When team rituals stop scaling: which meetings to cut first

A ritual designed for nine people becomes a broadcast at twenty-two, and broadcasts are the cheapest thing to delete. Our Friday demo was the first to go.

The demo that became a broadcast

At nine people the Friday demo was forty minutes and everyone showed something. At twenty-two it was fifty minutes and four people spoke.

Attendance had doubled and airtime had not. The other eighteen watched a screen, typed nothing, and learned about work they would never touch.

Nobody complained, because a demo is pleasant and complaining about one looks petty. Pleasant is exactly why it survived eleven months past usefulness.

Rituals are sized to a room

A ritual carries two jobs: transmitting information and building a shared picture. The first scales with a document, and the second does not scale at all.

Past about a dozen people, the shared picture breaks into subgroups. A meeting that ignores this becomes a performance for an audience.

The useful question is not whether a ritual still works. It is which of its two jobs it is still doing, and whether that job now needs a room.

Cut the ones that only transmit

We listed every recurring meeting and marked what it actually moved. The categories were a decision, a relationship, or information, and the information-only meetings went first.

Three of them disappeared in a week, replaced by a written note with the same content. Nobody asked for any of them back.

The test is blunt. If nobody has ever changed their mind in a meeting, the meeting is a distribution channel with a calendar invitation.

The standup that turned into a report

Ours had been fifteen minutes at eight people. At twenty-two it took twenty-five minutes and consisted of nineteen status updates read aloud in turn.

We replaced it with a written thread and kept a ten-minute call for blockers only. Attendance dropped by half and the blockers got solved faster.

The people who stopped attending were not disengaged. They had no blockers, and the meeting had been charging them for other people’s.

Keep the ones that decide something

One ritual survived untouched, and it is the Thursday scope call. Two people argue, one decides, and it has never exceeded thirty minutes.

It survived because it produces an artefact nobody can produce alone. A decision requires disagreement in the same room at the same time.

That is the line we now use. If the output can be written by one person afterwards, the meeting is optional by definition.

Put a length limit on the demo

We did not delete the demo immediately. We cut it to twenty-five minutes with four pre-booked slots and a hard stop at the bell.

The limit did most of the work. Presenters prepared, the audience read the notes first, and the questions got sharper because time was scarce.

It still died four months later, but by then the team had decided it, not me. A ritual removed by its own participants stays removed.

Split before you shorten

Shortening a meeting that has outgrown its purpose produces a rushed version of the same problem. Splitting produces two things that each work.

The demo became three team-level showcases of fifteen minutes, each with its own audience. Total time rose and total relevance rose faster.

This is the opposite of the usual efficiency move. Sometimes the fix is more meetings, held with fewer people and a clearer job.

Someone has to own the calendar

Recurring meetings accumulate because adding one costs nothing and removing one costs a conversation. That asymmetry explains most of the calendar debt in any team.

We gave one person a monthly half hour to review the whole set. Their only power was to ask the owner a single question: what does this decide?

Half the meetings had no answer at all. Their owners usually volunteered the deletion before the question had been finished, which told us they already knew.

The real cost of a recurring hour

A weekly hour across twenty-two people is twenty-two hours, which is half a working week. We had four such meetings and had never multiplied.

The cost is not the hour. It is the fragment it creates on either side, which is why the damage exceeds the arithmetic.

Once we wrote the number on the wall, the conversation changed from preference to budget. Nobody defends a meeting against a visible price.

Delete for two weeks and watch

We now suspend a doubtful ritual for two weeks instead of debating it. Nothing is cancelled, which makes the experiment cheap to approve.

If nobody notices, it stays deleted. If two people ask for it in the same week, that is the strongest evidence a meeting can produce.

Most suspended meetings were never mentioned again by anyone. One came back smaller, at eight people, and has run weekly for a year without complaint.

What people did with the time

The honest answer is that some of it went into other meetings. Calendar space fills itself unless someone decides what the space is for.

The better answer is that three people used the reclaimed Friday hour to write. Two of those documents are still the reference for onboarding.

We did not measure this and would not trust a measurement. We only know which documents exist now that did not exist a year ago.

The ritual we should have kept

We cut one thing too far: the monthly half hour where anyone could ask the leadership team anything. It looked like transmission and was not.

It was the only place where a junior person could hear a decision being reasoned about. We restored it at twenty people with a written summary afterwards.

The lesson was not to protect all rituals or to cut all of them. It was to ask what a ritual builds before checking what it announces.

A calendar with no empty space has stopped describing the work. Which meeting this week would nobody notice missing, and which one would somebody quietly miss?