Every restaurant operator has heard someone say "food cost should be thirty percent." That sentence is not wrong, but it is not right either. It depends on what you are running. A steakhouse doing $60 tickets and a burger counter doing $12 tickets are both restaurants, and they share almost none of the same P&L math.

What follows are the benchmarks I actually use when I walk into a group. Sixteen years, $54M in combined operating scope, everything from franchise units inside Walmart and Sam's Club to a Michelin-recognized Bay Area group. These are the ranges healthy units hit. If your unit is inside the range on every line, the P&L is not the problem. If you are outside the range on two or more lines, you know where to look.

Segment first: which one are you actually running?

Pick the segment that matches your service model and ticket average, not the one you wish you were.

Full-service casual

Table service, server takes the order, food runs out from the kitchen. Ticket average $18 to $45 per cover. Applebee's on the low end, Cheesecake Factory on the high end, most independents in the middle. Zareen's, the group I ran in the Bay Area, sits in this bucket at a $22 to $28 average.

Quick-service restaurant (QSR)

Counter order, food handed over the counter, minimal or no table service. Ticket average $8 to $15 per cover. McDonald's, Chick-fil-A, Taco Bell.

Fast-casual

Counter order but higher ingredient quality and slightly longer prep. Ticket average $12 to $22 per cover. Chipotle, Sweetgreen, Cava. The Hana Group units I ran inside Walmart, Sam's Club, Whole Foods, and Target sit inside this segment operationally, even though the retail context is unusual.

The benchmark ranges, line by line

All figures as a percent of net sales unless otherwise noted. Net sales means gross sales less comps, discounts, and tax.

Healthy benchmark ranges, percent of net sales Food cost Labor cost Prime cost Rent Controllable EBITDA (unit) 0% 20% 40% 60% 80% 28-32% 28-32% 58-62% 6-10% 68-74% 12-18% Ranges shown for full-service casual segment

Fig. 1 · Full-service casual benchmark ranges.

Food cost

  • Full-service casual: 28 to 32 percent
  • QSR: 28 to 33 percent (packaging is a bigger share)
  • Fast-casual: 28 to 34 percent (ingredient quality is usually higher)
  • Fine dining: 32 to 38 percent (the plate costs more)

If your food cost is 4 or more points above the top of the range, it is almost always a counting problem, not a buying problem. Something is not being logged: waste, comps, portion drift, or theft.

Labor cost (all-in)

Hourly wages, salaried management, payroll taxes, workers comp, benefits.

  • Full-service casual: 28 to 32 percent
  • QSR: 24 to 28 percent
  • Fast-casual: 26 to 30 percent

Labor 4 points above the range is almost always a schedule that does not follow demand, not a wage problem. Look at hourly sales by day part and overlay the schedule. The gap will be obvious.

Prime cost (food + labor)

  • Full-service casual: 58 to 62 percent
  • QSR: 55 to 60 percent
  • Fast-casual: 56 to 62 percent

Prime cost is the single most important number on the P&L. It is the one the operator controls directly. Rent, insurance, and utilities are largely fixed for the life of the lease. Prime cost is the number you fight for every week.

Rent (occupancy)

  • Full-service casual: 6 to 10 percent
  • QSR: 6 to 12 percent (often higher for street-front)
  • Fast-casual: 8 to 12 percent

Rent is not really negotiable mid-lease. If your rent is above 12 percent of sales, the lease was signed against a sales projection that never showed up, and you are eating that mistake until renewal.

Controllable cost

Prime plus supplies, uniforms, small equipment, credit card fees, direct marketing, R&M, third-party delivery commissions, cleaning.

  • Full-service casual: 68 to 74 percent
  • QSR: 65 to 72 percent
  • Fast-casual: 66 to 73 percent

Controllable is the general manager's total scorecard. If the general manager can name their controllable cost from memory within a percentage point, they are running the unit. If they cannot, the unit is running them.

EBITDA (at the unit level, before G&A)

  • Full-service casual: 12 to 18 percent
  • QSR: 15 to 22 percent
  • Fast-casual: 13 to 20 percent

At the enterprise level, after corporate G&A, these numbers drop 4 to 6 points. A five-unit group hitting 16 percent unit EBITDA will typically net 10 to 12 percent enterprise EBITDA after covering shared services.

Segment comparison at a glance

Segment comparison: healthy midpoint, percent of sales Full-service QSR Fast-casual Food Labor Rent EBITDA 30% 30% 8% 15% 30% 26% 9% 18% 31% 28% 10% 16% Midpoint numbers only. Actual healthy ranges span 4 to 6 points on each line.

Fig. 2 · Segment comparison, midpoint of healthy ranges.

What the ranges actually mean at different scales

A benchmark range is not the same across a five-unit group as it is inside a single independent. Inside a group, you get the benefit of shared purchasing, shared training, and pooled scheduling talent. Numbers should sit tighter to the middle of the range or slightly better. Inside a single independent restaurant, the ranges are wider because you cannot spread the cost of a bad week across other units, and you are more exposed to any single vendor or hire.

Two adjustments I make in practice:

  • Independents: allow yourself the full width of the healthy range on food and labor. A 32 percent food cost on an independent doing $2M a year is fine. That same 32 percent on a five-unit group doing $20M is a warning, because at scale you should be getting 1 to 2 points of vendor concession.
  • Multi-unit groups: the benchmark tightens as you add units. Three units should be inside the range. Five units should be at the midpoint or better. Ten units and above should be at the tighter end of the range, because the operational discipline is supposed to compound.

