When we say bring us a decision, the reasonable next question is: which ones qualify? Some decisions become excellent systems. Others should stay exactly where they are — in a person's hands. The difference comes down to five tests.
The five tests
- 1 · It recurs. Claims are submitted every day, supplier letters arrive every quarter, campaigns are reviewed every Monday. A decision that happens once — an acquisition, a relocation — is judgment's job, not a system's.
- 2 · It matters. Real money, real time, or real risk moves on the answer. If getting it wrong costs nothing, getting it right isn't worth building for.
- 3 · The choices are real. There's more than one defensible answer, and you can name the options in advance. If there's secretly only ever one answer, that's not a decision — it's a policy. Write it down; no system needed.
- 4 · The judgment can be said out loud. Your best person can explain, even roughly, how they'd decide: what they'd look at, what would tip them one way or the other. If the reasoning can't survive being written down, it isn't ready to be built — yet.
- 5 · The outcome shows up on a clock. Within weeks or a quarter, you can see whether the call was right. That feedback is what makes the tenth cycle sharper than the first.
What this rules out
The tests are useful mostly for what they exclude. One-off bets fail the recurrence test. Decisions that are really politics — where the answer depends on who's asking — fail the judgment test, because nobody will say the real rule out loud. Decisions whose outcomes can't be observed fail the clock test; the system would never learn.
They also catch the most common imposter: the approval that nobody ever declines. That's not a decision, it's a ritual. Replace it with a rule and reclaim the meeting.
The test that matters most
If we had to keep only one, it's recurrence — because recurrence is where the compounding lives. A recurring decision gets measured, corrected, and sharpened every cycle. That loop is the whole reason to build (and the subject of How Business Judgment Becomes Reusable).
Have a decision in mind that passes three tests and wobbles on two? Bring it anyway. The mapping conversation is exactly where that gets sorted — and it costs you one conversation to find out.