How to Build a Strategy — The Product Playbook

A strategy isn't a mood you set in a kickoff meeting. It's a sequence. You build one the way you build anything that has to actually work, in order, where each step only makes sense once you've done the one before it.

Here's the order I use. The bones come from Richard Rumelt. In Good Strategy / Bad Strategy he calls the core of any real strategy the kernel. Strip a strategy down to what actually matters and three things are left. A diagnosis, a guiding policy, and a set of coherent actions. I teach it as five steps, not three. I pulled two moves out and gave each its own line, finding the crux and adjusting as you go, because those two are where most people fall down.

Diagnose the challenge.
Find the crux.
Write a guiding policy.
Build coherent actions.
Implement and adjust.

Five steps. The rest of this page is what each one actually asks of you.

1. Diagnose the challenge

Start here. Always. Before you set a single target, figure out what's actually going on.

Not the symptom. The cause. "Revenue is down" is a symptom. "Our two biggest customers left because onboarding drags past ninety days and they ran out of patience" is a diagnosis. Most of the complaints on your list are downstream of one or two real problems. The whole job of a diagnosis is to find those one or two and name them plainly, so the mess gets smaller instead of longer.

This is the step everyone wants to skip, because it's the uncomfortable one. It's where you have to say out loud what's really broken, and that's rarely flattering to anyone in the room.

I gave this step its own page, because doing it well is a skill on its own. When you're ready to actually run a diagnosis, go to Diagnosing the Challenge. It walks the questions I ask.

2. Find the crux

Now narrow. Out of everything you just diagnosed, name your crux. The crux is the one pivotal obstacle, the single thing that has to move. Rumelt cared about this idea enough to write a whole second book on it, called The Crux. To qualify, it has to pass three tests at the same time:

Important. Fix this one and a stack of other problems ease up or fall away on their own.
Addressable. You can actually act on it with the resources and authority you have. Not "someone, somewhere, should."
Tractable. Effort actually moves it. Some problems you're allowed to work on but still can't budge. Not those.

The crux is not the biggest problem. It's not the easiest problem. And it's definitely not the goal. It's the obstacle in the way of the goal that's worth concentrating everything on.

Here's the test for whether you've actually got one. If you can't write your crux down in a sentence or two, you don't have a strategy yet. You have a wish list.

3. Write a guiding policy

A guiding policy is the overall approach you'll take to get past the crux. It's a compass, not a task list. It commits you to a direction without naming every move. "Win on reliability, not features" is a guiding policy. So is "get everything onto one platform before we build anything new."

The best ones lean on a real strength, some place where concentrated effort pays off more than spreading it thin.

And then the part almost everyone leaves out. Name what you will deliberately not do.

Strategy is choice, and choice means refusal. If your guiding policy forbids nothing, you haven't actually chosen anything, you've just written down a hope. "We're going to focus on enterprise" only means something when it's paired with "which means we're not chasing small business this year." The "not" is where the strategy lives.

4. Build coherent actions

Now the moves. A short set of actions that all point at the crux and reinforce each other instead of fighting.

Coherence is the test here. Read your actions together and ask one question. Do any two of them pull in opposite directions? Cutting price to win volume while repositioning as the premium choice is two actions at war. If they fight, that's not a strategy. It's a pile of activity pulling itself apart.

One of those actions is your pivot. The pivot action is the one that, once it's done, makes the rest easier or possible at all. If your crux is slow onboarding, the pivot might be rebuilding the setup flow first, because faster onboarding is what makes the new pricing, the sales pitch, and the support load all start to work. Find that one and sequence it first.

This is also where strategy hands the work to execution. In this playbook that handoff has a name and a home. The Opportunity Backlog, a running board of the problems you've chosen to go explore and validate. Every item on it traces straight back to your guiding policy and the crux it serves. If an idea doesn't serve the crux, it doesn't belong on the board yet.

5. Implement and adjust

No strategy survives first contact with the real world intact, and it isn't supposed to. You implement, you watch what actually happens, and you adjust. If you've worked in agile, it's the same instinct, ship something, learn from it, change course. Your diagnosis was the best read you had at the time. New information is allowed to sharpen it, change the actions, even send you back to re-cut the crux.

The one thing you don't get to do is quietly abandon the strategy and drift back into doing a little of everything. Adjusting is deliberate. Drifting is how you end up back at the wish list.

Now go build one

That's the method. Five steps, in order, each one earning the next.

You might have noticed none of the five is "set the goal." That's on purpose. The objective, the actual target you're aiming at, is the last thing you write, not the first. It's pulled straight from the diagnosis and the crux. Once you finally know what you're up against, you set a believable near-term number. Not before.

Two things will help you run it for real:

The Strategy Template is the one-page format I use to write a strategy down, section by section, in this exact order.
The worked examples show the whole thing filled in end to end. Once on a simple problem (winning close games at the free-throw line), and once on a heavier bet (Apple deciding to build its own chips).

So if you ever feel the pull to start with the objective, the 20% number, the shiny outcome you want to announce, stop. Go back to step one. The goal is the last thing you earn, not the first thing you pick.