Prime Cost Explained

By Marji · Published August 10, 2026 · Updated October 9, 2026 · 4 min read

Add your COGS and your fully loaded labour for the same week, then divide by that week's sales. A bistro at $14,880 of COGS and $16,320 of labour on $48,000 of sales runs 65.0%. Leaving out payroll tax, workers' comp and benefits — $3,120 in the week below — would understate prime cost by 6.5 points.

Formula

Prime cost % = (COGS + Total labour) ÷ Total sales × 100

Total labour means fully loaded: wages, salaries, payroll taxes, workers' comp and benefits. Not just gross wages. The labor cost guide lists what to include, tips and service charges among them.

Worked example · One week at a 90-seat bistro

Food sales: $38,400
Beverage sales: $9,600
Total sales: $48,000
COGS — food + beverage: $14,880
Wages and salaries: $13,200
Payroll tax, comp, benefits: $3,120
Prime cost in dollars: $31,200
Prime cost: 65.0%

Exactly at the 65% rule of thumb. Whether that is too high depends on what occupancy, other expenses and profit need — the section below derives it.

Why food cost alone lies to you

Food cost and labour cost trade against each other constantly, and looking at either alone hides the trade completely.

Hypothetical: the same kitchen, two purchasing decisions
	Butcher in house	Buy pre-portioned
Food cost %	29%	34%
Labour cost %	34%	28%
Prime cost	63%	62%
Prep hours per week	22	6

Read the food cost column and buying pre-portioned looks like a five-point catastrophe. Read prime cost and it's a one-point improvement that also gives you back sixteen prep hours. This is the single best argument for the combined number.

What belongs in each half

In COGS	In labour	In neither
Food inventory consumed	Hourly wages	Rent and occupancy
Beverage inventory consumed	Salaried management	Utilities
Packaging that goes out with the order*	Employer payroll taxes	Insurance
	Workers' compensation	Marketing
	Benefits and paid leave	Repairs and maintenance
	Contract labour	Third-party delivery commission

*Packaging is a judgement call — some operators put it in COGS, some in supplies. Either is defensible. What isn't defensible is moving it between categories, because that makes your trend line meaningless and the trend is the entire point.

The 65% rule of thumb, and your own ceiling

The 65% figure is a practitioners' rule of thumb — RestaurantOwner.com calls it "the generally accepted rule of thumb" for table-service restaurants — not a measured average. The National Restaurant Association's own accounting manual, the Uniform System of Accounts for Restaurants (8th edition, 2012, p. 107), says prime cost "typically consumes 60 percent to 65 percent of every restaurant sales dollar" — also a rule of thumb, not survey data. The one published survey median found is for limited-service restaurants: in its 2025 data abstract, the Association reported that "prime costs including food, beverage, and labor were a median of 65 cents of every sales dollar in the limited-service segment" (press release, August 20, 2025). For full-service restaurants, its public summaries give 2024 medians only for the largest line in each half: 32.0% of sales for food and non-alcohol beverages (alcohol cost is a separate line) and 36.5% for salaries, wages and benefits in full-service restaurants. Medians of separate lines cannot be added into a median prime cost, though. Derive your own ceiling from the bottom of the P&L up:

Worked example · Your actual ceiling (hypothetical P&L)

Occupancy — rent, utilities, insurance: 9%
Other operating expenses: 12%
Target owner profit: 8%
Everything below prime cost: 29%
Your prime cost ceiling: 71%

A restaurant with 6% occupancy has room for a 74% prime cost. One paying 14% needs it under 66%. This is why borrowed benchmarks mislead — and why a low-rent food truck can run a food cost that would sink a downtown bistro.

Both halves of that are a minute of arithmetic. The prime cost calculator does the weekly number with labour loaded properly; the break-even calculator derives the ceiling from your own fixed costs rather than from an industry average.

How to run it weekly

Same period boundaries for everything. Sales, inventory count and payroll all covering the same days. Mismatched weeks produce phantom swings and send you chasing nothing.
Count inventory weekly. Only what's expensive and what moves — proteins, seafood, dairy, alcohol. A partial count keeps it short enough to do every week.
Use loaded labour. Gross wages alone understate it — by 5.3 to 7.3 points in the labor cost guide's example — which is enough to make a problem invisible.
Watch the trend, not the week. A holiday, a large private event, or a month with three pay periods will each throw one week. Three weeks in a row is a signal.
Split it when it moves. A prime cost that jumped two points tells you something is wrong; the food/labour split tells you where.

Does prime cost include rent?

No. Rent, utilities, insurance and marketing sit below prime on the P&L. Prime cost is deliberately restricted to the two lines you can change this week — which is exactly what makes it actionable.

My prime cost is 58% — should I be worried?

Check it isn't the result of understaffing or buying below the quality your menu implies. Consistently very low prime cost with flat or declining sales sometimes means you're under-investing in the thing guests came for.

How do I calculate prime cost without weekly inventory?

You can approximate it using purchases instead of COGS, but only if your inventory level is genuinely stable week to week. It will be wrong in any week you stock up or draw down, which is most weeks. Use it as a rough signal, not a number you act on.

The weekly number, without the weekly spreadsheet.

Marji extracts invoice lines for your review. Confirming a checked line records its ingredient cost for recipe recosting. Review the recipe's quantities, units, yields and completeness too. Period food cost and prime cost still require your inventory, sales and labor records.

Start free →