Every Managerial Decision Is a Hypothesis

There are two ways to hold a decision, and the difference between them shapes an organization's whole relationship to evidence. You can hold a decision as a commitment: we have chosen, the matter is settled, now we execute, and to reopen the question later is to admit weakness. Or you can hold it as a hypothesis: given what we currently know, we predict that this action will produce this result under these conditions—and we are going to find out whether it does.

Managerial culture overwhelmingly favors the first. Decisiveness is celebrated; reversal is punished; leaders are praised for conviction and mocked for changing their minds. The trouble is that this framing quietly corrupts everything downstream of the decision. Once a choice becomes a commitment—and, worse, a matter of the decider's identity and standing—measurement stops being a search for the truth and becomes a search for vindication. The numbers are consulted to confirm the wisdom of the choice, not to test it, and any data that dissents is explained away.

The alternative framing is not a novelty; it has a long and respectable lineage. Deming's cycle of plan, do, study, act treated every action as a study designed to produce learning, not merely a task to complete. The broader movement from intuition toward evidence-based management rests on the same instinct. And McGrath and MacMillan's discovery-driven planning made the logic explicit for strategy under uncertainty: when you cannot know the answer in advance, you plan in order to learn, converting your riskiest assumptions into milestones that will test them as cheaply and as early as possible.

Treating a decision as a hypothesis changes concrete behavior in three ways. First, it forces you to state the prediction and its conditions in advance—"we expect this, by this much, within this time, because of this"—which makes the claim falsifiable and makes later self-deception far harder, because the goalposts are set before the outcome is known. Second, it reassigns the job of measurement: from justifying the choice to generating evidence about it. Third, and most importantly, it changes the meaning of being wrong. Under the commitment framing, a wrong decision is a personal failure to be hidden or spun. Under the hypothesis framing, a disconfirmed prediction is simply information—the experiment ran, the result came back, we know more than we did.

That last shift carries a dividend that is easy to underestimate. A great deal of organizational dishonesty about performance is downstream of the commitment framing. When a decision has become an identity, the numbers are required to agree with it, and so people manage the numbers rather than report them. Lowering the stakes of being wrong is therefore not softness; it is the precondition for getting honest data at all. An organization that can say "the hypothesis was reasonable and it turned out to be false" will learn things that an organization staking its leaders' egos on every call can never afford to see.

Honesty requires a caveat. Not every decision can or should be provisional. Some commitments are irreversible by nature, some require the full weight of conviction to succeed, and an organization that treats everything as a tentative experiment can dissolve into an inability to commit to anything—a pathology of its own. The claim here is narrower and, I think, hard to dispute: most decisions are held with far more certainty than the knowledge behind them can justify, and the gap between the two is where avoidable error lives.

The reason the idea is easier to state than to adopt is that the hypothesis framing cannot survive as a private virtue; it has to be protected by the culture around it. If a leader treats their own decision as a testable prediction while the organization continues to punish disconfirmation as failure, that leader has merely exposed themselves to no purpose. The framing works only when being wrong about a well-reasoned hypothesis is genuinely safe—when the question asked after a poor outcome is whether the reasoning was sound given what was known at the time, not whether the decider should now be blamed for the result. That safety does not arise on its own; it has to be built deliberately and defended consistently, because the natural gravity of any organization pulls relentlessly back toward vindication and blame. A single punished honest failure will teach everyone watching to stop stating falsifiable predictions and go back to defending commitments.

What is striking is how little the better framing costs. Treating a decision as a hypothesis requires no new technology, no reorganization, no budget—only the discipline to write down what you expect and why, and the humility to let reality answer. That so few organizations do it suggests the obstacle was never the idea. The obstacle is a culture that has learned to confuse confidence with competence, and to treat the willingness to be tested as a failure of nerve rather than the beginning of knowledge.