Rolling Forecasts: Replacing the Annual Budget Nobody Opens
- Why the annual budget stops working
- Build it on drivers, not lines
- The monthly routine
- Track the forecast’s own bias
- Connect it to cash
- Keep it small and owned
- Where this fits
A rolling forecast always looks the same distance ahead. At the end of each month you drop the month that has just finished, update the months that remain with what you now know, and add a new month at the far end, so the business always has a twelve-month view. Unlike an annual budget, which is set once and goes stale by the second quarter, a rolling forecast stays useful for decisions all year.
For a small business it does not need to be elaborate. A good one is built on a handful of drivers and takes an hour a month to update.
Why the annual budget stops working
An annual budget is prepared months before the year starts, using assumptions that are already out of date by the time the first actual figures arrive. By July, owners are comparing performance against a plan everyone knows is wrong, and the budget becomes a document for explaining variances rather than making decisions.
A budget still has a role as a target and a commitment. The rolling forecast answers a different question: given what we know today, where are we heading? Many businesses keep both: the budget for accountability, the forecast for decisions.
Build it on drivers, not lines
The mistake that kills most forecasts is detail. A forecast with two hundred expense lines is too tedious to update monthly, so it stops being updated. Instead, identify the few numbers that actually drive the business and forecast those.
A B2B services firm might use:
| Driver | Example value |
|---|---|
| Qualified enquiries a month | 400 |
| Conversion to paying client | 12% |
| Average first-year value per new client | ₹45,000 |
| Revenue from existing clients, monthly | ₹18,00,000 |
| Gross margin | 55% |
| Headcount and average monthly cost | 22 people, ₹60,000 |
| Other fixed overheads | ₹4,50,000 a month |
New-client revenue for a month is 400 × 12% × ₹45,000 = ₹21,60,000 of first-year contract value, which you then spread across the months in which it is billed. Everything else follows from the other drivers.
When something changes, you change one driver and the whole forecast updates. If conversion falls to 10%, new-client value drops to ₹18,00,000 a month, and you can see immediately what that does to profit and cash three and six months out.
The monthly routine
- Close the month. Replace the forecast for the month just ended with actual figures.
- Compare. Note how far the forecast you made three months ago was from what happened. This is the forecast’s own accuracy, which is different from a budget variance.
- Update drivers. Change only what the evidence supports: a lost client, a signed contract, a hire, a price rise.
- Extend. Add the next month at the far end, usually by carrying forward the drivers with any known seasonality.
- Decide. Look at the next three to six months in particular. What needs action now, while there is time?
Track the forecast’s own bias
The most valuable habit is measuring whether the forecast is systematically wrong in one direction. Record the revenue forecast for each month as it stood three months before, and compare it with the actual:
| Month | Forecast made 3 months earlier | Actual | Forecast above actual by |
|---|---|---|---|
| April | ₹38.9 lakh | ₹36.0 lakh | 8% |
| May | ₹40.0 lakh | ₹36.0 lakh | 11% |
| June | ₹39.2 lakh | ₹37.0 lakh | 6% |
Three months in a row, the forecast was too high by a similar margin. That is not bad luck. It is an optimistic assumption, usually in conversion rate or deal timing, and it should be corrected in the driver, not explained away each month. A forecast that is sometimes high and sometimes low is healthy. One that always misses the same way is misleading the business.
Connect it to cash
A profit forecast tells you whether the business is working. It does not tell you when the bank balance will be low. Add three simple conversions to turn the forecast into a cash view: when customers pay, using your typical collection period; when suppliers and salaries are paid; and when lumpy items such as tax, insurance and loan repayments fall due.
For the next thirteen weeks, a weekly cash forecast is more precise; for months four to twelve, the monthly rolling view is enough.
The combination is where the forecast earns its keep. A plan to hire three people in October may look affordable in the profit view and still push the bank balance below your reserve in November, because the new clients they serve pay 60 days after invoicing. Seeing that in July leaves time to stagger the hiring, negotiate terms or arrange a limit.
Keep it small and owned
- One owner updates it, on a fixed day each month.
- Drivers sit on one tab, clearly labelled, with the date they were last changed.
- Actuals come from the accounts, not re-typed estimates.
- Scenarios are copies of the drivers tab, not separate forecasts: a base case, and a downside where the two most uncertain drivers go wrong.
Where this fits
- The cash view for the next quarter: 13-week cash flow forecast
- Reviewing actuals against plan: budget vs actual
- How much buffer the forecast should protect: how much cash reserve a business should hold
- Tools: cash flow calculator · monthly and annual revenue converter