How Much Corporate Overhead Should Your Restaurant Group Carry? Run the Office Like a Unit

Sit in on your own Tuesday ops call. Unit by unit, the GMs answer for everything: prime cost up sixty basis points, a soft Friday, overtime on the line. The questions are specific, the numbers are a week old, and nobody leaves the call confused about what they owe. Then the call ends — and the one operation that was never reviewed is the most expensive one in the group.

Picture an eight-unit group doing $32 million in revenue with a 6 percent corporate load. The office above the stores costs $1.9 million a year — more than the four-wall profit of the best unit in the group. It has no weekly flash report. It is never ranked against anything. Nobody's bonus moves when it runs over. We see this constantly in groups between five and twenty units: the stores are run hard, and the office above them is run on trust.

We've written before about the mechanics of allocating corporate overhead — pick a driver that reflects consumption, show it below the four-wall line, keep it off the GM bonus. All of that still stands. This article is about the pool itself, because clean allocation of a bloated pool is just a well-organized leak.

Allocation is how the office hides

Keeping allocated G&A out of four-wall results protects unit economics and manager bonuses, and it is the right call. It also has a side effect nobody prices: it converts one $1.9 million decision into eight footnotes of $240,000 each. No GM fights an allocation that sits below their bonus line. The controller books it mechanically. The accountant audits the method, not the amount. The only person left with standing to challenge the pool is the owner — and every line in it is a person the owner hired, usually to get a task off their own desk.

That is the actual trap, and it is an agency problem rather than an accounting one. Unit spending is policed weekly, by named people with their compensation at stake. Corporate spending is policed by its own author. Each addition was individually reasonable — scheduling chaos produced an ops director, vendor sprawl produced a purchasing manager, a bad quarter of marketing produced an in-house marketer — and none of them ever had to pass the test a store passes every single week.

The percentage will flatter you while the pool gets fatter

Owners who track corporate at all usually track G&A as a percentage of sales, and the percentage is a poor guard dog. Open unit nine and the ratio falls even in the quarter corporate adds two salaries. A group growing revenue 10 percent a year can watch its G&A percentage drift down while the absolute pool compounds faster than any store's budget would ever be allowed to.

So track a second number: G&A dollars per unit supported, with three years of trend beside it. A pool that went from $180,000 per unit to $240,000 per unit is telling you something the percentage conceals — and when you ask what changed, the answer is usually a version of "we're building infrastructure for twenty units." Sometimes that's true. Often it is the sentence that retroactively justifies hires nobody scoped.

The backdrop makes slack above the stores expensive right now. The National Restaurant Association's 2026 State of the Restaurant Industry reports that 42 percent of operators said their businesses were not profitable in 2025, with food costs running more than 35 percent above pre-pandemic levels on federal producer price data. At compressed four-wall margins, a $1.9 million office is not funded by eight units equally; it is funded, in practice, by your two or three best stores. Which is also why a fat pool makes a consolidated P&L even more misleading — the blended statement nets strong stores, weak stores, and corporate into one number that nobody can act on.

Give corporate the review a store gets

The fix is a governance change, and it costs almost nothing: treat the corporate office as the group's extra unit, with the discipline that implies.

Give it a monthly P&L of its own, reviewed on the same calendar as the stores — not buried in the consolidation. Put a named person on the budget, and let a real part of their bonus move with it, the way a GM's does. The symmetry matters more than the amount; the point is that for the first time, someone in the room is spending their own money.

Then run the market test, service by service. Accounting, purchasing, marketing, training, HR: if your best GM could opt out and buy each one outside, would they buy yours at its allocated cost? You don't need literal chargebacks — the question alone does the work. Where the honest answer is no, you've found either a service run inefficiently or one that exists for the office's comfort rather than the stores' benefit.

Finally, map every corporate role into one of two buckets: work the units cannot do themselves — the consolidated close, the 13-week cash flow, lender reporting, the lease portfolio — or work that scale makes cheaper or better than eight separate versions of it, like purchasing. A role that fits neither bucket is usually a piece of the owner's old job, delegated and never re-examined.

Where this doesn't apply

Three honest exceptions. A deliberate build-ahead deserves its oversized pool — if units eleven and twelve have signed leases and committed capital, carrying an ops layer sized for twelve at eight units is a dated investment with a payback, and the per-unit trend will look bad on purpose. The test is whether the growth plan has signatures on it or is just a sentence.

Second, plenty of groups have the opposite problem: five or more units and no real finance function at all, with the owner approving invoices at midnight. Adding a genuine controller will raise G&A per unit, and it should. The goal is a governed pool, not a minimal one — under-building corporate is how groups end up with inventory nobody counts and prime cost nobody can state by location.

And below three or four units, skip the framework. Corporate is you, a bookkeeper, and a shared office; the market test takes thirty seconds and you already know the answer.

What we'd advise

At the next budget season, make the corporate office build its budget the way you'd make a new GM build one: zero-based, every role justified into one of the two buckets, G&A per unit trended on one page next to the percentage. Review it with the same teeth as a unit review, once a year at minimum. Our judgment: the group that does this rarely ever needs the dramatic corporate layoff, because creep gets caught at one salary instead of five. The group that never does it is usually about two years from a painful one — announced, inevitably, in the same meeting where the GMs get asked to tighten labor again.

Visions Alliance provides fractional CFO leadership to private clubs, hospitality businesses, and owner-led companies across South Florida.