Rumelt's kernel of good strategy.
Rumelt’s minimum viable structure for a real strategy: a diagnosis, a guiding policy, and coherent action. Three parts, in that order, and most strategies skip the first.
From Good Strategy/Bad Strategy (2011), the kernel is deliberately spare. A diagnosis: an honest, simplifying account of what is actually going on, the challenge named, including the crux, the pivotal solvable point where effort will actually cash out. A guiding policy: the overall approach chosen to deal with the diagnosed challenge, not a goal, not a vision, an approach that rules things out. And coherent action: coordinated steps that implement the policy and reinforce each other, rather than a list of independent initiatives that merely coexist in the same document.
The order is the discipline
Everything in the kernel flows from the diagnosis, which is why skipping it is fatal and why it is the part most commonly skipped, as catalogued in Bad Strategy. Without a named challenge there is no way to judge a proposed policy, no way to improve the strategy, and no way to know when it has worked. A strategy document that opens with ambitions instead of an account of the situation is answers to unasked questions.
The kernel and the cascade
Rumelt's kernel and Martin's Strategy Choice Cascade are complementary structures, not rivals. The cascade specifies the choices a strategy must make; the kernel specifies the argument a strategy must be. A cascade without a diagnosis is a set of confident answers to questions nobody asked; a kernel without Where to Play and How to Win choices is an argument that never commits. And coherence in action is Martin's integration wearing Rumelt's clothes: in both frameworks, the parts must need each other.
How this connects to OKRs
The seam with our methodology is structural: a healthy OKR pipeline runs kernel-shaped. The diagnosis draws on strategy, the KPI Scorecard and discovery findings; the guiding policy is the How to Win; and coherent action has a measurable spine, the quarter's OKRs, with the work captured as initiatives, process commitments and experiments. OKRs written without a diagnosis are bad strategy with better formatting, and no amount of from-X-to-Y discipline rescues a quarter aimed at the wrong challenge.
Our synthesis of published thinking, Richard Rumelt, Good Strategy/Bad Strategy (2011) and The Crux (2022), sources credited. Read the originals: they’re excellent.
Reading about method is not the same as running it. We install this system and build the capability that stays.
From diagnosis to executed strategy →