The 13-Week Rolling Cash Flow Forecast: What It Is, How It Works, and How to Run It Every Week

Almost every owner I have worked with has had the same week at least once. The month closed profitable. Yet the bank balance says something different. Payroll is Friday, a quarterly insurance renewal just cleared, and the two biggest invoices are still sitting in a customer’s approval queue. Nothing is wrong with the business. Everything is wrong with the timing. A 13-week rolling cash flow forecast exists for exactly this week.

The 13-week rolling cash flow forecast is the single tool that fixes that week, five weeks before it happens. It is the first model I build for a new client, ahead of the reporting package, ahead of the annual plan, ahead of anything else — not because cash is more important than profit, but because cash is the constraint that decides whether you get to act on anything else you have learned.

This is what it is, what makes one work, where they usually go wrong, and how to run it weekly without it becoming a project.

What is a 13-week rolling cash flow forecast?

A 13-week rolling cash flow forecast is a week-by-week projection of cash in and cash out over the next 13 weeks, rebuilt every week from actual bank balances. It is a cash model, not an accrual one, and it rolls forward continuously rather than being built once and left to age.

Three words in that name are doing all the work.

• Rolling means the window moves. Every week you drop the week that just closed, add a new week 13 at the far end, and reset the starting balance to what is actually in the account. The forecast is never more than seven days from reality.
• Cash means cash, not revenue and not profit. An invoice you sent is not in this model — the day a customer’s payment clears is. A prepaid annual insurance premium hits in full on the day the bank drafts it, not spread across twelve months the way your income statement correctly shows it. This model deliberately ignores accrual accounting, which is exactly why it needs to sit alongside GAAP financial statements rather than replace them.
• Weekly means weekly. Monthly cash forecasting hides the problem, because the two weeks that hurt are averaged against the two weeks that do not. Most cash crises are four-day events inside an otherwise fine month.

Why 13 weeks?

Thirteen weeks is one quarter — the longest horizon where you can still name most of the individual dollars, and the shortest one where decisions with real lead time still show up. Beyond a quarter, cash forecasting turns into assumption. Inside a few weeks, you are too late to change anything.

The practical test is lead time. A hire made today costs cash in about six weeks. Equipment ordered today gets invoiced in eight. A quarterly tax payment, an insurance renewal, an annual software bill, a note payment schedule: all of them sit inside a quarter, all are knowable to the day, and all are exactly the things that collide with payroll and surprise people.

It is not a magic number. Some businesses run a 16-week window because their collection cycle is longer; a few run eight because everything is card-swiped at the point of sale. What matters is that the window covers your longest normal collection lag plus your longest committed payment lead time, with a few weeks of margin on the end. For most owner-operated businesses, that lands on a quarter.

How is a 13-week cash flow forecast different from a budget or a P&L?

The P&L tells you whether you made money in a period that has already ended. Your annual budget lays out what you expect over the year. And the 13-week forecast tells you whether you can cover what is coming in the next 90 days, in cash, by week. They answer three different questions, and none substitutes for another.

This is why a profitable company can run out of cash and a break-even one can sit comfortably. Profit and cash are separated by timing: when you bill, when you collect, when you pay, when you buy inventory, when you service debt, and when the owner takes distributions. The P&L measures the first half of those. The forecast measures the second.

What does a 13-week cash flow forecast actually look like?

It is a grid: weeks across, cash lines down. Each week starts with the prior week’s ending balance, adds expected collections, subtracts expected payments, and checks the ending balance against a minimum cash buffer and flags it if it falls short. The value is not the total. It is which specific week goes red.

Here are the first eight weeks of an illustrative forecast — I invented the figures for this example, but the shape is one I see constantly. The minimum cash buffer here is $25,000.

Illustrative figures. Weeks 9 through 13 continue the same grid.

Read week 6. Two payrolls, a quarterly insurance renewal, and a note payment all landed in the same seven days, against the lightest collection week in the quarter. Nothing in that week is a surprise on its own — every payment was knowable in January. It only becomes a crisis because nobody had ever laid them on the same calendar.

