A One-Page Decision Log Beats Memory for Team Choices. Teams re-litigate choices because nobody wrote them down. One page is enough.
The five fields worth keeping
A decision log beats memory and slide decks because it records what you chose when context was fresh. Keep five fields only:
- Decision — one sentence, present tense.
- Date — when commit happened.
- Owner — who can reopen or revoke.
- Options considered — brief, including rejected paths.
- Why — the reason string future-you needs.
Add a sixth field only if you use it: review date for experiments. Skip tiny reversible toggles. Log choices that spend calendar, money, or trust. A missing log is how history becomes opinion and teams re-litigate for sport.
The log is not bureaucracy for small teams—it is external memory when people rotate, go on leave, or join mid-project. New contributors read logs and understand why the weird workaround exists without a oral history tour that never gets scheduled.
Same-day entry ritual
Write the entry the day you decide—not next week, not when someone asks. Link the entry from the project note so search finds it. Five sparse fields today beat a beautiful template filled never.
Ritual that fits real weeks: block five minutes before shutdown on decision days. Paste the template, fill bullets, send a one-line summary to attendees. Total time under ten minutes. If logging takes longer, your template is too heavy.
Store the page in the most boring shared place your team already opens—a doc linked from the project hub, not a new tool nobody checks. Tool novelty is not the bottleneck; the twenty-four-hour writing habit is.
Pair logging with shutdown on decision days: five minutes, same chair, same template. Habits chained to existing rituals survive busy weeks better than habits that require a new app notification.
One-line summary example: Decided pilot A through May 15; Priya owns; review May 16—full log linked. Attendees who will never open the log still get the contract; people who need detail have the link.
Sending the log before reopening debate
When someone reopens a settled topic, reply with the log entry before re-arguing options. Here is what we decided on April 3, why, and who owns review. That message ends many loops faster than charisma.
Reversible decisions still deserve review dates so experiments actually end. Without review dates, “pilot” becomes permanent limbo. Owner plus review date is how pilots graduate or die.
Start from today, not from company mythology. Backfills die as archaeology projects. After twenty entries, patterns appear—usually missing owners or missing review dates. Fix the template before giving culture speeches about accountability.
Link each log entry from the project note’s decision section—not buried in comment thread 47. Future search should answer what did we decide? in one click. Buried links recreate oral history tours for every new teammate.
What good entries look like in the wild
Decision: Pilot vendor A for beta feedback through May 15. Owner: Priya. Options: A, B, build-in-house (deferred). Why: export works day one; in-house cost > six weeks. Review: May 16 go/no-go on contract.
Decision: Delay public launch one week. Owner: Marcus. Options: ship on original date with known bug, delay, cut scope. Why: signup bug affects >30% mobile users; reputational cost > schedule cost. Review: none—irreversible public commit after fix verified.
Entries should be boring enough to paste into chat. Excited prose ages badly. Facts, dates, owners age well.
Bad entry: We decided to be agile and move forward boldly. Good entry: We will ship dark-mode beta to internal users only through June 1; owner Lee; rejected company-wide launch due to contrast bugs on Android. Boring entries win arguments later.
Checklist before you close the log entry
Run this checklist in thirty seconds before saving:
- Can a stranger tell what we chose without reading the whole project thread?
- Is there exactly one owner name—not a team name?
- Did we list at least one rejected option?
- If this is an experiment, is review date on the calendar?
- Did we link the entry from the project note?
If any answer is no, fix before sending summary to attendees. Incomplete logs create second meetings to reconstruct what the first meeting already decided.
Quarterly, scan logs for missing review dates on pilots. Promote or kill experiments that exceeded review without a note. Logs without review discipline become graveyards of “still testing” forever.
When teams argue that logging takes too long, time five entries. Most honest entries land under eight minutes once the template is muscle memory. Compare eight minutes to the sixty-minute relitigation meeting caused by a missing owner name—logging is cheap insurance.
Rotate log review in monthly retro: read three random entries aloud. Ask whether each still feels true. Stale logs with wrong owners are worse than no logs—they authorize confusion. Update or strike entries that no longer match reality.
Teach new leads that reopening a decision requires new evidence, not new opinions. Point to the log first. That norm alone can reclaim hours previously lost to circular debate every sprint.
Decision logs pair naturally with reversible labels: experiments without log entries and review dates are orphans waiting to become permanent. One page, five fields, one link—keep the stack thin enough to use daily.
If your team already uses project tickets, add a decision-log link field—not a duplicate essay inside the ticket. Tickets track work; the log tracks commits. Separation keeps both readable.
Log the next real decision tonight
The next time your team commits, fill five fields before you close the laptop and link the entry from the project note. Skip logging for reversible toggles; log anything that would hurt to forget in thirty days. General workplace documentation practice—not legal recordkeeping advice for regulated industries.