Definition
A rolling forecast is a forecast that always looks a fixed distance into the future, commonly twelve months, regardless of where the company sits in its fiscal year. Instead of forecasting January through December and then stopping, a rolling forecast for March would run through the following February, and the one built in April would run through March of the next year. Each time the forecast updates, usually monthly or quarterly, the oldest period drops off the front and a new period gets added at the back, so the company is never staring at a forecast that only covers the next six weeks.
The reason rolling forecasts exist is that an annual budget set once a year gets stale fast, especially as the year goes on. A budget built in November for the following January often looks reasonable through spring and increasingly disconnected from reality by autumn, because a full year is a long time for assumptions to hold. Worse, a company running only on an annual budget spends the last quarter of its fiscal year staring at a forecast horizon that keeps shrinking toward nothing, with no formal look-ahead into the next year until the whole budgeting cycle starts over. Rolling forecasts solve that by making sure there is always a meaningful window of visibility ahead.
A rolling forecast is not just an annual budget with the months relabeled or copied forward. Doing it properly means actually re-forecasting the remaining periods using the latest actuals and current assumptions each time it updates, not simply carrying old numbers forward and tacking a new month onto the end. A shallow version of this exercise, where the team just adds a month without revisiting whether the existing numbers still make sense, produces a forecast that looks continuously updated but is not actually more accurate than the stale annual budget it replaced. The real work is in the re-forecasting, not the rolling mechanic itself.
By 2026, rolling forecasts are common, though far from universal, particularly in technology and other fast-moving industries where a twelve-month-old assumption is close to worthless. Many companies run a hybrid, keeping the formal annual budget as the fixed target used for compensation and board commitments, while running a rolling forecast alongside it purely for internal decision-making and visibility. Planning software has made the mechanics of rolling a forecast forward much less painful than it used to be in spreadsheets, where dragging every model forward a period by hand was tedious enough that many teams simply avoided doing it as often as they should have.
This page covers how a rolling forecast actually works, how it compares to a fixed annual budget, how it differs from a one-off reforecast done mid-year, where it is worth the ongoing effort and where it is not, and how to run one well. The idea worth keeping is that a rolling forecast trades the comfort of a fixed annual target for a forecast that stays honest about what is likely to happen next, at the cost of doing the forecasting work more often rather than once a year and calling it done.
Key Takeaways
- A rolling forecast always extends a fixed number of periods into the future, adding a new period and dropping the oldest one each time it updates.
- It exists because a fixed annual budget goes stale as the year progresses and leaves shrinking visibility into what comes next.
- A real rolling forecast re-forecasts the remaining periods with current assumptions each cycle, not just copies old numbers forward and adds a month.
- By 2026 it is common, especially in fast-moving industries, often run alongside a fixed annual budget rather than replacing it entirely.
- The tradeoff is doing forecasting work more often in exchange for a forecast that stays realistic rather than one fixed and forgotten for a year.
How a Rolling Forecast Works
The mechanics start with picking a horizon, most often twelve months, though some companies run eighteen or twenty-four for longer lead-time businesses. Each forecasting cycle, typically monthly, the team drops the period that just closed, updates the remaining periods with fresh assumptions and any new information, and adds one new period at the far end to keep the horizon length constant. So a forecast built in June covering June through the following May becomes, in July, a forecast covering July through the following June, with every number in between reconsidered rather than left untouched.
Every cycle starts by pulling in the actuals for the period that just closed, since those actuals are the most reliable signal about how current assumptions are holding up. If actual revenue came in above forecast because a deal closed early, that is not just noted, it should change the assumptions feeding the remaining periods too, since it may signal something about the pipeline or the market that the old forecast did not account for. Ignoring what actuals are saying and mechanically rolling old assumptions forward defeats the purpose of doing this monthly instead of annually.
Most rolling forecasts are built on a driver-based structure underneath, since re-forecasting manually every month without formulas tied to real drivers would be exhausting and error-prone. When the underlying model is built around drivers like customer count, headcount, or unit volume, updating the forecast each cycle mostly means updating driver assumptions rather than rebuilding the whole model, which is what makes running this process monthly practical rather than a burden the team quietly lets slip after a few cycles.