And notice what the model gives you: five weeks of warning and three ordinary levers. Move the insurance renewal to its grace date in week 7. Call the two customers scheduled to pay in week 7 and ask for week 5. Defer one owner distribution by a month. Week 6 goes from negative $25,700 to comfortably positive, and nobody drew on a line of credit or had a hard conversation with a vendor.

That is the entire point of the tool. It converts a future emergency into three phone calls you make early.

What makes a 13-week forecast actually work?

Six things make it work: a control panel that re-times the whole model, a rolling weekly anchor, dates instead of averages for every payment, an effective date on every change, a minimum cash buffer instead of zero, and formulas that trace every figure back to a source.

• A control panel. One start date, one beginning cash balance, one minimum buffer. Change the start date and all thirteen weeks re-anchor; change the balance and the whole runway re-walks. If updating the model means editing thirteen columns by hand, it will not survive a busy week.
• A rolling weekly anchor. Weeks run from the actual start date forward in seven-day windows, not fixed to calendar months. Cash does not care that a month ended on a Tuesday.
• Landing dates, not monthly averages. Payroll lands on the 15th and the last day. Rent lands on the 1st. Spreading a monthly total evenly across 4.33 weeks produces a model that is roughly right in total and useless in every individual week — the only place the answer lives.
• Dated changes. A hire starting in week 7 should not draw cash in weeks 1 through 6. Every recurring item needs an effective date, or the model quietly lies about both ends of the quarter.
• A minimum cash buffer, not zero. Forecasting to zero means planning to be broke on schedule. Set a floor that reflects one payroll plus a bad week, and treat the flag as the alarm rather than the negative number.
• Formulas and an audit trail. Calculate every total; never type one in. Every inflow and outflow traces to a bank line, an invoice, or a signed agreement. If nobody can tie a cash model out, someone will argue with it the first time it says something inconvenient — exactly the moment it needed to be trusted.

Where do 13-week cash flow forecasts go wrong?

Six failure modes account for nearly all of the broken models that land on my desk.

• Averaging instead of dating. The most common one by a distance. Dividing a monthly total by four does not produce a weekly forecast.
• Forecasting revenue instead of collections. What you bill in week 3 is not cash in week 3. Model every customer’s payment behavior the way they actually pay, not the way the terms read.
• Double-counting the credit card. Model the categorized card total, or the payment to the card — never both. I have seen this overstate outflows by six figures over a quarter.
• Starting from a balance nobody tied out. Beginning cash is the one number the entire model rests on. It should tie to the bank, to the penny, including outstanding checks that have not cleared.
• Modeling to zero. A forecast that shows $600 in the account on Thursday is not telling you that you are fine.
• Building it once. A forecast nobody refreshes weekly is a document, and it starts decaying immediately. Most people build abandoned cash models to impress, not to update in twenty minutes.

How do you run a 13-week forecast every week?

Roll the start date forward one week, reset beginning cash to the actual bank balance, update collections and payments that changed, compare the week that just closed to what you had forecast, then look at the flagged weeks and decide. Twenty to thirty minutes once the model is built. This is the Execution discipline inside our HELM System — a weekly rhythm with a decision attached, not a report generated and filed away.

The comparison step is the one people skip, and it’s the one that makes the model get better. Every week, look at what you projected for the week that just closed against what actually happened, and ask which side missed — collections, almost always. Then ask which customer, and adjust that customer’s payment behavior rather than the total. Six or eight weeks of that and the forecast stops being an estimate and starts being close.

The output of the weekly run is not a report. It is a short list of decisions: which payables to release this week, whether a deposit on new equipment can go out before the 15th, whether the owner draw waits, whether anyone needs to make a collection call today. The forecast produces a payment queue in priority order, so the meeting spends its time deciding, not reconstructing.

What decisions does a 13-week forecast actually drive?

