Construction bookkeeping.
Know your margin per job, not per year.

A contractor's books can balance perfectly and still be useless. If every cost lands in one bucket called "materials" and one called "labor," your P&L tells you the company made money last quarter. It will never tell you that the school gym job carried the whole quarter while two residential builds quietly lost twelve thousand dollars each.

That's the difference between recording history and running a business. Contractors don't have one business, they have a portfolio of small ones, and each job needs its own scoreboard.

Then there's the timing problem. Construction revenue rarely matches construction cash. You bill on a schedule, costs land when they land, retainage sits with the customer for months after the work is finished. A cash-basis view of a percentage-of-completion business is a distorted picture, and it's the reason so many contractors feel busy and broke at the same time.

Under-billing is the version nobody notices. You've done $180,000 of work and billed $140,000, so you're effectively lending the customer $40,000 at zero interest while paying suppliers on thirty-day terms. A WIP schedule surfaces that in a fortnight. Without one, you find out when payroll is tight.

Job costing starts with cost codes — a standard list of categories, usually built around something like a simplified CSI MasterFormat structure, that every cost on every job gets tagged against: labor, material, equipment, subcontractor, within a division of work like sitework, concrete, framing or electrical. The same code list is reused across every job, which is what makes job-to-job comparison possible at all. Without it, "concrete" on one job and "foundation" on another are really the same cost hiding under two names, and nothing ever rolls up cleanly.

Committed cost is the piece most bookkeeping setups miss entirely. It's the cost you've obligated yourself to — a signed subcontract, an issued purchase order — but haven't actually paid or even been billed for yet. A job can look profitable on a cash or even accrual basis right up until $40,000 in committed subcontractor cost lands, because nothing recorded that commitment when it was made. Tracking committed cost alongside actual cost is what lets you see a problem while the job is still running rather than after it's closed out.

Cost-to-complete follows directly from the first two. The basic version is arithmetic: estimated cost to complete equals your total estimated cost for the job minus costs incurred to date. The more reliable version re-estimates remaining cost based on what's actually left to do, rather than just subtracting — because a job that's 60% through its budget isn't necessarily 60% through its actual work. Committed cost belongs in this number too: money you've already obligated is money you're going to spend, whether or not it's hit the books yet.

Retainage runs in two directions and both need their own line on the balance sheet, not a blend into ordinary receivables and payables. Retainage receivable is money owed to you, held back by the owner or general contractor until substantial completion — sometimes longer, if it's tied to a warranty or punch-list period. Retainage payable is the mirror image: money you're holding back from your own subcontractors under the same logic. Both age differently from ordinary invoices and release on project milestones rather than payment terms, which is exactly why blending them into regular AR and AP makes both reports harder to read and easier to lose track of.

The cash flow distortion follows from the timing. Under percentage of completion, you recognize earned revenue as the work happens — but a real slice of every dollar billed doesn't actually arrive until the retainage releases, often months after the work that earned it. A job can be genuinely profitable on paper and still strain cash for its entire duration, because the profit is real but the cash is deliberately delayed.

Retainage percentages aren't universal — they're set by contract, and increasingly by state law. California, for example, has capped retainage on public works at 5% since around 2012 under Public Contract Code §7201, and a new law — Civil Code §8811, added by Senate Bill 61 — extends a 5% cap to many private construction contracts entered into on or after January 1, 2026. Other states set their own caps, or none at all, so what applies to your contract depends on your state and your contract type, not a single national rule. Outside of statutory caps, 5% to 10% is the range we most commonly see negotiated in private contracts — often 10% on early progress payments, stepped down to 5% at substantial completion — which is a market observation from the contracts we work from, not a surveyed industry average.

Federally funded construction work is generally subject to the Davis-Bacon Act, which requires contractors and subcontractors to pay at least the locally determined prevailing wage for each trade classification and to submit certified payroll reports — historically on federal Form WH-347 — for every week worked on the project. Prevailing wage isn't one number; it's set per trade, per county, and can be paid partly as cash wage and partly as bona fide fringe benefits such as health insurance or pension contributions, which affects how payroll should actually be coded rather than just what it totals.

Many states run their own version of Davis-Bacon — sometimes called "little Davis-Bacon" acts — extending similar prevailing wage and certified payroll requirements to state- and locally-funded public work, with their own forms and thresholds. Whether a given job is covered depends on the funding source and your state's specific rules, which is worth confirming before bidding rather than after the first payroll audit.

Misclassifying a worker's trade, or splitting fringe benefit credit incorrectly between cash and benefits, is where most certified payroll problems actually originate — not underpayment on purpose, but a classification that doesn't match what the worker actually did on site that week.

Same rule as our HVAC bookkeeping guide: what follows is labeled as sourced, calculated from a sourced figure, or market observation — and where we couldn't confirm a number against a source we'd stand behind, we've left it out rather than guessed.

Labor cost basis. The U.S. Bureau of Labor Statistics puts the median wage for Construction Laborers and Helpers at $46,680 a year in its May 2025 Occupational Employment and Wage Statistics — about $22 an hour. That's base wage only; a true loaded rate adds payroll tax, workers' comp (which runs unusually high in construction relative to most trades, and varies enormously by state and by the specific work) and any benefits, and we don't have a reliable industry-wide burden multiplier to turn that into one number, so we're not inventing one.

Retainage percentage. Covered above: California's statutory caps are sourced and specific to that state; the commonly-negotiated 5–10% range elsewhere is market observation from contracts we've actually worked from, not a cited survey.

What we won't state as fact. Numbers claiming a "typical" WIP over- or under-billing swing, a standard percentage of margin lost to unapproved change orders, or industry-average gross margin by trade circulate widely in construction accounting content online. We looked for the original, checkable studies behind those figures and didn't find them — they trace back to marketing content citing other marketing content, not a named survey with a methodology. So we're not repeating them here.

By the second month you'll have a job cost report you actually read. Not a spreadsheet somebody maintains reluctantly — a monthly view showing which jobs are tracking to estimate, which are slipping, and by how much.

By month three you'll have enough history to bid differently. Most contractors discover the same two things: one category of work is far less profitable than they assumed, and one crew or foreman consistently comes in under estimate. Both are worth knowing before the next bidding season.

You'll also have a WIP schedule you can hand to a bonding agent or a lender without a week of scrambling. That alone has widened the jobs some of our clients can go after.

And you'll stop being surprised in April. Tax planning happens during the year — equipment timing, entity structure, how you pay yourself — so the bill is something you saw coming rather than something that happened to you. This is what our monthly bookkeeping is actually built to produce, month after month, not just at year-end.