How to write a decision your team can still find in three weeks
The point of recording a decision isn't bureaucracy. It's so the next person who asks "why did we do it this way?" gets an answer instead of a meeting.
Every team has a version of this conversation:
"Why are we doing it this way?" "I think we decided that a while back?" "Do you remember why?" "…no."
And so the question gets reopened, discussed for forty minutes, and resolved — usually the same way it was resolved the first time, occasionally the opposite way, which is worse, because now the team has two contradictory conclusions and no record of either.
What actually gets lost
It isn't usually the decision. People often remember what was chosen. What disappears is why — the constraint that made the alternative unworkable, the thing someone knew that nobody wrote down.
That's the part that makes a decision durable. A choice without its reasoning is just an assertion, and an assertion loses to whoever argues most confidently in the next meeting. A choice with its reasoning can be evaluated: has that constraint changed? If not, the decision stands and the conversation ends in ninety seconds.
The minimum viable record
Four fields. Any more and it won't get filled in.
The decision. One sentence, in the active voice, stating what will happen. Not "we discussed moving to monthly billing" — "We're moving to monthly billing."
The reasoning. Two or three sentences on why this rather than the alternative. Name the alternative explicitly. Most of the value of a decision record is in this field, and it's the one most often skipped.
The date. Decisions have shelf lives. A call made under last year's constraints deserves re-examination; one made last week usually doesn't. Without a date you can't tell them apart.
The people. Not for blame — so that the person with the missing context is findable. "Ask Sam, they made this call in March" is a fast path back to the reasoning when the written version is thin.
That's it. Anything beyond this — options considered, risks, rollback plans — is genuinely useful for large decisions and actively harmful as a default template, because a heavy format is a format nobody completes.
Write it where it will be found
A decision recorded in the middle of a meeting recap is filed under the meeting, not under the topic. Nobody looking for the billing decision six weeks later thinks "that would have been the Tuesday sync." They search for "billing."
So the record has to be addressable on its own terms: its own entry, its own title, findable by the subject rather than by the event that produced it. This is the single biggest practical difference between a decision log that gets used and a pile of well-intentioned meeting notes that doesn't.
Supersede, don't delete
When a decision gets reversed, the instinct is to fix the record. Resist it.
Mark the old one superseded, link to the new one, and leave the reasoning intact. A log where reversals have been tidied away implies the team has never been wrong, which is both false and useless. The genuinely valuable entries are often the ones where a decision was made for good reasons, the constraints changed, and the team changed its mind — that's the sequence a new joiner learns most from.
What this buys you
Three things, all of them boring, which is the point.
Meetings stop reopening settled questions, because "we decided this in March, here's why" is a complete answer that takes ten seconds to produce. New team members can reconstruct how the current state came to be without a scheduled brain-dump from someone senior. And when a decision genuinely does need revisiting, the conversation starts from what was previously known instead of from zero.
None of that is dramatic. It just removes a recurring tax that most teams have stopped noticing they pay.