Why your food cost is lying to you
The number on your weekly P&L is a story, not a fact. Here's how to read what it's really telling you about the kitchen, the suppliers, and the menu.
Most operators look at one food-cost percentage every Monday morning. It comes out of the accounting software, lands in an email, and becomes the headline number for the week. If it's up, somebody's in trouble. If it's down, everybody relaxes. Both reactions are usually wrong · because that single percentage is a blend of three or four very different stories, and you can't fix what you can't separate.
The number nobody questions
Food cost as it appears on most P&Ls is roughly: opening stock + purchases − closing stock, divided by sales. It's a clean equation. The problem is that every input is fragile. Closing stock is whoever's count was used. Purchases include things you haven't sold yet. Sales include comps, voids, and that one party that paid in cash. The percentage is real, but it isn't truth.
Three stories inside one number
When the weekly food cost moves, ask which of these three things actually moved · because the fix for each is completely different.
1. Theoretical vs actual gap
Theoretical food cost is what your menu mix should have produced this week given recipe yields and selling prices. Actual is what the kitchen actually consumed. The gap between them is waste, over-portioning, comps, theft, or recipe drift · operational, not strategic. Close the gap with prep sheets, line checks, and a serious approach to portioning.
2. Supplier price drift
A 3% move in beef prices that you didn't reprice for is invisible until the period closes. By then the damage is done. Track three to five "watch list" items every week against the price you costed the menu at, and flag anything past a 5% move. Don't wait for the P&L to tell you.
3. Mix shift
The same menu sold at the same prices can produce wildly different food-cost percentages depending on what guests order. A weekend with a lot of high-margin cocktails and a Tuesday full of pasta orders are not the same business · even if the total is identical. Menu engineering is not a poster on the wall; it's the only honest way to read mix.
You don't have a food-cost problem. You have a theoretical-vs-actual problem, or a supplier problem, or a mix problem · and the three want completely different things from you.
What to do on Monday morning
- Before opening the P&L, look at the gap between theoretical and actual food cost for the week. That's the operational number.
- Check your watch list of five high-volume ingredients against last month's costed price. Anything over 5% is a repricing conversation.
- Pull the top ten dishes by quantity and the top ten by gross margin. If those lists don't overlap meaningfully, the menu is doing the wrong work.
- Then, open the P&L. The single number will now read like a sentence, not a sentence with three words missing.
The point of the exercise
The point isn't to drive food cost down. The point is to know · every Monday · exactly which of three levers moved and why. Once you can read the number, you can stop being surprised by it. That's a quiet kind of control, and it's what the rest of the operation is built on.