Why GCs Lose Money Self-Performing Scopes Without Realizing It

The decision to self-perform a scope usually starts with a number that looks good on paper. The sub prices are too high, or coverage is thin, or the PM says the crew has bandwidth. So the GC absorbs the work. The estimate gets a line item that reflects direct cost, maybe a modest markup, and the project moves forward. What rarely gets built into that line item is the real cost of doing the work yourself.

This is where GCs bleed money without connecting the loss to the decision that caused it.

The Estimate Captures Labor Hours. It Misses Everything Else.

Direct labor is the easy part. An estimator can price man-hours for a concrete scope, a rough carpentry package, or a selective demo package using historical production rates and current wage data. The number looks defensible. What doesn’t get captured is the loaded cost of that crew: supervision time, payroll burden, tools and equipment that aren’t sitting on a rental line, small material runs that don’t hit a PO, and the hours a foreman spends coordinating something that a sub’s PM would have handled.

At a GC with $45-to-$55-per-hour craft wages, payroll burden alone runs 28 to 35 percent on top of that. Liability, workers’ comp, and benefits add another layer. A scope priced at $80,000 in direct labor often has a true loaded cost north of $110,000 before a single supervisor hour is accounted for. Most estimates don’t model this, they use a blended crew rate that was built two years ago and hasn’t been updated since the last union negotiation.

The irony is that a sub’s bid would have included all of this automatically. Their price looked high because it was complete.

Scope Changes Are Invisible When Your Own Crew Does the Work

When a sub does the work, scope changes generate RFIs, change orders, and a paper trail. The sub protects their margin by documenting deviations. A GC’s own crew doesn’t operate that way, the foreman just does the work. If the tile scope expands by 400 square feet because field conditions required it, that cost gets absorbed into payroll and materials without anyone flagging it as a change. The field log shows the hours. The budget analysis doesn’t explain them until the job is 80 percent complete and the self-performed scope is tracking 22 percent over.

This is one of the real costs of self-performing that almost never gets discussed: the absence of a contractual boundary. With a sub, the contract defines the scope. With your own crew, the scope is whatever the foreman was told to do, plus everything that got added verbally on a Tuesday.

The Overhead Math That Quietly Stops Working

Self-performed work draws on GC overhead in ways that bid estimates don’t model. The estimating department spends time pricing the work. The PM spends time managing the crew. The superintendent spends time coordinating the self-performed scope alongside the rest of the project. Safety compliance, scheduling, quality control, punch list, and closeout documentation all require internal resources.

None of that gets billed to the project as a direct cost. It gets absorbed into general conditions or it simply doesn’t get captured at all. A GC that wins a job assuming 6 percent overhead and then self-performs three scopes is probably running closer to 9 or 10 percent by the time the final cost report is reconciled. The difference doesn’t appear on the self-performed scope line, it appears in the project’s overall margin, which is already being scrutinized for something else.

There’s a broader preconstruction discipline question embedded in this problem. What preconstruction looks like at high-performing GCs by 2030 includes sharper cost modeling on self-performed work as one of the structural changes that separates intentional preconstruction from reactive estimating.

Self-Performing Is Not the Mistake. Doing It Without a Real Cost Model Is.

Some GCs self-perform concrete, framing, or sitework as a genuine competitive advantage. They have the crew depth, the equipment, the supervision structure, and enough historical cost data to know what those scopes actually cost them. That’s a real capability. The mistake isn’t choosing to self-perform. It’s choosing to self-perform because coverage was thin, the bid looked high, or no one modeled what it would actually cost.

Before absorbing a scope, the estimate needs to reflect the loaded crew rate, not the bare wage. It needs a supervision allocation and a scope boundary that would survive a field change without silently expanding. Someone also needs to track actual cost against the self-performed line with the same discipline applied to any major sub. The ASA has documented how scope ambiguity drives cost overruns in subcontracted and self-performed work alike, and the pattern holds: undefined scope is expensive regardless of who’s doing the work.

Comms Center tracks bid coverage across every trade, which means estimators can see exactly which scopes came in with thin or no competition and flag them for closer cost review before absorbing the work. When a self-performed scope gets added to a project, the communication and follow-up around that decision stays in one place alongside every other bid conversation. Learn more at commscenter.com.