How to calculate food cost percentage, step by step

Food cost percentage is the share of your food sales that went on the food itself. If a kitchen sold 10,000 in food last week and used 3,000 of ingredients to do it, its food cost percentage was 30. That is the whole idea, and it is simple enough to fit on a beer mat. The difficulty is not the arithmetic. It is that almost every input to the sum can be wrong in a way that looks perfectly reasonable on paper.

This guide walks through the calculation the way it has to be done in a real kitchen: with a stock count, a clean cut-off date, and three corrections that most venues skip until a strange month forces the question. There is a worked example with a full week of numbers, a section on what the figure should be for your menu rather than somebody else's, and the order in which to investigate when the percentage moves.

If you only want the formula, it is in the next section. If you want a number you can make decisions with, read on.

The figure answers one question: of every unit of currency a guest spent on food, how much went back out of the door to pay for the ingredients? It does not measure labour, rent, packaging or anything else. It does not tell you whether a dish is profitable, because a dish with a high percentage can still earn more cash than one with a low percentage. It is a ratio, and ratios are good at showing change over time and poor at telling you what to do about it.

That is why the number is most useful when it is calculated the same way every period. A kitchen that counts carefully one month and guesses the next will see its food cost swing by several points for reasons that have nothing to do with cooking. A kitchen that counts the same way every time will see real movement, and real movement is what you want to catch early.

It also helps to be clear about what is in and what is out. Food cost usually covers everything that goes into a dish the guest pays for: proteins, produce, dry goods, dairy, oils, garnish. Drinks are normally tracked separately as beverage cost, because their margins behave very differently. Cleaning products, disposables and staff food are not ingredients in a dish a guest bought, and putting them in the same line hides more than it shows. We come back to staff food later, because it causes more trouble than its size suggests.

The food cost formula is:

The top half of that fraction, the cost of goods sold, is where the work sits. You do not get it from your invoices alone, because some of what you bought this week is still on the shelf, and some of what you cooked was bought last week. So cost of goods sold is calculated from stock, not from spend:

Opening stock is the value of the food you had at the start of the period. Purchases are everything that arrived during it. Closing stock is what is left at the end. The difference is what was used, whether it was sold, wasted, eaten by the team or walked out of the back door. The accounting idea behind it is general rather than specific to restaurants, and the cost of goods sold is the same line any business that holds stock has to calculate.

Opening stock should be last period's closing stock, carried over unchanged. If you recount at the start of a period you introduce a second count and a second set of errors. One count, at one moment, closes one period and opens the next.

Purchases means everything delivered and accepted in the period, at the price you were actually charged. Credits for short or damaged deliveries come off. A delivery that arrived on the last day of the month belongs to that month, even if the invoice is dated the next.

Closing stock is a physical count of everything in the walk-in, the freezers, the dry store, the bar fridges that hold food and the prep containers on the line, valued at a consistent price. It is the slowest part of the job and the one that decides whether the final figure means anything.

Food sales are what guests paid for food, after discounts and before tax. If your till reports sales including tax, take it out first, or your percentage will look better than it is by exactly the rate of tax.

Here is the full method, in the order it should be done. Each step is short. Skipping any of them is where the trouble starts.

Most bad food cost figures trace back to the first step. If the count happens on Sunday night but the sales report runs to Monday lunchtime, you have half a day of sales with no matching stock movement, and your percentage drops for no reason. If a Monday morning delivery is counted in the closing stock but its invoice falls into the next period, you have inflated closing stock and flattered the number again. The cut-off is a promise that stock, purchases and sales all describe exactly the same stretch of time.

A full count does not have to take a full shift. Count in the order the shelves are laid out rather than alphabetically, so nobody walks back and forth. Use the same units every time: if olive oil is counted in litres one week and bottles the next, the total is fiction. Split the venue into zones and give each pair of counters one zone. Our guide on how to run a restaurant stock count properly covers the routine in detail.

The numbers below are illustrative, built for a mid-sized restaurant doing lunch and dinner six days a week. They are not a benchmark, just a way to see the arithmetic work.

Cost of goods sold is 6,200 plus 9,450 minus 5,850, which comes to 9,800. Divide that by 27,600 and multiply by 100, and the food cost percentage for the week is 35.5.

Now suppose the head chef says the number feels high. Before anybody changes a recipe, check the inputs. In this example, three things turn up:

Correct those and the week reads very differently. Staff food comes out of cost of goods sold and into its own line, the Monday delivery comes out of closing stock, and the transfer comes off purchases. The recalculated cost of goods sold is 9,800 minus 420, plus 600 because closing stock falls by that amount, minus 310: 9,670. The food cost percentage becomes 35.0. Not a dramatic change in this case, but in a week where the errors all run the same way, a two or three point swing is easy to produce from nothing but timing.

When a food cost figure looks wrong, it is usually one of three things, and they are worth checking in this order.

The most common error is treating each period as if it starts from nothing. If you only add up purchases and divide by sales, you are measuring buying, not using. A month in which you stock up for a holiday looks terrible, and the month after looks wonderful, and neither number tells you anything about the kitchen. The fix is simple: every period opens with the last period's closing count, unchanged.

A team meal uses real food. If it is not recorded, it sits invisibly inside cost of goods sold and makes the kitchen look careless. If it is rung through the till at full price as a sale, it does the opposite and flatters the figure. Neither is right. Record staff food at cost, separately, and take it out of the food cost calculation so the percentage describes food sold to guests.

