General Contractor Profit Margins — 2026 Benchmarks by Revenue Size

The 2026 numbers

Two numbers in that table explain most GC financial misery: the 20–30% going in and the 1.4–2.4% coming out. Everything between them has a name and an invoice. (This page is the GC-wide view; for kitchen-and-bath remodeling specifically, see what is a good profit margin for a remodeling contractor.)

Where 20 points become 2: the waterfall

Markup arithmetic gives back 5 points immediately. 25% markup = 20% margin — $25 added to $100 of cost is $25 of $125. Shops that quote markup but think margin are short five points before work starts. (Calculator.)

Overhead takes 10–20 points — more than most GCs believe. The folk-wisdom “10 and 10” assumes 10% overhead; measured reality for most contractors is 25–45% of revenue at small scale once insurance, vehicles, software, estimating time, and warranty work are counted. A GC pricing with 10% while running 13.5% (the case study) donates 3.5 points of revenue on every job. Measure yours: overhead percentage calculator, and the full breakdown in overhead and profit in construction.

Change orders leak $8,000–$22,000 a year. Scope grows; paperwork doesn’t. Work performed without a signed, priced change order is margin transferred to the client — the mechanics are in change order profit leakage.

Sub risk is the GC’s signature exposure. You mark up sub costs 10–25%, but you warrant their work. One re-done tile job or schedule cascade can consume the markup on every sub for the quarter. The markup has to price the risk, not just the coordination.

The owner works for free. A GC owner doing PM + estimating is a ~$100k+ cost center priced at $0 in most books, which flatters net by several points and falsifies every bid. The fix — market salary in overhead — is the subject of how much should a contractor pay themselves.

Benchmarks by company size

• $500k–$2M (owner-operator GC): The best percentage-net opportunity in the industry — 10–15% is achievable — if the owner’s labor is priced and change orders are billed. Most underperform it for exactly those two reasons.
• $2M–$5M (first employees layer): PM and estimator salaries move from “owner’s unpaid weekend” to real overhead; nets typically compress to 5–10% during the transition. The shops that keep 10%+ are the ones that raised prices when they added structure, instead of absorbing it.
• $5M+ (professional GC): Bigger competitive jobs compress gross; structure grows; 3–8% net is the common band, and working-capital discipline starts mattering as much as margin. 2026 adds a squeeze: smaller firms are carrying just 5.8 months of backlog, so there’s less cushion to outrun a bad job.

Contract model shifts the curve too: fixed-price carries the most margin and the most risk; cost-plus 10–20% caps upside but transfers cost risk to the owner. The tradeoffs are mapped in fixed price vs cost-plus vs T&M.

The 8% floor, and getting above it

Treat 8% net as a floor, not a target — below it, one bad job, one slow quarter, or one sub failure puts the year in the red. The path from 2% to 8%+ is unglamorous and entirely measurable:

1. Quote margin, not markup — divide by (1 − margin).
2. Re-measure overhead annually (2026 insurance and wage drift guarantees yours moved) and price with the measured number.
3. Bill every change order, signed, before the work — that alone recovers $8k–$22k/year for a typical shop.
4. Put your own labor in overhead at market rate.
5. Track each job’s costs against budget while it runs — the 2% GCs discover their numbers at year-end; the 12% GCs discover them on day 10, when the overrun can still be managed.

Bottom line

GC economics are a waterfall: 20–30% enters as markup, and arithmetic confusion, unmeasured overhead, unbilled changes, sub risk, and free owner labor each take their cut until the industry average reads 1.4–2.4%. None of those cuts is mandatory. A GC who prices at margin with measured overhead, bills scope as it grows, pays themselves on the books, and watches job costs in real time doesn’t beat the average by being lucky — they beat it because the average is mostly leaks, and leaks are fixable.