How to Reduce Food Cost in a Restaurant: 14 Practical Steps

To reduce food cost without shrinking the plate, find out what each dish really costs, stop paying more than you should for ingredients, stop food leaving without a sale, and steer the menu toward the dishes that earn the most per plate. Start with an accurate food cost percentage, built from the food you actually used rather than what you bought, because every decision after that depends on it.

Food cost is one half of prime cost, the number at the center of the guide to reducing restaurant costs without hurting service. It is also the half that moves without warning. The USDA's Economic Research Service forecasts wholesale beef prices up 8.5% in 2026 and farm-level egg prices down 79.2%. A recipe card costed at January's prices is already wrong, in one direction or the other.

How to calculate food cost percentage

Formula: Food cost percentage = cost of food used during the period ÷ food sales for the same period × 100.

The cost of food used is not the same as what you bought. Purchases include food that is still on the shelf at the end of the period, and they leave out food you used from stock bought earlier. An inventory count at each end of the period corrects for both:

Formula: Cost of food used = beginning inventory + purchases − ending inventory.

Dividing purchases alone by sales only works when inventory didn't change, and it rarely holds week to week: a big delivery on the last day of the month inflates this month and flatters the next. Three more rules keep the number honest:

• Use the same period for everything, a week or a month, and count inventory at the same point each time, before the day's deliveries arrive.
• Keep food and beverage apart. Food cost divides the cost of food used by food sales. Give beer, wine, spirits and soft drinks a beverage cost of their own, or a profitable bar will hide a kitchen problem.
• Net out credits, and log staff meals and comps. Vendor credits come off purchases. Staff meals and comps leave inventory without a sale, so record them separately or they look like waste.

Worked example: Say a restaurant starts the month with $8,000 of food inventory, buys $26,000 of food and ends with $7,500 on the shelves. Cost of food used is $8,000 + $26,000 − $7,500 = $26,500. Food sales were $85,000, so food cost is $26,500 ÷ $85,000 = 31.2%. Dividing purchases alone by sales would have given $26,000 ÷ $85,000 = 30.6%, because $500 more came off the shelves than went on. At this volume, each point of food cost is worth $850 a month, or $10,200 a year.

Those are the starting values in the food cost calculator, so you can check the example and then replace them with your own numbers. Add your own target to see the dollar gap for the period. Its second tab works out the cost of one plate and the menu price that hits your target.

What is a good food cost percentage?

There isn't one number, and the ranges you'll read online don't all measure the same thing. Software companies publish targets. The National Restaurant Association and accounting firms publish what operators actually report, often on a different ratio. Here is who says what.

Three things to take from it. First, even one publisher's ranges differ: Restaurant365 gives 28 to 33%, 28 to 35% and 28 to 30% on three pages, so treat them as a sanity check, not a target. Second, the NRA and Citrin figures divide by total sales, drinks included, so a restaurant with a busy bar can't compare its kitchen's food cost with them directly; the NRA itself says its data is "not intended to represent standards or goals for individual restaurants."

Third, food cost tracks profit and scale. In Citrin Cooperman's 2025 benchmarking report, the most profitable full-service group ran food, beverage and disposables 4.57 points of revenue below the least profitable, and in the NRA data, full-service operators with sales of $2 million or more ran a median 2.7 points lower than smaller ones. Set your own target from your concept, your prices and your best recent months.

14 steps to reduce food cost

1. Cost every recipe against this month's invoices

Write a recipe card for every menu item: each ingredient, the quantity on the plate, the price on your latest invoice and the yield after trimming and cooking (step 2). Add them up and you have the plate cost. Recost your best sellers whenever a big invoice price moves, and the whole menu at least once a quarter.

Worked example: Say a chicken plate uses $3.20 of protein, $0.60 of starch, $0.90 of vegetables and $0.45 of sauce and garnish, $5.15 in all. On an $18 menu price, its food cost is 28.6%.

2. Yield-test your proteins and produce

The invoice price is for what you buy, not what you serve. A yield test measures the difference: weigh what comes in, trim and portion it the way the recipe says, weigh what's usable, and divide.

