Revenue is not the same thing as good work.

A project can be huge, exciting and completely wrong for the business. It shows up in the revenue line looking like a win, it gets mentioned in the all-hands, and the team that killed themselves to deliver it gets thanked. Then the same kind of work gets bid again the following quarter, because nobody ever established what it actually made.

Load in the labor hours that got burned rather than the hours that got quoted. Add the materials that got re-ordered because the first spec changed. Add the freight that got expedited. Add the subcontractor who came in over. Add the change orders nobody priced because the relationship mattered more that week. Add the management time that never gets coded to anything because it belongs to everyone. Now tell me what the job made.

The three places margin leaks

In most project-based businesses the leak is in one of three places, and it is usually the same place every time.

The estimate. The bid was built on a template that was accurate three years ago, with labor rates that have not moved and a contingency that gets negotiated away in the last conversation before signing.

The scope boundary. Work gets added without a change order because saying no felt expensive in the moment. Each individual concession is small. In aggregate they are the difference between a good year and a flat one.

The overhead absorption. Management time, shop time, engineering time and rework all land in a general bucket, so the jobs that consume the most attention look exactly as profitable as the jobs that run themselves.

What good looks like

• Every cost type ties back to a job, including burden and overhead
• Budget versus actual is visible while the job is still open, not after it closes
• Estimate-to-complete gets updated by people who are actually on the work
• Change orders are priced before the work happens, every time, including for good customers
• Post-mortems change the next estimate, not just the file

That last one is where most companies stop. The post-mortem happens, everyone agrees the job was harder than expected, and the estimating template stays exactly as it was. A cost report that does not feed back into pricing is a history lesson.

The uncomfortable output

Do this work honestly and you will find a category of revenue that is not worth having. Often it is a familiar customer, a signature capability, or the kind of project the company is known for. That is exactly why it survived unexamined for so long.

You do not have to fire the customer. You do have to price the work like it costs what it costs. The companies that do this stop chasing volume and start choosing it, and the difference shows up in cash long before it shows up in the income statement.

The goal is not prettier cost reports. It is knowing which work deserves more capital, and which work you should politely stop bidding.