Hiring timing, equipment purchases, owner distributions, which payables get released and when, whether to draw on a line of credit, whether to accept work at a discount to pull cash forward, and how to sequence tax payments, renewals, and debt service so they never land in the same week as payroll.

The pattern across all of those is the same: the forecast rarely changes what you decide. It changes when you decide it, and that is usually worth more.

A manufacturer I work with builds to order with customer deposits funding production. Cash arrives on milestones, not on delivery, so we build the forecast off the production schedule rather than the sales pipeline. When a unit slips two weeks in the shop, the model shows immediately that a milestone payment moved into the following month and which payables have to move with it. Without that, a routine production delay becomes a cash surprise four weeks later that gets blamed on collections.

Who should own the 13-week forecast?

The bookkeeper or controller supplies the actuals and keeps the beginning balance tied to the bank. Ownership of the model, the assumptions, and the read belongs to the CFO — this is core to what a fractional CFO does inside the finance function. The owner owns the decisions the flagged weeks force. If one person does all three, people will maintain the model but not use it, or use it but not trust it.

The failure I see most often in owner-operated businesses is a forecast that whoever is best at Excel builds, and that nobody reviews. Someone updates it for a while, then it drifts, and people quietly stop opening it. Ownership is not about who types. It is about whose meeting the flagged weeks show up in.

The short version

A 13-week rolling cash flow forecast is the shortest distance between your bank account and a decision. It works when you date every payment to the day it lands, when the whole grid re-times from one control panel, when there is a real buffer instead of zero, and when somebody runs it every week and compares it to what actually happened.

This will not make you more profitable. What it will do is make sure the profit you earn is still there when the payment comes due, and that you saw the tight week coming with enough runway to handle it with three phone calls instead of a loan.

Frequently Asked Questions

How accurate is a 13-week cash flow forecast?

The first few weeks should be close to exact, because you already know most of it. Accuracy decays toward the back of the window, where collections are estimates. Weekly variance review tightens it fast. Judge the model by whether it correctly identified the tight weeks, not by whether the week-13 balance was right to the dollar.

What is the difference between a 13-week cash flow forecast and a statement of cash flows?

The statement of cash flows is a GAAP financial statement that explains where cash went in a period that has closed. A 13-week forecast is a forward-looking management tool with no reporting standard behind it. One is history and compliance, the other is planning. Neither replaces the other.

How often should you update a 13-week cash flow forecast?

Weekly, on the same day each week. Roll the window forward, reset beginning cash to the actual bank balance, update any collections or payments that changed, and compare the week that just closed to what you had projected. A forecast updated monthly is a monthly forecast with extra columns.

Do I need a cash flow forecast if my business is profitable?

Yes — profitable, growing businesses are the most common place cash runs short. Growth consumes cash before it produces it: you pay labor, materials, and equipment ahead of collecting. Profit tells you the work was worth doing. The forecast tells you whether you can fund it between now and getting paid.

Can you build a 13-week cash flow forecast in QuickBooks?

QuickBooks supplies the inputs — A/R aging, A/P aging, recurring bills, bank balances — but its built-in cash flow tools work in monthly buckets and cannot date payments the way a weekly model needs. Most working forecasts are a spreadsheet or a dedicated tool that pulls from accounting data, not a report inside it.

What is a minimum cash buffer and how do I set it?

It is the floor your ending cash should never fall below, and the number the forecast flags against. A common starting point for owner-operated businesses is one full payroll cycle plus one week of ordinary operating outflows. Set it deliberately, put it in the model, and revisit it as fixed costs change.

Where to Start

Westport Financial builds and runs 13-week rolling cash flow forecasts as part of the finance function for owner-operated businesses between $1 million and $10 million in revenue. We drive the model with formulas, tie it to the bank to the penny, and refresh it every week, so you always have a current answer instead of an old one to rebuild.

Schedule a free Financial Health Evaluation and we will show you which of the next 13 weeks is the one to watch.