Worked example: Say a whole beef tenderloin costs $14.00 a pound, or $140 for a 10-pound case. After trimming, 6.5 pounds is usable for steaks, a 65% yield. The usable cost is $140 ÷ 6.5 = $21.54 a pound, so an 8-ounce filet costs $10.77, not the $7.00 the invoice price suggests. Costed at the invoice price, the plate looks $3.77 cheaper than it is.

Repeat it whenever the supplier, the cut or the butcher changes.

3. Price from plate cost, and reprice when it moves

Formula: Menu price at your target = plate cost ÷ target food cost percentage.

Worked example: If the protein in the step 1 plate rises from $3.20 to $3.84, the plate costs $5.79 and its food cost is 32.2% of the $18 price. To get back to 30%, the price would need to be $5.79 ÷ 0.30 = $19.30.

You don't have to put the whole increase on the price. Re-spec a garnish, change the side, or spread small increases across items guests don't price-compare.

4. Check vendor prices every week, line by line

Price creep arrives a few cents at a time. Compare each invoice with the last one for the same item and flag anything that moved.

Worked example: Say chicken thighs go from $2.40 to $2.64 a pound, a 10% increase, and you use 300 pounds a week. That's $72 a week, or $3,744 a year, on one line nobody looked at.

Watch for falling prices too. When a commodity drops, ask your supplier to pass it on, because contract and list prices don't always follow on their own.

5. Get competing bids on your biggest-spend items

Rank your purchases by dollars and start at the top. Get quotes on those lines from at least two approved suppliers every month, compared on the same spec, pack size and grade. A cheaper case with a lower yield (step 2) can cost more per usable pound.

Citrin Cooperman's high performers "often maintain relationships with multiple vendors to ensure competitive pricing and reduce vulnerability to commodity swings."

6. Check every delivery before you sign

Weigh proteins, count cases, check temperatures and dates, and match the invoice to what you ordered. Get a credit for anything short, damaged or substituted while the driver is still there. A receiving error you sign for is food you paid for and never had.

7. Portion to spec with scales and sized tools

Worked example: Say a burger spec calls for 6 ounces of beef at $5.00 a pound, about $1.88 a patty. If cooks free-hand 7 ounces, each burger carries an extra $0.31 of beef. At 1,200 burgers a month, that's $375 a month, or $4,500 a year, and no guest asked for it.

Give every station a scale, scoops and ladles sized to the recipe, and a photo of the finished plate. Check portions during service, not only in training.

8. Count inventory weekly and compare actual with theoretical food cost

Theoretical food cost is what your recipes say you should have used for what you sold. Actual food cost is what the count says you did use. Restaurant365 describes theoretical cost as "what you expect to spend based on standard recipes and portion sizes" and actual cost as "the real amount spent, including any waste or mistakes."

Worked example: Take the restaurant from the formula example. Its sales multiplied by its recipe costs give a theoretical food cost of $24,650, or 29.0% of $85,000. Actual cost of food used was $26,500, or 31.2%. The gap is $1,850 a month, or $22,200 a year.

A gap like that is a question, not a verdict. Before you put it down to waste or theft, reconcile the paperwork, because some of the gap may not be lost food at all:

1. Recheck the counts. Look for an item counted in the wrong unit (cases against pounds), a missed shelf or freezer, or a price on the count sheet that doesn't match the latest invoice.
2. Match invoices to the period. An invoice entered twice, missing, or dated in the wrong week moves the cost of food used by its full amount. Make sure vendor credits were recorded.
3. Log staff meals, comps and voids. Food that left inventory for a recorded reason is a cost, but it isn't waste or theft.
4. Check the recipe cards and POS. A recipe card with an old quantity, or a POS button mapped to the wrong item, makes theoretical cost wrong, not actual cost.

What's left is the real leak. Use the waste log (step 10) and portion checks (step 7) to find out whether it's overproduction, spoilage or over-portioning, and treat theft as something to investigate with evidence, not assume.

Count everything once a month. Counting your high-cost items, such as proteins, cheese and seafood, every week is a useful early warning, and Citrin Cooperman notes that "top operators track inventory daily or weekly, not just monthly." But a weekly figure built from a partial count is an estimate: count the high-cost items, use the week's purchases for the categories you didn't count, label the result an estimate, and check it against the monthly full count. What prime cost is works through a weekly estimate with a worked example.

