Food cost percentage is the number that quietly decides whether a busy restaurant makes money. Here’s how to calculate it the right way — and the handful of levers that actually move it.
July 27, 2026
Food cost percentage is the most-quoted number in the restaurant business and one of the most misunderstood. Operators trade it like a batting average — “we run a 28” — but the way it gets calculated varies so much from kitchen to kitchen that two restaurants quoting the same number can be in completely different financial shape. Calculated properly, food cost percentage tells you whether your menu pricing, purchasing, and kitchen discipline are holding together. Calculated loosely, it hides five figures a year of drift. This guide walks through the formula the way an accountant would want it done, the difference between theoretical and actual cost, what a reasonable target looks like for different concepts, and the handful of levers that reliably bring the number down without touching quality.
The formula itself is short: food cost percentage equals cost of goods sold divided by food sales, over the same period. The mistake is in the numerator. Cost of goods sold is not “what we spent on food this month” — it is beginning inventory, plus purchases, minus ending inventory. Skip the inventory counts and just divide purchases by sales, and a big delivery on the 30th makes a good month look terrible while a lean ordering week makes a bad month look fine. The purchases-only version swings with your ordering schedule; the true version measures what actually left the shelves. Run it on food sales only — pull alcohol out of both sides and track beverage cost separately, because bar margins behave differently and blending them flatters the food number.
Period matters too. Monthly is the standard for the P&L, but a weekly calculation is where the operational value lives: a spike shows up while the invoices and the prep schedule that caused it are still fresh enough to explain. Weekly numbers wobble more — one big order lands inside a single week — so read the four-week trend, not any one week in isolation.
Theoretical food cost is what your menu should cost to produce: every recipe costed ingredient by ingredient at current invoice prices, multiplied by what you actually sold. Actual food cost is what the formula above says really happened. The gap between the two is the most useful number in your kitchen, because it has only a few causes: over-portioning, waste, spoilage, unrecorded comps and staff meals, vendor price creep that has not made it into your recipe costs, and theft. A tight kitchen keeps the gap to a couple of points; a widening gap tells you exactly where to look while the month can still be saved. You cannot see the gap at all unless recipes are costed and inventory gets counted — which is why the discipline below matters more than the arithmetic.
There is no single correct target — the right food cost depends on what your concept spends everywhere else. A steakhouse runs a high food cost and a comparatively modest labor cost; a made-from-scratch café inverts that. Treat the commonly cited ranges below as orientation, not grading.
| Concept | Commonly cited range | Why it lands there |
|---|---|---|
| Pizza and bakery | Low to high 20s | Flour, cheese, and sugar are cheap relative to menu price |
| QSR and fast casual | High 20s to low 30s | Competitive menu pricing offsets inexpensive ingredients |
| Casual full service | Around the low 30s | Broad menus carry both high- and low-cost items |
| Steakhouse and seafood | Mid 30s and up | Protein-heavy menus; margin is made on volume and the bar |
The more useful comparison is you versus you: this month against last, this location against your other one, actual against theoretical. Seasoned operators and lenders tend to watch prime cost — food and labor together — and we cover how food cost fits that bigger picture in restaurant profit margins.
Half of food cost control is not cost at all — it is pricing. Once recipes are costed, every menu item has a contribution margin, and the pattern is always uneven: some items earn their space, some sell well but earn little, and a few do neither. Rebalancing the menu around what each item actually contributes — nudging a price, trimming a portion that mostly comes back on the plate, retiring the dish nobody orders — moves the blended food cost without any kitchen heroics. That exercise has a name and a method, and we walk through it in menu engineering. One caution: chasing a low percentage for its own sake backfires. A higher-percentage dish that sells in volume can contribute more dollars than a “well-costed” dish nobody orders. You bank dollars, not percentages.
Food cost percentage only counts what leaves the walk-in, but margin leaks happen after the plate is made, too. Third-party delivery commissions never show up in food cost — the food number can look perfect while a large slice of every delivery ticket goes to the marketplace. Moving regulars to commission-free direct online ordering protects the margin your kitchen discipline just earned. And if you run more than one store, compare food cost side by side across locations — the same menu running three points apart between stores is a management finding, not a math error. A multi-location platform makes that comparison a glance instead of a spreadsheet project.
Everything above is possible with a spreadsheet and a clipboard, and almost nobody sustains it that way through a busy season. The realistic version is a POS with inventory built in: recipes costed once and kept current as invoices land, theoretical usage depleted automatically with every sale rung at the register, counts entered on a tablet standing in the walk-in, and variance reported by item. That turns food cost from a monthly autopsy into a weekly habit — the setup is covered step by step in our inventory management guide.
Divide cost of goods sold by food sales for the same period. Cost of goods sold is beginning inventory plus purchases minus ending inventory — not just purchases. Exclude alcohol from both sides and track beverage cost separately. Most operators run the calculation monthly for the books and weekly for operations.
It depends on the concept. Pizza and bakery concepts are commonly cited in the 20s, casual full service around the low 30s, and protein-heavy concepts like steakhouses in the mid 30s or above. The more meaningful benchmarks are your own trend over time and your actual cost against your theoretical cost — a widening gap matters more than the absolute number.
Theoretical food cost is what your sales mix should have cost, based on costed recipes at current invoice prices. Actual food cost is what inventory movement says you really used. The difference is the variance — driven by over-portioning, waste, spoilage, unrecorded comps, price creep, or theft — and shrinking that variance is where most food cost improvement comes from.
A full count monthly, at minimum, so the P&L is accurate. Add a short weekly count of your highest-value items — ten to fifteen products usually cover most of the dollars at risk. Weekly counting sounds like overhead, but with a POS inventory tool and a tablet in the walk-in it takes well under an hour and catches problems while they are still cheap to fix.