Meeting Minutes vs a Decision Log: Why One Rots and the Other Is Your Institutional Memory

Two colleagues leave the same meeting. One writes: "Attendees: eight. K. presented the vendor comparison. Discussion followed on integration costs. S. raised the EU-hosting concern. Agreed to proceed with Vendor A. Meeting closed 15:40." The other writes an entry in a different document: "Decision: Vendor A over Vendor B and in-house build. Decisive argument: EU hosting requirement eliminated B; build cost two quarters. Accepted trade-off: 12-month lock-in. Owner: K. Review: contract renewal, March."

Same meeting, same decision, both documents accurate. But eleven months from now, when someone asks "why are we on Vendor A?", the first document is nearly useless — you'd have to know which meeting to look in, and the answer you'd find is "discussion followed." The second document answers the question in one paragraph, findable by searching the vendor's name.

That is the entire difference between meeting minutes and a decision log — and because the two sound interchangeable, most teams write the first believing they're getting the second. This post draws the boundary precisely: what each document is for, where minutes genuinely win, and how to run the log without adding a bureaucratic layer. (Boundaries in the other directions: the task level is covered in action items vs. decisions, and the full record-writing practice in how to document decisions.)

What Meeting Minutes Are — and Where They Fail

Minutes are the older instrument by centuries: a chronological account of a single meeting — who attended, what was discussed, what was moved, seconded and resolved. In formal governance they are not optional record-keeping but the record itself: boards, associations and public bodies are often legally required to keep them, and parliamentary procedure has precise conventions for what they contain — actions taken rather than verbatim discussion.

For their designed purpose, minutes work. The failure begins when teams use them as their only written memory — because minutes have three structural properties that make them a poor decision store:

What a Decision Log Is

A decision log inverts the organizing key: one entry per significant decision, whichever meeting, thread or hallway produced it. Each entry carries the fields that make it useful cold: the question, the options considered, the arguments that decided it, the decision, the owner, and a review date — the seven-field record covered in full in the how-to.

The model is proven at scale in engineering: Architecture Decision Records, from Michael Nygard's 2011 essay, are exactly this — one lightweight record per significant decision, kept where the team works. The insight that made ADRs spread industry-wide applies to every department: the record is organized by the decision because that is how the future asks its questions.

Two properties follow that minutes cannot have. The log is searchable by topic — "Vendor A" finds the decision in seconds. And it is reviewable — entries carry review dates, so outcomes get compared against reasoning, which is the feedback loop that makes decision quality improve rather than merely accumulate. That loop is the memory layer of decision intelligence.

Side by Side

The boundary, compressed:

When You Still Want Minutes — and When the Log Wins

This is not a case for abolishing minutes. Where the record of proceedings is itself the obligation — board meetings, member associations, works councils, anything with statutory or bylaws requirements — minutes are the correct instrument, and disputes about what was said are settled by them, not by a log. If your context requires minutes, keep writing them properly.

The point is that minutes and the log answer different questions, so for most teams the answer is both, with clear jobs: minutes where governance requires them; the log for every significant operational decision, wherever it was made. What fails is the middle path most teams actually walk — writing informal pseudo-minutes for every meeting and treating them as institutional memory. That produces documents with the obligations of neither and the benefits of neither.

A practical migration for a team drowning in unread meeting notes: keep whatever notes you like, but the moment a meeting produces a significant decision, its record goes into the log — written in the room, in the last five minutes. The notes stay as context; the log becomes the memory. If you want a meeting ritual that ends by producing that entry rather than leaving it as homework, SPADE closes on E — Explain: the decision and the reasoning written down and circulated while the room is still in it. That is the log entry, made at the only moment it is cheap to make.

"Our Minutes Already Capture the Decisions"

The strongest form of the objection: disciplined minutes DO record every resolution, ours are well-kept, and a second document means double bookkeeping — the same decision written twice, drifting apart over time. Better one imperfect store than two contradicting ones.

The double-bookkeeping worry is legitimate — and the answer is that the two documents don't duplicate, because they carry different content. The minutes record that the decision happened; the log entry carries what minutes structurally omit: the rejected options, the decisive arguments, the accepted trade-offs, the review date. Writing "see decision log, entry 47" in the minutes is one line, and the drift problem disappears because each fact lives in exactly one place — proceedings in the minutes, reasoning in the log.

And if your minutes genuinely do carry options, rationale, owner and review date per decision — then you are already keeping a decision log; it is merely filed under the wrong key. Reorganize the same content by decision instead of by date and you've adopted the practice without writing a word more.

How Argumentree Produces the Decision Log Automatically

The log's one real cost is discipline: someone must write the entry, every time, with the reasoning intact. Argumentree removes that step by making the deliberation itself structured: the question is explicit, options and their pro/con arguments form a rated tree, and when the decision lands, the entry exists already — question, options, arguments, decision, rationale, owner. The log is the deliberation's residue, not a transcription task.

For meetings specifically, meeting intelligence closes the loop from discussion to logged decision — so the answer to "why are we on Vendor A?" is always one search away, with the full argument tree attached. The fastest way to feel the difference is to run one real meeting through it — start free and log the next decision instead of minuting it.

File Decisions Under the Question, Not the Date

Documents shape what an organization can remember. Minutes remember meetings — which is sometimes legally necessary and rarely operationally sufficient. The log remembers decisions, with their reasoning, filed under the question future colleagues will actually ask.

Keep minutes where governance demands them. But the next time a meeting settles something significant, spend the last five minutes writing the log entry — and give the two colleagues from the opening one shared answer that outlives them both.