9. Store first in, first out, and date everything

Rotate new deliveries behind old stock, label every container with what it is and when it was made, and set up the walk-in so the oldest product is the easiest to reach. For prepared food, section 3-501.17 of the FDA Food Code requires refrigerated ready-to-eat food that needs time and temperature control for safety and is kept more than 24 hours to be date-marked for use within 7 days at 41°F or below, counting the prep day as day 1; your state or county's adopted code is what applies. Every item thrown out on its discard date is food cost with no sale against it.

10. Keep a waste log

Keep a log by the bins and record everything before it's thrown out: the item, how much, why (spoilage, trim, overproduction or plate waste) and what it cost. Once food goes into a mixed bin, you can weigh the bin but you can't tell what was in it. Measuring is where waste programs start. In a review of 114 restaurants in 12 countries for Champions 12.3, restaurants cut kitchen food waste by 26% on average within a year. The guide to reducing food waste in a restaurant shows how to run a two-week waste audit and turn it into dollars.

11. Engineer the menu by contribution margin, not food cost percentage

Food cost percentage tells you how efficient a dish is. Contribution margin, the menu price minus the plate cost, tells you how much money it makes. Menu engineering sorts each dish by popularity and contribution margin, and the quadrant it lands in tells you what to do with it.

Worked example: Say four dinner entrées sell 1,010 plates a month, about 252 each on average, at an average contribution margin of $16.16 a plate (total contribution ÷ plates sold). Above both averages is a star; popular but below the margin average is a plowhorse; above the margin average but unpopular is a puzzle; below both is a dog.

In this example, the dish with the worst food cost, the steak at 37.5%, makes the most money per plate. The dish with the best food cost, the flatbread at 20%, makes the least and barely sells. Push guests toward low food cost dishes and your food cost percentage falls, but so does your gross profit. The better move is the plowhorse: $1 more on the pasta adds $500 a month if it keeps selling 500 plates.

12. Shorten the menu and share ingredients across dishes

An ingredient used in a single dish is one more thing to spoil. Cut dishes that rarely sell and build the rest around shared ingredients, so the same proteins, herbs and sauces turn over across several plates. The NRA suggests the streamlined menus many full-service restaurants have kept since the pandemic "may have led to some purchasing efficiencies."

13. Run specials that use trim and overstock

The trim from step 2 is already paid for. Tenderloin tips become a steak sandwich, vegetable trim becomes stock and soup, day-old bread becomes croutons or bread pudding, and a case that came in too big becomes this week's special. Plan specials around what's in the walk-in, not around what the chef wants to buy. Use only trim that was handled and stored as food, inside its date marks.

14. Prep to a forecast, and make the team part of it

Prepping the same batch every day builds overproduction into the schedule. Set prep pars by day and daypart (breakfast, lunch, dinner) from your sales history and check them weekly against what sold; the food waste guide linked in step 10 works through a daypart forecast. Then share the food cost number with the kitchen, so the people who portion and prep can see what their work does to it.

What not to cut

Some savings cost more in sales than they save in food. Leave these alone:

• Portions on signature dishes. The dish regulars come back for is the last place to trim. Fix the price and the spec first.
• The ingredient the dish is named for. Swap the commodity in a sauce or side before you downgrade the protein guests ordered.
• Food safety. No discount on near-date stock is worth a temperature or date-mark violation.
• Training time. Portioning and yield only hold if someone teaches them and checks.

Food cost also has to be read next to labor. A cheaper ingredient that adds an hour of prep a day can raise prime cost, the total of food, beverage and labor, even as food cost falls. The guide to reducing labor cost in a restaurant covers the other half.

Where software and AI help

Every step above depends on steady tracking, which is where busy operators fall behind. Kutlerri.ai's Prime Cost Optimization Agent, currently in beta, takes on that tracking as a managed service: it prices menu items off what they cost to make, tracks food and labor cost, flags vendor price increases and recommends recipes to drop or replace.