Building a rolling 12-month forecast

Almost every small business forecasts to December 31. In January you have a twelve-month view. By July it is five months, and by November you are making decisions about a business you can only see six weeks into.

The horizon shrinks exactly as year-end decisions get bigger. A rolling forecast fixes it by adding a month every time one closes, so the view ahead never contracts.

How it works

Close a month, drop it from the front, add a new month at the back. In September you forecast through the following August. In October, through the following September. The window is always twelve months, and the year boundary stops being a cliff.

Keep it to about fifteen lines

The reason rolling forecasts get abandoned is that people build them at the same detail as their accounts. Forty expense rows re-forecast monthly is a job nobody sustains.

Revenue by stream, cost of sales as a margin percentage, people, and four or five grouped expense categories is enough. You are forecasting a shape and a scale, not a chart of accounts, and the precision you lose by grouping is far smaller than the precision you lose by not doing it at all.

Forecast the drivers, not the totals

Revenue built as "last year plus 8%" cannot be interrogated. Built from your engine — customers and churn, or capacity and utilization, or spend and conversion — every month becomes a set of assumptions you can argue with.

That is what makes the monthly update fast. You are not re-typing twelve columns of numbers; you are adjusting four or five assumptions and letting the arithmetic follow.

Compare each month to what you said

Before you revise anything, look at how last month came in against what you forecast for it. That is the calibration step, and it is the entire reason the exercise compounds.

Within a few months you will know your own bias — almost everyone forecasts revenue optimistically and costs conservatively, in fairly consistent proportions. Once you can name yours, you can correct for it, and the forecast becomes genuinely useful rather than aspirational.

It does not replace the budget

Keep both. The budget is the fixed commitment you measure against; the rolling forecast is your live view of where things land. Reporting actual, budget and rolling forecast side by side answers "are we on track" and "where does this end up" at the same time.

The rolling version also makes next year's budget dramatically easier. By the time you sit down to build it, you have already forecast most of those months several times and know which assumptions have held.