Richard Rumelt's test for whether you actually have a strategy: a real one has a diagnosis, a guiding policy, and coherent action. Most “strategies” are just goals in a nice font.
▸ Try the interactive toolRichard Rumelt's Good Strategy / Bad Strategy makes a simple, devastating claim: most of what organisations call “strategy” isn't strategy at all. It's a list of goals, a set of vague aspirations, or a deck full of buzzwords. Real strategy is rarer than it looks.
Rumelt's test is his kernel: a genuine strategy contains three things — an honest diagnosis of the challenge, a guiding policy for dealing with it, and a set of coherent actions that follow from that policy. If any of the three is missing, what you have is “bad strategy” — not because it's poorly executed, but because it isn't strategy in the first place.
Diagnosis — an honest, specific account of the challenge and what makes it hard
Guiding policy — the overall approach chosen to address the diagnosis
Coherent action — the concrete, mutually-reinforcing steps that carry out the policy
Bad strategy isn't just incomplete — it's actively harmful, because it lets a team feel aligned while pointing nowhere in particular.
| The shortcut | What it costs | What it gives you instead |
|---|---|---|
| Goals dressed as strategy | “Grow 30%, delight customers, lead the market” — targets with no method. | The kernel demands a diagnosis and a policy, not just a wish list of outcomes. |
| Fluff and buzzwords | “Synergistic, customer-centric, best-in-class” — words that survive deletion without loss. | A real diagnosis is specific enough that a competitor's diagnosis would differ. |
| Avoiding the hard choice | Strategy that says yes to everything makes no choice at all. | Guiding policy is a choice — it rules things out, which is what makes it strategy. |
| Incoherent actions | A pile of initiatives that don't reinforce each other. | Coherent action means the steps fit together and point the same way. |
State the single most important challenge and — crucially — what makes it hard. Be specific enough that the diagnosis itself is a non-obvious insight, not a restatement of the goal.
Pick the overall approach to the challenge. This is where the hard choice lives: a guiding policy rules some options out. If it rules nothing out, it isn't one.
List the concrete steps that carry out the policy — and check they reinforce one another rather than scatter. Coordination is the point.
Read it back for the four tells: fluff, failure to face the challenge, mistaking goals for strategy, and bad objectives. Any one means rewrite.
A team presented a “strategy” that read: become the category leader, achieve best-in-class retention, and delight every customer. Asked what the actual challenge was and how they'd address it, they had no answer — the document was three goals with no diagnosis or method.
Reworked through the kernel, the diagnosis became specific (new users churned before reaching value, because activation depended on a step most never completed), the guiding policy was a real choice (redesign around that one step rather than add features), and the actions all reinforced it. The same ambition, but now it was strategy.
The deliverable is three short paragraphs — diagnosis, guiding policy, coherent action — that pass the bad-strategy test.
| Part | The test it must pass | Fails when… |
|---|---|---|
| Diagnosis | Is it a specific insight about what makes this hard? | It just restates the goal |
| Guiding policy | Does it rule something out? | It's compatible with every option |
| Coherent action | Do the steps reinforce each other? | It's a scattered list of initiatives |
| Overall | Is it free of fluff and real about the challenge? | It survives any buzzword deletion unchanged |
The most dangerous strategy document is the one that feels good and commits to nothing. Bad strategy is comfortable precisely because it skips the hard diagnosis and the choice that rules options out.
Rumelt's deepest point is that the hard part of strategy is the diagnosis, not the ambition. Anyone can set a stretch goal; the work is naming the specific obstacle and choosing an approach that sacrifices other approaches. A PM who insists on a real diagnosis before any roadmap is built saves the team from confidently executing a wish.
Targets aren't strategy. Add the diagnosis and the method, or it's just a scoreboard.
If it's compatible with every action, it isn't a choice. Strategy requires sacrifice.
Jumping to actions without naming the challenge produces motion without direction.
If a sentence survives deleting its adjectives with no loss of meaning, it was fluff. Cut it.
The business-strategy rung of the Stack should pass the kernel — if it's just goals, the whole stack rests on sand.
A sharp diagnosis usually comes from Five Forces, PESTLE, and competitive analysis — they supply the specifics.
A guiding policy that creates uncontested space (Tool 11) is often stronger than one that fights head-on.
The checklist re-tests the kernel each quarter as the diagnosis ages.
Find a strategy statement — your company's, or a public one. Try to locate its diagnosis, guiding policy, and coherent action. Underline each.
If you can only find goals and adjectives, you've found bad strategy. Try writing the missing diagnosis: what's the actual challenge, and what makes it hard?
Most published “strategies” fail the test — which is exactly why the ones that pass tend to win.