The Restaurant Inventory  Guide

By Rob Bruni, founder of OpsEdge Hospitality, twenty years running restaurants, bars, clubs and resorts. Updated November 2026.

Free printable: Year-End Inventory Count checklist (PDF)

Restaurant inventory is the weekly count of everything in the walk-in, the freezer, the dry storage and the line, valued at what you paid for it. It is the only way to get a true cost of goods sold, the only way to see waste and theft, and the only sane basis for ordering. Most operations that say they "do inventory" count monthly, late, and do nothing with the number. This guide covers the COGS formula, how to set up a count sheet that takes under an hour, a worked example, and how the count feeds food cost, variance and the order.

Why count: the number you cannot get any other way

Invoices tell you what you bought. The POS tells you what you sold. Neither tells you what you used. The difference between purchases and usage is the change in what is on the shelf, and the only way to know that is to count it. Without a count, a month where you stocked up looks like a terrible food cost and a month where you ran the walk-in down looks like a great one, and both are wrong.

The second reason is variance. Once you know what you used, you can compare it to what the POS says you sold (converted to ingredients through recipes). The gap is waste, over-portioning, spoilage, theft and unrung comps. A restaurant doing $100,000 a month in food sales with a three point gap between theoretical and actual food cost is losing $3,000 a month of product, and most that have never counted carefully are in that range.

The COGS formula

Beginning inventory is last week's ending count. Purchases are every invoice dated in the period, including the ones that arrived without a signature. Ending inventory is this week's count. Run it separately for food and beverage, and inside food by category if you want to know where the problem is (proteins, produce, dairy, dry, frozen).

Two adjustments keep it honest. Transfers: product moved to the bar for garnish or to a second outlet is not usage here, it is usage there. Employee meals and comps: either ring them through the POS at cost so they show in sales, or log them and subtract from COGS. Pick one and never change it.

Setting up the count sheet

The count sheet is a list of every product you stock, in the order you walk past it, with a unit and a cost. Build it once and the weekly count becomes mechanical.

• Walk order, not alphabetical. Walk-in left to right top to bottom, then freezer, then dry storage, then the line and reach-ins. The sheet should match the walk.
• One line per product and pack. A case of chicken breast and a 10-pound bag of the same are two lines with two costs.
• A counting unit that matches how it sits on the shelf. Cases for sealed cases, each for open units, pounds for bulk protein and produce, portions for pre-portioned items. Note the conversion (one case equals six #10 cans) on the sheet so nobody does math on the floor.
• Current unit cost from the most recent invoice. Update it when a price changes, or at least monthly. A count valued at last year's prices gives you last year's food cost.
• Category tags (protein, produce, dairy, dry, frozen, paper) so you can roll up by category.
• Par and order columns, covered below.

Expect 150 to 350 lines for a full service kitchen. If it is 600, you have a SKU problem, not a counting problem.

How to count in under an hour

1. Same day, same time, same people. Sunday night after close or Monday morning before deliveries. Consistency matters more than the day.
2. Nothing comes in during the count. If a delivery arrives mid-count, it sits in the office until the sheet is done.
3. Two people. One counts and calls, one writes. It halves the time and most of the errors. The person who orders should not be the only person who counts.
4. Count open containers by tenths or by weight. Half a case is 0.5. A bulk tub of sauce goes on the scale with the tare written on the sheet. Do not skip the open ones; they are where the money is.
5. Count everything, including the line. Product in the reach-ins and on the steam table is still product. Count the ugly stuff at the back of the walk-in too.
6. Enter it the same day. A count on a clipboard for three days is a count that gets lost.
7. Review before you order. The count feeds the order. If something is short, find out why before buying more of it.

A 250-line kitchen with two people takes 40 to 55 minutes. Entering and reviewing is another 15 to 20. The first two weeks take longer while the sheet gets fixed. After that it is routine.

A worked example

Theoretical food cost from recipes and the POS mix came to 29.4 percent. The gap, 3.7 points, is $1,920 of product this week that was bought and not sold. The category roll-up showed proteins at 2.4 of those points. That is where to look first.

From the count to food cost, variance and waste

The count gives you actual usage. Three things come out of it.

• Actual food cost by week and by category, as above. Compare to the benchmark for your concept (28 to 34 percent for most full service, see the food cost guide) and to your own trend.
• Variance: actual usage versus theoretical usage from recipes and sales mix, by product. Sort by dollars. The top ten products will hold most of the gap, and they are almost always proteins and the one expensive produce item.
• Waste tracking. A waste log (what, how much, why, who) during the week explains the variance. Without it you know there is a gap, not where it went. Our Waste, Spoilage and Comp Tracker is built for this.

Inventory value itself is worth watching. Most full service restaurants hold five to nine days of food on hand. Fifteen days means cash tied up and spoilage waiting to happen.

Ordering from the count: pars and order guides

A par is what you need on hand to get to the next delivery with a buffer. Set it from usage, which you now have from the count, not from memory. Order quantity is par minus the count. Review pars quarterly and whenever the menu changes.

Ordering from pars does two things: it stops the walk-in filling with slow movers that spoil, and it makes stock-outs of fast movers rare because the par is built from how fast they actually move. It also removes the Thursday panic order at full price from the local market. Our Restaurant Inventory and Ordering Guide has the count sheet, usage, variance and par ordering in one workbook.

Mistakes that make counts useless

• Counting on different days. A Sunday count one week and a Wednesday count the next makes usage meaningless.
• Stale unit costs. The quantities are right and the dollars are wrong.
• Skipping the line, the reach-ins or the second storage area. Product not counted looks like product used.
• Not recording transfers and employee meals. They appear as variance and the kitchen gets blamed for the bar's garnish.
• Letting the person who orders count alone. It is not about trust. A control checked only by the person it controls is not a control.
• Counting and never comparing. A count entered and never set against sales is an hour wasted every week.
• Chasing a half-point with sloppy counts. Tenths and estimates are fine for finding a three point problem. Weigh the top products before you accuse anyone of anything.

The bar inventory guide covers the same routine for liquor, draft and wine, where variance by bottle is the goal. The Prime Cost System page covers the help available when you want the first count and variance report built with you.

Related guides and tools

Free tool: the par level calculator sets pars from usage and the delivery schedule and tells you what to order. No signup.

FAQ