Food Cost Percentage, Restaurant Prime Cost and Labor Cost Percentage: Formulas and Benchmarks

Food cost percentage is cost of goods sold divided by net sales for the same period, expressed as a percentage. For a restaurant that used $31,000 of food against $100,000 in net food sales, food cost is 31 percent. Most US restaurants run between 28 and 35 percent depending on format, and the figure is only meaningful when the cost side comes from a physical inventory count rather than a stack of supplier invoices.

That last sentence is where most of the trouble lives. Below is the formula with a worked period, what the 2026 benchmark ranges actually are by restaurant format, how labor cost percentage is calculated alongside it, how the two combine into prime cost, and the six reasons a food cost percentage comes out wrong when the arithmetic is right.

How do you calculate food cost percentage?

The formula has two halves and people usually get the second one right and the first one wrong.

Cost of goods sold for a period is beginning inventory, plus purchases during the period, minus ending inventory. It is a physical calculation, not an invoice total. Beginning inventory is whatever you counted at the end of last period, purchases is everything that arrived and was invoiced during this one, and ending inventory is what you counted at the close. What that subtraction does is charge the period only for what you actually consumed, which is the entire point.

Net sales is gross sales minus comps, voids and discounts, and it excludes sales tax. Sales tax is money collected for the state and it was never revenue. Leaving it in inflates the denominator and makes food cost look better than it is, typically by two to three points depending on your local rate. Comps left in the denominator do the same thing.

Worked through a period:

Food cost percentage, worked for one four-week accounting period

Line	Amount	Note
Beginning food inventory	$18,400	Counted at close of prior period
Food purchases	$62,900	All invoices dated in the period, received or not
Ending food inventory	$21,100	Counted at the same day and time
Cost of goods sold, food	$60,200	18,400 plus 62,900 minus 21,100
Gross food sales	$203,500	Point of sale gross ring
Comps and discounts	$7,200	Subtract
Net food sales	$196,300	Sales tax already excluded
Food cost percentage	30.7%	60,200 divided by 196,300

Note that the inventory build of $2,700 is what separates this from the naive version. Divide purchases straight into sales and you get 32.0 percent, a point and a third worse, purely because a delivery landed before the count. Over a year of doing it that way the number swings enough that nobody trusts it, and an untrusted metric stops getting managed.

What is a good food cost percentage?

There is no single good number, because format drives it. Published 2026 industry references put quick service around 25 to 30 percent, casual dining roughly 30 to 34 percent, and fine dining 34 to 40 percent, where higher food cost is a deliberate trade for a higher check and a menu built on expensive proteins. Pizza and bar-led concepts sit lower still. Beverage runs on its own scale entirely: liquor pours near 18 to 20 percent, beer around 24 percent, wine typically 30 to 35 percent, which is why food and beverage cost of goods sold belong on separate P&L lines.

Two cautions on benchmarks. First, they moved. Food costs sit materially above pre-pandemic levels, so a target quoted from a 2019 article will read as easily achievable and is not. Second, your own trend is a far better comparison than any published range. A concept running 33 percent stably and profitably does not have a problem; a concept that ran 29 percent in the spring and 33 percent now has a specific problem that happened at a specific time, and the trend is what tells you where to look.

How do you calculate labor cost percentage?

Labor cost percentage is total labor cost divided by net sales. The word that matters is total. Hourly wages for kitchen and front of house, salaried managers and chefs, employer payroll taxes, workers compensation premium, health benefits and any contract or agency labor all belong in it. Tips passed through to staff do not, because that money was never yours, but the employer payroll taxes on tipped wages are.

Restaurant labor typically lands in the mid twenties to mid thirties as a percentage of sales, and 2026 data on full service restaurants shows labor alone trending toward a median in the mid thirties, with the more profitable operators holding it a couple of points below that. Quick service runs leaner. The reason a labor number reported without benefits and payroll taxes is close to useless is that the loaded cost is roughly 15 to 25 percent above the wage line, so an unloaded 26 percent is really something like 31, and a manager optimizing against the wrong figure will schedule to a target that does not exist.

What is prime cost in a restaurant?

