Why Contractors Bill Millions and Collect Thousands Less

The $280 Billion Problem

Rabbet's 2024 Construction Payments Report estimated that slow payments cost the U.S. construction industry $280 billion per year. That's not a typo. It's also not some abstract macro number, it reflects financing cost, higher bids, delayed work, and cash trapped after contractors have already paid for labor and materials.

In that August 2024 online survey, 82% of respondents (93% of the sample were general contractors) reported payment delays longer than 30 days, versus 49% in its 2022 report. These are respondent shares from different survey years, not an invoice-dollar share or a matched panel. CFMA's January 2025, Siteline-sponsored collections guidance separately repeats a 5% on-time-payment figure without publishing the underlying survey denominator on the page. Rabbet's report and that guidance are external sources, not Level Index metrics; do not merge their percentages into one population.

But here's what I've learned from reviewing contractor financials across PE due diligence on contractor roll-ups, building financial tooling used by 1,000+ commercial contractors at BuildOps, and now running profitability audits at Level: check two separate things: customer payment timing and whether your team reconciles its own collection rate.

What Collection Rate Actually Means

Collection rate = total cash collected / total revenue billed. Simple math. Devastating when you see the variance.

In the Level Index, collection rate is cash collected divided by billed revenue for the same company and measurement window, across 464 contractors. The bands below are read from those measured percentiles. The right-hand column is an operator checklist, not a measured cause.

Here's the precise percentile breakdown from 464 contractors:

The gap between the median and the top decile is large. Take a fictional contractor billing $10M over a year. At 85.1% it collects $8.51M; at 96.0% it collects $9.60M. The difference is $1.09M ($10M x 10.9 percentage points). That difference is not automatically recoverable cash. Some of it may be retainage not yet due, credits, disputed scope, receipts that land just outside the measurement window, or payments recorded somewhere else. Reconcile the aging report first. The collectible remainder is money you already earned that is still sitting in someone else's account.

And look at the bottom 10%: a measured ratio below 38.8%. Do not read that as work given away until you test the measurement. Check these in order:

• Did the window catch large billings without their later receipts?
• Were payments deposited or recorded outside the accounting system?
• How much is retainage not yet due?

Only after those checks is the remainder a question of true non-payment.

The Four Layers of the Collection Problem

Separate collection into four checks. And each layer has a different fix.

Layer 1: Slow Invoicing

The most fixable and most neglected.

Measure billing speed, the number of days between completing a job and sending the first invoice. Progress billing can put the first invoice before job completion, producing negative billing days. But the often-cited 1-day median is misleading, it includes those progress billers. In the Level Index billing-speed metric (n=733 companies, days from job completion to invoice), the median is 1 day when progress billers are included. Among post-completion invoicers, the adjusted median is 7 days. The slowest decile of the cohort waits 30 days or more. Level does not publish a 75th-percentile figure for this metric, so measure your own spread.

Every day you delay an invoice is a day you delay getting paid. If your payment terms are Net 30, and you invoice 10 days after completion, your effective payment cycle is 40 days. At Net 30 with same-day invoicing, it's 30 days. Fictional arithmetic: $10M of annual billings is about $27,400 billed per day ($10M / 365). Ten extra days of invoice lag therefore holds roughly $274K more in unbilled or unpaid work at any time. Shrinking the lag releases that balance once. It is not recurring profit.

The fix: Invoice the same day the job closes. If your field service software supports it, automate the invoice trigger on job completion. For commercial projects, progress-bill monthly on percentage of completion. For project work, distinguish first progress billing from the final closeout invoice. Moving the first invoice earlier does not prove that retainage or the final balance will arrive sooner.

Layer 2: Retainage

We've written a full guide on how retainage kills contractor cash flow, so I won't repeat it all here. The key numbers:

Fictional example: an $8M contractor with 60% commercial work bills $4.8M a year subject to 10% retainage, so $480K is withheld over the year. If each holdback sits about 9 months before release, the average retainage balance is about $360K ($480K x 9/12). That balance persists as long as the work mix and release timing persist. It is a working-capital level, not lost revenue, and it falls when release terms or mix change.

The strategies: forecast it explicitly, negotiate it down (5% instead of 10%, reduction at 50% completion), know your state's retainage laws, and build the financing cost into your pricing.

Layer 3: Weak AR Follow-Up

This is one place where process can separate a median collector (85.1%) from a top-decile collector (96.0%). The Level Index does not isolate follow-up cadence as the cause.

Most contractors send an invoice and wait. Maybe a reminder at 30 days. Maybe a phone call at 60 days. By 90 days, the invoice is "old" and the urgency to collect drops, which is exactly backwards.

A follow-up cadence to test (an operating playbook, not a measured top-collector practice):

• 7-day reminder, automated, friendly, just a nudge
• 21-day escalation, direct contact with the AP person, not the project manager
• 30-day flag, check the contractual due date before marking it overdue and assign any follow-up to an owner
• 45-day stop work consideration, for repeat offenders, stop scheduling new work until the balance is current, where the contract allows it. Contract terms and state prompt-payment, notice and lien deadlines override this cadence, so calendar those dates first
• Weekly AR aging review, every Monday, someone looks at the aging report and takes action

Level has not published a measured comparison of weekly versus quarterly AR review. The operating logic is simpler: a weekly review surfaces a 30-day invoice while there is still time to act on it. To test it, track the share of each month's billings collected within 30 days, before and after you start the weekly review.

Layer 4: Structural Revenue You'll Never Collect

Some uncollected revenue isn't a collections problem, it's a billing problem or a contract problem.

Underbilling: If you're consistently completing more work than you're billing (common on T&M and change order work), you have revenue that was never invoiced. It's not in your AR because it was never billed. This is invisible in your collection rate but very real in your bank account.

