Your food cost is a guess: why monthly inventory catches problems six weeks too late
October 13, 2026
Ask most operators their food cost and you will get a confident number. Thirty-one percent. Twenty-nine and a half. Ask where that number came from and the answer is usually the P&L, which means it is somewhere between four and six weeks old by the time anyone looks at it.
That lag is the problem. Food is one of your two biggest costs, alongside labor, and for most independents it is the only big one managed with a number that describes last month. This post is about why that number misleads you, what the honest version looks like, and what changes when you can see it every week instead of every quarter.
4–6 wks
how old your P&L food cost number is by the time anyone reads it
Under 2%
the gap between theoretical and actual food cost in a well-run kitchen
$40,000
what a 4% variance costs a restaurant doing $1 million a year
The once-a-month count tells you what already happened
A monthly inventory count produces a food cost percentage that is accurate for the period it covers and useless for fixing anything in it. By the time the count is done, the invoices are entered and the P&L closes, the month is over. The over-portioning, the spoiled case of chicken, the vendor who quietly raised the price of your most-used product all happened weeks ago and were paid for weeks ago.
Then the number comes back high. Thirty-four instead of thirty-one. So you have a meeting. Everyone agrees to tighten up. Nobody can say which of the hundred things that happened in those four weeks caused the three points, so the fix is a general one: watch portions, check the walk-in. The next count comes back at thirty-three and a half and the cycle repeats.
This is the same pattern behind the back-of-house bottlenecks that eat margins quietly. The problem is not that you do not care. It is that the feedback arrives too late to act on.
Your plate costs were true in March
Most recipe costing lives in a spreadsheet built once, usually by whoever cared most about it at the time, and updated when someone remembers. The ingredient prices in that spreadsheet were accurate the week it was built. Produce has moved since then. Protein has moved. Say the case of avocados cost $42 when the sheet was built and costs $61 now: your guacamole is still priced as if it were $42.
The result is a menu where the margins you think you have and the margins you actually have drift apart one item at a time. A dish that was a 28% food cost on paper is a 36% food cost on the plate, and because the spreadsheet never changed, nothing flagged it. You find out when the P&L food cost drifts up, and the P&L, again, cannot tell you why.
Vendor price creep is the sneakiest version of this. A 3% increase on one line item does not trigger anyone’s attention. Twenty line items each moving 3% over six months does, but only when it lands as a point and a half on the P&L.
Food cost percentage is a scoreboard, not a map
Food cost percentage tells you that you lost money. It does not tell you where. It is one number that blends every ingredient, every dish and every mistake of the month into a single figure, which is exactly why it feels precise and is almost impossible to act on.
A high food cost can mean any of the following, and the percentage cannot tell them apart:
- The kitchen is over-portioning a few key items
- Something is spoiling before it sells and nobody logged it
- A vendor raised prices and the menu did not follow
- Product is leaving the building without being sold
- The menu mix shifted toward lower-margin items
Each of those has a different fix. The first is a training problem. The second is a par and prep problem. The third is a pricing problem. The fourth is a controls problem. The fifth is a menu engineering problem. Treating all five with “watch portions” fixes, at best, one of them.
The two numbers that matter
The honest version of food cost is a comparison between two figures. The first is theoretical usage: what your sales should have consumed, based on your recipes. The second is actual usage: what your counts say actually left the shelf, which is your opening count plus what you received, minus your closing count.
The gap between them is variance, and variance is where the money is. Take one ingredient for one week:
Theoretical vs. actual: ground beef, one week
| Amount | Where it comes from | |
|---|---|---|
| Burgers sold | 140 | Your sales |
| Theoretical usage | 52.5 lb | 140 burgers × a 6 oz patty |
| Actual usage | 61 lb | Opening count + received − closing count |
| Variance | 8.5 lb | Beef you paid for and did not sell |
8.5 pounds of beef is a specific question with a specific answer. "Food cost was high" is not.
Across the whole menu, variance is usually expressed as a share of sales, and industry benchmarks are fairly consistent about what the bands mean:
What your variance is telling you
| Variance (share of sales) | On $1M in sales | What it usually means |
|---|---|---|
| Under 2% | Under $20,000 | Tight operational control |
| 2–3% | $20,000–$30,000 | Acceptable for most concepts; watch the trend |
| 3–5% | $30,000–$50,000 | Common, and a sign of room to improve |
| Above 5% | Over $50,000 | Something systematic is wrong |
On a restaurant doing $1 million a year, a 4% variance is roughly $40,000 in product you paid for and never sold. Cutting that in half is worth $20,000, and it does not require a single new guest.
Where the gap hides
Variance is almost never one big leak. It is a handful of small ones that you can only find item by item.
What changes when variance is a weekly number
Weekly variance turns food cost from a quarterly surprise into a Monday morning conversation. A manager who sees that ground beef ran 14% over theoretical last week can walk into the kitchen that day, pull a patty, weigh it and fix the problem before it costs another week of margin. The same manager looking at a monthly P&L sees three points of food cost and a hundred possible causes.
It also changes what you count. When variance is visible by item, you count the twenty or so ingredients that drive most of your cost every week, and everything else monthly. Proteins, cheese, seafood and anything with a short shelf life get a weekly count. Dry storage does not need one.
The number is only useful if it arrives while you can still do something about it.
That is the same shift the manager log made for shift reviews, and the same logic the labor cost post applied to scheduling. Your POS already knows exactly what sold, down to the modifier, and your kitchen display sees every ticket the line fires. Theoretical usage is sitting in that data. The only missing piece is connecting it to what is on the shelf.
The question to sit with this week
What would you do differently if you knew your real food cost every Monday? Not the P&L number. The item-by-item version, with the gap between what you should have used and what you did.
Most operators answer quickly: fix portions on two or three items, reprice a couple of dishes, have a word with a vendor. That is the whole point. The fixes are not hard. Seeing them in time is.
Next week we will show you how NX Restaurant closes that loop without a separate inventory system.
Frequently asked questions
How often should a restaurant count inventory?
Weekly for high-cost, short-shelf-life items such as proteins, seafood and dairy, and monthly for a full count of everything else. The 20 or so ingredients that drive most of your food cost deserve a weekly number. Counting everything weekly burns labor without adding much insight, and counting everything monthly means problems sit for weeks.
What is a good food cost variance?
Under 2% of sales between theoretical and actual food cost is considered tight control. Two to three percent is acceptable for most concepts. Three to five percent is common but signals room to improve. Anything above 5% points to a systematic problem worth investigating immediately, whether that is portioning, unlogged waste, pricing drift or shrink.
Why is my food cost percentage high when my recipes are costed correctly?
Because the recipe cost is a theoretical number and the P&L is an actual one. If recipes are costed correctly and food cost is still high, the gap is in execution: over-portioning, waste that was not logged, product that left without a sale, or a menu mix that shifted toward lower-margin items. Variance by ingredient tells you which.
Find out where your food cost is leaking
If your food cost is a number you see once a month, a 20-minute conversation with an authorized NX Restaurant partner is a good place to start. Bring your last P&L and we will talk through where the gap is most likely hiding.
Book a Demo