Your margin is decided in week one, not at invoice time

Ask a team why a project lost money and you will hear about scope creep. Scope creep is real, but it is almost never the first cause. It is the symptom you notice last.

Projects go underwater in the first week, in three small moments that nobody logs and nobody argues about at the time.

The three moments

1. The estimate that was never really agreed

Someone puts a number in a proposal. The client says “great, let’s go.” Nobody writes down what that number covered, who committed to it, or what happens if it turns out to be wrong. Six weeks later two people have completely different memories of what “the redesign” included, and both of them are being honest.

An estimate should be a record with a state, not a sentence in an email. Draft, submitted, approved, or rejected with a reason. When it comes back for a re-edit, the note travels with it. That is not bureaucracy. It is the difference between “we agreed” and “we both remember agreeing to different things.”

2. The first week nobody tracks

Kickoff week is discovery, setup, environment access, three calls, and a lot of reading. Almost nobody logs it, because it does not feel like the work yet. It is often ten to fifteen percent of the budget, spent before the first commit, and it is invisible in every report you will look at later.

You do not fix this with discipline lectures. You fix it by making logging a five-second act tied to the thing being worked on, so that the ambient cost of tracking is lower than the cost of remembering.

3. The favour

Somewhere in week one, the client asks for something small and you say yes, because saying yes is how relationships work and because it genuinely is small. The problem is not the favour. The problem is that the favour is never written down, so it never shows up against the estimate, and by month two there have been nine of them.

Why the end-of-month view is too late

The classic agency reporting rhythm is monthly. Hours come in, someone builds a sheet, and the sheet says the project is at 78% of budget with 40% of the work left. That report is accurate and useless, because the decisions that would have changed the outcome happened five weeks ago.

A margin report you read once a month is a post-mortem. A remaining balance you can see today is a steering wheel.

The fix is not more reporting. It is shortening the distance between the estimate and the hours. When logged time is attached to the estimate it was quoted against, “how much is left on this” stops being a monthly exercise and becomes a number that is simply true at all times, per person.

A one-week protocol

None of this requires a methodology. It requires four habits in the first week:

1. Get the estimate approved before work starts. Not “verbally agreed.” Approved, by a named person, with a date. If a client will not approve an estimate, that is information worth having on day one rather than day forty.
2. Log kickoff week like it is real work. Because it is. Discovery, access wrangling, and calls all belong against the project.
3. Give every favour a task. Thirty seconds. It does not have to be billed. It has to be visible, so that when there are nine of them you can point at nine of them instead of a feeling.
4. Look at remaining balance on Friday. Once a week, per developer, per estimate. Five minutes. If burn does not match progress in week one, it will not correct itself in week four.

What “in progress” hides

Most tools tell you a project is in progress. That is technically true and practically empty. It does not distinguish between work being done, work waiting on a client’s feedback, and work blocked on an approval that is sitting in someone’s inbox. Those three states have wildly different implications for margin, and lumping them together is how a project spends eleven days idle inside a green status.

We wrote more about that in why “waiting on client” deserves its own state.

Common questions

Should we track non-billable time too?

Yes, at least at the project level. Non-billable time is not free time; it is margin. Teams that only log billable hours are systematically flattering themselves.

What if the client will not sign off on an estimate?

Then start smaller. Quote and get approval for a discovery phase on its own. An unapprovable estimate usually means the scope is not understood yet by either side, and no amount of tracking fixes that.

Does this work for fixed-price projects?

It matters more for fixed price. On hourly work, overruns are partly the client’s problem. On fixed price they are entirely yours, which is exactly why the remaining balance needs to be visible weekly rather than monthly.

You cannot manage margin at the end. You can only report on it. The work is in week one.