13-Week Cash Flow Forecast for Construction Companies
It's Monday morning. The owner of a $24M civil contractor pulls up his bank balance on his phone. Looks healthy. Smiles. Heads into a 9am with his estimator.
Tuesday a $180K subcontractor draw hits.
Wednesday it's bi-weekly payroll. $310K out the door.
Thursday the $220K retainage release he was counting on gets pushed three weeks because the GC's PM is on vacation.
By Friday afternoon, the line of credit is back at the cap. Again. He's not sure why. It feels like this happens every six weeks.
It doesn't feel like a cash flow problem. It feels like bad luck that keeps repeating itself.
Here's the truth: cash doesn't lie. Your P&L might. And the 13-week rolling cash flow forecast is the single tool that would have shown him this exact week was coming. Six weeks before it arrived.
Let's dive in.
Why construction cash flow is uniquely brutal
Every business has cash flow concerns. Construction has a structural cash problem that almost no other industry shares.
Six things are working against you simultaneously:
Retainage. 5-10% of every invoice gets held back for months. Sometimes a year on slow GCs. On a $5M job, that's $250K–$500K in cash you've earned but can't touch.
Progress billing. You bill once a month, on a schedule of values, with approval cycles that can run 30-45 days before the check actually shows up.
Mobilization costs. Big jobs need cash up front for crews, equipment, materials, bonds. Well before the first pay app gets cut.
Pay-when-paid. If you're a sub, you're financing the GC. If you're a GC, the same delays from the owner cascade into your subs.
Seasonal swings. Civil and exterior work has revenue mountains and revenue valleys depending on geography. Payroll doesn't get smaller in February.
Change order lag. Work performed today might not get billed for 60 days and not get paid for another 60. Meanwhile the labor and materials are out the door.
CFMA and FMI industry data has consistently shown cash flow, not bad work, not market conditions, as the single biggest cause of contractor business failure for over a decade. More contractors get put out of business by cash than by competition.
That's not because contractors are bad at construction. It's because construction cash flow is uniquely brutal, and most contractors are running it on a spreadsheet that gets opened twice a month, or not at all.
What a 13-week rolling cash flow forecast actually is
Strip away the jargon.
A 13-week rolling cash flow forecast is a weekly-updated table that shows two things:
- Cash in. Every dollar you expect to collect, by week, for the next 13 weeks
- Cash out. Every dollar you expect to pay out, by week, for the next 13 weeks
The difference is your weekly cash position. Stacked across 13 weeks, it shows you exactly when you'll have a cash trough, and how deep it'll be. Before it arrives.
It is not a P&L. Profit and loss is about whether the business is profitable on the accrual side. Cash flow is about whether the bank account stays positive in actual dollars. A business can be profitable and bankrupt at the same time. Construction is the textbook industry for this.
It is not a budget. A budget is annual, set at the beginning of the year, and updated at most quarterly. The 13-week forecast is weekly. It updates every Monday.
It is the treasury tool every other industry has been using for 40 years. And most contractors still aren't running one.
Why 13 weeks (not 4, not 26)
Why this specific window? It's a deliberate choice. 13 weeks is exactly one quarter. Enough runway to see trouble coming, short enough to stay accurate.
- Shorter than 13 weeks (4-6 weeks) and you're reactive. You can see the cliff but you can't change direction in time.
- Longer than 13 weeks (26+ weeks) and the accuracy decays. Too many unknowns. Forecasting six months of construction cash is fortune telling. We don't do fortune telling.
13 weeks gives you about 60-75 days of meaningful warning on the back half of the window. That's enough time to chase AR, accelerate billing, defer a payment, draw on a line, push out a hire, or have an uncomfortable conversation with a customer about an unpaid change order. All actionable. None reactive.
What goes on the forecast
Here's the structural skeleton of what a construction 13-week forecast actually looks like. Specific numbers in the tool itself. Here's the structure.
Starting cash position (top row, week 0. What's in the bank Monday morning)
Cash inflows by week:
- Progress billings net of retainage (by job)
- Retainage release (modeled by expected release date)
- Change orders (only ones with high probability of collection in the window)
- T&M billings
- Deposits or mobilization payments
- Customer prepayments
- Insurance reimbursements / other one-offs
- AR collections from prior-month invoices
Cash outflows by week:
- Net payroll (the payroll dates, not the wage period)
- Payroll taxes and benefits (separately. They hit different dates)
- Subcontractor draws (by job, on actual pay date)
- Vendor AP (broken down by terms. 30, 45, 60)
- Equipment payments (leases, finance, fuel, repair)
- Rent and utilities
- Insurance (workers comp, GL, umbrella, equipment)
- Taxes (quarterly estimates, sales tax, payroll tax remittances)
- Debt service (interest and principal)
- Line of credit interest
- Owner distributions (planned)
- Other recurring overhead
Net weekly cash position (inflow minus outflow)
Cumulative cash position (starting cash + running net)
Minimum cash threshold (a horizontal line you don't want to cross. Usually two weeks of operating cash for most industries; we'd argue 3-4 weeks for construction because of the structural lumpiness)
Variance to plan (what you forecasted last week vs what actually happened. This is how the model gets sharper over time)
That's the structure. The hard part isn't the structure. The hard part is feeding it the right inputs every Monday.
The construction-specific lines most templates miss
This is where most free 13-week templates fall apart for construction. They were built for SaaS or distribution. They miss the lines that move construction cash:
Retainage release, modeled by expected release date. Most templates treat retainage as a footnote. In construction it can be 5-10% of every invoice, sitting somewhere, for months, with a specific release trigger (substantial completion, final completion, lien releases). Each chunk of retainage needs its own expected release week. Otherwise the back half of the forecast is fiction.
