Food cost percentage calculator

The formula

Food cost % = (Beginning inventory + Purchases − Ending inventory) ÷ Food sales

The part in parentheses is your cost of goods sold: what came off the shelves during the period. Buy $13,000 of product and finish the month holding more than you started with, and some of that money is still sitting in the walk-in. It belongs to next month.

Two counts make the whole thing work, one at the start of the period and one at the end. Ending inventory rolls forward as the next period’s beginning inventory, so once you count on a schedule you only ever do one count per period.

A worked example

Four weeks at a café doing $42,000 in food and non-alcoholic drink sales. Beginning inventory came to $8,400. Invoices received during the period added up to $13,100. The closing count came to $7,900.

COGS = 8,400 + 13,100 − 7,900 = $13,600
Food cost % = 13,600 ÷ 42,000 = 32.4%

Purchases on their own would have said 31.2%. The counts are what make the number true, and counting is the step most people skip.

Plate cost is a different question

Plate cost is what one item costs to make divided by what you sell it for. A 12 oz latte with $0.94 of milk, beans, cup, and lid, sold at $5.25, runs 17.9%. The second mode above does that math.

Food cost for the period tells you how the month went. Plate cost tells you which items are dragging. They rarely agree, and the gap is the useful part. If your recipes price out at 28% and the period lands at 34%, six points of product left the building without becoming a sale.

What a normal number looks like

For 2024, the National Restaurant Association reported a median food and non-alcohol beverage cost of 32.0% of sales for full-service restaurants and 32.4% for limited-service. Volume matters: full-service operators above $2M in annual sales ran 31.0%, and those under $2M ran 33.7%.

Cafés don’t get a free public number. The NRA does survey a coffee-and-snack segment, but that breakout sits inside a paid report, so anyone quoting you an exact coffee-shop food cost percentage is quoting a blog quoting another blog. Trade estimates land somewhere between 25% and 35%, which is too wide to steer by. Your own last three periods are a better benchmark than anybody’s median.

Where the number goes wrong

• Alcohol in the denominator. A bar-heavy week will make your kitchen look brilliant. Split the two.
• Counting an invoice that hasn’t landed. Purchases means product you received in the period. A delivery that shows up on the 1st belongs to the new month even if you ordered it on the 28th.
• Counting the same item two different ways. One case of oat milk counted as “1” in January and “12” in February shows up as a cost swing that never happened. Fix the unit.
• Treating waste, comps, and staff meals as invisible. They pull product out of inventory without producing a sale, so they land in your food cost anyway. Track them on their own line and you can tell a portioning problem from a pricing problem.
• Skipping the physical count. Theoretical usage out of a POS is a forecast. Shrinkage only shows up when you compare it against something you counted with your hands.

Questions we get

Other free calculators

• Prime cost calculator — COGS plus fully loaded labor against sales — the number that decides whether a month works.
• Recipe cost calculator — Ingredient rows with yield adjustment, cost per portion, and the price that hits your target.
• ShelfCount pricing — one plan, $50 a month, every feature included.

Sources

• National Restaurant Association, “Restaurant operators kept food cost ratios in check in 2024” (published September 2025)
• National Restaurant Association, “Higher volume restaurants reported lower food-cost ratios in 2024” (published October 2025)
• Penn State, Introduction to Food Production and Service (open textbook), Ch. 12: Analyzing and Managing Food & Beverage Expense

Written by the ShelfCount team. We run Stella Coffee in Los Angeles and built these for our own back office first. Last checked August 2026.