Food cost percentage: the formula, worked examples and how to actually control it

Ask a restaurant owner what their food cost percentage is and you usually get one of two answers: a number copied from an industry article, or a shrug. Both are expensive. Food is one of the two biggest lines in a restaurant's cost structure, roughly a third of every dollar of sales according to the National Restaurant Association's analysis of restaurant profitability, and it is the line you can influence the fastest: not by cutting quality, but by measuring it properly.

This guide covers the whole subject in plain terms: the food cost percentage formula, a worked example in dollars and one in pounds (the two are not calculated on the same tax basis, and that trap alone skews plenty of UK numbers), the difference between theoretical and actual food cost, how often to count inventory, and the honest answer to "what is a good food cost percentage", including the cases where a high one is exactly what you want.

In short:

• food cost percentage = (beginning inventory + purchases - ending inventory) ÷ food sales × 100, calculated over a defined period;
• the subtlety is inventory: purchases alone tell you what you bought, not what you used;
• theoretical food cost is what the period should have cost given your recipes and sales mix; the gap to your actual number is the variance, and that is where waste, portioning drift and shrinkage hide;
• count inventory at least monthly, weekly if you can; a number you calculate once a year is trivia, not management;
• there is no universal "good" percentage: judge dishes on gross margin in dollars or pounds, not on the ratio alone.

What is food cost percentage?

Food cost percentage is the share of your food revenue that goes back out the door as ingredient purchases. If your food cost percentage is 32%, then 32 cents of every dollar of food sales pays for the raw materials on the plate, and 68 cents remain to pay the team, the rent, the utilities and, at the end of the line, you.

Two clarifications before the formula, because they cause most of the confusion:

• It is a period metric, not a per-dish metric. You calculate it over a week or a month, across everything the kitchen used. The per-dish version exists too (that is recipe costing, covered below), but the headline number is periodic.
• It is calculated on food, against food sales. Beverages have their own ratio (beverage or pour cost) with very different economics. Mixing bar purchases into your food cost, or dividing by total revenue including drinks, produces a flattering and useless number.

One scoping note before the math: this whole article works the cost side of the fraction. The revenue side, keeping the dining room full, is a different job, and it is ours: ViteUneTable takes online bookings for free with 0% commission, which keeps food sales, the denominator of this ratio, healthy. Cost control and full tables are the two halves of the same margin.

The food cost percentage formula

The formula everyone quotes is the right one, as long as all three inventory terms are in it:

Food cost % = (beginning inventory + purchases - ending inventory) ÷ food sales × 100

The top of the fraction is your cost of goods sold (COGS): not what you bought during the period, but what you actually consumed. That distinction is the whole reason inventory appears twice. If you stocked up heavily just before the period ended, your purchases look enormous but your ending inventory is high too, and the formula corrects for it. Skip the inventory counts and divide purchases by sales, and your "food cost" will swing wildly with every big delivery while telling you nothing about the kitchen.

Worked example in dollars

A US bistro closes the month with these figures:

Food cost % = (8,000 + 22,000 - 7,200) ÷ 70,000 × 100 = 32.6%

Note that food sales here means food revenue excluding sales tax, which is straightforward in the US because the menu price is already tax-free: the tax is added to the check (or the bill, as UK readers would say) at the end.

The same example in pounds, and the VAT trap

A UK gastropub runs the same calculation:

Food cost % = (5,500 + 14,800 - 5,100) ÷ 45,000 × 100 = 33.8%

The trap: UK menu prices legally include VAT, and eat-in restaurant meals carry the standard 20% rate (VAT Notice 709/1, gov.uk). The £54,000 on the till is not your revenue; £9,000 of it belongs to HMRC. Divide by the gross figure and this pub would believe its food cost is 28.1%, more than five points better than reality. Always compute on sales net of VAT, and be suspicious of any benchmark that does not say which basis it uses.

Theoretical vs actual food cost: the variance that eats your margin

The formula above gives your actual food cost: what the period really consumed. There is a second number worth knowing: your theoretical (or ideal) food cost, which is what the period should have consumed if every dish had gone out exactly as its recipe card specifies.

