Profitable businesses fail by running out of cash. A rolling forecast is the only thing that shows the gap coming while there is still time to do something about it.
A cash flow forecast projects the money actually entering and leaving your bank account over a coming period. It is not a profit forecast and the difference is the whole reason it exists: profit is recorded when a sale is earned, cash arrives when the customer pays, and businesses fail in that gap while looking profitable on paper.
The version that works for small businesses is a rolling thirteen-week forecast — a quarter, updated weekly. Thirteen weeks is long enough to see a problem while there is still time to act on it, and short enough that the numbers are real rather than aspirational.
Forecast when customers will actually pay, not when the invoice is due. This single adjustment is what separates a forecast that works from one that is politely wrong every week. If a customer has paid at forty-five days for two years, they will pay at forty-five days again, and a forecast built on thirty-day terms will show cash you do not have — in the week you were relying on it.
Three neighbouring documents worth separating. The cash flow forecast is forward-looking, short-horizon and updated constantly. Financial projections are longer-term and usually built for a plan or an investor, covered in financial projections template. The cash flow statement is historical, part of the annual accounts, and explains where cash went last year rather than where it is going — covered in cash flow statement. Only the first will tell you about the week after next.
The forecast showing a gap is the forecast doing its job, and the actions available are more numerous twelve weeks out than two. Chase receivables specifically rather than generally. Agree longer terms with a supplier before you need them, which is a normal conversation in advance and an awkward one when a payment has already been missed. Delay discretionary spending. Talk to the bank early. And check the assumptions once more — a forecast built on optimistic payment dates produces false alarms, which is its own failure because it teaches you to ignore the next one.
Ettex Sheets is the natural home: a thirteen-column grid you update weekly, with the arithmetic visible rather than hidden. The actuals come from Ettex Books and the outstanding invoices from Ettex Invoices, and the receivables position behind the money-in column is covered in accounts receivable.
Being clear about limits: there is no bank feed, no automatic import of transactions and no forecast generated from your history. The weekly update is manual, which is a genuine cost — and also the reason the person doing it knows what is coming, which is most of the value.
Thirteen weeks for the working version, updated weekly. Longer horizons become projections rather than forecasts and are useful for different questions.
Profit records revenue when earned and costs when incurred; cash flow records money when it moves. A profitable business with slow-paying customers can still run out of money.
From their history, not from your terms. If a customer consistently pays at forty-five days, forecast forty-five days regardless of what the invoice says.
Act while the horizon is long: chase specific receivables, agree extended terms with suppliers before you need them, delay discretionary spending, and talk to the bank early rather than in the week of the gap.
Break-even analysis is three inputs and one division. The reason it gets skipped is not difficulty — it is that the answer is often uncomfortable, and finding out early is the entire point.
A media plan template is a grid: channel, dates, budget, what you expect back. Its value is not the forecast in the last column — it is that everything competing for the money is visible on one page before any of it is committed.
Supply chain management sounds like something that needs a platform. At the scale most companies operate, it is five lists kept honestly — suppliers, lead times, orders placed, stock on hand, and what went wrong last time.