strategyplaybook.aiSqwyz
Sign in
← All 124 tools
Choice · Diagnose, Choose

Rumelt's kernel

Three parts, one page. The fastest way to find out whether what you have is a strategy or a list of aspirations with numbers attached.

The printable canvas and the handoff into Claude Code are part of a paid plan. See what a plan includes. It needs the Claude desktop app on this machine, and your team and the prompt library already installed in that project — we cannot see your disk, so open it there.

Stage
01 Diagnose, 03 Choose
Works at
All levels
Maturity
Any
Time
90 minutes, and the first hour will be uncomfortable

What it is

Richard Rumelt's stripped-back definition of what a strategy actually is: three parts, and nothing else. Everything that is not one of the three — the vision statement, the values, the financial targets — is decoration. The kernel is diagnostic as much as constructive: run an existing plan through it and you find out very quickly whether there is a strategy inside it.

Diagnosis
One paragraph naming the core challenge. The obstacle, not the goal. It simplifies an overwhelming situation down to the thing that actually matters, and in doing so rules out other explanations.
Guiding policy
The overall approach chosen to overcome that obstacle. Not a plan and not a target — a direction of travel that makes some responses obvious and rules others out. If it rules nothing out, it is not a policy.
Coherent actions
A small set of moves that follow from the policy and reinforce one another. "Coherent" is the load-bearing word: actions that could each belong to a different strategy are the commonest failure mode in the whole discipline.
The mistake people makeRumelt's companion idea is "bad strategy", which has four tells: fluff, failure to face the challenge, mistaking goals for strategy, and bad strategic objectives. If a plan sets targets without naming an obstacle, it is a budget with adjectives.
What it’s forYou need to strip a sprawling plan back to the actual strategy: the problem, the approach, the actions.
What it’s not forYou need buy-in more than clarity. The kernel is brutally short and can read as dismissive of prior work.

How you run it

  1. Write the diagnosis first, and only the diagnosisOne paragraph naming the core challenge. Not the goal — the obstacle. Most teams cannot do this, and discovering that is the point.
  2. Test it against the symptom trapIs what you have written the problem, or a symptom of it? "Sales are down" is a symptom. Keep going until you reach something structural.
  3. Write the guiding policyThe overall approach to overcoming the obstacle. It rules things out. If it rules nothing out, it is not a policy.
  4. List the coherent actionsA handful of moves that reinforce one another. Actions that could each belong to a different strategy are the commonest failure.
  5. Read it aloud to someone outsideIf they cannot repeat the challenge back, the diagnosis is not clear. Rewrite it, do not explain it.

The prompt

Two ways to run it

Run this tool in your own Claude

The short prompt starts your partner against the library on your disk. The long one carries everything with it and needs nothing installed.

🔒 Prompt locked

Everything above is free — what this tool is, what it is for, what it is not for, and how to run it as a workshop. Three tool prompts a month are free with an account; beyond three, and for the printable canvas, it is the paid part.

See plans — from £19 a month

Your playbook

It lands in the earliest stage this tool suits. Move it on the Playbook page.

You’ll need

  • Whatever diagnosis work you have done
  • Your existing plan, if there is one
  • A tolerance for finding out it is not a strategy

You’ll end up with

  • A one-paragraph diagnosis naming the obstacle
  • A guiding policy — the approach, not the goal
  • A short set of actions coherent with each other

Runs well with

Acquisition screen and synergy thesisAI opportunity mapAI risk classification and governanceAnalysis plan