Reversible or not: the only sorting that survives a deadline

Most decision frameworks are built in calm conditions and tested in calm conditions. Then a real deadline arrives, and the two-by-two goes in a drawer. Not because it was wrong, but because using it requires more time than the situation has, and under pressure people fall back on instinct and seniority.

One distinction survives, in my experience, and it survives because it can be applied in about fifteen seconds: can we undo this, and how much would undoing it cost?

Why this one holds

It works under pressure because it needs no data you do not already have. You do not have to estimate a probability, model a scenario or agree a weighting. You have to know what it would take to go back, and that is usually a matter of contracts, notice periods, public statements and how many people you would have to tell twice.

It also allocates the scarcest thing in the room, which is not information. It is deliberation. Groups have a fixed amount of serious attention, and the failure mode of most leadership teams is spending it evenly. Reversibility tells you where to concentrate.

Slow down for the doors that lock behind you. Everything else, decide now and watch what happens.

The two failure modes

Both are common and they look nothing alike.

Treating a reversible decision as irreversible is the expensive one, and it is epidemic. A pricing experiment gets six weeks of analysis and three committee slots when it could have been run on a segment for a fortnight and read from the data. The cost is not the six weeks. The cost is that the six weeks came out of the budget for something that genuinely could not be undone.

Treating an irreversible decision as reversible is the dramatic one. Redundancies, public commitments, an acquisition, letting a senior person go badly. These come dressed as reversible because in principle you could rehire, retract, restructure. In practice you cannot restore the trust, and the reversal costs several multiples of the original decision.

Making it usable

Three practical notes, because the idea is easy and the discipline is not.

Ask about the reversal cost, not reversibility in principle. Almost everything is technically reversible. The useful question is what the reversal would cost in money, time, credibility and people, and whether you would actually pay it.

Watch for the decisions that are reversible individually and irreversible in aggregate. Any single hire is reversible. Forty hires into a culture is not. Any one discount is reversible. A pricing reputation is not. This is where the distinction most often gets misapplied, and it is worth naming explicitly when a decision is the fifth of its kind.

Put it on the agenda as a field. Not a discussion, a field: reversible, cost to reverse, who decides. Two minutes at the top of a paper. It changes how the rest of the hour is spent, which is the entire point.

What it does not do

It will not tell you which option is right. Nothing will. It sorts the decisions by how much care they deserve, which is a smaller claim than most frameworks make and considerably more useful on a Thursday afternoon when three things are due.

That is the test I would apply to any decision tool: does it still work when you are tired, late and mildly under-informed. Almost nothing passes. This does.