If you learn the margin after the project closes, you learned it too late.

Most project-based companies have accurate numbers. They just have them three weeks after the only decision that mattered. The close is clean, the reports reconcile, and none of it arrives in time to change anything.

A project creates revenue in month one and reveals its economics in month five. If the reporting cadence matches the accounting calendar instead of the build calendar, management is permanently reacting to work it can no longer influence. The team gets very good at explaining variances and never gets a chance to prevent them.

Two calendars, one company

Every project business runs on two clocks. The finance clock closes monthly and looks backward. The operating clock moves daily and looks forward. Job costing sits at the intersection, and in most companies it has been quietly assigned to the finance clock because that is who owns the software.

The consequence is that the estimate-to-complete becomes a month-end accounting entry rather than a weekly operating judgment. That single distinction explains most of the surprise in project-based businesses.

A project can create revenue without creating value. The job is to see the economics while there is still time to change the outcome.

The cadence that fixes it

• A weekly job review with the project leads, not the accountants, in the room
• One number per job that must move: cost to complete, updated by the person closest to the work
• A trigger threshold, so a job that moves more than an agreed percentage escalates the same week
• A short written note on why the number moved, because the reason is the useful part

None of this requires new software. It requires someone with authority to run the meeting and hold people to the estimate they gave last week. Companies that install this cadence usually discover the same thing: the data was already there, and the problem was that nobody was accountable for interpreting it before the month ended.

What changes when it works

Jobs still go sideways. They just go sideways in week three instead of week eleven, when there is still room to re-sequence labor, renegotiate a change order, escalate to the customer, or decide to eat the loss deliberately rather than discover it.

Moving that visibility forward is usually not a systems problem. It is a discipline problem: who updates the estimate-to-complete, how often, and what happens in the meeting when the number moves.