Organizational Memory
Decision logs: the highest-leverage document your team isn't keeping
Decision-record UI: one selected decision, three linked constraints, and a revisit trigger in the margin.
Elena Marsh
Co-founder & CEO
24 JUN, 2026
9 min
The meeting where it happens twice
Every mature team has lived this meeting. Someone proposes an idea. Someone else frowns and says, "didn't we look at this in 2024?" A third person half-remembers a spreadsheet. Twenty minutes of collective archaeology later, the group agrees to "dig into it" — which means one person will spend a week rediscovering conclusions the company already paid for.
The decision wasn't lost. The reasoning was. The minutes said "we're not doing usage-based pricing this year." Nobody wrote down the three constraints that drove it — one of which quietly expired last quarter.
What a decision log actually is
Not meeting minutes, and not a project tracker. A short structured record with five parts.
The decision, in one sentence.
Written so someone with no context understands it. "We will keep seat-based pricing through FY25" — not "pricing: keep as is."
The alternatives seriously considered.
Two or three, each with one line on why it was plausible. This proves the road not taken was seen, not missed.
The constraints that forced the choice.
The most valuable and least recorded element. Constraints expire. "Our billing system can't meter usage" is a fact with a shelf life — when it changes, the decision deserves a revisit. If you record nothing else, record the constraints.
Who decided, and how contested it was.
Not for blame — for calibration. A unanimous decision that failed teaches something different from a 4–3 decision that failed.
The revisit trigger.
One sentence: "Revisit if billing migration completes, or if two enterprise deals cite pricing as the blocker." A decision with a trigger is a living instrument. Without one it's history.
Why teams don't keep them
Three honest reasons. Writing feels like overhead when the decision is already made. The document belongs to no project and every project, so it has no natural home. And the payoff goes to a future team, months away, who may not know the log exists.
That last one is the killer. A decision log's value depends entirely on being findable at the moment of doubt — which is precisely when nobody remembers to look for it.
"A decision log that nobody can find is indistinguishable from a decision that was never explained."
Making the log findable
Mechanics that don't require any particular software: keep every log in one predictable place; give each a stable ID (`DEC-2026-014`) people can cite in threads; link the log wherever the debate happened, so the trail leads back; and put revisit triggers on an actual calendar, because a trigger nobody checks isn't one.
Where a system like Noema changes the economics: the log stops depending on someone remembering it. When a related question comes up, the record surfaces itself, reasoning intact, with a flag if any constraint has changed since. The write-up you did in March defends its decision in November without you in the room.
Start with the next decision
Don't backfill — backfilling is a graveyard project. Take the next meaningful decision your team makes and spend fifteen minutes writing the five parts. Then do it again. A quarter from now you'll have a dozen, and the meeting that happens twice starts happening once.
Elena Marsh
Co-founder & CEO
Formerly product and ops. Author of the 400-page answer doc.
One essay a month
Product
Overview
Features
Solutions
Pricing
Integrations
Noema for Slack
Company
About
Careers
Contact
Press Kit
Status
Help Center
Resources
Blog
Changelog
Documentation
Security
Legal
Privacy Policy
Terms of Service
Data Processing
Cookie Settings
© 2026 Noema Systems B.V. All rights reserved.
All systems operational
X
GitHub
RSS