Nobody designed the set. It accumulated.

The rhythm is the mechanism

Week 9 ended on a planning engineer's observation about what actually survives on a project. The things that last have a clear owner and attach to a daily or weekly workflow. What gets left to evaluate separately later disappears.

Read that back and notice what it makes the meeting calendar. It is the project's weekly rhythm — the only one it has. Every decision that needs a heartbeat to stay alive gets attached to something in that calendar or it gets attached to nothing.

Which reclassifies this from a housekeeping subject to a load-bearing one. An unexamined meeting set isn't merely wasteful. It is an unreliable carrier, and everything the last six weeks asked you to produce is riding on it.

It also explains a pattern that otherwise looks like weak discipline. A project can hold a decision perfectly for six weeks and lose it in the seventh, and the seventh week is the one where the meeting that carried it was cancelled for a visit. The decision was never being held by anybody's memory. It was being held by a recurring appointment, and the appointment moved.

Every meeting has an origin

Before a set can be improved it has to be sorted, and the useful sort is by where each one came from.

Some are contractual: the contract names them, states a frequency, and sometimes says who chairs. These can't be removed and their agendas are partly fixed. Some are organisational: somebody senior wants a view of something, and the meeting exists to give them that view. Some are remedial: they were created in response to a specific failure.

The third kind is where the waste concentrates, because a remedial meeting outlives the problem it was created for. Nothing in the process closes it. The problem gets solved, the meeting stays in the diary, and two years later nobody in the room can say what it is for — which is a different thing from it being useless, and needs a different response.

Three things a meeting can produce

The second sort is by output, and there are only three.

A decision: something is settled that wasn't settled before, and the settling required the people present. A transfer: information moves from people who have it to people who need it. A record: the meeting produces a document that will be relied on later, and the meeting exists as much for the minute as for the discussion.

Only the first genuinely needs everybody in a room. A transfer can be a document, and as a document it can be read at the reader's speed, kept, and searched later. A record can be produced by two people and circulated.

The trouble is that nothing obliges a meeting to announce which of the three it is, so everybody attends everything at full strength. Naming the output on the agenda — this session decides, this session transfers — costs nothing and changes who has to be there.

The ones that move blame

There is a fourth category that doesn't appear on any list of meeting types, and it is worth being able to recognise.

Some meetings exist so that a decision can be attributed to a room. The decision has already been taken. It is presented, discussed, and adopted, and the discussion could not have changed it. What the meeting produces isn't the outcome but the attribution of the outcome to a group.

The test is one question, asked of yourself rather than out loud: was any outcome other than the announced one available in that room? Where the answer is no, you are in a meeting whose function is distributive, and treating it as a decision meeting will cost you preparation time for nothing.

This isn't always cynical. Collective adoption is how an unpopular decision survives contact with the people who have to carry it out, and a decision announced by one person is a decision one person can be talked out of. What is corrosive isn't knowing which kind of room you are in, because you argue in the wrong one and go quiet in the wrong one.

What Phase B has actually produced

Seven weeks, and each has asked for one page. It is worth laying them out together, because the pattern is only visible from the end of the phase.

A date register and the basis of the programme. The open list from the kick-off. The compliance list out of the execution plan. The access register. The identifier map. The permanence list. Six pages, none of which takes more than an afternoon, and every one of which exists because something downstream will otherwise have to be reconstructed from memory.

Every one of them decays without a cycle. The access register goes stale unless something makes you look at it monthly. The open list becomes a historical curiosity unless it appears on an agenda. The identifier map is accurate on the day it is written and drifts from that day forward. That is the whole argument of this week applied to the phase it closes: producing the record is the easy half, and attaching it to a rhythm is the half that decides whether it was worth producing.

System design

Row three is the row that closes Phase B. Every other record this phase has asked for is an artefact; this one is the index that says which of them is still alive, and it is the only entry in the whole dictionary whose subject is other entries.

Record	Produced by	Required quality	Verified against	Feeds
Meeting inventory	You, in twenty minutes	Origin and output stated for each, not just a name and a time	What each meeting actually produced last month	Who needs to attend what
Standing agenda	Whoever chairs	Names the output the session is for, so attendance can follow it	The meeting’s origin	Whether the right people are in the room
Record-to-cycle map	You, closing this phase	Every record from Phase B against the meeting that keeps it current	Whether the record is still current	Whether any of Phase B survives
Standing item for open questions	The chair, permanently	Carried forward until closed, rather than re-raised when remembered	The open list it draws from	Whether deferred items ever return
List of meetings whose absence changes nothing	The inventory itself	Written down, because the finding evaporates otherwise	One month of not holding them	Everybody’s available hours

Row four is the direct answer to week 9. A deferred item returns when a standing agenda carries it forward and stops returning when it depends on somebody remembering — which is the same distinction, in the one place where it can be built.

Practical insight

Make an inventory of your recurring meetings. It takes you twenty minutes and you will be the only person on the project holding one.

One row per meeting. Five columns: what it is called, where it came from — contractual, organisational or remedial — what it produces, who genuinely has to be there for that output, and what would happen if it didn't occur next week.

Your last column does the work. Where your honest answer is nothing, you have found either a meeting you can leave or a meeting whose real output is attribution — and both are useful things to know about your own diary before somebody else fills it.

Then do the part that matters more. Take your seven pages from this phase and write against each one which meeting keeps it alive. Access register to the weekly internal. Open list to the monthly with the client. Permanence list to nothing, because it is written once. Any page that ends up against no meeting won't survive the year, and you now know which ones those are while there is still time to attach them.

Key takeaways

• A project’s meeting calendar accumulates rather than being designed, because each meeting was added by a different person for a different reason.
• What survives on a project has an owner and attaches to a daily or weekly rhythm, and the meeting set is the only rhythm a project has.
• So an unexamined meeting set is an unreliable carrier, and everything this phase asked you to produce is riding on it.
• Meetings sort by origin into contractual, organisational and remedial.
• Remedial meetings outlive the problem that created them, because nothing in the process closes a meeting.
• They sort by output into three: a decision, a transfer of information, or a record.
• Only a decision needs everybody present; a transfer works better as a document that can be kept and searched.
• Naming the output on the agenda costs nothing and changes who has to attend.
• A fourth kind exists to attribute an already-taken decision to a room, and the test is whether any other outcome was available.
• That isn't always cynical, but not knowing which room you are in means arguing in the wrong one and going quiet in the wrong one.
• Producing a record is the easy half. Attaching it to a cycle decides whether producing it was worth anything.

Records born here. The meeting inventory, with each meeting’s origin and output · the standing agenda for each recurring meeting · the attendance each output actually requires · the mapping from every record this phase produced to the cycle that keeps it alive · the list of meetings whose absence would change nothing.

What is coming next

Phase B ends here. The project is set up, the structures exist, and the rhythm that carries them has been looked at once.

What starts now is the work itself, and it doesn't start on site. It starts in an engineering office producing documents on a list, and in a procurement department turning a line on one of those documents into an order with a delivery date. Those two run in parallel and feed each other, and between them they decide when construction can begin — because nothing gets built from a drawing that hasn't been issued, using material that hasn't arrived.