enterprise when the actual issue was that their per-seat model made the CFO's eye twitch. Same revenue target. Completely wrong axis.
Each one right about their slice and wrong about the whole.
Why the first cut sticks
You could argue this is fine — test the hypothesis, learn, adjust. But decomposition is sticky in ways that most strategic choices aren't.
Once you've sliced along an axis, the organizational machinery locks in behind it. Budgets get allocated. Teams get formed. KPIs get set. Dashboards get built. Three months later, when someone asks "wait, are we sure the problem is channel reach?" — the honest answer is "we've spent $400K assuming it is, so let's not revisit that."
This is Martin's cascade in reverse. In theory, choices flow downward: where to play → how to win → capabilities → management systems. In practice, the existing management systems — org chart, budget lines, quarterly review rhythm — flow upward and quietly determine the strategic choices.
Companies with a reseller team decompose by channel. Companies with a product org decompose by feature. Not because that's where the constraint is, but because that's where the budget lines are. This isn't strategy. It's inertia with a growth target attached.
(The tell is when someone's "strategic priority" maps perfectly onto their team's existing responsibilities. Convenient, that.)
The question before the question
So how do you choose the right first cut?
I'm increasingly convinced the answer starts somewhere uncomfortable. Not "which decomposition is correct?" — but "where is our greatest ignorance, or our tightest bottleneck?"
If you're entering a new segment and don't understand who you're talking to — start with audience. The risk of getting the "who" wrong dwarfs everything else. Optimizing channels for a segment you don't understand is like perfecting your delivery route before you know what's in the truck.
If you have a mature product in a known market and growth has plateaued — start with constraint diagnosis. Not "who should we target?" (you know), but "what's actually preventing further penetration?" Maybe it's pricing. Maybe it's the product itself. Maybe it's that your channel mix was built for a market that's already moved.
If people who encounter you convert well, but not enough encounter you — start with distribution.
If people try and leave — start with product.
If people want it but the deal structure kills the conversation — start with business model.
The principle underneath: decompose along the axis of your greatest uncertainty or your tightest constraint. Not along the axis of your org chart. Not along the axis of your comfort. Along the axis of the thing you understand least, or the thing that's most stuck.
Which raises the obvious next question. How do you identify the constraint before you've started analyzing? If diagnosis requires a lens, and the lens is what you're trying to choose — aren't we stuck in a loop?
Turns out, no. But it requires a different kind of thinking. Less strategy, more vitals.
That's Part 2.