The five marks

1. The stakes are real and reversal is expensive. If you could cheaply undo it, you'd learn more by deciding than by deliberating. The reversible version of almost any question is an experiment, and experiments want speed, not ceremony. This mark also decays: a decision that was reversible in March stops being reversible by October, as customers, contracts, and identity pile up against undoing it. Re-run the classification; never cache it.

2. Analysis alone won't close it. If one more spreadsheet or one more customer call would produce the answer, go get it. That's an operational decision wearing a costume, and the costume comes off the moment the data arrives. On a true hard call you already know what you need to know. What's missing is honesty, not information, which is why the fifth spreadsheet feels so productive and changes nothing.

3. It's yours alone. You can't delegate it, and consensus can't make it. Everyone available to advise you is compromised somewhere: investors have a portfolio, employees have a paycheck, your spouse has a stake in your stress. Their input is data, often good data. The call is still yours, and the loneliness you feel about it isn't a bug in your support network. It's the mark itself.

4. It cuts across currencies. The business case says one thing, and the same call also pays and charges in your money, your people, your health, and your identity. Decision frameworks call that contamination and try to model it away. This system calls it the terrain: the cross-currency tension isn't noise around the decision, it usually is the decision.

5. There's tension. It looks wrong from the outside, feels wrong from the inside, or both. The board hates it, or your gut does, or the two disagree with each other. If nothing resists, nothing needed deciding.

Zero to two marks: decide now and move. At 51%, today, logged in one line. The cost of ceremony on a reversible call is speed you never get back, and speed compounds.

Three marks: the middle. Usually one honest missing fact away from resolving in either direction. Name the fact, get it by a date, decide on arrival.

Four or five: you're holding a hard call. It deserves the full machine: currencies priced, ruin line checked, criteria signed before options argue, evidence interrogated, a premortem, a memo with tripwires. Twenty to forty minutes, once, instead of six weeks of ambient dread.

The count matters more than it looks like it should, because the two failure modes it prevents are both expensive. Founders who treat everything as a hard call burn their decisiveness on inventory software and have nothing left for the cofounder conversation. Founders who treat nothing as one run the biggest decisions of their lives on the same autopilot they use for lunch. The marks are how you spend heaviness only where heaviness pays.

The most common sorting error runs the other way: an operational decision dressed up as a hard one, because dressing it up defers it, and deferring it feels like diligence. "Should we change pricing" carries weeks of dread until the marks get counted: reversible for new customers (mark one fails), closable with a test (mark two fails), delegable to data (mark three fails). Two marks at most. That's a fast-lane decision that's been renting hard-call anxiety for a month. The sort hands the month back.