What Should Labor Cost Be as a Percentage of a Job?
Owners ask for the right labor percentage hoping for a number to steer by, and the honest answer is that the number worth steering by is your own. Trade, service mix, and materials share move the 'right' percentage so much that borrowed benchmarks mostly mislead. Here's how to compute yours from the last ten jobs, and the four levers that actually move it.
JULY 17, 2026 · BY THE CLOX TEAM
Somewhere in every contractor forum, someone asks what percentage of a job labor should be, and someone else answers with a confident number. The number is different in every thread, and everyone in the thread is sure. That's not because half of them are wrong; it's because the question has no universal answer, and pretending it does is how owners end up chasing someone else's business model.
Here's why the benchmark varies so much, how to compute the only version that matters, yours, and what to do about it once you have it.
01
Why borrowed benchmarks mislead
Labor's share of a job is mostly determined by things that have nothing to do with how well you run a crew:
- Materials intensity. A flooring job where materials are half the invoice will show a low labor percentage even on a sloppy week. A demo job is nearly all labor and will show a high one on a great week. Neither percentage says anything about efficiency until you hold the mix constant.
- Service versus install. Service work bills for expertise and response, so labor dominates the invoice and the percentage runs high on purpose. That's the business working, not failing.
- Subs. Whether you self-perform or sub out concrete moves labor's share wildly without changing your profit a dollar.
- Whose labor number. Some shops count the wage; some count the burdened rate, which runs 20 to 40 percent higher. Two identical companies can report percentages ten points apart on accounting choices alone.
What a borrowed benchmark is worth
So when a benchmark says trades labor "should" land somewhere in the 20-to-40-percent-of-revenue neighborhood, treat that the way you treat a weather forecast for the whole state: not useless, but not what you build the day around.
02
Compute your own in an afternoon
The number you can steer by comes from your own closed jobs. Take the last ten, or a quarter's worth, and for each job compute:
burdened labor cost for the job, divided by the revenue for the job.
That 32.5 percent is not a target
It's one made-up job, shown only to make the division concrete. The only benchmark worth steering by is the average across your own last ten jobs, and specifically the ones that hit their margin.
Burdened labor cost is actual tagged hours times burdened rates, exactly as worked through in How to Calculate Labor Cost per Job. Then look at the jobs in two groups: the ones that hit their margin and the ones that missed. The average labor percentage of the jobs that worked is your benchmark. Not the industry's, yours, for your mix, your market, and your crew.
Now the number has teeth. A live job trending five points above your own good-job average is a specific alarm, and week-by-week labor percentages are visible mid-job if hours are being tagged as they happen, which is the whole premise of the job costing method.
03
The four levers that move it
When the percentage runs hot, the useful move is knowing which lever is loose, because each has a different fix.
Leakage. Hours getting paid that never reach an invoice: untracked drive time, yard mornings, the off-the-clock work you owe for whether you record it or not. Fix is capture: punches at the first work act, a travel code, a shop code.
Callbacks. Warranty and go-back hours booked against jobs that were already closed, or worse, smeared invisibly across open ones. Fix is a callback code and a monthly look at what it collects, as in Cost Codes for Small Crews.
Estimating drift. Bids built on the wage instead of the burdened rate, or on hope instead of history. Fix is the closeout habit: estimated hours next to actual, every job, and bids that read the file, per Bid Jobs Accurately With Labor Data.
Price. Sometimes labor percentage is high because the market price of the work moved and your rates didn't. No timesheet fixes that one, but honest job-level numbers are what give you the nerve to reprice, because you can see exactly which work no longer pays.
Notice what is not on the list: pushing the crew to hurry. At 3 to 25 people, the money is almost never in shaving effort; it's in hours that were never captured, work that was never billed, and bids that never learned.
04
Where the data comes from
Every computation above assumes one thing exists: hours tagged to jobs at the moment they happened. That's the habit Clox is built around: crews clock in from their phones, each hour lands on a job and task, overtime and breaks calculate automatically, and Friday's report already contains hours by job, ready to multiply by your burdened rates.
Clox is free for 14 days with no credit card, with a 30-day money-back guarantee, on one plan at $29 a month including the first 3 users, then $6 per user. Start a free trial, close out ten jobs with real numbers, and you'll never need to ask a forum for a benchmark again.