Six weeks ago your team decided to stop doing the Thursday sync. Last Tuesday someone asked whether the Thursday sync should come back. Nobody said “we decided this in September” — not because they are being difficult, but because they genuinely could not remember whether it was decided, floated, or just discussed until everyone ran out of energy.
This is the most expensive thing that happens in recurring meetings, and it almost never gets named. A team can absorb a bad decision; bad decisions produce results, and results teach you something. What a team cannot absorb is a decision that never finishes — one that gets made, unmade, and remade until the subject itself becomes exhausting and people stop raising it at all.
Why do decisions come undone?
Because most of what a group calls a decision is really a moment when the conversation stopped. Someone made a strong point, nobody objected, the meeting was nearly over, and the group moved on. That feels identical, in the room, to agreement. It is not. Silence at the end of a long discussion usually means “I have run out of ways to say this” rather than “yes”.
So the decision leaves the room with no owner, no date, and — critically — no record of why it went that way rather than the other way. Six weeks later the why is gone and only the what remains. A conclusion without its reasoning is indistinguishable from an arbitrary rule, and groups reopen arbitrary rules. They should: an arbitrary rule is exactly the kind of thing that ought to be challenged.
What's the difference between revisiting and relitigating?
Revisiting is what a healthy team does when the facts change. Relitigating is reopening something because nobody can reconstruct why it was settled. They look the same from inside the meeting and they are completely different activities, which is why the distinction is worth making out loud.
The test is simple. A legitimate revisit can name what is new: a customer churned, the hire fell through, the cost doubled. A relitigation cannot — it opens with “I've been thinking about the Thursday sync again”, and it has nothing new, only renewed discomfort.
Teams that cannot tell these apart end up in one of two failure modes, both bad. Either they reopen everything, and nothing becomes settled ground to build on. Or they treat all reopening as disloyalty, and genuinely wrong decisions calcify because challenging them has become socially expensive.
What does a decision that sticks actually contain?
Four things, and the one everyone skips is the one that does the work.
- The decision itself, in a sentence, phrased as a change in behaviour rather than an intention. “We are cancelling the Thursday sync” sticks. “We should probably look at the meeting load” does not.
- Who decided, by name — not to assign blame later, but to make it possible to ask a specific person a question in six weeks instead of asking a room.
- The date. This is what turns “didn't we talk about this?” into “we decided this on 18 September”, which ends a certain kind of argument instantly.
- The reasoning — and specifically, what you traded away. This is the part that gets skipped, and skipping it is why decisions come undone.
That last one deserves a full sentence in the record: “we know this means product and support will go two weeks without talking, and we are accepting that risk because the sync was costing six people an hour a week and the last four had no agenda.”
That sentence does something a bare conclusion cannot. When the Thursday sync comes up again in November, the group does not have to reconstruct its old thinking from memory. It can read what it was worried about, look at what actually happened, and ask a much better question: did the thing we were afraid of occur? If product and support have been fine, the decision holds and the conversation takes ninety seconds. If they have drifted badly, the group has learned something specific, and the revisit is legitimate — it has new information, which is exactly the standard.
How do you know whether you've actually decided?
Say it out loud before the meeting ends, and watch what happens.
State the decision back to the room in its finished form — “so: we're cancelling Thursday, Priya's telling the team on Monday, and we'll check in four weeks whether handoffs got worse” — and then stop talking. Not “does that sound okay?”, which invites a nod. Just stop, and leave the silence there for a few seconds.
Real agreement survives that pause. Fake agreement does not: someone will fill the silence with the objection they had been carrying for ten minutes but could not find room to say. This costs about fifteen seconds, and it is the highest-return habit available to anyone running a recurring meeting — an objection surfaced now costs fifteen seconds, and the same objection surfaced in six weeks costs the whole discussion again.
One caution: this only works if the objection is actually welcome when it arrives. If the reflex response to “wait, I'm not sure” is visible impatience, people learn within two sessions that the pause is ceremonial, and you are back to collecting nods.
What about the decisions nobody makes?
Some of the most consequential decisions a team makes are never discussed at all. The hire you keep not making, the customer you keep not firing, the integration everyone has quietly concluded is a mistake but nobody has said so. These get decided by default, through drift, and because they were never decided they can never be revisited — there is nothing on record to revisit.
A useful habit for any group that meets regularly: once a quarter, write down the three biggest things you are not deciding, and for each one say explicitly whether you are deferring it — until when, and pending what — or avoiding it. The distinction matters enormously and takes about ten minutes to draw.
Deferral with a date is a legitimate decision and should be recorded as one. Avoidance is a decision too; it is just one the group has not admitted it is making, which means it gets none of the scrutiny it deserves and all of the consequences anyway.
Who should hold the record?
Someone with nothing at stake in what it says. This is harder than it sounds, because the obvious candidate is whoever runs the meeting, and that person is frequently the one whose decision is being recorded.
The problem is not dishonesty. It is that a person with a stake writes “agreed to revisit in Q1” when what actually happened was that the room could not reach agreement and ran out of time. Both descriptions are defensible. Only one of them will still be true in January, and the softer one is always the one that gets written, by people acting in complete good faith.
If a neutral human is available, use them — a rotating scribe works, and the rotation matters less than the neutrality. What matters most is that the record gets written in the meeting, while the reasoning is still in the room. Reconstructed afterwards from memory, it reliably becomes tidier and more confident than the discussion actually was, and the doubts — the most valuable part — are the first thing to evaporate.
Common questions
Isn't this a lot of process for a small team?
It is four lines per decision, and only for decisions that change what someone does on Monday. Small teams feel this pain more sharply than large ones, not less, because they have no formal record anywhere else — no project tool, no written approvals, nothing but what people remember.
What if we genuinely disagree and can't decide?
Record that. “We could not agree; Devika thinks X, Tomas thinks Y; revisiting on 4 December with the churn numbers” is a real entry and a real decision about how to proceed. It is far better than a vague consensus nobody believes, which will come apart anyway, just later and with less clarity about why.
We write decisions in our meeting notes already — why doesn't that work?
Because notes record what was said, and decisions need to be findable. If answering “what did we decide about the Thursday sync?” means scrolling three months of notes, the information exists but is functionally gone. It needs to be one list, in one place, that someone reads before the discussion rather than after it.
Doesn't recording the reasoning make people defensive about changing their minds?
The opposite, in practice. People resist reversing decisions when reversal looks like admitting the original call was stupid. A recorded trade-off reframes it: the decision was sound given what you knew, and now you know something new. That makes changing course a sign of a team paying attention rather than one that got it wrong.
How far back should the record go?
Further than feels necessary. The decisions that come undone are rarely last month's — they are the ones from long enough ago that everyone who was in the room has a slightly different memory of it, which is roughly the six-month mark for most teams.
I'm building Facilitable, a facilitator for groups that meet regularly. It keeps the decision record across sessions — what was decided, by whom, and what the group was willing to trade away — and reads it back when the same subject comes round again, so nobody has to remember. It isn't ready yet. If you'd like to see it when it is: