Cash Flow Forecasting

Building a Rolling Cash Flow Forecast

A rolling forecast never has a fixed end date. Every week you close, you add a new week thirteen weeks out. The window always stays thirteen weeks long.

Start with your actual bank balance today. That's your only hard number. Everything after it is an estimate, and estimates need labels, not blind faith.

Each week gets three lines. Opening balance, net cash movement, closing balance. Once you build the chain, it carries forward on its own.

Cash movement splits two ways. Inflows are collections, new sales, and financing draws. Outflows cover payroll, rent, and supplier payments.

Build it in a spreadsheet. Not your head. A mental forecast feels accurate, right up until it isn't.

The Small Business Administration's cash flow management guidance treats regular tracking as a survival habit. It recommends comparing projected cash flow against actual results often. That comparison is exactly what a rolling forecast forces.

Update it every single week. A forecast built once in January and never touched again isn't a forecast. It's a fossil.

Ownership matters more than the tool. A bookkeeper can build the spreadsheet. But someone with real spending authority has to own the weekly review.

If nobody owns it, nobody updates it. A stale forecast is worse than none at all. It creates false confidence right when you need the opposite.

Small businesses often skip a rolling forecast because it looks like a big lift. It isn't. The first build takes an afternoon, and every weekly update after that takes fifteen minutes.

The trigger that gets most owners started is usually a scare, not discipline. A missed payroll date. A bounced supplier payment.

A tight month that came out of nowhere. Any one of those is reason enough to build the forecast that week.

Inputs: MRR, Churn, AR Aging, and Fixed Costs

A forecast is only as good as what feeds it. Four inputs matter most. Each one behaves differently depending on your model.

MRR, or monthly recurring revenue, is the cleanest input for subscription businesses. It's predictable, and lenders weigh it heavily. Net out expected losses before treating it as guaranteed cash.

Churn is what erodes MRR. Lose 3% of accounts a month and your base shrinks. Forecast churn against your trailing three-month average, not your best month ever.

AR aging matters most for invoice-based businesses. Not every invoice pays on time, and some never pay at all. Bucket receivables into 0-30, 31-60, 61-90, and 90-plus days.

Apply a realistic collection rate to each bucket. Don't assume 100% pays on schedule. That single assumption breaks more forecasts than any other.

Check your own history. Pull six months of invoices. See how many cleared inside 30 days, and how many drifted past 60.

Most owners are surprised by the gap. Memory runs optimistic. The paper trail almost always disagrees with it, sometimes by a wide margin.

Fixed costs are the easiest input, yet people still get them wrong. Payroll, rent, insurance, and software rarely move week to week. List each one by exact due date, not a monthly total divided by four.

Variable costs scale with revenue instead. Cost of goods, commissions, and hourly labor swing with volume. Tie them to your sales forecast, not a flat number.

Combine all four inputs and you get one weekly net figure. That figure feeds your working capital position, and by extension your runway. Watching runway shift is really the point.

Track that figure long enough and it doubles as an early warning on efficiency. See our working capital turnover ratio guide for how to turn those weekly numbers into a single benchmark.

If overstocked inventory is driving the weekly outflow line, that's a separate lever worth pulling. See how optimizing inventory management improves cash flow alongside the forecast itself.

Seasonality changes how much weight each input gets. A restaurant forecasting December leans on historical sales patterns more than AR aging. A B2B contractor forecasting the same month leans on AR aging almost entirely.

That's because most contractor revenue sits in unpaid invoices, not daily walk-in sales. Know which input drives your model before you build it.

One-time cash events deserve their own line. A tax refund or an equipment sale is real cash, but it isn't recurring. Mixing it into your regular revenue line inflates the baseline.

That inflated baseline makes next month's forecast look worse by comparison. Keep windfalls separate and label them clearly.

13-Week Cash Flow Forecast

Thirteen weeks is the standard window for a reason. It's long enough to see a slow season coming. It stays short enough to keep estimates grounded in reality.

A monthly forecast hides weekly volatility. Payroll doesn't care. A big payment can still land in the same week as a slow collection stretch.

Structure the model as a grid. One row per cash category, one column per week. Payroll, rent, debt service, and taxes each get their own row with exact due dates.

Collections and new sales get their own rows too, broken out from recurring revenue. That separation shows exactly which input drove a bad week.

Run it every Monday, before that week's payments go out. That gives you the full week to react.

Once a week closes, compare its forecast to the actual bank activity. That gap is your forecasting error. Tracking it sharpens every future estimate.

Most operators watch their error shrink fast. By week eight or nine, projected and actual numbers usually land close. Often within 5 to 10%.

A quick example makes the math concrete. A landscaping company opens week one with $22,000 in the bank. Payroll and fuel take out $14,000.

Collections bring in $16,000 that same week. The week closes at $24,000, and that number becomes next week's opening line.

Roll that math forward for thirteen weeks. Add one slow week where a big invoice runs late. The model shows exactly which week the balance gets tight.

Cash Flow Projection Template

A working template needs five sections, laid out left to right across your 13 weeks. Skip any one of them and the forecast loses its ability to explain itself.

• Opening balance — the actual bank balance carried in from the prior week's close
• Cash inflows — collections, recurring revenue, new sales converted to cash, financing draws
• Cash outflows — payroll, rent, debt service, supplier payments, taxes, one-time expenses
• Net change — inflows minus outflows for that single week
• Closing balance — opening balance plus net change, which becomes next week's opening line

Add a sixth row underneath everything: your minimum operating cash threshold. This is the balance you can't safely run below. It should sit as a flat line across every week.

Watch for the week your balance dips under that line. There it is. That's the week to plan around, not the week it happens.

Color-code it if that helps. Green above threshold, yellow within two weeks, red below it. The visual cue matters more than it sounds like it should.

Keep a separate tab for assumptions. Churn rate, collection rate by AR bucket, new-sales close rate. When a forecast misses, check assumptions first.

Date the filename. A stale file named forecast-final is a trap. Someone always opens the wrong version during a real crisis, at the worst possible moment.

Start with a spreadsheet. It works fine for most businesses under $2 million in revenue. Past that scale, dedicated software adds automated bank feeds.

The five-section structure stays identical either way. Protect the formulas from accidental overwrites. One deleted cell can silently break every balance downstream of it.

Using the Forecast to Time a Financing Decision

The single best use of a cash flow forecast isn't prevention. It's timing. A shortfall flagged eight weeks out gives you eight weeks to act, not to panic.

Financing applications take time, and that time varies by instrument. Revenue-based structures often move in 24 to 72 hours once documentation is in. Bank and SBA products can take weeks.

Match your application timing to that speed. Say your forecast shows a gap in week six, and your channel is slow. Start now, not in week five.

This is where forecasting and financing connect. Lenders underwriting revenue-based structures typically size an advance off trailing MRR or deposit history. Your forecast just tells you which week to start.

A forecast also helps you right-size the request. A projected gap of $18,000 across four weeks doesn't need a $60,000 advance sitting idle.

See our full cash flow financing resource hub once a gap shows up. It breaks options down by operator type.

Revenue-based financing is a common instrument for bridging a forecasted gap. It's sized against revenue, not personal credit or collateral.

Lead time matters even more with stacked documentation. Bank statements, a P&L, and an AR aging report all take days to assemble.

A forecast that flags the gap early gives you room to gather that paperwork calmly. Nobody wants to scramble the same week payroll is due.

The forecast also protects against the opposite mistake, applying too early. Some owners see one tight week and panic-apply for financing they don't need. That capital then sits idle, accruing cost.

Often the real gap was three weeks off and fixable with a short AR push. Timing discipline cuts both ways. A forecast tells you when to move, and when to wait.

Common Forecasting Mistakes

Most forecasting failures aren't math errors. They're behavioral. The model is usually fine, and the inputs feeding it aren't.

Treating best-case sales projections as certain is the most common mistake. A forecast built on your best month ever, repeated thirteen times, is a wish list.

Assuming every invoice pays on its due date is a close second. Real AR aging always includes a slice that pays late. Some of it never pays at all.

Forgetting irregular expenses ranks third. Quarterly insurance, annual software renewals, and estimated taxes don't show up weekly. They're easy to leave out entirely.

Building the forecast once and never updating it kills accuracy fast. A forecast is a living document. Stale assumptions don't reflect today's churn or today's AR balance.

Confusing profit with cash is the deepest mistake, and it catches profitable businesses off guard. Profit is an opinion. Cash is a fact.

Overcomplicating the model ranks high too. Some owners build forty rows and twelve tabs before forecasting a single real week. Start with the five sections, then add detail once it proves useful.

Finally, ignoring the forecast once it's built. A model nobody checks on Monday is a spreadsheet gathering dust. Weekly discipline is the whole point.

Frequently Asked Questions