Change order timing. Change orders performed but unbilled. Change orders billed but disputed. Change orders approved but priced at a margin you'd never accept on a base bid. We covered the full breakdown in our change orders and profit fade post. For the forecast, the question is simple: when will this actually convert to cash? If you don't have a clean answer, don't put it in the inflows.
Pay-when-paid lag. If you're a sub on a public job, you might not see cash for 60-90 days after billing. If the GC is in dispute with the owner, you might not see it at all. Forecast realistically, not optimistically.
Seasonal labor ramp. Civil and exterior contractors carry a different payroll number in May than in February. Your forecast needs to reflect that. Don't average your annual payroll across 52 weeks and call it good. You'll be wildly off in both directions.
Mobilization cash on new jobs. Every new job that's about to mobilize has a cash chunk going out before any cash comes back in. Insurance, bonds, materials, mobilization-period labor. Add those before the first pay app, not at the same time.
Equipment financing payments by asset, not by lump sum. Equipment lines come from different vendors with different terms. Lumping them into "equipment expenses" loses precision on the weeks where two big payments overlap.
These are the lines that turn a generic forecast template into a construction forecast.
How to actually run it weekly
The forecast is worthless if it sits in a tab nobody opens. Here's the rhythm that makes it work:
Monday morning, 30 minutes:
- Update last week's actuals. What actually came in last week vs what you forecasted? What actually went out? The variance is gold. It tells you where the model needs to get sharper.
- Roll the bottom 13 weeks forward. Move the next week into "this week," and add a new week 13 at the bottom.
- Update inflows. Any AR that came in or got delayed. Any new change orders approved or billed. Any retainage that released. Any new contracts where the first pay app is now scheduled.
- Update outflows. Any new vendor commitments. Any payroll changes (new hires, raises, layoffs). Any equipment buys. Any tax payments coming due.
- Flag any week below your minimum cash threshold. If a week dips below your floor, that's the action week.
- Decide one action per flagged week. Push a vendor payment. Chase an AR. Accelerate a billing. Pull on a line. Have the uncomfortable conversation. Do something. The forecast is only as useful as the actions it triggers.
Done. 30 minutes a week. Owned by the controller, the bookkeeper, or the Fractional CFO. Depending on your setup.
What it surfaces in the first 90 days
Here's what happens when a contractor actually runs a 13-week forecast for the first time. We see the same pattern over and over:
Trapped cash gets found. Almost every contractor we've worked with has $200K–$800K stuck somewhere they hadn't seen. Slow change orders that weren't invoiced, retainage that should have been pushed for, AR that nobody was chasing. Not new cash. Cash that was always there. The forecast forces visibility, and the visibility finds it.
Patterns get exposed. The contractor who realized he was financing two jobs out of operating cash. The civil contractor whose Q4 was going to be 35% short on cash before he saw it 11 weeks out. The mechanical sub who'd been quietly underbilling on his largest project for two months. These aren't catastrophes. They're patterns that the forecast made visible in time to fix.
Distributions get sized correctly. Owners stop taking distributions based on instinct and start taking them based on what the cash position can actually support. Sometimes that means smaller distributions in the short term. It also means no surprise tax bills and no "should we pull from the line for our own paycheck" conversations.
The line of credit gets used as a tool, not a lifeline. Forecasting in advance is the difference between strategically drawing on a line to bridge a known cash gap and reactively drawing on a line because Friday's payroll is going to bounce. Same draw, two completely different stories to your banker.
The owner stops worrying about Friday. This sounds soft but it's the biggest one. When the cash picture is visible 13 weeks out, the constant low-grade anxiety about payroll, AP, and the bank balance goes away. The owner gets sharper on the strategic decisions because his head is no longer half-occupied with whether next Friday will clear.
Frequently asked questions
Can I run a 13-week forecast in Excel, or do I need software?
Excel works. Most of our clients run their forecast in Excel or Google Sheets. There's good software for it (Float, Cashflow Frog, even Jirav and a few construction-specific tools), but software doesn't replace the discipline of the Monday update. Get the discipline first. Add the software later if it earns its place.
Who owns the forecast. Me, my controller, or my Fractional CFO?
The bookkeeper or controller usually owns the mechanics. Updating the model with last week's actuals and rolling it forward. The Fractional CFO owns the interpretation. What's the model telling us and what action are we taking this week. The owner reads the output and signs off on the actions. That three-seat structure works.
How accurate should the forecast be?
Weeks 1-4: aim for ±5%. Weeks 5-8: ±15%. Weeks 9-13: ±25%. The accuracy decays as you push out. That's expected. Anyone telling you they can forecast week 11 within ±5% is lying or guessing.
What's the minimum cash threshold?
Two weeks of total operating cash is the textbook answer. We'd argue 3-4 weeks for construction because of the structural lumpiness. One bad week of payroll plus subcontractor draws plus an insurance payment can eat a month of "normal" cash. Build an honest floor, not a comfortable one.
Do I really need this if my books are healthy?
If you've never had a cash surprise that scared you, maybe not yet. If you've had even one in the last 12 months, you needed this six months ago. The cost of running it is negligible. The cost of not running it shows up in a single bad week.
Cash doesn't lie. Your P&L might.
If you've been running your business by checking your bank balance and hoping next Friday cooperates, you don't have a cash flow problem yet. You have a visibility problem. The cash flow problem comes next.
At Civil CFO, the 13-week rolling cash flow forecast is one of the first things we install during onboarding. Typically delivered in draft inside the first week. We work exclusively with $10M–$70M construction companies trying to move from 1-3% net margin to 10%+. If that's the gap you're trying to close, you'll know what to do.