What food cost percentage actually measures
Food cost percentage answers one question: of every dollar a guest spent on food, how many cents did the food cost you? At 30%, thirty cents of each sales dollar bought ingredients and seventy cents were left to pay cooks, rent, gas, insurance and, eventually, you.
The number is a ratio of two things that must describe the same period. The numerator is not what you bought — it is what you used. Buying four cases of chicken on the last Friday of the month does not make the month expensive if three cases are still sitting in the walk-in on Sunday night. That is exactly what the two inventory counts do: they convert purchases into usage by adding what you started with and removing what you still hold.
The denominator is net food revenue. It excludes beverage, sales tax, service charges and the value of comped checks. Bar sales carry a completely different cost structure — a 20% pour cost against a 30% food cost — so blending them produces a figure that hides both.
Two subtractions distinguish a real calculation from a rough one. Food you send to the bar for garnishes, infusions or a bar-snack menu is a cost of the beverage department, not the kitchen; food eaten by staff or given away is a labour or marketing expense. Both leave the kitchen without generating a food sale, and both belong out of the numerator if you want a percentage that reflects how the line is cooking.
The formula, term by term
Start with the identity that governs any stock account. Whatever you had, plus whatever you brought in, minus whatever you still have, is what left. In a kitchen the things that leave are sold as food, transferred to another department, or eaten by staff:
COGS = opening inventory + purchases − closing inventory − transfers out − staff meals
Opening inventory is last period's closing count, unchanged. If you re-price it, you have quietly moved cost between two periods.
Purchases are invoices dated inside the period on an accrual basis — the delivery date governs, not the payment date. Credits for short deliveries and returns belong here as negatives.
Closing inventory is a physical count, extended at the price you paid, of everything edible on the premises: dry store, walk-in, freezer, line, prep fridges. Half-used cases count. Stock you have prepped into a sauce still counts, at ingredient cost.
Transfers net out. Food sent to the bar reduces kitchen COGS; limes and juice sent from the bar to the kitchen increase it. Enter the net figure, negative if more arrived than left.
Staff meals and comps come out at cost, never at menu price. A $28 steak that a manager comped cost you perhaps $8; subtracting $28 would understate food cost by $20 and flatter the kitchen for a decision the front of house made.
Divide the result by food sales and multiply by 100. That is the whole calculation. The difficulty of food costing has never been the arithmetic — it is getting two honest counts.
Worked example: one week at a 90-seat bistro
The kitchen counted $18,500 of stock on Monday morning before the first delivery. Across the week the invoices totalled $42,000. The following Monday, counting at the same hour and before any delivery, the stock came to $19,200. The bar took $1,200 of food at cost for its snack menu, and staff meals cost $850. Net food sales for the week were $132,000. The target is 30%.
- Add what was available. $18,500 + $42,000 = $60,500.
- Remove what is still on hand. $60,500 − $19,200 = $41,300.
- Remove what left without a food sale. $41,300 − $1,200 − $850 = $39,250. That is cost of goods sold.
- Divide by food sales. $39,250 ÷ $132,000 = 0.29735.
- Express as a percentage. 0.29735 × 100 = 29.73%.
- Compare with target in dollars. A 30% target allows 0.30 × $132,000 = $39,600. Actual was $39,250, so the week came in $350 under the allowance — a variance of −0.27 percentage points.
- Gross profit on food. $132,000 − $39,250 = $92,750, or 70.27% of food sales.
- Inventory turns. Average inventory is ($18,500 + $19,200) ÷ 2 = $18,850, so $39,250 ÷ $18,850 = 2.08 turns in the week.
Note how small the percentage variance looks and how ordinary the dollar variance is. A quarter of a point sounds like rounding; on this volume it is $350 a week, or roughly $18,000 a year. That is the reason to convert every percentage into dollars before you decide whether it matters.
How to read the result
There is no universally correct food cost percentage, and chasing a low one is a good way to go out of business. A steakhouse at 38% and a pizzeria at 22% can be equally healthy, because the steakhouse sells a $54 entrée with a $20 plate cost and the pizzeria sells an $18 pizza with a $4 plate cost. The steakhouse keeps $34 of gross profit per cover, the pizzeria $14. Gross profit dollars pay your bills; percentages do not.
What you can judge is the gap between your actual percentage and your ideal percentage — the figure your costed recipes say the mix you actually sold should have cost. That comparison is the only benchmark that controls for your menu, your prices and your market. A gap of about one point is normal noise from counting error and portioning drift; a persistent gap of three points or more is money leaving the building through over-portioning, waste, spoilage, theft, or prices that have not moved since the last supplier increase.
Convert the gap to dollars immediately. One percentage point on $1,000,000 of annual food sales is $10,000 — which is the arithmetic that decides whether a scale on the sauté station is worth buying.
