Restaurant Prime Cost Benchmarks: Why Your Group Average Is Hiding Your Problem Store

A five-unit fast casual group in Chicago sent us a year-to-date P&L last winter with a note that said, more or less, we're fine. On the page they were: revenue up 7%, no covenant issues, and a blended prime cost of 60.5% — comfortably inside every published restaurant prime cost benchmark you can find. The founder wanted help with a lease renewal, not a diagnosis. It took two hours to split the statement into five columns and find unit four running in the high 60s since April, eleven points worse than the best store on nearly identical volume. Nobody had misreported anything. The average had quietly absorbed a store that lost money every week for seven months.

That is the problem with prime cost as most groups report it. Below: what belongs in the calculation, benchmark ranges by service model, how to read theoretical versus actual food cost, and when chasing prime cost is the wrong fight entirely.

What is a restaurant prime cost benchmark, and what belongs in the formula?

Prime cost is cost of goods sold plus total labor, divided by net sales, and "total labor" means fully burdened labor, not the wage line on your payroll register. The restaurant prime cost formula is simple; the discipline underneath it is where groups lose comparability. A restaurant prime cost benchmark is worth only as much as the definition behind it.

Prime Cost % = (COGS + Total Labor Cost) ÷ Net Sales

COGS is food plus beverage, and for fast casual and quick service it usually includes paper and packaging. Whether paper sits in COGS or operating supplies is your call; what matters is that the treatment is identical in every unit and matches the benchmark you compare against.

Total labor is the half that gets truncated. All of this belongs:

Hourly wages, front and back of house, including overtime and premium pay

Salaried management wages, allocated to the unit they run

Employer payroll taxes: the 7.65% employer share of Social Security and Medicare, plus FUTA and a state unemployment rate that follows your own claims history (IRS Topic 751)

Workers' compensation premium, no rounding error at restaurant class codes

Health, dental, vision, and any employer retirement match

PTO, vacation, and bonus accruals, recorded in the period earned

Contract and agency labor

Leaving out taxes, workers' comp, and benefits is the single most common error in restaurant prime cost reporting, and it is not small. If labor runs 30% of sales and your all-in burden is 12% to 18% of gross wages, the omission understates prime cost by 3.6 to 5.4 points. A group that thinks it runs 58% is running 62%.

If you take a tip credit, compare yourself only to operators in the same wage structure (U.S. Department of Labor, FLSA). Locking these definitions into the chart of accounts is unglamorous work and the foundation of any useful multi-unit restaurant financial reporting package.

Prime cost benchmark ranges by service model

Prime cost targets sit roughly between 55% and 65% of net sales for most models, but the useful answer is a range by service model, not a single number.

Treat these as directional. Any restaurant prime cost benchmark moves with market, menu, and wage environment. A unit in a $20-plus minimum wage market runs labor several points above a suburban Southeast unit with an identical operation, and that gap is not a management failure. The target is a range for a second reason: the two halves trade against each other. A scratch kitchen buys lower food cost with higher labor; a heat-and-serve model does the reverse. Two units at 60%, one at 32/28 and one at 26/34, can both be right. Watch the sum, the trend, and the gap to your own best unit.

Why your group average is not a benchmark

A blended prime cost tells you almost nothing about whether any store is healthy, because averaging is the one operation that destroys the signal you need. Here is that Chicago group's year, anonymized and rounded.

The blended number looks fine. Unit four sits 8.3 points above it, which on $2.6 million of sales is roughly $216,000 a year of margin that is not there. The average lies in both directions: strip unit four out and the other four stores blend to 58.8% — a strong operation reporting itself as average.

The fix is not complicated. Report prime cost by unit, every period, ranked worst to best, with two added columns: gap to the best unit, and the dollar impact of closing it at current volume. Ranking creates the conversation. Dollarizing the gap decides which store your director of ops visits Monday.

Theoretical vs. actual food cost: what the variance actually tells you

Theoretical food cost is what your recipes and actual sales mix say you should have spent. Actual is what you did spend. The variance between them is the only number that points at a cause; food cost percentage alone never will.

Read the size first. Under 1 point is noise: count timing, rounding, ordinary trim. One to two points is worth pulling a category to item-level detail. Three points or more is structural, meaning something is broken rather than drifting.

Then read the shape:

Portioning. Variance concentrated in a few high-cost items, consistent week over week, worse on shifts with newer staff. Scale-and-spec problem.

Waste and spoilage. Variance spikes on prep-heavy items and worsens in slow weeks. A par level problem, not a people problem.

Theft. Variance in the highest-value, easiest-to-move categories: liquor, center-of-plate protein, to-go packaging. Clean counts, no waste log, no portioning story.

Menu mix. If theoretical is calculated against actual item-level sales, mix is already neutralized and cannot be your explanation. If it isn't, you are not measuring anything.

Stale recipe costing. The most common cause, the least investigated. If your recipe file carries last year's invoice prices, the variance is your purchasing, not your kitchen.

Re-cost your top 20 items quarterly at minimum, monthly for volatile proteins. Plenty of groups spend a season disciplining a kitchen over a variance an afternoon of re-costing would have explained.

