Rolling Forecasts vs. Annual Budgets: Why Multi-Location Operators Need Both

If you read enough FP&A content, you will eventually be told that the annual budget is dead and the rolling forecast has replaced it. This is a good example of an FP&A idea that is half right, and the half that is wrong is the half that matters.

The annual budget and the rolling forecast are not competing tools. They are complementary instruments that do different jobs. Operators who replace one with the other almost always regret it within two quarters. Operators who run both — with clear, separate purposes — outperform on both cash management and accountability.

What an annual budget is actually for

The annual budget is not a forecast. It never was. It is a commitment. Specifically, it is a set of measurable promises the operating team makes to the ownership team (and to itself) about what performance will look like over the next twelve months, against which compensation, credit, and capital decisions are benchmarked.

Three jobs the annual budget does that no other artifact does as well:

• Sets accountability. A branch manager’s bonus is measured against the budget, not the forecast. The budget must be fixed at the start of the year or the accountability breaks.
• Anchors debt covenants. Bank covenants are negotiated against annualized budget numbers. A moving forecast cannot satisfy a fixed DSCR covenant test.
• Forces real trade-offs in Q4. The annual planning cycle is the one time a year every department head argues for their resource allocation against every other department head. That argument, held once a year, is useful even when the output is imperfect.

Remove the annual budget and you lose all three. The operators who tried to replace it with a rolling forecast in the late 2010s reported the same pattern: accountability decayed, covenant monitoring got chaotic, and resource allocation turned into a series of monthly skirmishes with no forcing function.

What a rolling forecast is actually for

The rolling forecast is not a commitment. It is a projection — the team’s best current estimate of what is going to happen over the next 12 or 18 months, updated each month with the most recent actuals.

Three jobs the rolling forecast does that an annual budget cannot:

• Updates with reality. When a branch opens earlier than planned, or a hurricane closes two locations for a month, the budget is wrong. The forecast reflects the new reality.
• Extends beyond the fiscal year. In month 10, your budget is only two months of runway. The rolling forecast is always looking 12 to 18 months out, which is the horizon you actually make hiring and lease decisions on.
• Pressures assumptions. The act of updating the forecast monthly forces the team to examine each assumption monthly. The budget is set once and forgotten.

The two-line framework

Run both, side by side. In every meeting. Every variance analysis. Every board pack.

The budget is the line you committed to.
The forecast is the line you now expect.
The gap between them is the conversation.

When budget = forecast, the team is on plan. When forecast > budget, the team is beating plan and should know where the upside is coming from (so they can extend it). When forecast < budget, there is a gap, and the conversation is not “ the number is wrong” — it is “which driver drifted, who owns it, what is the recovery plan.”

The driver-based structure we describe in the driver-based budgeting framework makes this conversation trivial. When every line of the P&L rolls up from three revenue drivers, six cost drivers, and a handful of step-cost triggers, the gap between budget and forecast decomposes cleanly into which drivers moved.

Cadence: how to run both without the team drowning

Running two sets of numbers sounds like twice the work. It is not, if the cadence is clean.

Annual budget cycle — October to December

Ten weeks. One lock. Builds driver-by-driver with branch owner input. Signed by ownership on December 20. Does not change again until next October. Bonuses and covenants anchor here.

Rolling forecast cycle — monthly close + 3 days

Two hours of controller work on day 3 after month close. Replace the frozen budget driver with the actual driver for the closed month, re-project the next 12 months using a trailing-three- month weighted average for each driver. Push the forecast line out one more month so the horizon is always 12+. Publish in the same BvA view alongside the budget.

Variance meeting — monthly

Forty-five minutes. Three questions only: (1) Which drivers are off budget, by how much, and why? (2) Which drivers are off forecast, by how much, and why? (3) What are we doing this month to narrow the gap? Assign exactly two action items per meeting with two-week deadlines.

Common mistakes

Collapsing budget and forecast into one line

The single most common mistake. When the forecast and budget are merged, accountability disappears and the “number” becomes whatever is defensible that month. Do not do this. Keep them as two distinct series in every view.

Updating the budget mid-year

The budget is a commitment. Commitments that are rewritten mid-year are not commitments. If conditions change enough to warrant a budget reset — a major acquisition, a force-majeure event — either do not reset it and explain the variance, or fully re-anchor with ownership approval. There is no middle option.

Forecasting too far out

Eighteen months is plenty. Anyone telling you that a multi-location SMB needs a 36- or 60-month forecast is selling something. Past 18 months, the driver assumptions are noise, and the time spent maintaining them is better spent on the next two quarters.

How Aziell handles budget and forecast

Both live as first-class series in the data model. When you create a plan, you are creating the budget. On day 3 after month close, Aziell pre-computes the forecast by replacing the closed month’s drivers with actuals and re-projecting. You review and publish; you do not recompute from scratch. Variance views always show three series — budget, forecast, actual — so the monthly review meeting has the exact artifacts it needs.

If you are still running one spreadsheet with a single series, the hidden-cost analysis is the right next read. If you have the budget in place and want to add the forecast layer properly, the scenario planning guide covers how to extend from single-series forecast to three-scenario planning.