Rolling Forecast

A rolling forecast always looks the same distance ahead. When January closes, the forecast drops January, adds the next January, and keeps a 12 to 18 month window in front of the business. An annual budget does the opposite: it covers a fixed calendar year, and by October it describes only the last two months of that year.

An FP&A lead at a growing company usually adopts a rolling forecast after one specific failure. The budget was set in November, a major customer churned in April, and by June nobody trusted the plan. Every review turned into an explanation of why the budget was stale instead of a decision about what to do next.

The mechanics are short. At each close, the team replaces the forecast for the closed month with the actual result from the general ledger. It then revisits the assumptions for the months ahead, adds the new final month, and publishes the revised view. Most teams update monthly, and fast-moving businesses update every week for cash.

The forecast rests on drivers, not line items. Revenue comes from bookings, renewal rates, and billing terms. Headcount cost comes from hires by start date. Collections come from DSO applied to the billing schedule. When a driver changes, one assumption changes and the model recalculates, which keeps the update to hours instead of weeks.

A rolling forecast does not remove the budget. The budget stays the commitment the board approved, and the forecast says where the business is heading against it. The pair gives a reviewer three numbers per line: plan, actual, and forecast. Variance analysis then explains why the forecast has moved away from the plan.

Three habits keep the forecast useful. Limit the number of drivers to the ones that move the result, because a model with 400 inputs gets updated by nobody. Record each assumption change with a date and a reason, so next month's reviewer can see what moved. And measure forecast accuracy over time. A team that never compares last quarter's forecast with what happened cannot tell whether the method improves.

The risk is a forecast that updates to match the actuals every month and never gets tested. If the forecast for next quarter always lands within 1% of the last month's run rate, it is a trend line, not a forecast. The value is in committing to a view of customer, cost, and timing changes and learning from the misses.

The monthly refresh is mostly plumbing: pulling closed actuals, replacing the forecast months, and checking the roll-forward ties. Adopt's agents prepare that update from the ledger and flag the drivers that moved, so the FP&A lead spends the time on assumptions. You review every change before it goes into the model. Sign up free.