Cadence and discipline matter as much as the model itself. A rolling forecast that is supposed to update monthly but actually gets refreshed every other month, or gets rushed through without real scrutiny of the assumptions, loses most of its value, since the whole point is currency. Successful rolling forecast processes usually have a fixed calendar slot each cycle, clear ownership of who updates which driver, and a short review with business leaders before the numbers go out, so the update does not slip when everyone gets busy.
A Rolling Forecast Compared to an Annual Budget
An annual budget is set once, usually in the last quarter of the prior fiscal year, and typically stays fixed for the year as the reference point for performance, compensation targets, and board commitments. Its strength is stability: everyone knows what number they are being measured against, and that number does not move just because conditions changed, which keeps incentives clean and comparisons consistent across the year. That predictability matters most in setting targets that people are expected to be held to.
A rolling forecast trades that stability for currency. It updates continuously, so it reflects what is actually likely to happen rather than what was assumed many months ago, but that same flexibility means it is a poor tool for holding anyone accountable to a fixed number, since the number itself keeps moving. Using a rolling forecast as if it were the budget, changing the target every month, tends to quietly erode accountability even as it improves accuracy.
Because of that tradeoff, most companies that use rolling forecasts run both rather than replacing one with the other. The annual budget stays fixed as the yardstick used for accountability and target-setting, while the rolling forecast runs alongside it purely to give leadership a current, honest read on where the business is actually headed, something the fixed budget increasingly cannot provide as the months pass and conditions change. The two numbers diverging over the year is not a failure, it is the rolling forecast doing exactly what it is meant to do.
The other real difference is effort and cadence. An annual budget concentrates a large amount of planning work into a few weeks once a year, then mostly goes quiet until the next cycle. A rolling forecast spreads a smaller amount of work across every month of the year, which suits some teams and burns out others, depending on how well the underlying model is built to make each update fast rather than a mini version of the full annual process repeated twelve times.
What Makes a Rolling Forecast Different From a Reforecast
A reforecast is a one-off update to the annual plan, usually triggered at a specific point, mid-year being the most common, where the company looks at actuals so far and updates its expectation for the rest of the year given what has happened. It happens once, or maybe twice a year, and it stays tied to the original fiscal year boundary rather than extending past it. Many companies run this informally, without ever calling it a reforecast by name.
A rolling forecast does something structurally different. It is not a single course correction but an ongoing process that happens every month or quarter, and its horizon extends past the current fiscal year rather than stopping at year end. Where a mid-year reforecast answers 'given what we know now, how will the rest of this year land,' a rolling forecast is always answering a slightly different, always-forward-looking question, regardless of where the fiscal year happens to end.
The two are not mutually exclusive. A company can run periodic reforecasts, say quarterly, without ever adopting a true rolling structure, simply updating the same fixed-year budget a few times as the year goes on. A rolling forecast, by contrast, effectively subsumes the reforecast idea into a continuous habit rather than a scheduled event, which is more work but also removes the gap between reforecasts where the plan can quietly go stale again.
The confusion between the two mostly comes from the fact that both involve looking at actuals and updating a forecast, which is the entire visible mechanic to someone outside finance. The practical distinction that matters is the horizon: if the update stops at the same fiscal year end every time, it is a reforecast. If it keeps extending forward past that boundary every cycle, it is a rolling forecast, and that difference in horizon is what actually changes how the company plans.
Where a Rolling Forecast Fits and Where It Does Not
Rolling forecasts fit well in businesses where conditions change quickly enough that a forecast more than a few months old is not very trustworthy, fast-growing companies, businesses in cyclical or volatile markets, and anything where a single external shock, a supply disruption, a rate change, a competitor move, can meaningfully shift the outlook within a quarter. In these settings, the extra effort of forecasting more often pays for itself in decisions made with a more current picture.
