Count weekly, the night before order day. Lay out the sheet in shelf order, one storage area at a time, with one fixed counting unit per product. Split the storage areas between two or three people counting into one live session, and use OK/LOW/OUT checks for cheap items. A full count then fits in under an hour.
Inventory night is the shift everyone dreads. One person, a clipboard, three hours alone in the walk-in after close — and the numbers still come back wrong. So the count gets skipped, then skipped again, and pretty soon you're ordering by gut and pricing your menu off numbers from last quarter.
Counting doesn't have to be slow, and it doesn't have to be a solo punishment. This guide covers how often to count, how to run a count that finishes in under an hour, and how to split the work across your team without wrecking accuracy.
Why most inventory counts fail
Counts rarely fail because someone can't count. They fail because of how the count is run:
- Inconsistent frequency. Weekly counts give you comparable data points. Counts taken "whenever we get around to it" don't, because you can't compare two snapshots taken at random intervals.
- No standard process. If the sheet is ordered differently than the shelves, every product becomes a scavenger hunt.
- One exhausted person after close. Tired people skip shelves, transpose digits, and estimate. Nobody catches it because nobody else was there.
- Mixed units. One week flour is counted in bags, the next in pounds. The numbers look fine and mean nothing.
The cost of a bad count isn't the wasted evening. It's everything downstream. Your food cost percentage is built on beginning and ending inventory — garbage counts in, garbage food cost out. If your food cost has been drifting and you can't explain why, an unreliable count is suspect number one. The food cost guide walks through that connection in detail.
How often should you count?
Weekly is the answer for most kitchens. It's frequent enough to catch a problem — spoilage, theft, a portioning issue — within days instead of letting it burn cash for a month. Monthly counts tell you something went wrong; weekly counts tell you while you can still do something about it.
Not every product needs the same cadence, though. A practical hybrid:
- Daily: high-value perishables. Proteins, seafood, anything that walks away or spoils fast.
- Weekly: most of the kitchen. This is your food-cost backbone.
- Monthly: stable dry goods. The case of napkins isn't going anywhere.
One more rule: count before you order, not after the truck arrives. A count taken right after a delivery shows you flush on everything and teaches you nothing about what you actually need. Counting the night before order day means your order is built on real numbers, hours old.
Your accountant has a stake in this too. If you keep a running book inventory, IRS Publication 538 says you must take a physical inventory at reasonable intervals and adjust the book amount to agree with what's actually on the shelf.
How to run an efficient count, step by step
Organize by storage area, not alphabetically
Your count sheet should mirror your shelves. Walk-in top to bottom, then freezer, then dry storage — one pass through each room, no backtracking. An alphabetical sheet sends the counter sprinting between the freezer and the dry shelf eleven times. This one change cuts more time off a count than anything else on this list.
Use consistent units
Pick one counting unit per product and never change it. Mixing units is a common source of bad numbers, and it's invisible: "12" looks like a fine count whether it meant bottles or cases. Write the unit on the sheet, on the shelf label, wherever it takes.
Handle cases and loose items separately
Real shelves hold 3 full cases and 14 loose portions, and forcing someone to do case math in their head mid-count is where errors breed. Count cases, count loose, record both. Let the math happen on paper — or better, let software do the conversion the moment you key it in.
Don't round, don't estimate
"Call it 20" feels harmless. It isn't, because rounding errors don't cancel out — they pile up. A $0.50 average error across 200 products is $100 of phantom inventory in a single count, and it lands directly on your food cost line as money that appears to exist and doesn't. Count what's actually there, including the open container. Half a case is half a case.
Read the date marks while you're there
Every container in the walk-in passes through your hands during a count, so read the labels as you go. The FDA Food Code, the model state and local regulators use for their own food safety rules, calls for date marks on ready-to-eat, perishable food prepped in-house and held cold, so everyone knows when it must be used or thrown out. Pull anything past its date, log it as waste, and count what's left.
How do I split a count between three people?
Give each person their own storage areas and have everyone count into the same live session, so nothing needs merging at the end. One takes the walk-in, one takes the freezer, one takes dry storage, and a three-hour count becomes three short parallel counts.
The catch is how teams usually do this: three clipboards, then someone merges the sheets at the end. That's where it falls apart. Handwriting gets misread, a page gets wet, two people count the same shelf, nobody counts the low-boy because everyone assumed someone else had it.
Real-time collaborative counting removes the merge step. Everyone counts on their own phone in the same live session, progress is visible as it happens, and there's nothing to merge: when the last storage area turns green, the count is done. A quieter benefit: every count is stamped with who entered it, so when a number looks off next week, you know who to ask instead of interrogating the whole crew.
In Rinvy, collaborative counting is on the free plan, which covers a 50-product kitchen — small teams splitting the count is the core of how the product works, not an upsell. Paid plans (Standard at $29/month, Pro at $69/month) add larger catalogs and deeper analysis on top.
Precision counts vs. status checks
Not everything deserves a decimal. Nobody should be weighing the salt or counting ketchup packets.
Use two modes:
- Precision counts for anything that drives food cost: proteins, dairy, produce, expensive dry goods. Exact quantities, cases plus loose.
- Status checks for the rest: a simple OK, LOW, or OUT. A few seconds per item, and it still catches the thing that matters — running out mid-service.
Sorting your catalog into those two buckets once will permanently shrink your count time without giving up the accuracy where accuracy pays.
Common counting mistakes
- Counting right after a delivery. Everything looks full; you learn nothing. Count before order day.
- Counting at inconsistent times. Tuesday close one week, Friday mid-service the next — the two snapshots aren't comparable.
- Skipping storage areas because "nothing changed." The skipped shelf is exactly where shrinkage hides, because everyone knows it's never checked.
- Not recording who counted what. Without attribution, every odd number becomes an unsolvable mystery.
- Paper. It gets wet, lost, and misread, and then someone retypes it into a spreadsheet — a second chance for errors.
From counting to action
A count is only worth the hour if the numbers get used. Three things should come out of every count:
- Your food cost percentage. Beginning inventory plus purchases minus ending inventory, divided by food sales. Weekly counts give you a weekly food cost, so problems surface within days instead of at month-end.
- A shrinkage check. Compare what the count says against what should be there given purchases and sales. A gap means waste, over-portioning, or theft — and now you know which products to watch.
- Your order. Tonight's count should directly shape tomorrow's order — par levels and usage rates turn counted stock into suggested quantities, so ordering stops being a memory exercise.
A disciplined count pays off in every decision you make this week. What to order, what to 86, and what to reprice all rest on numbers you can trust.