The Aristotle Distinction
The philosophical foundation of Martin’s whole approach: the world divides into what cannot be otherwise and what can, and business education teaches the wrong domain.
Cannot be otherwise: physics, chemistry, gravity. The future is identical to the past, so data from history legitimately predicts tomorrow, and the scientific method is the right tool. Can be otherwise: markets, customers, competitors, technology. The future is routinely different from the past, so historical data is not a representative sample of tomorrow, and statistical inference is formally invalid for precisely the decisions that matter most.
The error and its products
Business education's fundamental error, in Martin's view, is teaching analytical, data-based decision-making as if commerce lived in the first domain. It produces excellent analysts and poor strategists: people trained to demand proof in a domain where proof is unavailable in advance, who therefore either delay until the opportunity has passed or dress conviction in analysis and call it rigour. The confident regression on last year's market is not wrong arithmetic; it is arithmetic aimed at the wrong universe.
The consequences
Four follow directly. Strategy is imagination disciplined by logic, Strategy as Logic Not Analysis. You cannot prove a strategy in advance; you reason to the most plausible bet and test its conditions, What Would Have to Be True. Judgement is built through practice and reflection rather than acquired from frameworks, Strategy as Craft. And moonshots are legitimate acts of imagination, wrongly pursued analytically, BHAGs through Martin's lens.
How this connects to OKRs
This distinction is why our methodology treats a Key Result as a hypothesis carrying a confidence level, not a commitment carrying certainty. Weekly check-ins grade beliefs against arriving evidence because in the can-be-otherwise domain that is the only honest epistemics available. It is also the philosophical parent of outcome thinking: you cannot prove the future, but you can say what would be different in it, and then look.
One line to keep: In the can-be-otherwise world, a Key Result is a hypothesis with a confidence level, not a forecast.
Our synthesis of Roger Martin’s published work, sources credited. Read the originals: they’re excellent.
Why can't strategy be proven with data? +
Because it lives in Aristotle's “can be otherwise” domain, the part of the world that is not fixed. You can analyse what is invariant; you cannot analyse into existence a future that depends on your own choices. You reason to the most plausible bet and test its conditions.
What does this mean for Key Results? +
A Key Result is a hypothesis carrying a confidence level, not a commitment carrying certainty. Weekly check-ins grade beliefs against arriving evidence, which is the only honest epistemics available in a can-be-otherwise world.
Reading about method is not the same as running it. We install this system and build the capability that stays.
Turn strategy into executed strategy →