You’re viewing a shared decision brief. Run your own →

Before you decide anything, check whether you’re deciding at all.

Cassie Kozyrkov’s Decision Intelligence: a decision is only made once an irrevocable allocation of resources takes place — and half the “decisions” on your calendar fail that test. Five checks, then a framed brief you can actually act on.

Built on Cassie Kozyrkov’s “Introduction to Decision Intelligence” — ex-Google Chief Decision Scientist.

0Name the situation

One line. The thing you’ve been calling a decision.

1Decision or calculation?

In decision-analysis terms, a decision is only made once an irrevocable allocation of resources takes place. As long as you can change your mind for free, no decision has been made yet.

It moves resourcesMoney, headcount, or time gets committed and can’t come back
Nothing moves yetI can still change my mind for free

2The lamp-post test

“He uses statistics as a drunken man uses lamp-posts — for support rather than illumination.” Name one still-unknown fact that would change your choice. If there isn’t one, you’ve already decided.

3Frame before facts

Set your decision criteria before you look at the information — the way you’d decide the most you’d pay for a ticket before looking up the price. Surprise information ambushes people.

4Fact fit

Do you have the facts your framed decision needs — or are you making leaps beyond the data? Data science gets interesting exactly there. Just don’t fly too close to the sun.

I have the factsExecute the pre-framed decision. No fancy modeling needed.
Partial factsLeaps required: inference, updating opinions, or a model. Mind the Icarus splat.
No facts yetGo collect them first — or admit this is a low-stakes call.
Low stakesLunch-tier. Decide sloppily on purpose — just don’t lean on it like it’s sturdy.

5Who owns it?

A computer can execute a decision. It can’t be the decision-maker — that responsibility rests on the humans who built it. Name yours.