Job Costing for Contractors (2026): Budget vs Actual, by Category, Within 2%

Your price is your markup times your job cost. The markup covers overhead and profit; the job cost is everything else. Get the job cost wrong and the price is wrong, and the miss comes straight out of profit.

Profit is the thin layer on top. The NAHB's 2026 study puts the average remodeler's net at 6.3%. A 4% miss on job cost would have taken nearly half of it.

Ask one question of every dollar: would this cost exist without this job? Materials, labor, subs, the permit, the dumpster, the toilet on site: yes. Your insurance premium, the truck, the office, your pay for running the business: no. Those are overhead, and the markup pays for them.

Break out more when a category keeps missing. Setup and cleanup on an occupied house can be a day on each end. If that's where the hours go, give it its own line so you can see it.

The biggest structural miss is pricing labor at the base wage. A $30 an hour carpenter costs you $30 plus Social Security and Medicare, unemployment tax, workers' comp, paid time off and any benefits.

Loaded cost, by role

A role's loaded cost in Clamp: the wage, each burden line as a percent, and per-hour allocations for costs not already in overhead. Hours logged under the role cost this, never the base wage. Recreated from the app with sample data.

Consistency beats precision. The goal is a job cost you can trust from one estimate to the next, so the variance means something.

The habit that breaks small contracting businesses: the owner swings a hammer, runs every job and the business, and takes whatever cash is left. That hides three different things in one number.

Field hours and salary, apart

Owner Labor in Clamp: your field hours cost the job at a rate you set, and your admin time is covered by the salary line in overhead. Recreated from the app with sample data.

The IRS draws the same line for S corporations: a shareholder who works in the business must be paid reasonable compensation before distributions. A sole proprietor's draws aren't a deductible expense at all.

The budget is the estimate linked to the signed proposal. It changes for one reason: a signed change order. Allowances that run over are recovered the same way: under AIA's A201 (section 3.8), the contract sum is adjusted by change order when the cost differs from the allowance.

Budget from the signed estimate

The job costs card late in the job: budget from the estimate linked to the signed proposal, spent so far by category, labor already over with hours against the estimate's range. Owner and admins only. Recreated from the app with sample data.

Risk doesn't go in the lines. If the kitchen is a 1960s house and you don't know what's behind the walls, price it as a high-risk job and raise the profit, or carry a contingency the customer can see. Padding the tile hours to cover the unknown wrecks the one number you're trying to learn.

The Oak Street bathroom sold at $17,698.76 on an $11,799.17 estimate, a 1.50 markup. It closed at $11,972.65. The total passed the 2% test. One category didn't.

Estimated vs actual, by category

The final review in Clamp: estimated against actual by category with the variance, profit before overhead, and a prompt to write down why each category missed. Sample data.

A 1.5% total hides a 4.6% labor miss because materials came in under. The whole point of categories is to catch that. The question at close is why: did it take longer, or did the hours cost more than the rate?

Job costing pays on the next job. The hours tell you whether the estimate's labor was wrong or the crew was slow. The receipts tell you what tile really costs now.