Your Cart
Loading

“We came in about where we thought.” Check.

Margin does not usually collapse. It erodes — a few points a job, from phases that quietly ran over while others quietly ran under and made the total look fine. Put one finished job in below and see what actually happened to the margin you bid.

$

Including approved change orders.

$

What you bid it at.

What it actually cost

$

Final, including everything you forgot to code to the job.

jobs
Margin you actually made
16.6%
you bid 23.1% — 6.5 points gone
Estimated margin
23.1%
what you sold
Cost variance
$25,000
8.4% over estimate
Gross profit
$64,000
bid at $89,000

An 8% cost overrun does not cost you 8%. It comes entirely out of gross profit, which is a much smaller number — so it takes 28% of the profit on this job.

$150,000
gross profit a year, if every job runs the same way
$675,676

Where the points actually go

CauseWhat it looks like on the jobWhere to catch it
Scope creepSmall extras done to keep the client happy, never written up.The single biggest source, and the easiest to fix.Change order log
Labor overrunHours run past the estimate on one phase and nobody notices until closeout.Weekly labor vs budget
Sub buyout gapThe sub bid more than the number you carried, and you ate the difference.Buyout log at award
ReworkDoing it twice. Rarely coded anywhere, so it hides inside labor.A rework cost code
Uncoded costDump fees, fuel, small tools, the trailer — charged to overhead instead of the job.Makes the job look better and overhead look worse. Both numbers end up wrong.Job cost coding

A total that lands close to the estimate is not evidence of a good estimate. Phases cancel out. Framing runs 15% over, sitework runs 12% under because you did not need the rock excavation, and the total looks like a 1% variance. You learn nothing, and you carry both errors into the next bid. Variance is only useful at the phase level.

The point of tracking this is not the post-mortem. It is that after eight or ten jobs you stop guessing — you know which phases you habitually under-bid, and by how much, and you can fix it in the estimate instead of discovering it at closeout.