Actual vs Theoretical Food Cost Calculator
Find out whether you're spending more on food than you planned — and what the gap is costing you over a year. You need three numbers off your stock count, your food sales, and the food cost percentage you aim for. No recipe costing required.
The trade calls this your actual vs theoretical (AvT) variance. It's simpler than it sounds, and it's usually the fastest way to find money leaking out of a kitchen.
1 What you spent on food
Three numbers from your stock count and your invoices.
Value of the food in your fridges, freezers and dry store on day one.
Add up every supplier invoice for the period.
The same count again, on the last day of the period.
2 What you sold
Total food sales for the same period, before tax. Leave out drinks if you track them separately.
3 What it should have cost
If you cost your recipes, use that figure. If you don't — and most kitchens don't — use the food cost percentage you aim for.Your budget, your accountant's number, or what you ran in a good year.
The percentage of sales you budget for food.
Your answer appears here
Once the three steps are filled in, you'll see:
- • How much more (or less) you spent than planned
- • What that adds up to over a year
- • Whether it's normal or worth chasing
Everything is worked out in your browser. Nothing you type is sent anywhere or stored.
How actual vs theoretical food cost is calculated
Two numbers, worked out separately, then compared. Most operators track one of them and assume the other.
Actual cost of goods sold
Opening stock + Purchases − Closing stock
Divide by net food sales for your actual food cost percentage. This number is only as good as your stocktake.
Theoretical cost of goods sold
Σ (recipe cost of each dish × quantity sold)
What perfect execution would have cost, given exactly what you sold. Needs costed recipes and sales quantities by item.
Variance
Actual − Theoretical
Expressed in currency, and in percentage points of sales. Points make periods comparable when sales move.
What counts as a normal variance?
There is no universal benchmark, and be sceptical of anyone who gives you one as a hard figure — it varies enormously by cuisine, menu complexity, prep method and how honestly waste is recorded. A kitchen doing butchery in-house will run a wider variance than one buying portioned protein, and neither is wrong.
As a working rule of thumb, many operators treat under a percentage point as ordinary noise, one to two points as a pattern worth chasing, and anything beyond that as structural. But your own trend is a far better benchmark than any industry figure. A variance of 2 points that has been stable for a year is a costing problem. A variance of 0.8 points that doubled last month is an operational problem, and it is the more urgent of the two.
What causes the gap, in order of likelihood
Portioning drift
The most common cause by a distance, and the least dramatic. A recipe says 180g and the line serves 210g. Nobody is doing anything wrong on purpose — but on a high-volume dish that single habit can account for most of a variance on its own.
Waste that never gets recorded
Spoilage, prep trim beyond the assumed yield, dropped plates, over-production on a slow night. If it is not written down it does not reduce theoretical cost, so it lands entirely in variance.
Recipes that no longer match the plate
The kitchen improved the dish, added a garnish, changed a supplier spec — and nobody updated the recipe. Your theoretical cost is now describing a dish you stopped serving months ago.
Receiving errors
Short deliveries signed for in full, price rises absorbed without anyone noticing, credits never chased. This shows up as purchases you paid for and never received.
Comps, staff meals and unrecorded transfers
Food that left the kitchen without a sale attached. Legitimate, but if it is not captured it looks identical to theft in the numbers.
Shrinkage
Last on the list deliberately. It is real and it happens, but most first-time variances turn out to be the five causes above. Rule those out before you start suspecting people.
Before you act on the number, check it is real
Most first AvT calculations are wrong, and they are wrong in the direction of looking alarming. Rule these out before you change anything or accuse anyone:
- The stocktake was rushed or inconsistent. If opening and closing counts were taken by different people using different assumptions about part-used containers, the variance you are looking at is measurement error.
- Invoices landed outside the period. A delivery received in the period but invoiced after the cut-off inflates variance now and deflates it next period.
- Transfers between outlets or bar and kitchen were not recorded. One site shows a loss and another an unexplained gain.
- Recipe yields are theoretical rather than measured. If your yields came off a supplier spec sheet instead of your own scales, theoretical cost is understated and every dish inherits the error.
How often should you run it?
Monthly is the realistic minimum, weekly is where it becomes an operational tool rather than a post-mortem. The value is not in any single number — it is in the trend, and in being able to tie a movement back to something that changed. Run it once and you have a statistic. Run it every week and you have a control.
The practical constraint is usually the stocktake, not the maths. Weekly AvT means weekly counts, which is only sustainable if counting is quick and the item list is disciplined.
Want this calculated for you, every week?
This page needs you to bring the numbers. FullGuest already has them — costed recipes, live supplier prices, sales by item and stock movements in one place — so the variance is worked out continuously, per outlet, and broken down to the dishes driving it.