The Margin Was Gone Before the Job Closed

No contractor loses a job's profit in a single visible moment. The margin bleeds out through variances too small to trigger a meeting: a crew that runs four days on a two-day phase, a material price that drifted between estimate and PO, a fix for work that failed inspection, an extra mobilization nobody priced.

Each is survivable. What kills is that no one is comparing them to the estimate while they happen. The estimate — the document that defines what "on track" even means — gets filed the day the job is won and consulted again at closeout, when the books finally catch up to reality and the post-mortem begins.

The scale of the bleed

The industry numbers on this are blunt. The Construction Industry Institute puts direct field rework at an average of about 5 percent of total project cost, with the research range running from 2 to 20 percent. PlanGrid and FMI tied more than $31 billion a year in U.S. rework to poor communication and missing project data specifically — and found project teams spending roughly 14 hours per person per week on conflict resolution, mistakes, and hunting for information. On a job bid at 10 percent margin, the average rework experience consumes half the profit before a single other variance lands.

The instinct is to treat these as field problems. They are measurement problems first. A crew cannot manage to a number nobody shows them, and a PM cannot catch a drift that only accounting will reveal in ninety days.

Why weekly tracking never happened

Estimate-versus-actual is not a new idea; it is in every construction management textbook. It failed in practice for the same reason most field data goes unused: the clerical cost. Weekly job-level tracking meant someone coding every timecard to the right phase, chasing every receipt, allocating every invoice — for every open job, every week. That is a full-time role no small contractor could fund, so the discipline was reserved for megaprojects with cost engineers.

This is exactly the labor that has now collapsed in price. Extracting costs from documents as they are photographed, coding hours against phases, aligning spend to estimate lines, and producing a variance summary is preparation work — squarely inside the boundary this series drew. What lands on the PM's desk Monday morning is one page per job: three lines drifting, two ahead, one that needs a conversation with the foreman today rather than a write-off in the fall.

What the Monday page changes

Three specific behaviors change when variance arrives weekly:

1. Drift gets a conversation instead of a closeout line. A phase running 30 percent hot in week two is a coaching moment, a re-sequencing, or a change-order signal. The same variance discovered at closeout is just a smaller check.
2. Scope creep becomes visible while it is still billable. Hours accumulating against work the estimate never included is the classic signature of an unbilled change — next week's post is entirely about that leak.
3. The estimate itself starts learning. Every closed job becomes calibration: which assemblies you consistently underestimate, which crews outperform the standard, which job types deserve a higher bid or a polite decline. That feedback loop — actuals informing the next estimate — is the learning system doing its job, and it is impossible without the weekly discipline that makes actuals trustworthy.

The four-system frame said the leaks live in the handoffs. This one lives in the handoff between operations and control: the field produces the truth daily, and the company hears it quarterly. Closing that gap does not require better builders. It requires the estimate and the actuals to finally meet each other while the job can still be steered — and for the first time, the meeting costs almost nothing to arrange.