Watch the trend more than the level. Weekly figures bounce because inventory counts contain error and because a large delivery landing on either side of the count date shifts cost between periods. A four-week rolling average of the same calculation strips most of that out. If you only ever look at one number, look at the four-week average against your ideal cost.
Food cost planning ranges by segment
| Segment | Typical planning range | Why it sits there |
|---|---|---|
| Pizzeria | 20–28% | Flour, tomato and cheese are cheap relative to the price a whole pizza carries. |
| Quick service, sandwiches | 28–33% | Low labour per cover permits a higher food cost while still covering overhead. |
| Casual full service | 28–35% | The broad middle: mixed protein and produce, moderate labour. |
| Steakhouse, fine dining | 33–40% | Expensive centre-of-plate protein, but very large gross profit dollars per cover. |
| Café and bakery (food only) | 25–35% | Cheap ingredients, but high waste on anything baked fresh daily. |
| Catering and events | 25–35% | Quantities are known in advance, so waste is controllable and yields are planned. |
| Bar food programme | 30–35% | Often deliberately generous because the beverage margin carries the check. |
Ranges commonly taught in hospitality costing courses and used as budget starting points. They are not survey results. Your own costed ideal food cost, weighted by the mix you actually sell, always outranks a published benchmark.
Seven ways this number comes out wrong
- Counting at different points in the week. Count before Friday's delivery one week and after it the next, and the two periods swap several thousand dollars of cost. Fix the count time and never move it.
- Leaving beverage revenue in the denominator. This is the single most common error and it always makes food cost look better than it is. Split the revenue at the POS level.
- Valuing comps at menu price. A comp costs you what the ingredients cost, not what the guest would have paid. Menu-price comping understates food cost badly.
- Ignoring the walk-in floor and the line. Prepped sauces, open cases and mise en place on the line are inventory. Skip them and your closing count is low, which inflates the period you are closing and deflates the next one.
- Mixing invoice dates and payment dates. Purchases belong to the period the food arrived. Paying a March invoice in April does not move March's cost.
- Comparing your percentage with someone else's. Different menus, different prices, different markets. Compare against your own ideal cost and your own trend.
- Reacting to a single week. Counting error alone moves a weekly figure by a point in either direction. Act on a four-week trend, not on one bad Monday.
Where the definitions come from
The account structure this calculator follows — cost of sales built from opening inventory, purchases and closing inventory, with transfers between departments treated as adjustments, and food and beverage reported separately — is the one set out in the Uniform System of Accounts for Restaurants (USAR), published by the National Restaurant Association and taught throughout hospitality accounting. Using the USAR layout matters if you ever want to compare your statement to anyone else's, or hand it to a lender.
USAR also fixes the definition of prime cost as cost of sales plus total labour, which is the figure most operators actually manage. See the restaurant prime cost calculator for that layer.
Ideal food cost, actual food cost, and what to do with the gap
Your actual food cost is what this calculator produces from counts and invoices. Your ideal food cost is what the same period should have cost if every plate had been made to the recipe and nothing had been wasted: take each dish sold from the POS mix report, multiply by its costed plate cost, and divide the total by the same food sales. Build those plate costs with the recipe cost per serving calculator.
The gap between the two is where management happens. Ordered roughly by how often each turns out to be the cause: portioning drift on the highest-volume dishes; yield loss on trimmed proteins and produce that the recipe assumed away; spoilage from over-ordering; and only then theft. Test yields on your top five ingredients with the food yield percentage calculator before you accuse anyone of anything — a recipe written against as-purchased weight when the cook works from edible-portion weight will open a two-point gap all by itself.
If the ideal cost itself is too high, the answer is pricing rather than policing. Reprice the offending dishes with the menu price calculator, or reweight the menu toward what already sells profitably. And remember that food cost is only one leg of prime cost: a kitchen at 33% food cost with disciplined labour is in far better shape than one at 28% that needs three extra prep hours a day to get there.
Key terms
- Cost of goods sold (COGS)
- The cost of the food actually used in a period, derived from two inventory counts and the purchases between them. Not the same as what you bought.
- Ideal (theoretical) food cost
- What the period should have cost if every plate matched its recipe: the POS sales mix valued at costed plate costs, divided by food sales.
- Transfer
- Product moved between departments at cost, most often food to the bar or juice and citrus from the bar to the kitchen. Transfers keep each department's cost percentage honest.
- Extension
- Multiplying each counted quantity by its unit price to get an inventory value. Errors in extension are the most common source of a strange weekly percentage.
- Prime cost
- Cost of sales plus total labour including taxes and benefits, expressed as a percentage of total sales. The figure most operators manage weekly.
- Inventory turns
- COGS divided by average inventory for the period. It tells you how many times you sold through your stock; low turns tie up cash and raise spoilage risk.