Theoretical COGS = Σ (dishes sold × recipe cost per dish)

Take your sales report for the period, multiply each dish's sales count by its costed recipe, add it up, divide by food sales. If the bistro above sums its recipe cards against its sales mix and gets a theoretical COGS of $20,300, its theoretical food cost is 29%. Actual: 32.6%.

That 3.6-point gap is the variance, and it is the most valuable number in this whole article, because it is pure loss: food that was bought and paid for but never became a sold plate. On $70,000 of monthly food sales, 3.6 points is $2,520, every month.

Where does the variance come from? Almost always some mix of:

• Waste. Spoiled produce, over-prepping, trim that could have been used, and plates coming back half-eaten. The sums are not anecdotal: WRAP estimates food waste costs the UK hospitality and food service sector £3.2 billion a year, around £10,000 per outlet on average, and ReFED's US data shows that at foodservice level about 70% of the waste is plate waste, food guests were served and did not eat. We wrote a full guide on cutting food waste in a restaurant, and every ton of it lands in this variance first.
• Portioning drift. The recipe says 6.5 oz (180 g) of chicken; the pan gets 8 oz on a busy night. Nobody notices, the guest is happy, and the variance quietly absorbs the difference. A scale at the prep station and honest recipe cards fix most of it.
• Shrinkage and theft. Unpleasant to think about, real everywhere: a steak that leaves by the back door, a bottle poured for friends. A tight variance is also your best detection tool, because theft shows up as a gap no waste log explains.
• Untracked comps and staff meals. Every dish that leaves the kitchen without ringing a sale inflates the variance unless it is recorded. Track them; you do not have to stop them.
• Receiving errors. Paying for 10 kg and getting 9, or accepting a delivery at a higher price than quoted. The variance catches what the invoice check missed.

The discipline that follows is simple: calculate both numbers monthly, and investigate the gap instead of shrugging at it. A variance under one point is excellent; a variance that jumps is a signal something specific changed this month, and specific things can be found.

How often should you count inventory?

The formula only works if the inventory figures are real, which raises the practical question: how often do you count?

• Monthly is the minimum. It aligns with your accounting and makes the food cost figure meaningful. Once a quarter is bookkeeping; once a year is archaeology.
• Weekly is the professional standard for the key items. A full count every week is heavy for a small team, but a weekly count of your 20 most expensive SKUs (proteins, seafood, cheese, anything portioned) catches drift within days instead of weeks. Many operators do exactly this hybrid: full count monthly, top items weekly.
• Count consistently. Same day, same time, ideally before the big delivery day, by the same person or pair. An inventory taken Monday morning one month and Friday night the next makes the two figures incomparable.
• Count at cost, at current prices. Your inventory sheet needs the latest purchase price per unit, not last year's. Supplier prices move constantly, and stale prices corrupt both the COGS and the variance.

One habit multiplies the value of all of this: track your food cost as a trend line, not as isolated data points. A single week at 36% might be a big delivery or a holiday menu. Six weeks drifting from 31% to 34% is a real signal, and you will only see it if the number exists every week.

Recipe cards: the per-dish side of the same coin

Everything above works at the level of the whole kitchen. The per-dish tool is the costed recipe card: every ingredient, the exact quantity per portion, the current purchase price, a total at the bottom. That total is the plate cost, and plate cost ÷ menu price is the dish's individual food cost percentage.

Recipe cards earn their keep three times over. They are the raw material of the theoretical food cost above; they standardize portions, which shrinks the variance; and they are the foundation of pricing, because you cannot price a dish you have not costed. The pricing side, from the cost-plus formula to menu engineering and price psychology, is a subject of its own that we covered in our complete guide to restaurant menu pricing.

If keeping cards current sounds like a chore, cost the top 20% of your menu first. In most restaurants, a small set of dishes drives most of the volume, and those cards deliver most of the control.

What is a good food cost percentage?

Here is the honest answer most articles avoid: there is no universal good number, and any site telling you "keep it between X and Y" without a source is repeating folklore.