Labor is a function of demand, not a percentage

Restaurant labor cost percentage is an outcome, not a control, and managing to a percentage target produces bad schedules whenever volume swings. A Tuesday lunch doing $900 on a two-person minimum crew posts an ugly percentage even when it was scheduled perfectly, because you cannot staff below one cook and one cashier. Cutting to hit the target means closing the daypart, which is a strategy decision, not a scheduling one.

Manage these instead:

Sales per labor hour, by daypart. Not a weekly average. Many full-service units land between $50 and $90 depending on check average; find yours and hold each unit to its own trend.

Guests, entrées, or transactions per labor hour for back of house, where dollars distort with pricing.

The fixed/variable split. Managers, openers, closers, and minimum crew are fixed; everything above flexes to forecast. Fixed labor dollars per week per unit is what tells you a store's break-even volume, and the same input drives a credible new restaurant unit pro forma.

Forecast accuracy. A sales forecast 15% off produces a schedule 15% off. Scheduling discipline downstream of a bad forecast is wasted effort.

Run prime cost weekly, not at period close

Prime cost reported at period close is history, and history is not a decision. Weekly is the only cadence at which the number can still change an outcome. A unit running three points over on $60,000 a week bleeds $1,800 a week; across a four-week period that is $7,200 you learn about after it has been spent.

Weekly requires five things: a fixed week-ending day used by every unit; counts on the categories that move (protein, seafood, produce, alcohol), with a full count monthly; purchases captured by invoice date; labor from scheduling and payroll with a standard burden factor; and one page per unit in the operator's hands by Tuesday.

That weekly number will not tie exactly to the P&L, because the burden factor is an estimate. Fine. Direction weekly, precision at close. If the page is impossible to assemble each week, the constraint is your reporting infrastructure, not your operators, and that is a financial reporting problem with a known fix.

When chasing prime cost is the wrong fight

Prime cost is the wrong fight when the number is already inside range and the money is going somewhere else on the P&L. Four situations where operators grind on portion control while the real problem sits untouched:

Occupancy. Rent, CAM, taxes, and insurance above roughly 10% of sales is a real estate problem. A unit at 60% prime cost and 14% occupancy is not an operations story; it is a lease negotiation, a sublease, or a closure decision.

AUV. Fixed costs need volume. A store doing $1.4 million in a box built for $2.4 million looks bad on every percentage on the page. The fight is traffic, dayparts, catering, and delivery mix, not spec cards.

Menu price architecture. Manage food cost percentage without watching contribution margin per item and you win the percentage while losing the dollars. A dish at 38% food cost carrying $14 of margin beats a 24% dish carrying $6.

Below-the-line costs. Prime cost healthy in every unit, four-wall margin healthy, and the group still short on cash usually means G&A, royalties, or debt service. That is a capital structure conversation.

Timing matters too. The National Restaurant Association's 2026 State of the Restaurant Industry report projects sales of $1.55 trillion on 1.3% real growth, with 42% of operators reporting their restaurant was not profitable and 60% reporting softer traffic. In a soft-traffic year, cutting the labor hour guests actually feel is how a prime cost win becomes a comp sales loss two quarters later.

And the honest version: sometimes prime cost is fine everywhere and the real problem is that nobody can produce a reliable unit-level P&L within ten days of close. That is not a prime cost project. That is a finance function that needs rebuilding, and the tell is usually a controller carrying work no controller should carry.

Where we land

Prime cost is the right number and the group average is the wrong lens. Get the definition complete, especially the burdened labor half, because omitting payroll taxes, workers' comp, and benefits understates the number by three to five points. Then report it by unit, weekly, ranked, with the dollar gap to your best store next to each line. The table ranges are directional; your own best unit is the honest standard. And know when to stop: when occupancy, volume, or menu architecture is the binding constraint, prime cost discipline is effort in the wrong place.

Common questions

What is a good prime cost percentage for a restaurant?

Roughly 55% to 65% of net sales for most models, with full service running higher and bar-led concepts lower. Quick service and fast casual typically land between 55% and 63%. Treat any published range as directional, since wage rates, menu design, and service model shift it by several points. Your best-performing unit is a more useful target.

What is the restaurant prime cost formula?

Prime cost equals cost of goods sold plus total labor, divided by net sales. COGS covers food, beverage, and usually paper and packaging. Total labor covers hourly and salaried wages, overtime, employer payroll taxes, workers' compensation, benefits, and accrued PTO and bonus. Use net sales, after discounts and comps.

Do payroll taxes and benefits count in prime cost?

Yes, and leaving them out is the most common error in restaurant prime cost reporting. Employer FICA, FUTA, state unemployment, workers' compensation premium, health benefits, and retirement match all belong in the labor half. Omitting them understates prime cost by roughly three to five points, enough to make a struggling unit look acceptable.

How often should we calculate prime cost?

Weekly, by unit. At period close the number is history and cannot change an outcome. A weekly page needs a fixed week-ending day, counts on high-movement inventory categories, purchases logged by invoice date, and labor with a standard burden factor applied. Reconcile to the P&L at close and explain variances above one point.

What does a gap between theoretical and actual food cost mean?

It tells you where the money went. Under one point is noise. One to two points warrants a category-level look. Three points or more is structural: portioning, waste, or theft, each with its own signature in the data. Before blaming the kitchen, check that your recipe costing reflects current invoice prices, the most frequent cause.