We'll sort the paperwork later.

The team builds it. And the paperwork never happens, because it never does.

Here's what that costs you — and it isn't what you think.

The work that earns nothing

Remember the formula from Week 18: earned value is physical percent complete, multiplied by the budget for that work.

The plant room has no budget. It isn't in the baseline, because nobody put it there.

So when your crews build it, the model gives you nothing. Zero earned value — there's no budget line to earn against. But the money is entirely real: wages, plant, materials, all landing in your actual cost.

Spend a hundred thousand on the plant room and your cost efficiency drops from 0.79 to 0.62. You look twenty-one percent worse than last month.

You built exactly what the client asked for. The model charged you for it.

And it gets worse, because that number goes in a report. It goes to your board. It goes into the forecast. Someone senior looks at a collapsing efficiency index and concludes your site team can't build — when what actually happened is that they built something excellent that nobody wrote down.

When scope moves and the baseline doesn't, every number you produce becomes a lie about you.

That's the real reason change control exists. Not bureaucracy. Not paperwork. It's the thing that keeps your own measurements honest — and, two years from now when the delay claim lands, it's the only thing that can prove which delays were yours and which were theirs.

Two doors, and only one is optional

Changes arrive from two directions, and they are not the same animal.

The client's change is contractual. Someone outside the project wants something different, and it costs money and time that belong to them. This one is absolute: the contract is modified before the work starts on site. Not after. Not "in principle." The instruction, the price, the time impact — on paper, signed, before a single tool comes out.

Do it in the other order and you have done the work, spent the money, and kept the risk. You are now negotiating from the weakest position it is possible to occupy: the work is already built, and the only person who can pay for it has no reason left to hurry.

Your own change is different. Your engineers found a better sequence, a cheaper detail, a smarter buildability fix. That's valuable, and it should be encouraged — but it still goes through the door. It is proposed, priced, and offered to the client as a decision, not slipped into the work because it seemed obviously better.

Because a change nobody approved is not an improvement. Back in Week 10 we called it what it is: gold plating — work you'll never be paid for, on a date you'll still be held to.

Triage, or the process dies

Now the practical failure mode. Most change control systems don't collapse because they're too weak. They collapse because they're too heavy.

If moving a socket requires the same forms, the same meeting and the same three signatures as relocating a core wall, then within a month nobody uses the system for anything. Everything becomes a verbal agreement at the site gate, and you're back where you started — except now you also have a folder that makes it look like you had a process.

So sort changes by one question: how hard is this to undo?

Small and reversible — you decide, you log it, you carry on. Touches something critical but recoverable — price the ripple, tell everyone it touches, get the right sign-off. Hard or ruinous to reverse — the full process, the client's signature, before anyone digs anything.

Speed for the small stuff is what buys you rigour on the big stuff. A change process that everyone actually uses beats a perfect one that everyone routes around.

The window that closes

One more thing about timing, and it's the argument that actually works on people.

While a change is still on paper, undoing it costs an afternoon of someone's time. The same change, after the slab is poured, costs demolition, rework, delay and an argument.

The cost of changing your mind is not a straight line. It's flat, and then it's a cliff. And every project has a moment where each decision crosses that line — usually quietly, usually while everyone is busy.

This is why "we'll sort it later" is so expensive. Later is on the wrong side of the cliff. And it's why the most valuable thing a planner does with a proposed change isn't pricing it — it's saying, clearly, we need a decision by the 14th, because after that it costs ten times more.

People will make a fast decision when you show them what waiting costs. Almost nobody makes one because you asked nicely.

The baseline has one enemy: you

Finally, the temptation this whole series has been building toward.

Your project is over budget and behind. Every report is uncomfortable. And someone suggests the obvious fix: re-baseline. Set a new plan that reflects where we actually are, and start measuring from there.

It sounds so reasonable. It's the single most destructive thing you can do.

The baseline moves for one reason only: approved scope change. Not because you're late. Not because you're over. Re-baseline to escape a bad variance and you have deleted your own evidence — the variance history vanishes, the trend disappears, and next month's report starts from zero with a plan built around your own failure.

You have also, quietly, destroyed the only defence you had. Two years on, in a delay claim, the question will be: which of these delays was caused by the client's changes, and which were yours? The answer to that question lives entirely in a baseline that moved only when the scope did, with a record of every time it moved and why.

Protect that, and you can prove what happened. Lose it, and you're just two parties with different memories and expensive lawyers.

Every measurement in the last four weeks — every index, every forecast, every honest amber on a report — is measured against the baseline. It is the fixed point the entire discipline stands on. And it survives only if one person is willing to be slightly annoying about paperwork on a Friday afternoon.

Practical insight

Open your project's change log. If you don't have one, that's the finding.

Then count: how many changes have been built on site that never went through it? Take the biggest one and work out what it cost — then check whether that cost is currently sitting in your variance, being blamed on your team.

That's the number to take to your next meeting. Not "we need better change control." Something much harder to ignore: we are being measured as 21% less efficient than we are, because of work we were told to build and never got paid for.

Key takeaways

✔ Unbudgeted work earns zero value while consuming real cost — it wrecks your indices.
✔ When scope moves and the baseline doesn't, every number you report is a lie about you.
✔ Client changes are contractual: the contract moves before the work starts on site.
✔ Your own improvements still go through the door — an unapproved change is gold plating.
✔ Triage by how hard it is to undo, or the process becomes too heavy to use.
✔ The cost of reversing a decision is flat, then a cliff. Force decisions before the cliff.
✔ The baseline moves for approved scope change — and for nothing else, ever.
✔ Re-baselining to escape a bad variance deletes the only evidence you have.

What's coming next

The baseline is protected, the changes are on the record, and the numbers are honest.

So we have the machinery. Next week we step back and ask what it's actually for — because a project generates a flood of numbers, and most organisations drown in them while still missing the thing that mattered.

The control philosophy: which signals to watch, how to read a variance as a symptom rather than a verdict, and how to tell the difference between a project that is genuinely off course and one that is simply having a bad week.

We've built the instruments. Now we learn to fly on them.