It also fits well anywhere leadership needs continuous visibility into something specific, cash position, hiring capacity, or production capacity, where waiting for the next annual cycle to update the view would leave a real blind spot. A software company watching burn rate closely, for instance, benefits a great deal from always having a rolling twelve-month cash view rather than one anchored to a fiscal year that resets every January. Waiting for the annual cycle to catch that kind of shift would mean acting months too late.
It fits poorly in stable, slow-moving businesses where conditions genuinely do not change much month to month, a small business with steady recurring revenue and predictable costs, for instance. There, the ongoing effort of a monthly re-forecast produces very little new information relative to the annual budget, and stretching a lean finance team thin on a process that is not adding much value is a real cost worth weighing honestly.
It also fits poorly at a company that does not yet have the discipline or the underlying data quality to update a forecast credibly every month. Rolling the forecast forward on autopilot, without real scrutiny of whether the driver assumptions still make sense, produces something that looks current but is not actually more accurate than a once-a-year budget, just more frequently wrong in slightly different ways. Fixing the underlying forecasting discipline is a better first step than adopting the rolling mechanic on top of a shaky process.
How to Run a Rolling Forecast Well
Pick a horizon length that matches how far ahead the business actually needs to see, and hold it steady. Twelve months is the common default, but a business with long sales cycles or long lead times on inventory might need eighteen or twenty-four to be genuinely useful. Changing the horizon length frequently, or letting it drift, makes it hard to compare one cycle's forecast to the next and undermines the whole point of having a consistent rolling structure.
Build the underlying forecast on drivers rather than static line items, since re-forecasting a driver-based model each cycle mostly means updating a handful of assumptions instead of rebuilding formulas from scratch. Teams that try to run a rolling process on a spreadsheet full of hard-typed numbers usually find the monthly workload unsustainable within a couple of quarters and quietly let the cadence slip. The upfront work of building the driver structure pays for itself within a few cycles.
Protect the cadence like a deadline that matters, because a rolling forecast that slips from monthly to whenever someone gets around to it stops delivering the currency that justifies the extra work in the first place. Put it on the calendar right after the monthly close, build the process to run quickly using data that is already available by then, and treat a missed cycle as a real problem rather than a minor scheduling slip.
Keep the fixed annual budget and the rolling forecast clearly separate in how they are used and communicated, so nobody confuses a moving target with the number people are actually accountable to. Label them distinctly in reporting, and be explicit in meetings about which one is being discussed, since mixing the two up is one of the more common ways a rolling forecast process causes confusion rather than clarity for the people relying on it.
When presenting each cycle's update, spend more time on what changed since the last forecast and why than on the fresh set of numbers itself. Leadership usually already has a rough sense of the current numbers, what they actually need from a rolling forecast is a clear read on what shifted, a slower sales cycle, a delayed hire, a cost increase, and what that shift implies for the months ahead, which is the entire reason to run this process every month instead of once a year.
Best Practices
- Pick a horizon length that matches how far ahead the business genuinely needs visibility and keep it consistent across cycles.
- Build the forecast on driver-based formulas so each update mostly means changing assumptions rather than rebuilding the model.
- Protect the update cadence, since a rolling forecast that slips from monthly to occasional loses the currency that justifies the extra work.
- Keep the fixed annual budget and the rolling forecast distinct in reporting so nobody confuses a moving number with a fixed target.
- Lead each update with what changed since the last cycle and why, not just the refreshed numbers.
Common Misconceptions
- A rolling forecast is not the same as an annual budget updated occasionally; it extends the horizon forward continuously rather than stopping at fiscal year end.
- It is not just the annual budget with a month copied forward; every remaining period should be re-forecast with current assumptions, not left untouched.
- A rolling forecast does not replace the fixed budget used for accountability; most companies run both for different purposes.
- It is not free to maintain; it trades a smaller, more frequent workload for the comfort of forecasting once a year and being done.
- A rolling forecast is not automatically more accurate than an annual budget; it is only as good as the discipline behind each update.
