Restaurant food cost: why the number always arrives too late

Food cost arrives too late because it is calculated after the books close, from an inventory counted in a hurry. The formula is simple: cost of food consumed divided by sales excluding tax. To act in time, match it every week against purchases, recipes and sales, and above all explain the gap with the theoretical figure.

The 8th of the month, a number that no longer explains anything

A fictional example. The accountant sends last month's results. Food cost at one of the three sites is higher than expected. The owner calls the head chef: "What happened?" The chef no longer remembers which week was involved. A supplier had changed its prices, a preparation had been portioned more generously, a batch of fish had been thrown out. All of that happened more than three weeks ago. The number is correct, but nobody can act on it anymore.

Why the number arrives late

• It depends on the inventory, so on the close. Until the closing stock is counted, the cost consumed does not exist.
• Purchase prices move between two inventories. The cost of a recipe changes during the month without the recipe card being updated.
• The theoretical figure does not exist. Without up-to-date recipe cards, you cannot say what you should have consumed, so you cannot isolate the gap.
• Waste is not logged when it happens. It reappears at month end in the difference between two stock counts.
• The data lives in four places: POS, invoices, recipe cards, inventory spreadsheet. Nobody puts them side by side before the close.

Two formulas to know

Actual food cost:

Cost of food consumed = opening stock + purchases - closing stock Actual food cost = cost of food consumed / sales excl. tax

Theoretical food cost:

Theoretical cost = sum over dishes sold of (quantity sold x recipe card cost) Theoretical food cost = theoretical cost / sales excl. tax

The gap between the two is not a judgment: it is the list of questions to ask (unlogged waste, portions, possible theft, inventory errors, outdated prices). Look at it by product family, not as a block.

Moving from a monthly number to a weekly reading

1. Choose a few sentinel products. The most expensive or fastest-moving ones (meat, fish, dairy). They explain most of the gap without requiring a full inventory.
2. Count these products on the same day, at the same time, every week. Owner: the head chef or sous chef. Tool: a sheet or a fixed list.
3. Match against the week's purchases. Invoices and delivery notes are enough. The goal is to calculate the weekly consumption of each sentinel.
4. Compare with expected consumption. The week's sales multiplied by the quantity on the recipe card give theoretical consumption.
5. Explain the gap before correcting it. Price, portion, waste, counting error: sort the cause. A gap with no named cause cannot be corrected.
6. Update the recipe card when the price changes. Show the update date on the card, otherwise the theoretical figure lies.
7. Decide on an action with a name and a date. Renegotiate, review the portion, log waste, train on trimming.

The overview: monthly versus weekly

Criterion	Month-end food cost	Weekly reading of sentinels
Delay before seeing the gap	After the accounting close	A few days
Scope	All products, as a block	A few products that carry weight
Identifiable cause	Rarely, memory is blurry	Often, the week is still fresh
Load on the team	One big inventory	A short, regular count
Use	Observation	Decision

Mistakes that create the delay

• Calculating food cost on sales including tax instead of excluding tax.
• Counting closing stock on a date different from the end of the sales period.
• Never updating recipe cards after a price increase.
• Comparing two sites without checking that they count the same way.
• Demanding a full inventory every week: it will be rushed within three weeks.

What to measure

• The gap between actual and theoretical food cost, by product family. It steers the discussion toward the right cause.
• The age of the last recipe card update compared with the last purchase price change.
• The delay between the end of the week and the discussion of the gap with the chef. If it goes beyond the following week, the reading is already useless.

Where Tsuno comes in

Tsuno keeps dated recipes and standards, and matches facts coming from your tools: sales and invoice exports can be imported, with human validation of sensitive matches. You keep your POS and your purchasing software. You can ask "What deserves my attention today?" and get a short answer, with the facts behind it. When data is missing, for example an inventory that was not counted, Tsuno says so instead of making it up. It prepares the action, you confirm. The integrations page explains how it works alongside your tools.

Key takeaways

Food cost is accurate but late. Count a few products, often, and explain the gap between theoretical and actual before trying to reduce it.

Going further: why owners discover cost drift too late, what waste really costs, the supplier price list that stays accurate and prime cost.