How to calculate food cost percentage for a period
Per-dish food cost tells you what a plate should cost. Period food cost tells you what the kitchen actually used, and the two are rarely the same. The period figure comes from a stock count: whatever food you started with, plus everything you bought, minus whatever is left, is what went out of the door — on plates, in the bin, as staff meals or in someone's bag. Divide that by food sales and you have your actual food cost percentage.
COGS (actual usage) = opening inventory + purchases − closing inventory − staff meals & comps Food cost % = COGS ÷ food sales × 100 Variance = actual food cost % − theoretical food cost % Variance in money = variance points × (sales ÷ 100)Example: $5,800 of stock at the start, $13,100 of invoices, $5,600 at the end and $400 of staff meals gives $12,900 of usage. On $38,500 of food sales that is 33.5%. If your recipes, multiplied by what the POS says you sold, add up to 30%, the 3.5-point gap is $1,350 that left the building without a guest paying for it — in one month.
What a healthy food cost percentage looks like
| Concept | Rule-of-thumb range |
|---|---|
| Quick-service / fast casual | 25–30% |
| Pizza | 25–30% |
| Casual dining | 28–32% |
| Full-service restaurant | 28–35% |
| Fine dining | 30–38% |
| Café / bakery | 20–30% |
| Bar (food) | 20–25% |
Against real results rather than targets: the National Restaurant Association's 2025 Restaurant Operations Data Abstract (2024 data) reports a median food and non-alcoholic beverage cost of 32.0% of sales for full-service restaurants and 32.4% for limited-service, where paper goods are counted in. Published vendor benchmarks differ from each other and from these ranges: Restaurant365 (2026) gives quick service 27–32%, fast casual 26–30%, casual dining 28–33%, fine dining 28–35% and pizza 22–28%, while Toast (2025) gives quick service 28–32%, casual dining 32–36% and fine dining 28–32%. The table uses the rules of thumb most operators work to; the café, bakery and bar-food rows have no association or vendor source.
These are ranges, not targets. A steakhouse at 36% with a $60 check can be far more profitable than a café at 24%; what matters is the gross profit in money and whether labour and overheads fit inside what is left. Use the range to spot drift, then use the variance to find out why.
Actual versus theoretical: where the money goes
Theoretical usage is what your costed recipes say should have been used for the dishes the POS recorded. Actual usage is the count. The gap has only a few causes:
Waste and spoilage
Over-ordering perishables, poor rotation, trim that could have been used. Days of stock on hand (in the results) is the early warning: more than 7–10 days of food on the shelf usually means spoilage is coming.
Over-portioning
A ladle that is 10% generous on a sauce used in half the menu is a point of food cost on its own. Scales and portion tools fix it; costed recipes with portion weights make it visible.
Theft and unrecorded food
Staff meals and comps that are not written down look exactly like theft in the numbers. Record them so the real variance stands out.
Stale recipe costs
The most common cause is also the least visible: the theoretical number is wrong because recipes were costed on last year's invoices. When supplier prices rise 8% and the recipe file does not change, "variance" appears that is really just inflation. Re-costing recipes from current invoices removes it — and shows which dishes need a price change.
How often to run this
Weekly counts of the expensive items (proteins, dairy, oil) and a full monthly count is the usual rhythm. The shorter the period, the sooner you see a problem and the smaller it is when you find it.
Figures last reviewed .
Sources: NRA, 2025 Restaurant Operations Data Abstract (food cost), Restaurant365, Food cost benchmarks by concept (2026), Toast, How to price food to sell (2025).