Ticket band also matters. A full-service unit doing $18 tickets and one doing $42 tickets are both "full-service casual" but the food cost math is different. The $18 concept has less absolute dollars per plate to absorb ingredient inflation, so a 4 percent supplier hike hits harder. The $42 concept has more headroom on the plate but more expensive ingredients that spike faster.

Below the controllable line: the ones you cannot fight for

Everything below controllable cost is where operators lose room to move. These are the line items that get set once (rent, insurance, franchise fees, financing) and mostly do not move week to week. If they are in range at signing, they stay in range. If they are out of range, you are usually stuck with the mistake for the life of the lease or the loan.

Occupancy total (rent plus CAM plus utilities pass-through)

  • Full-service casual: 8 to 12 percent all-in
  • QSR: 8 to 14 percent
  • Fast-casual: 10 to 14 percent

Insurance (property, liability, workers comp not in labor)

  • All segments: 1.5 to 3.5 percent of sales

Franchise fees (royalty plus marketing fund)

  • Typical franchise contract: 5 to 8 percent of sales combined
  • Higher-end brands: up to 10 percent

Add franchise fees, occupancy, insurance, and debt service to controllable cost and you get to total operating cost. What is left is EBIT, and after debt and depreciation you get to net. On a healthy franchise unit doing 16 percent unit EBITDA, expect net income after fees, debt, and depreciation to land at 6 to 10 percent. On an independent with no franchise royalty but higher marketing spend, the shape is similar.

The two per-cover numbers that reveal what percentages hide

Percent-of-sales lines are useful and they also lie. A ticket average lift can hide a rising cost per plate. A price increase can hide a drop in productivity. The two numbers below cut through that noise.

Sales per labor hour

Net sales divided by total labor hours worked in the same period, including salaried hours (convert to hours worked, usually 45 to 50 per week per salaried manager).

  • Full-service casual: $85 to $110 per labor hour
  • QSR: $90 to $130 per labor hour
  • Fast-casual: $95 to $140 per labor hour

This single number tells you more about scheduling quality than the labor percentage does. Labor cost as a percent of sales is distorted by the ticket average. Sales per labor hour is not, because it is a productivity number. If your sales per labor hour is climbing week over week, you are getting more sales out of every scheduled hour, which is the actual goal.

Cost per cover

Total food cost divided by covers served in the same period.

  • Full-service casual: $6 to $10 per cover
  • QSR: $2.50 to $4.50 per cover
  • Fast-casual: $4 to $7 per cover

Cost per cover reveals leaks the food cost percentage hides. If you raise prices and your food cost percentage stays flat, the operator's instinct is to say the unit is fine. But if cost per cover crept up over the same period, ingredient cost per plate rose too, and the price increase just covered it. That is a warning, not a win.

Percent-of-sales lines tell you where you are. Per-cover numbers tell you where you are going. Track both. The gap between them is where the diagnostic sits.

How to use these numbers on a Monday morning

Pull your last 90-day trailing P&L on one page. Roll up food cost, labor cost, prime cost, rent, controllable, and EBITDA as percent of net sales. Add the two per-cover numbers underneath.

Then, line by line, plot your number against the segment range above. Highlight anything more than 2 points off the healthy midpoint. That is your work list, ranked by dollar impact, not by percentage gap.

Example, on a full-service unit doing $200K a week:

Food cost:      33.5% (bench 30%) → gap 3.5pts × $200K × 52 = $364K/yr
Labor cost:     34.0% (bench 30%) → gap 4.0pts × $200K × 52 = $416K/yr
Prime cost:     67.5% (bench 60%) → composite of the above
Rent:            8.2% (bench 8%)  → in range, no action
Controllable:   77.3% (bench 71%) → gap 6.3pts, mirrors prime
EBITDA:          9.1% (bench 15%) → gap 5.9pts × $200K × 52 = $614K/yr

This unit has $780K a year sitting in the gap on food and labor alone, and the EBITDA gap confirms that number. The work list writes itself: food cost first (the counting problem), labor second (the schedule problem). Everything else is in range.

The three ways operators misuse benchmarks

Benchmarks are a tool. Like every tool, they can be used badly. Three specific misuses I see often, and how to avoid each.

Comparing your unit to the wrong segment

A fast-casual unit doing $18 average tickets is not a QSR, and if you benchmark it against QSR numbers your labor line will look terrible when it is actually healthy. Pick the segment your unit actually is (service model plus ticket range) and use those numbers. If your concept sits between segments, average the two ranges rather than picking the more flattering one.

Optimizing for a single line at the expense of the others

Cutting food cost by 3 points is easy if you also degrade labor by 2 points and controllable by another 1 point through spec changes that push work back onto the kitchen team. The individual line looks great and the P&L is worse. Prime cost is the composite for a reason. Watch the composite, not the individual line, when you are moving something.

Treating a benchmark hit as the finish line

Hitting the middle of the range is not winning. The middle of the range is normal. The operators who compound value over five and ten years land in the tighter end of the range on prime cost and controllable cost, which usually means 2 to 4 points better than the segment average on EBITDA. The gap between average and top-quartile at unit level is often the difference between an enterprise worth a 3x EBITDA multiple and one worth a 6x. Same segment. Same address. Different operating discipline.

The point

Benchmarks are not judgment. They are a common language for operators who have to know quickly where their unit is bleeding and where it is not. Full-service casual, QSR, and fast-casual each have their own healthy shapes. Prime cost is the number that decides most restaurants. Sales per labor hour and cost per cover reveal what the percentages hide.

Pull the numbers, plot them against the range, work the biggest gap first. Do it Monday morning, in ten minutes, with the general manager in the room. Do it every week for a quarter and the P&L stops surprising you.

That is the whole method.