Food cost percentage: the formula, and why yours keeps moving

The formula

Food cost percentage is cost of goods sold divided by food sales. Cost of goods sold is not simply what you bought — it is what you actually used:

Beginning inventory + purchases − ending inventory = cost of goods sold.

So a period that opened with $8,000 of stock, bought $19,000 and closed with $7,000 used $20,000 of food. Against $62,000 in food sales, that is 32.3%.

Why purchases alone will mislead you

Dividing invoices by sales is quicker and is wrong in a specific, recurring way: it charges a heavy delivery to the week it arrived rather than the weeks it fed. A big Thursday order makes that week look terrible and the next week look excellent, when nothing about your operation changed.

If you are not counting inventory, at least count the few categories that move most in value — proteins, dairy, alcohol. Partial counts beat none, and they beat a purchases-only number that swings with the delivery schedule.

Theoretical versus actual

Theoretical food cost is what your recipes say you should have spent, given exactly what you sold. Actual food cost is what you did spend. Both are useful, but the number that teaches you something is the gap between them.

Theoretical cost assumes every portion was measured, nothing spoiled, nothing was comped, and nothing walked out the door. Reality does not, and the distance between the two is the size of that difference in dollars.

What the gap is made of

A persistent gap between theoretical and actual almost always decomposes into a short list:

• Over-portioning — the most common and the least visible, since nothing is thrown away
• Spoilage and waste, especially on items ordered on a standing schedule rather than to demand
• Prep loss and trim that recipes assume away
• Comps, staff meals and remakes that never reach a ticket
• Receiving errors: paying for a case, receiving less, and never reconciling

What counts as normal

Ranges depend heavily on menu, format and pricing, so use these to orient rather than to judge:

• Full-service restaurants: roughly 28–35%
• Coffee shops and cafés: often lower on drinks, 20–30% blended once food is included
• Bars: spirits frequently under 20%, beer higher, blended by mix
• Quick service: roughly 28–33%

The number that actually pays

Overall food cost tells you whether there is a problem. Per-item margin tells you where it is. Two products can sell equally well and keep very different amounts of the dollar, and the weaker one is usually invisible precisely because it sells fine.

That is where price moves earn the most: a small increase on a high-volume, low-margin item almost always beats a large increase on something nobody orders. Rank by contribution — margin multiplied by volume — rather than by margin alone.

Watching it over time

One period's food cost is nearly meaningless on its own, because inventory timing, a single heavy delivery, or one bad count will move it a full point. A trend over eight or ten weeks is what separates a genuine drift from ordinary noise.

When the trend does move, check price before blaming the kitchen. Supplier increases arrive quietly and rarely announce themselves, and a menu priced against last year's invoices will lose margin without anything going wrong on the line.