Actual vs. Theoretical Food Cost, Explained

Theoretical food cost is what your ingredients should have cost, given what you sold and what your recipes say each item uses. Actual food cost is what they did cost, from your purchases and your counts. The gap between the two is variance: waste, over-portioning, theft, sales that weren't rung, or recipes that don't match the kitchen.
Most restaurants track one of these numbers. The useful number is the difference.
Actual usage: what left the stockroom
Actual usage comes from counts and invoices:
Actual usage = opening count + purchases received โ closing count
If you move stock between locations, transfers in are added and transfers out taken away. Multiply usage by cost and you have actual food cost for the period.
Theoretical usage: what sales say you used
Theoretical usage comes from sales and recipes:
Theoretical usage = for each item sold, the quantity sold ร the recipe quantity of each ingredient
A pizzeria that sold 300 large pizzas with 10 oz of cheese each should have used 3,000 oz of cheese on them. Add every item that uses cheese and you have the theoretical cheese usage for the week. This is why recipes have to be written down and kept honest: theoretical usage is only as good as the recipes behind it.
Variance: the gap between them
Variance = actual usage โ theoretical usage, in units and in dollars.
A worked example for one week of mozzarella, with made-up numbers:
- Opening count 80 lb, purchases 240 lb, closing count 60 lb: actual usage 260 lb.
- Sales: 300 large at 10 oz and 100 medium at 8 oz is 3,800 oz: theoretical usage 237.5 lb.
- Variance: 22.5 lb. At $2.40 a pound, that is $54 for the week, about 9.5% more cheese than the sales account for.
Fifty-four dollars of cheese doesn't sound like much until you multiply it by fifty-two weeks and by every other ingredient.
Where variance comes from
- Over-portioning. Twelve ounces where the recipe says ten. The most common cause, and the easiest to fix with scales and portion charts.
- Waste. Dropped, burned, expired or remade food that was never logged.
- Unrung sales and comps. Food that left the kitchen without a ticket, or staff meals that weren't recorded.
- Recipe drift. The kitchen changed how it makes something and the recipe on file didn't follow.
- Counting and receiving errors. A case counted twice, or an invoice entered for stock that never arrived.
- Theft. Real, but usually smaller than the four above. Rule them out first.
How often to count
Counting everything every week is a lot of work for low-value items. A practical rhythm is a short hot list of high-value, high-variance items (cheese, proteins, liquor) counted weekly or even daily, and a full count monthly. Count at the same time of day, before deliveries or after them but always the same, so the numbers compare.
What a "good" variance is
There is no universal number. What matters is your own trend: variance that is steady and small is a kitchen under control; variance that jumps is a question to ask this week. Track it per item and per location, because a total across a group hides the one site with a problem.
How CashSheet does the arithmetic
In CashSheet Back Office, recipes explode each item sold into its raw ingredients, so theoretical usage comes straight from the till's sales. Counts, including hot-list counts, give actual usage, and the variance report shows the gap by item, store and period. If you run a commissary, transfers carry their cost to each store, so variance lands on the site that caused it. Each week the variance also posts to your accounting, so it shows on the P&L instead of hiding in cost of goods.
Costing a single item first? Try the food cost calculator, or read how to cost a pizza.