What is documented: the National Restaurant Association's pre-pandemic baseline puts food at about 33 cents of every dollar of sales for a typical US restaurant, alongside roughly 33 cents of labor and 29 cents of other expenses, leaving about 5 cents of pre-tax profit. That tells you where the average sits, not where you should sit: a pizzeria, a sushi counter and a steakhouse can all be healthy at wildly different ratios.

The number that actually decides whether you make money is not food cost alone but food cost plus labor: the prime cost. A kitchen that buys pre-prepped at higher ingredient cost but lower labor can be as profitable as one that fabricates everything in-house with a bigger brigade. In the NRA baseline above, those two lines together absorb about two thirds of every dollar; your target food cost falls out of your own rent, payroll and concept, not out of anyone's benchmark table.

The steakhouse math: when a high percentage is exactly right

The clearest proof that the percentage is not sacred: compare two dishes.

The rib-eye "fails" the ratio test and beats the chicken by $15.20 of margin every time it sells. You bank dollars and pounds, not percentages: the rent is paid in currency, and a steakhouse running a 38% food cost with $25 margins per cover can be far healthier than a soup bar at 25%. Judge the menu on contribution per plate and let the blended percentage land where your concept needs it. The percentage is a thermometer for drift, not a grade for individual dishes.

Fix the ratio, then fill the room

Everything in this article works the cost side of the fraction. The other side is food sales, and it deserves one honest remark: the fastest way to make every ratio look worse is an empty dining room, because your kitchen still preps, stocks and wastes against covers that never arrive. Predictable, well-filled services make prepping accurate and inventory calm.

That side is our trade. ViteUneTable's free version takes bookings around the clock through your own link and widget, sends automatic confirmation emails, and charges 0% commission and no per-cover fees, forever, with no commitment. When volume justifies it, the Standard pack (€29 excl. VAT per month) adds automatic email reminders and Reserve with Google, and the Standard + Anti No-Show pack (€49 excl. VAT per month) protects big tables with credit card holds. Let's be honest: no reservation tool counts your walk-in or prices your rib-eye. But steady, predictable covers are what turn a well-measured food cost into an actual profit.

Frequently asked questions

What is the food cost percentage formula?

Food cost percentage = (beginning inventory + purchases - ending inventory) ÷ food sales × 100, over a defined period. The top of the fraction is your cost of goods sold: what the kitchen actually consumed, not just what you bought. Use food sales excluding sales tax (US) or net of VAT (UK), and keep beverages in their own separate ratio.

What is the difference between theoretical and actual food cost?

Actual food cost comes from the inventory formula: what the period really consumed. Theoretical food cost is what it should have consumed, calculated by multiplying each dish's sales count by its costed recipe card. The gap between the two is the variance: waste, over-portioning, shrinkage, untracked comps and receiving errors. Calculate both monthly and investigate the gap.

How often should a restaurant count inventory?

Monthly at minimum, so the food cost figure aligns with your accounts. The stronger practice is a hybrid: full count monthly plus a weekly count of your most expensive items (proteins, seafood, cheese), always on the same day and time, valued at current purchase prices. Weekly visibility catches drift in days instead of discovering it in the annual accounts.

What is a good food cost percentage for a restaurant?

There is no universal target. The National Restaurant Association's baseline puts food at about 33 cents of every dollar of sales for a typical US restaurant, but healthy concepts range widely around that. Your real target comes from your own rent, payroll and pricing: what matters is that food cost plus labor (prime cost) leaves room for operating costs and profit, and that the trend is stable.

Why is a high food cost percentage sometimes fine?

Because you bank money, not ratios. A rib-eye costing $15 and selling at $40 runs a 37.5% food cost yet contributes $25 of gross margin per plate, far more than a 30% chicken dish contributing $9.80. Expensive proteins usually carry a higher percentage and a much higher margin in dollars or pounds. Judge dishes on contribution per plate; use the overall percentage to detect drift over time.

Do UK restaurants calculate food cost differently from US restaurants?

The formula is identical; the sales basis is not. UK menu prices include 20% VAT on eat-in meals, so food sales must be converted to net of VAT before dividing, or the percentage looks about five points better than reality. US menu prices exclude sales tax, so the till figure is already the right basis. Never compare UK and US percentages without checking the tax basis.