The core wall moves 150mm on Tuesday.
You know straight away what it touches. The engineer is working off the old dimension. The fabricator priced the screen against it. The services layout assumed the old face. And the client approved a plan that showed the wall where it used to be.
That list is complete and it's correct. It exists in your head for about ten minutes.
You tell the engineer, because he rings you about something else that afternoon. You mean to tell the rest. Then Wednesday happens.
Three weeks later the fabricator's shop drawings come in at the old dimension.
Look at what actually went wrong there, because it wasn't the drawing. The revised plan was issued the same day it changed. It went out properly, with a revision number, to the distribution list. The information moved.
What didn't move was the list of people the change had consequences for.
Those are two different jobs, and only one of them has ever been solved. Sending a drawing to eleven people is a solved problem and has been for twenty years. Knowing which of those eleven now have to do something, and which can carry on, is a separate question. It gets answered once, by one person, in their head, in the ten minutes after they make the change.
And that answer is not admin. This is the part worth being clear about.
You knew the fabricator mattered because you know how he sequences his work and that he prices early. You knew the services layout was affected because you remembered where the duct run sits. Nobody taught you that in a process document. It's the same expertise that makes you worth hiring, applied to a question that looks clerical and isn't.
Which is also why it can't be written into a protocol. A rule that says notify affected parties is not a rule, it's a restatement of the problem. Working out who is affected is the work.
The arithmetic is what changes as a firm grows. On a four person house, a change like that has three recipients and you can hold three. On a commercial job you have four consultants, a contractor, a fabricator, a client's project manager and your own team leads. The same single change now has eleven people attached to it, and the method for working out which ones to tell is exactly what it was when there were three. Recall, then one conversation at a time.
And most of that list doesn't work for you. The engineer, the fabricator, the contractor's quantity surveyor and the client's project manager all sit in other offices, on their own systems, with their own reasons to be paying attention that week or not. No shared workspace solves this, because they were never going to be in your workspace. Whatever holds the list has to work from your side of the project.
The usual answers are real attempts and both of them fall short in the same place.
A transmittal records what was sent and to whom. That's useful, and it's a record of distribution rather than of consequence. The fabricator was on the distribution list for that revision. He received it. He didn't open it, because he'd priced the screen a fortnight earlier and had no reason to think the plan affected him.
A group chat does the opposite. Tell everyone and you've told nobody in particular. Eleven people see the message, each of them makes a reasonable private judgment about whether it concerns them, and the one who judges wrong is usually the one furthest from the conversation.
So here's the shape of the cost, and it isn't the delay.
The people you told act on it the same afternoon. The people you didn't find out later, from the artifact itself, when their own work arrives in the wrong place. A set of shop drawings at the old dimension. A consultant's layout that no longer coordinates. A client noticing in month seven that the plan isn't the one they remember approving. Every one of those is the most expensive possible moment to learn something, because by then somebody has already done the work.
A missed name on Tuesday is twenty seconds of inattention. The same missed name in week four is a fortnight of fabrication.
What a well run project does differently isn't more diligence. It's that the change and the people it reaches are recorded together, once, at the moment the change is made, while the list is still in somebody's head and still correct. Not who was emailed. Who has to act, and whether they have.
You can try this on the next change you make, and it costs nothing. Before you tell anyone, write the names down. Four names, twenty seconds. Then at the end of the week, look at the list and check which of the four have actually done something about it. The gap between the names you wrote and the people who acted is the thing this is all about, and most firms have never seen it written down in one place.
Projectsmate holds that list against the project. Not as a message sent to eleven people, most of whom will never open your software, but as your own record of who a change reaches and whether they've been told yet. The list stops being something one person remembers on Tuesday and becomes something the project carries.
A revision has been easy to send for twenty years. Knowing who it lands on is still being done from memory, once, by whoever made it.