Disputed invoices: Some portion of outstanding AR is genuinely disputed, scope disagreements, quality issues, change orders that weren't approved before work started. These need to be resolved, not just collected. The fix is upstream: better change order documentation, clear scope agreements, and pre-approval before additional work begins.

Bad debt: Some customers won't pay. The question is how fast you identify them and stop extending credit. If you're still doing work for a customer who owes you 120+ days, you're financing their business with your cash.

The Real-World Impact

Let me walk through what this looks like for a specific contractor profile:

Fictional $12M annual-billing contractor (illustrative assumptions):

• Collection rate over the trailing 12 months: 82% (below the Level Index median of 85.1%)
• Simplifying assumption: the uncollected 18% is all still open receivables at year end, so outstanding AR = $12M x 18% = $2.16M
• DSO = $2.16M / $12M x 365, about 66 days

If they reach 92% collection (just below the Level Index 75th percentile of 92.7%):

• Outstanding AR = $12M x 8% = $960K
• One-time cash released: $2.16M minus $960K = $1.2M
• DSO = $960K / $12M x 365, about 29 days

That $1.2M is a one-time working-capital release, not new annual revenue or recurring profit. It only materializes for the part of the gap that is collectible. In a real aging report, some of the 18% will be retainage not yet due, credits, disputed change orders or true bad debt. Those dollars do not convert just because follow-up improves. The release does not require a new customer or a new job, which is why it is worth reconciling first.

Do not compare the $1.2M directly with annual net profit: profit recurs, and the release happens once. The recurring benefit is the financing cost avoided. For example, that is about $108K a year if the cash would otherwise sit on a line of credit at a fictional 9% ($1.2M x 9%). When the median contractor holds just 21.4 days of cash, every dollar stuck in AR isn't just a line item, it's a survival risk.

When the Collection Problem Isn't Really a Collection Problem

Here's the contrarian take: sometimes poor collection is a symptom, not the disease.

If you're over-collecting on some customers and under-collecting on others, you have a customer quality problem. Some customers are slow-pay by nature, government entities, certain GCs, under-capitalized developers. If you know this going in, price for it (add a financing cost to the bid). If you didn't know, you've learned.

If your collection rate is declining over time, it might be a growth problem. Rapid revenue growth with the same AR staff means invoicing gets sloppy, follow-up gets delayed, and aging balances pile up. Before hiring an AR specialist, run a payback test. Compare the fully loaded annual cost of the hire with two things: the financing cost avoided on the balance you expect to release, and any write-offs avoided.

If your collection rate is fine but cash is still tight, the problem might be overhead allocation or pricing, you're collecting what you bill, but you're not billing enough to cover your true costs.

And if a few customers explain most of the balance, start with named follow-up and resolve their specific blockers. If the gap spans many customers or recurring invoice errors, add a shared aging queue and owner. Revenue alone does not determine whether a collection system is needed. The fractional CFO conversation starts when the problem is systemic, not individual.

---

The Bottom Line

The median contractor collects about 85% of what they bill. The top 10% collect 96%+. On a fictional $10M of annual billings, the measured gap between those ratios is about $1.09M. Only the reconciled, collectible part of it is cash you can pull forward. Pulling it forward is mostly a one-time release with a smaller recurring financing benefit, achieved without adding a single new customer.

Revenue and profit reports do not substitute for a reconciliation from bills issued to cash applied. Keep a named queue for the gap: not due, retainage, disputed, credit due, unrecorded payment or truly delinquent.

Q: How does Level help improve collection rates? A: We build an AR review from available connectors or reviewed exports of QuickBooks and field service data, with a visible refresh time. Invoices are mapped by job, customer, aging bucket, and expected payment date. We flag overdue invoices, calculate your actual collection rate by customer, and identify the specific process gaps (invoicing delay, retainage, follow-up cadence) that are costing you cash. The first audit is free.

Q: What's a realistic collection rate target? A: As an illustrative planning goal, a mixed commercial-and-service contractor might test 90-92% over six months after confirming its starting cohort, measurement window, retainage and disputed balances. That is an assumed operating goal, not a measured Level outcome, a universal target or a promised result. Test progress billing, invoice triggers and weekly AR review against your own invoice cohort; Level has not isolated those practices as the cause of a 95%+ ratio. In the Level Index cohort, the top decile reached 96.0%.

Q: What tools do I need? A: QuickBooks plus your field service software (ServiceTitan, Jobber, Housecall Pro) gives you the data. What most contractors lack isn't the tool. It's a weekly billing and collections review where someone owns the number. That's the CFO function, and it's a different job than keeping the books.

Construction DSO is useful for the balance-sheet view, but the weekly owner should also track invoice timing, collection rate, and the dollars aging past each promised payment date.

Source and claim note

The $280 billion estimate and 82% slow-payment figure come from Rabbet's 2024 Construction Payments Report, whose primary PDF was checked October 7, 2026. The report describes an August 2024 online survey; its industry cost estimate is not a measured Level client loss. The 5% on-time subcontractor figure is also reported in CFMA's subcontractor collections guidance. Collection-rate percentiles and the 464-company metric definition are available in Level's downloadable contractor benchmark data, drawn from Level's original historical research universe of 2,242 contractors representing $13.25B in job revenue. That broader revenue total is not the collection metric's sample denominator. The $8M, $10M and $12M examples are fictional scenario arithmetic, not measured client outcomes. Use the DSO calculator with your own aging report before setting a collections target.

Compare your result with the Level contractor collection benchmark. If billing, retainage, and AR ownership break across systems, Level's contractor finance team can trace the cash handoff and build the weekly review.

Related reading

• Public contractor DSO benchmarks
• Why the 30-day collection rate can be more actionable than DSO
• How fast contractors should invoice