Suppliers invoice on their own schedule. A delivery that arrives on the 31st with an invoice dated the 1st will drift into the wrong month unless somebody moves it deliberately. Over a year this evens out, but over a single period it can move the percentage by several points. Date purchases by when the goods arrived and were accepted, not by the paperwork.

Transfers between venues, if you run more than one, need recording at both ends or one kitchen looks generous and the other mean. Tax-inclusive sales are the other: if the sales figure includes tax and the purchases do not, the percentage is understated by the whole tax rate.

You will hear target figures quoted with great confidence. Treat them with suspicion. A steakhouse, a pizzeria, a café and a tasting-menu restaurant all have different natural food costs, because what they sell, how they price it and how much labour sits in each plate are completely different. A single target borrowed from another venue tells you very little about yours.

A better way to set a target is to work it out from your own menu. If every dish has a costed recipe, you can multiply each dish's cost by the number sold and divide by the sales it brought in. That gives a theoretical food cost: what the percentage would be if every plate went out exactly to spec and nothing was wasted. The gap between that and your actual figure is where the money is, and it is the subject of our guide to actual vs theoretical food cost.

The theoretical figure moves when your menu mix moves. If guests start ordering more of your expensive-to-make dishes, your theoretical food cost rises even though nobody in the kitchen did anything wrong. That is the strongest argument for knowing it: without it, you will blame the kitchen for a change that the dining room made.

Food cost also sits beside labour. The two together are often called prime cost, the two largest controllable costs in most restaurants, and a change in one frequently shows up in the other. A kitchen that buys in prepared ingredients will see food cost rise and labour fall. The definition of prime cost is a useful reminder that a better food cost bought with more prep hours is not always a saving.

Monthly is the minimum, because it is when the accounts close. It is also too slow to act on. By the time a monthly figure tells you something went wrong, it went wrong four weeks ago and has been going wrong ever since.

Weekly is the rhythm most kitchens can sustain. A full count every week is real work, so many venues do a full count monthly and a short count of their twenty most valuable lines every week. Those twenty lines usually cover most of the money, which means a change in the percentage shows up in them first.

Daily food cost is possible, but only if you can estimate usage without a daily count, usually from theoretical usage based on what was sold. It is useful for very high-volume operations and for kitchens chasing a specific problem. For most restaurants it produces a lot of numbers and not many decisions.

Whatever rhythm you choose, keep it fixed. A number calculated the same way each week is worth more than a more precise one calculated differently each time.

Some weeks will always look odd, and it helps to know which ones before they arrive. A week that ends just before a bank holiday weekend carries a large closing stock, because the kitchen has bought ahead, so its figure looks unusually good and the following week's looks unusually bad. A week with a large private event can do the same in either direction, depending on whether the event's food was bought in advance or on the day.

The answer is not to adjust the figure by feel. It is to note the reason beside it. A short line in the margin of the weekly sheet, such as "bought ahead for the long weekend" or "two buyouts, one prepped Friday", turns a confusing number into an explained one. Over a year, those notes become a guide to your own calendar: which weeks distort the figure, by roughly how much, and which way.

If you run more than one venue, calculate each one separately before you add them together. A group figure hides the kitchen that is drifting behind the kitchen that is doing well, and the conversation that fixes it can only happen with the venue that owns the problem.

A rise of a point or two is a question, not a verdict. Work through the causes in this order, because the cheap checks come first:

Working in this order protects the kitchen from being blamed for an accounting problem, and it protects the business from missing a real one. When the cause does turn out to be the kitchen, our guides on reducing food costs without a new menu and reducing food waste cover what to change first, and managing suppliers covers the price side.

Food cost percentage is a stock calculation, and stock is counted in the kitchen. Where software helps is on the sales side of the sum and in the detail behind it.

In servearound, every dish carries its own record across every menu it appears on, and once a dish's cost is set, the menu report in Analytics shows its units sold, its revenue and its margin from what actually went out during service. That is the raw material for a theoretical food cost: you can see which dishes carry your sales and which of them carry your margin. Because orders and checks are recorded as the service runs, the sales figure for a period comes from the same orders the kitchen cooked, on the venue's own clock, rather than from a report pulled at a slightly different time.

The same records help with two of the mistakes above. Comps and voids are logged against a member of staff, so you can see food that left the kitchen without a sale. And when a dish is marked sold out, it comes off every menu at once, which stops the kitchen cooking food nobody can order. If you want to see how the menu report reads for a venue like yours, talk to us.

Usually not. Drinks have different margins, different waste patterns and different controls, so most venues calculate beverage cost separately. Combining them gives one figure that describes neither well.

Not in food cost. Packaging is a real cost of delivery and collection orders, but it is not an ingredient. Track it as its own line, ideally per order, so you can see what each takeaway costs to send.

Pick one method and keep it. Last invoice price is simplest. An average of recent prices smooths out swings. What matters is consistency, because changing method between periods moves the figure on its own.

That is recipe costing rather than food cost percentage: the cost of everything on the plate divided by the dish's selling price. Our guide to recipe costing works through one dish from box price to plate, and our food cost calculator guide shows how to lay one out in a sheet.

To calculate food cost, take your opening stock, add purchases, take away closing stock, and divide the result by food sales for exactly the same period. The arithmetic is easy. The discipline is in the cut-off, the count and the three corrections: carry opening stock over, keep staff food out, and date deliveries by when they arrived.

Do it the same way every week and the number becomes a reliable early warning. Compare it with the theoretical figure from your own menu and it becomes a map of where the money goes. That is worth far more than hitting a target somebody else chose.