Prime cost is cost of goods sold plus fully loaded labor, as a percentage of net sales. It is the sum of the two largest cost blocks a restaurant has and the two that a general manager can genuinely move inside a week. Common 2026 targets are roughly 55 to 60 percent for quick service and fast casual, 60 to 65 percent for full service casual dining, and higher for fine dining.

The reason operators watch prime cost rather than food cost alone is that the two components trade against each other. Buying pre-portioned proteins raises food cost and lowers prep labor. Butchering in house does the reverse. Judged on food cost alone, in-house prep always wins, and that conclusion is wrong about half the time. Prime cost is the number that makes the trade visible, and it is the subtotal that belongs on every restaurant profit and loss statement, reviewed weekly rather than waiting for a month end close.

Why is my food cost percentage wrong?

When the arithmetic is right and the number still looks implausible, it is almost always one of six things.

No inventory count. Covered above and by far the most common. Without a count, cost of goods sold is an invoice total and the percentage moves with your delivery schedule rather than your consumption. If a full weekly count is unrealistic, count proteins, seafood and liquor. That is usually 70 percent of the value and nearly all of the variance.

Sales tax left in sales. Inflates the denominator and flatters the percentage by roughly the local rate. Check that the sales figure feeding the calculation is net of tax, particularly if it was pulled from a bank deposit total rather than the point of sale.

Third party delivery booked net. Delivery marketplaces keep a commission of roughly 15 to 30 percent and deposit the remainder. Booking that deposit as sales understates the denominator, so delivery-heavy periods show a food cost percentage that looks impossible. Book gross sales with the commission as a separate expense line. It takes reading the platform statement instead of the bank feed, and it is the only way to answer whether the channel earns its place.

Category mix drift. If food and beverage are combined, a month where the bar outperformed will show an improvement in cost of goods sold that has nothing to do with purchasing. Separate lines fix it permanently.

Invoices in the wrong period. A supplier invoice dated the 29th that gets entered on the 3rd of the next month moves cost between periods and creates a good month followed by a bad one. Cut off purchases on the same date you count. Keying delivery invoices by hand is where this usually breaks down, and it is worth using something that can pull the line items straight off scanned supplier invoices so the entry is fast enough to happen on time rather than in a Friday catch-up.

Waste, theft and over-portioning. This is the residual, and it is what the whole exercise exists to surface. When counts are clean, sales are net, periods are cut properly and the percentage is still three points above where the menu says it should be, the gap between theoretical and actual is real product leaving the building without being sold.

Theoretical food cost versus actual

Theoretical food cost is what your recipes say the period should have cost: every item sold, multiplied by its costed recipe, summed. Actual food cost is what the count says it did cost. The variance between them is the operating question.

A variance under about one percentage point is normal noise. Two to three points is a specific problem with a specific cause: portioning drift on a high volume item, a supplier price increase that never made it into the recipe cost, spoilage on a slow-moving product, or theft. The value of the comparison is that it points at a line rather than telling you to try harder. It does require recipe costing maintained against current prices, which is a real ongoing commitment and the reason many operators run actual food cost only and manage against their own trend instead. That is a defensible choice, as long as the trend is measured honestly.

Getting the numbers out of the books each period

All of this depends on a P&L that separates the categories properly: food sales and beverage sales apart, food cost of goods sold and beverage cost of goods sold apart, hourly labor apart from management labor, with payroll taxes and benefits shown so labor can be loaded correctly. A generic chart of accounts collapses most of that, which is why so many restaurant operators calculate these percentages in a spreadsheet on the side instead of reading them off the statements.

Fixing it is a mapping job done once. Split the sales and cost of goods sold accounts by category in the chart of accounts, keep employer taxes and benefits out of the general overhead bucket, then produce the statements in restaurant format with a percentage of sales column and a prior period next to it. From there the percentages are read, not recalculated, and the same set feeds the monthly reporting package a lender or investor asks for. The broader operating metrics that sit alongside these are in financial ratios small business owners should track, and if the books live in QuickBooks, how to read a QuickBooks profit and loss covers reading the export you will be starting from.