Food cost creeps up one over-portioned plate at a time. Here’s a practical inventory routine — counts, pars, recipe costing and variance — that catches it early.
July 2, 2026
Food cost rarely blows up in one dramatic event. It creeps — a heavier pour here, a case of chicken that went up a dollar there, a prep cook who trims a little too aggressively, a delivery that was short two cases and nobody noticed. By the time the month-end numbers land, the damage is done and you cannot reconstruct where it went. Inventory management is how you catch that drift while it is still fixable: a routine of counting what you have, costing what you sell, and comparing what you should have used against what actually disappeared. This guide walks through a practical system a busy independent restaurant can actually sustain — no clipboard martyrdom required — and shows where your POS should be doing the heavy lifting for you.
Most operators can name their food cost percentage from the last P&L, but far fewer can say why it moved. The usual suspects are mundane: over-portioning (the most common and least visible), kitchen waste that never gets logged, supplier price creep that outpaces menu prices, spoilage from over-ordering, short deliveries that get signed for anyway, and — less often than owners fear but more often than zero — theft. Each of these erodes margin a fraction of a point at a time. None of them show up in your sales reports, because sales reports only tell you what you sold, not what you used. The gap between the two is where your money leaks, and the only way to see that gap is to count.
Your food cost percentage is the cost of the food you used divided by the food sales it generated. The formula is simple: take your starting inventory value, add purchases for the period, subtract your ending inventory value — that is your cost of goods sold — then divide by food sales for the same period. What counts as a “good” number depends heavily on your concept: pizza and pasta concepts can run lean because flour and cheese are cheap relative to menu price, while steakhouses and seafood-driven menus run structurally higher and make it back on the check size. Rather than chasing a universal benchmark, establish your own baseline over a few periods and manage against the trend. A number that jumps two points month over month is telling you something specific happened — and with a weekly routine, you will know what.
The biggest reason inventory systems fail is that they are designed for a spreadsheet enthusiast, not a restaurant. The fix is to count at different frequencies depending on how fast an item moves and how expensive it is. Full monthly counts satisfy your accountant; weekly counts of your high-value, high-theft, high-spoilage items are what actually protect your margin.
Counting tells you what you used; recipe costing tells you what you should have used. Every item on your menu should have a costed recipe: each ingredient, the exact portion, and its current cost, adding up to a plate cost you can compare against the menu price. This is tedious to build the first time and enormously valuable forever after. It tells you which items are quietly unprofitable, which deserve a feature on your online ordering home screen, and exactly how much a supplier price increase should change your menu. It also unlocks the single most powerful report in inventory management: theoretical versus actual usage. If your costed recipes say Tuesday’s sales should have consumed forty pounds of chicken and your counts say fifty pounds left the walk-in, you have a ten-pound variance with a short list of explanations — portioning, waste, or shrinkage — and a specific conversation to have at the next pre-shift.
Here is a realistic cadence for an independent restaurant. It front-loads the work into small, repeatable blocks instead of one dreaded monthly marathon.
| When | Task | Why it matters |
|---|---|---|
| Delivery days | Check invoices against what actually arrived; note price changes | Short deliveries and price creep get caught at the door, not at month end |
| One fixed day weekly | Count proteins, seafood, liquor and other high-value items | Your riskiest dollars get watched fifty-two times a year, not twelve |
| Same day, after the count | Run theoretical vs actual variance on those items | Turns a vague “food cost feels high” into a named item and a cause |
| Before ordering | Order to par using current on-hand counts | Stops over-ordering, which quietly becomes spoilage |
| Month end | Full count and inventory valuation | Clean cost of goods sold for your books and your baseline trend |
Everything above is possible on paper, and generations of operators did it that way. But the reason inventory management has gotten dramatically easier is that a modern POS can automate the tedious half. When your recipes live in the same system that rings in every sale, the platform depletes inventory in real time as orders come in — every burger sold pulls a bun, a patty and two ounces of sauce from theoretical stock, whether the order came from the counter, your online ordering, or the kitchen display queue. Purchases entered against invoices keep costs current, so recipe costs update when your suppliers move prices. Variance reports stop being a spreadsheet project and become a report you open. Low-stock alerts fire before the Saturday rush instead of during it. If your current system cannot do this — or charges a stack of third-party integration fees to bolt it on — factor that into the real cost of your POS.
If you run more than one site, inventory discipline pays twice — but only if the system is built for it. You want one shared item catalog and recipe library so “chicken breast, 6 oz” means the same thing everywhere, with per-location counts, pars and supplier costs layered on top. That structure lets you compare variance across sites — if one location runs chicken variance three points higher than its siblings on the same recipes, you have found either a training gap or a shrinkage problem — and it makes commissary or central-kitchen transfers trackable instead of invisible. A multi-location native platform gives you that roll-up in one dashboard; stitching it together from per-site spreadsheets is how the picture stays blurry.
Novaryq builds inventory into the same all-in-one platform that runs your registers, kitchen and online ordering — recipes deplete stock in real time from every channel, counts and pars live next to your sales data, and variance is a built-in report rather than a weekend spreadsheet project. It is multi-location native with one shared catalog and per-site counts, offline cash is Beta on supported, prepared terminals with the workflow enabled; inventory depletion during an interruption is not implied, commission-free on direct orders, and available across the US and Canada with a month-to-month option. If you are starting from scratch, our guide for new restaurants covers how inventory fits into the rest of your opening stack, or you can explore the platform to see how the pieces connect.
Count high-value, fast-moving items — proteins, seafood, liquor — weekly, and do a full wall-to-wall count monthly for accurate books. Weekly counts of your riskiest items are what actually catch portioning, waste and shrinkage problems while they are still fixable; the monthly count is for your accountant.
It depends on your concept. Cheap-ingredient, high-markup menus like pizza can run lean, while protein-heavy concepts like steakhouses run structurally higher and recover it on check size. Rather than chasing a universal number, establish your own baseline over a few periods and investigate any move of a point or more.
Theoretical cost is what your sales should have consumed based on costed recipes — every burger sold depletes a bun, a patty and a set portion of sauce. Actual cost is what your physical counts say you really used. The gap between them, called variance, points directly at over-portioning, unlogged waste, or shrinkage.
Yes, if recipes live in the same platform that rings in sales. The system depletes theoretical stock in real time from every order channel, keeps costs current as you enter invoices, and turns variance into a built-in report. What it cannot automate is the physical count — you still need a weekly routine to compare against.