The 13-Week Cash Flow Forecast: The One Report That Keeps Subcontractors Alive

The subcontractor business model has a structural cruelty at its center: you buy labor weekly and materials on 30-day terms, then sell the result on 60-to-90-day cycles with 5-10% held back until someone else's project finishes. Growth makes it worse, not better — every new job front-loads cash out before cash in, so the reward for winning work is a deeper hole. The industry-wide bill for this timing gap was an estimated $299 billion in 2025, and it lands hardest at the bottom of the chain.

That's why the P&L can say you're winning while the bank account says you're dying. Margin is an opinion updated monthly; cash is a fact updated Friday. The tool that bridges them is the rolling 13-week cash flow forecast — standard practice in turnaround work precisely because it's what creditors demand when they stop trusting the story — and the healthiest subs run it before anyone makes them.

A quarter is the natural planning horizon of this business: long enough to contain a full billing cycle (work performed → pay app → approval → payment → retainage math), long enough to see next quarter's payroll problem while options still exist, short enough that you can forecast receipts individually instead of statistically. Beyond thirteen weeks you're modeling; inside it you're scheduling actual invoices against actual pay cycles.

The rolling part is the discipline. Every week, the week that just ended drops off, a new week-13 appears, and — critically — you record what actually happened against what you forecast. That variance column is where the forecast earns its keep: it teaches you, receipt by receipt, that this GC pays in 52 days, not the 30 the contract says; that this owner funds the 25th regardless of when the pay app landed; that your 'conservative' assumptions were optimism in a spreadsheet costume.

Receipts get built bottom-up, project by project. For each job: what have we billed that's unpaid, what will we bill in each coming week, and — the actual forecast — when will each dollar arrive, based on this customer's demonstrated behavior, not the contract's promise. This is where your contract terms stop being legal abstractions: net-45 versus net-30, pay-if-paid conditions, retainage percentage and release timing, and pay-app procedure traps that add a rejection cycle are each worth a column note, because each one moves real dollars across week boundaries.

Score each receipt's confidence honestly — contracted and approved, billed but unapproved, forecast work not yet billed — and let the low-confidence rows haircut the plan. Retainage deserves its own line per project with a realistic release date (which on many jobs means 'substantial completion plus a fight'). If a receipt depends on a change order that isn't signed yet, it isn't a receipt. It's a hope with a number on it.

Disbursements are easier because you control more of them, but the forecast forces you to see their hierarchy. Payroll and payroll taxes are immovable — they anchor every week. Then insurance and bond premiums, union benefits if applicable, rent, and debt service on their calendar dates. Then the discretionary tier: supplier payments, sub-tier payments, equipment, owner draws — the tier where timing decisions actually get made.

Two warnings belong in bold on this side of the page. First, if you're in a trust fund state, project money is legally earmarked — the forecast can't treat Project A's draw as Project B's payroll without personal exposure, so build the forecast project-aware, not just company-total. Second, sub-tier and supplier payments are your own payment reputation in the making: the same 91% of GCs who grade owners on payment behavior are hearing about you from their other subs. Stretching vendors is borrowing against your own bid list.

The output that matters is one line: projected cash position, by week, thirteen weeks out. Where that line dips below your floor — payroll plus taxes plus a survival buffer — you have found a future crisis while it's still cheap to fix. Eight weeks of warning turns a shortfall into options: accelerate a billing, press a slow receivable with lien leverage, draw the line of credit before you're desperate (banks price desperation accurately), slow a discretionary disbursement, or decline the new job that would deepen the dip — the capacity math nobody runs in bid season.

The cadence is the whole trick: same day every week, thirty minutes, controller drives, owner reads. Log last week's actuals, roll the window, update the receipt dates that slipped, and flag any week that breaches the floor. A forecast maintained sporadically is worse than none — it produces confidence without accuracy. Maintained weekly, it quietly rewires the company: estimating starts asking about payment terms because the forecast made slow-pay visible; the WIP schedule and the cash line start explaining each other; and payroll Friday stops being an adrenaline event.

Don't wait for software. Columns: this week plus twelve. Rows, top block: each project's expected receipts, one line per expected payment, dated by demonstrated customer behavior. Middle block: disbursements — payroll (with taxes), fixed obligations on their dates, then supplier/sub-tier planned payments. Bottom: net for the week, running cash position, and your floor. Conditional-format anything below the floor red. That's the entire tool; a competent controller builds it in an afternoon from the AR aging, the payroll register, and the AP calendar.

Version two adds the learning loop: a variance tab (forecast vs. actual by receipt), per-customer average days-to-pay computed from your own ledger, and retainage tracked as its own aging. Version three connects it upstream — every new contract's payment terms get entered the week it's signed, which means someone has actually read them; the automated first pass that flags a pay-if-paid clause is, among everything else, feeding your forecast's assumptions. Cash discipline and contract discipline turn out to be the same discipline, seen from two ends.

You can't manage what you can't see, and monthly financials show you cash in the rearview mirror. The 13-week forecast points the lens forward: every receipt dated by how customers actually behave, every disbursement ranked by what actually can't move, and a red cell appearing eight weeks before the crisis instead of eight hours.

It costs thirty minutes a week and a spreadsheet. It repays you with cheaper financing (arranged early, not desperately), calmer vendor relationships, bid decisions that respect working capital, and the specific serenity of a sub who knows — not hopes — that the next thirteen payroll Fridays clear. In a business where profitable companies die of timing, that's not an accounting nicety. That's the survival instrument.