A rolling forecast is the answer to a problem every once-a-year plan shares: it ages. A budget set in January says almost nothing in November about the weeks ahead — the look-ahead has shrunk to nothing and the assumptions are out of date. A rolling forecast fixes that by letting the horizon travel: as each period closes at the front, a new one is appended at the back.
This guide covers what “rolling” actually means, how a rolling forecast differs from a static annual budget, how the roll-forward works week by week, and why the plan-versus-actual comparison surfaces variances so early.
What “rolling” means in cash forecasting
Rolling means the outlook window stays a constant length while moving forward through time. You are not planning a fixed period up to a cut-off date once; you are always planning the next n periods from today. When a period ends it drops off the front and a new one joins at the back. The horizon rolls along with the calendar.
For the standard format of cash forecasting, the 13-week window, that means you see exactly one quarter ahead at any moment. In the first week you are looking at weeks 1 to 13; a week later, at weeks 2 to 14 — the 13 weeks of look-ahead are preserved. That is precisely the difference from a rigid plan whose remaining visibility shortens with every month gone.
The rolling approach applies to any horizon. For short-term operational control the 13-week view is the usual format; for the medium term many finance teams also run a rolling 13-month or 18-month forecast. The principle is identical — only the cadence and the granularity differ. The medium-term variant is covered in the guide to 13-month planning.
Rolling forecast versus static annual budget
The classic annual budget is a point-in-time document: set once, perhaps revised through the year as a forecast update, but fixed in essence. Its outlook window shortens systematically — by December it reaches only a few weeks ahead. Rolling planning inverts that behaviour: the look-ahead stays constant because new periods are continuously added at the back.
| Feature | Static annual budget | Rolling forecast |
|---|---|---|
| Horizon | Fixed year end | Travels — constant look-ahead |
| Look-ahead through the year | Shortens continuously | Stays the same length |
| Updating | Once, rarely revised | Fixed rhythm (e.g. weekly) |
| Purpose | Budget, target setting | Control, early warning |
The important qualification: a rolling forecast does not replace the annual budget. The budget remains the yardstick you measure against. The rolling forecast supplies the matching, continuously updated view of what is actually coming — and makes the gap between the two visible early. How the short-term format is built in detail is in the guide to the 13-week cash forecast.
The roll-forward in practice, week by week
The roll-forward follows the horizon: weekly for a 13-week forecast, monthly for a 13-month one. The routine is the same either way and comes down to three steps:
- Close. The period just ended is switched from plan to actual — real bank movements replace the estimate.
- Append. A new period joins at the back so the look-ahead stays constant (weeks 1–13 become weeks 2–14).
- Adjust. The assumptions in the remaining periods are updated against the new actuals — receipts that slipped, new open items, changed dates.
A simplified example with generic illustrative values (not real figures) shows how the window travels. In week 1 you plan weeks 1 to 13; a week later the former week 1 has become an actual week, the window starts at week 2, and a new week 14 has appeared at the back:
| Cut-off | Window | Newly closed | Newly appended |
|---|---|---|---|
| Monday, week 1 | Weeks 1–13 | — | — |
| Monday, week 2 | Weeks 2–14 | Week 1 (actual) | Week 14 |
| Monday, week 3 | Weeks 3–15 | Week 2 (actual) | Week 15 |
For the roll-forward to work reliably it needs a fixed rhythm and a fixed cut-off — first thing every Monday, say. Once the cadence becomes irregular the forecast loses its main advantage, which is being current. This weekly routine is exactly where manual spreadsheet models break down, because the re-sorting, re-keying and re-categorising costs time every single week. The data side is easier than it looks if the source is right: the open items list you already pull for collections carries most of what the roll-forward needs.
LiquidityLens rolls the 13-week outlook forward automatically from your real bank movements — the window moves on every week with no manual re-sorting.
See the Liquidity Control SprintPlan versus actual: seeing variances early
The real value of rolling planning shows up in the plan-versus-actual comparison. Because a plan week becomes an actual week every week, a comparison arises automatically: what did you expect for this week, and what actually moved? That difference is the early warning.
An example with generic illustrative values: if a planned customer receipt of €90,000 fails to arrive and only €40,000 comes in, the variance is immediately visible. More importantly, the rolling forecast carries that shift forward into the following weeks by itself. The missing receipt does not just lower the week that has closed — it moves the whole balance path down, bringing a possible low point into view that a one-off annual plan would never have shown.
That is why rolling forward is more than refreshing numbers. It connects actual performance to the outlook, so every variance is immediately worked through into the future. Run that comparison weekly and you see squeezes weeks ahead, while there is still room to act — which is what turns a forecast from a reporting artefact into a control instrument. Rolling is the cadence rather than the method itself; the method is set out under cash flow forecasting, and what to do once a squeeze is in view is covered in the guide to a cash flow shortage.
A weekly grid with roll-forward logic and plan/actual columns — ready to use as an entry point to rolling planning.