Changes arrive daily. The board meets monthly.

The decision is made in the paper

A board knows what it is told. Its members weren't present when the change arose, have other jobs, and read the paper shortly before the meeting.

That is the correct arrangement. A board of people who had been present would be a board of people too close to the item to weigh it, which is the thing a board exists to avoid.

So everything that will be decided has already been shaped: which options were listed, how each was described, which consequences were quantified and which were mentioned in a sentence. The person who wrote the paper made all of those choices, and they made most of them without noticing that they were choices.

None of that is manipulation, and treating it as such gets you nowhere. Writing a paper requires selecting, and selecting is deciding. It does mean that the room where the decision looks like it happens is downstream of the room where it actually did.

A board can only pick from what it was shown

Follow that further and it produces the sharpest claim of the week.

A board can approve, reject, defer, or ask for more. What it can't do is choose an option nobody wrote up. The alternative that would have been better, but which the author considered and dismissed on a Tuesday, isn't available in the room — as far as the meeting is concerned it never existed.

And a board that asks for more information gets more information about the options in front of it. The question that would surface the missing one — what else did you look at — has to come from somebody who suspects there was something else.

Which places the real authority with whoever decides what reaches the agenda and in what form. That person can be junior, isn't thinking of themselves as exercising authority, and has no delegation limit governing what they do. The formal structure from last week measures the value of the decision and says nothing at all about the framing of it.

Seen one at a time, experienced together

Each paper covers one change. Each change is assessed on its own merits, and a change reaching a board has already survived somebody deciding it was worth writing up, which is a filter of its own.

The site doesn't experience them one at a time. Twelve approved changes land in the same three months, in overlapping areas, against a programme that absorbed each one on the assumption that the others weren't happening.

Each of those assumptions was reasonable when it was made. All twelve of them together were never assessed by anybody, because there was no document in which all twelve appear.

No paper covers more than one item, so nothing in the process holds the aggregate. This is the same blindness week 24 found in splitting beneath a threshold, arriving for a different reason: there it was the shape of a limit, here it is the shape of an agenda.

And the aggregate is the thing that matters. Individually approved changes are how a programme becomes undeliverable without any single decision being wrong.

Approving isn't absorbing

A board approves scope and cost. It has no mechanism for the other half, and nothing in the paper tells it there is one.

Absorbing a change means resequencing work, revising a package boundary, re-briefing a crew, moving an inspection, possibly re-erecting a scaffold. That work is real, it lands on site, and it carries no paper because the paper stopped at the approval.

It also lands on people who had no part in the decision. The crew being re-briefed wasn't consulted, the supervisor resequencing the week wasn't in the room, and neither of them has a mechanism for saying what it cost.

So the cost recorded against a change is the cost of the change. The cost of accommodating it appears somewhere else entirely, as productivity, and it arrives without a label.

Which is why a project can approve twelve modest changes and find its rates deteriorating for reasons the cost report attributes to labour. The money went where the paper pointed. The disruption went where nothing was pointing.

The other body in the same layer

Alongside the change board runs the project's risk review, and Risk Week 16 takes that apart properly — what makes a trigger visible, how lines actually get closed, and why releasing money is a decision in its own right.

What belongs here is only the relationship between the two. One body decides about things that have happened; the other about things that haven't. They sit in the same governance layer, share members, and are connected by nothing.

Which is worth noticing because the traffic between them is real. An approved change alters the risk profile. A risk that materialises becomes a change. Two bodies, both correct, and the item crosses between them by being raised again from scratch.

The cost of that is a fortnight and a repetition, and it falls on items that are by definition the significant ones — a risk large enough to have been on a register, now real.

System design

Row three is the entry nobody expects to see in a dictionary of records, and it is the one this week argues for hardest. Closing an agenda decides which changes get decided this month, which is authority by any definition, and it is currently exercised without a name attached.

Record	Produced by	Required quality	Verified against	Feeds
Board terms of reference	The project, at start-up	States what the board may decide and what it may not	The delegation schedule above it	What has to go higher
The paper	Whoever prepares it	Lists the options considered, not only the one recommended	What was actually available	Everything the board can choose from
Agenda decision	Whoever closes the agenda	Recorded, because it is an exercise of authority nobody calls one	What was raised but held over	Which changes get decided this month
Cumulative sheet	You, after each board	Cost and programme effect per change, sorted by area	The approved papers	Whether the aggregate is deliverable
Absorption work	Site, after approval	Booked against the change rather than against productivity	What actually had to be redone	The true cost of the change

Row two asks papers to carry the options that were rejected. It costs three lines and it is the only way a board can tell whether it is choosing or ratifying — and the only way anybody later can see what was available at the time.

Practical insight

Keep the cumulative sheet yourself. One page, every approved change on it, updated the week after each board sits.

Four columns: what it was, the approved cost, the effect on your programme in days, and the area or package it lands in. Your first two come off the paper. Your second two are yours alone, and nobody else on the project is producing them.

Your fourth column pays before the third does. Sort by area and you see in one glance that six of the last nine changes land in the same part of the works — a fact about your job that no individual paper could have contained, and one that changes how you sequence.

Then take your sheet to the board once, without a request attached to it. Not to object to anything and not to ask for anything. A board that has approved twelve changes and never seen them on one page will react to seeing them, and that reaction is worth more than any argument you could make about a single item.

Key takeaways

• The site generates changes continuously and the board processes them in monthly batches, which is a rate mismatch rather than a fault.
• Meeting more often doesn't remove it, because it comes from putting a batch process in front of a continuous one.
• A board knows what it is told, and everything decided was shaped when the paper was written.
• Writing a paper requires selecting, and selecting is deciding, so the room where the decision looks like it happens is downstream of where it did.
• A board can approve, reject, defer or ask for more, and can't choose an option nobody wrote up.
• So the real authority is in who decides what reaches the agenda, and that person has no delegation limit governing it.
• Each paper covers one change and each is justified on its own merits.
• The site experiences twelve of them together, in overlapping areas, against a programme that absorbed each assuming the others were not happening.
• Nothing in the process holds the aggregate, which is how a programme becomes undeliverable without any single decision being wrong.
• A board approves scope and cost; absorbing the change is separate work with no paper, and it arrives later as productivity.
• The change board and the risk review sit in the same layer, share members, and are connected by nothing.

Records born here. The board’s terms of reference and its cycle · each paper, with the options considered and not only the one recommended · who set the agenda · the decision and its date · the cumulative sheet of approved changes, by area · the absorption work each change created, and where it was booked.

What is coming next

Every one of those approved changes moves a number, and once a month somebody assembles all the numbers into a statement of where the project will end up.

That statement starts as an estimate and becomes something else. At some point in the life of a project the forecast stops being a prediction the team makes and becomes the figure the team is measured against, and nobody announces the moment it changes.