Most budgets are built the wrong way. Someone opens last year’s P&L, adds 7% to revenue, adds 5% to every expense line, and sends the file around for approval. The output looks like a budget. It is not. It is a wish with tabs.
“A driver-based budget forecasts the operational quantities that produce dollars — not the dollars themselves.”
A driver-based budget is the honest alternative. Instead of forecasting dollars, you forecast the operational quantities that produce dollars. If you run a dental DSO, your revenue driver is not “last year plus 7%.” It is chairs × patient visits per chair per day × operating days × average service value. Each of those numbers has a floor, a ceiling, and a team member accountable for it. When the plan misses, you know exactly which driver drifted.
This is the complete framework we use with multi-location service operators. It is opinionated, it assumes you have 4 to 50 locations, and it assumes you would rather be right than comprehensive.
Step 1: Pick the three revenue drivers
Every multi-location service business has exactly three revenue drivers. Only three. More than that and you are building a model that no operator will maintain. Fewer than that and you are hiding something. The three are:
- Capacity. How many units of service can you produce in a day? Chairs, trucks, rooms, stations, members served at full utilization.
- Utilization. What percent of that capacity do you actually sell? Seat fill rate, truck dispatch rate, chair hour utilization.
- Average ticket. What is the revenue per unit sold? Average check, average repair ticket, average monthly member revenue.
Revenue = Capacity × Utilization × Average ticket × Operating days. That is the whole equation. Every other “driver” you might be tempted to add is a sub-driver of one of these three. Resist the temptation. The power of driver-based planning is in its thinness.
Examples across verticals
| Vertical | Capacity | Utilization | Avg ticket |
|---|---|---|---|
| Dental DSO | Operatories | Chair hour fill rate | Revenue per visit |
| HVAC/trades | Service trucks | Dispatch rate | Ticket avg |
| Restaurant group | Seats | Turn rate | Check avg |
| Fitness / studio | Max members | % of cap | Monthly ARPU |
| Childcare | Licensed slots | Enrollment rate | Weekly tuition |
Step 2: Decompose cost into variable, step, and fixed
Costs do not behave the same way. Lumping them together is the single biggest reason budgets miss. Split every cost line into three bins:
- Variable. Scales with volume. Food cost, product cost, transaction fees. Model as a % of revenue (or per unit).
- Step. Adds a block of cost at a threshold. A second assistant when chair utilization exceeds X. An additional truck when dispatch fill exceeds Y. Model as
IF volume > threshold then add block. - Fixed. Rent, base management salary, software. Model as a flat monthly number with an annual escalator.
Labor is usually 60 to 80% of cost in a service business. It is almost never purely variable. Treat it as step cost with small variable overlays (overtime, bonus). Treating labor as a flat percent of revenue is the most common modeling error I see, and it is the error that drives the worst forecast miss.
Step 3: Make branches first-class citizens
The single feature most spreadsheet budgets lack is a clean way to fork branches. You want each branch to inherit the company-level assumptions (food cost target, labor standard, overhead allocation) and override the ones that are genuinely different (lease, local wage, payer mix).
The right mental model is a tree. Company at the root. Region under that. Branch under that. Assumptions cascade down; overrides travel up in consolidation. If your tool cannot do this, you will end up with one workbook per branch and no way to consolidate cleanly. This is exactly the problem we built Aziell to solve — see the branch benchmarking guide for how a clean branch tree enables comparison.
Step 4: Build the forecast layer on top
A driver-based budget is frozen at the start of the year. A driver-based forecast updates monthly with actuals and pushes drivers forward. You need both. The budget is the accountability line. The forecast is the steering wheel.
In practice: each month you close, you replace the budgeted driver with the actual driver for that month, and you re-project the next twelve months based on the new run rate. Some teams call this a rolling 18 (6 actual + 12 forecast). For an extended treatment of when to use budget vs. forecast vs. both, read rolling forecasts vs. annual budgets.
Step 5: Instrument for accountability
A budget without an owner per driver is a wall decoration. For each of your three revenue drivers and your top three cost drivers, assign exactly one person. Capacity usually belongs to the owner or real-estate lead. Utilization belongs to the ops lead. Average ticket belongs to the revenue lead or director of marketing. Labor cost belongs to the GM of the branch.
When the plan misses, you do not ask “why was EBITDA short.” You ask “which driver drifted, and who owns it.” That is the entire point of driver-based planning: it converts a finance artifact into an operating system.
Common traps (and how to avoid them)
Trap 1: Modeling too many drivers
If your driver tree has more than 15 input cells, no one will maintain it. Cut aggressively. The three-revenue-driver structure above is sufficient for 90%+ of multi-location service businesses.
Trap 2: Benchmarking drivers against external averages
Industry benchmarks are noisy. Your best branch is a better benchmark than any industry report. If branch 3 hits 78% chair utilization, the relevant question for branch 7 is not “what does the ADA say about the median?” It is “what is branch 3 doing that branch 7 is not.”
Trap 3: Skipping the scenario layer
A single-scenario budget is a point estimate. Reality is a distribution. At minimum, build three scenarios — base, upside, downside — by perturbing your three revenue drivers ±10%. The downside scenario is the one that prevents unforced errors in hiring and lease commitments. Our scenario planning guide walks through a three-scenario build end to end.
Where to go from here
If you have never built a driver-based budget, start with one branch, three drivers, and a 12-month horizon. Get the math right on one location, then fork it to the others. Once the skeleton is stable, layer in the forecast and the scenarios. Do not build all three at once; it will collapse under its own weight.
If you want the mechanical implementation — how to actually wire this up in Aziell so it stays alive after you close the laptop — the getting-started guide walks through a full setup in under 30 minutes.
The Aziell CFO Desk is a collective byline for posts covering driver-based planning, capital-stack optimization, and operating-level scenario work. Posts under this byline draw on the day-to-day practice of fractional CFOs serving multi-location operators; every post is reviewed by at least one practicing CFO and one member of Aziell's product team before it ships. Individual contributor names appear on posts they specifically authored when that contributor is a public voice.
See the math run on your own books.
Connect QuickBooks, map your branches, and let the CFO Copilot surface your first recommendation set overnight.