Write down the kill signal
A decision is not finished when it is made. It is finished when you have written down what would prove it wrong.
Every product team makes decisions. Very few write down how they would know they were wrong.
That is the whole gap. A feature is a bet: we spend effort because we believe it will change a behaviour that is worth money. The bet only pays if someone checks. Almost nobody checks, and there are three reasons.
- The belief was never written in a checkable form. “Improve engagement” cannot be killed. Nothing can prove it wrong, so nothing can prove it right.
- The check date has no owner. By the time the data exists, the team has moved on to the next fire.
- Nobody wants to read the answer. A killed bet feels like a failure, so the readout never gets written.
What a decision record needs
I write every decision in the same shape. It fits on one screen.
| Field | The question it answers |
|---|---|
| What | What did we decide, in one sentence? |
| Why | What do we believe will happen? |
| Evidence | What settled it, and how good is that evidence? |
| Options | What else was on the table, and why not? |
| Kill signal | What number, by what date, would tell us we were wrong? |
| Dissent | Who disagreed, and what did they say? |
The kill signal is the field people skip, and it is the one that matters most.
A good kill signal
A good kill signal has a number, a date and an action:
If fewer than three of five venues use the new flow unprompted in the second week, we stop the rollout and redesign the entry point.
A bad one has none of them: “If adoption is low, we will revisit.” Low compared with what? Revisit when? That sentence was written to be forgotten.
Annotate, never delete
When a decision is superseded, I strike it through and write the replacement beside it, with the date and the reason. I do not delete it.
It feels untidy. It is the most useful habit I have. Six months later, the question is never “what did we decide?” It is “why did we change our minds?” A struck-through line answers that. A clean page does not.
Why this matters for leaders
A killed bet is a win. It stopped more effort going into the wrong place. Teams that can say “here is a bet we killed, and here is how we knew” are teams that learn faster than the ones that only ever report wins.
It is also the most honest answer in any interview room. Everyone has shipped something. Far fewer can tell you what they expected it to do, and whether it did.