Have you ever had to explain yourself more than once? Or worse, revisit a decision that was already made? At Stripe, we had a simple solution that worked beautifully for this. The Decision Log. A meticulously maintained document that answered four questions for every decision:
What did we decide?
Who decided this?
Why did we decide this?
When did we decide this?
We even had a gavel emoji to stamp a final decision.
It worked magically well, even though it was tedious to maintain. Over time, multiple decision logs popped up across teams. There was no single source of truth. The entire success of the system was held together by a small number of OG Stripe engineers and PMs who deeply cared.
Last week, Lilian our designer brought up a problem: we were spending too much time tracking decisions—and sometimes re-litigating them entirely because we forgot the original decision. The fragmentation was showing up in our product and in our velocity.
Not everything is a decision, and not everything should be logged.