Inventory Shrinkage Calculator
Enter what your recipes say you should have used and what your stock count says you actually used, to see how much of the gap is explained — and how much is simply missing.
What shrinkage actually measures
Shrinkage is the difference between the stock you should have used and the stock you did use. It is not a category of cost in its own right — it is a gap, and the whole point of measuring it is to find out what is sitting inside the gap. Most operators discover shrinkage the same way: food cost comes in three points higher than expected, nobody changed a price, nobody changed a recipe, and no single invoice explains it.
Two numbers define the gap. Actual usage comes from your stock count and is a fact of arithmetic: whatever was on the shelf, plus whatever you bought, minus whatever is still on the shelf, must have left the building somehow.
Theoretical usage comes from the other direction: every item sold in the period, multiplied by the costed recipe behind it. If you sold 310 burgers and the costed recipe consumes $3.40 of ingredients, theoretical usage for that dish is $1,054. Add every dish together and you have what the period should have consumed.
Subtract one from the other and you have the variance. Then strip out the part you can already account for — waste you logged, meals you comped, staff food, samples poured for a supplier tasting — and whatever survives is unaccounted shrinkage.
That final figure is the one worth acting on. A large total variance with a good waste log behind it is a measurement success, not a crisis. A small total variance with no log at all tells you almost nothing, because you have no idea whether the offsetting errors cancelled out by luck.
Where the missing stock goes
Unaccounted shrinkage has four common sources, and they need completely different responses. Ranking them by likelihood before investigating saves an enormous amount of wasted suspicion.
Over-portioning
By far the largest contributor in most kitchens, and the most boring. A recipe specifying 140g of protein that goes out at 165g is a 18% overrun on the single most expensive line of the dish. Nobody is doing anything wrong on purpose; scales get put away during a rush and the hand takes over. Over-portioning is invisible to every system you own, because the stock left legitimately and the sale was rung correctly.
Unlogged waste
Product that spoiled, burned, or came back from a table and went into the bin without anyone writing it down. This is not really shrinkage — it is waste wearing a disguise. Improving the log converts it from an unexplained gap into an explained one, which is the entire value of running a waste cost calculation alongside this one.
Receiving and invoice errors
You were charged for 20kg and delivered 18kg. You were invoiced at $9.20 a kilo after agreeing $8.40. A credit note was promised and never arrived. These errors run in one direction far more often than chance would predict, and they inflate purchases without inflating the stock that arrived — which lands directly in your variance.
Theft
It exists, it is usually the smallest of the four, and it is the first one most people jump to. Investigate it last, not because it is impossible but because acting on the assumption of theft when the real cause is a 25g portion overrun damages a team for nothing and fixes none of the money.
Reading the number
| Unaccounted shrinkage as % of food sales | What it usually means |
|---|---|
| Under 0.5% | Within counting noise. Your recipes and count are broadly trustworthy. |
| 0.5–1.5% | Normal for an operation without portion controls. Mostly portioning and unlogged waste. |
| 1.5–3% | Systematic. Something specific is wrong — usually one department or one high-value item. |
| Over 3% | Either a broken count, a badly out-of-date recipe file, or a genuine control failure. |
Before treating a large figure as a loss, check that it is real. A variance above 3% is more often a data problem than a stock problem. The usual culprits are recipes that were costed two price rises ago, a stock count taken on a different day from the sales period, sub-recipes double-counted, or a new dish that was never added to the recipe file at all — which quietly removes its ingredient usage from the theoretical side while the actual side keeps counting it.
A negative result deserves the same scepticism. If actual usage came in below theoretical, you have either overvalued the closing count, missed an invoice that lands next period, or you are carrying recipe yields that are more pessimistic than reality. Money does not appear on its own.
Getting a theoretical figure worth comparing against
Shrinkage analysis is only as good as the theoretical number, and the theoretical number is only as good as your recipe file. There is no shortcut here: a variance calculated against stale recipes measures the age of your costings, not the health of your stock room.
- Cost every recipe at current prices. A file that has not been re-costed in six months will manufacture variance out of nothing. Rebuild costs in the recipe costing calculator whenever a major ingredient moves more than 10%.
- Use net yields, not purchase weights. If you cost a fillet at its as-purchased weight, every trim loss in the kitchen becomes phantom shrinkage. Apply a proper yield percentage to anything you butcher, peel or reduce.
- Include every sales channel. Delivery, catering, function menus and staff-facing tills all consume stock. Leave one out of the sales extract and its usage lands in your shrinkage figure.
- Match the dates exactly. The count and the sales report must cover identical periods. A count taken on Monday morning against sales running to Sunday night is fine; a count taken mid-service is not.
- Count the same way every time. Consistency matters more than precision. A count that is always slightly generous still produces a usable trend; a count that varies by counter produces noise you will misread as loss.
Once those five things are in place, the variance becomes a genuine measurement. Until then, treat the output as a rough direction rather than a dollar figure to chase.
Narrowing it down
A single site-wide shrinkage number tells you that money is leaving but not where. The way to find it is to stop calculating one variance and start calculating several.
Split by category first: proteins, produce, dry goods, dairy, beverage. Shrinkage almost never spreads evenly, and one category usually carries most of it. Beverage in particular behaves differently enough that it deserves its own analysis through beverage cost percentage rather than being folded into food.
Then within the worst category, run the variance on your ten highest-value items by usage. Those ten will typically account for a majority of the exposure, and because they are individually expensive, a small percentage error on any one of them shows up clearly. This is where over-portioning becomes visible: a protein running 12% over theoretical while everything around it is flat is a portioning finding, not a mystery.
Running counts more frequently on that short list also tightens the loop. Weekly counts on ten items are far more useful than a monthly count on four hundred, because a four-week-old signal has already lost the context that would let you explain it. Faster counting also feeds better ordering, which is what par levels and stock turnover both depend on.
Finally, close the loop with your headline number. Whatever you recover here shows up as a lower food cost percentage within one or two inventory cycles. If it does not, the shrinkage you thought you fixed was a measurement artefact.
Frequently asked questions
What is a normal shrinkage percentage for a restaurant?
Unaccounted shrinkage under 0.5% of food sales is tight, and 0.5–1.5% is common in operations without formal portion controls. Above 3% is usually a data problem — stale recipes or a mismatched count — before it is a loss problem.
Is shrinkage the same thing as waste?
No. Waste is food you know you threw away and can point to. Shrinkage is the part of your usage that nothing accounts for. Good waste logging shrinks the shrinkage figure by moving loss from the unexplained column into the explained one, which is progress even though the total variance has not changed.
How often should I calculate this?
Monthly is the minimum for a meaningful trend, and it matches most accounting cycles. If you are actively chasing a problem, run weekly counts on your ten highest-value items instead of full counts — you get a faster signal on the stock that carries most of the money.
My theoretical usage is higher than actual. What does that mean?
Usually an overvalued closing count, an invoice that has not landed yet, or recipe yields that are more conservative than what the kitchen really achieves. Treat a negative variance as a measurement error until you can prove otherwise rather than as a saving.
Should I include staff meals and comps in shrinkage?
Enter them separately, as this calculator does. They are real consumption with no matching sale, so they belong in the variance, but they are authorised and known. Keeping them visible stops a growing staff-meal programme from looking like a control failure.
Does shrinkage affect food cost percentage?
Directly. Any stock that leaves without a sale lands in cost of goods sold exactly like stock you sold, so it inflates food cost with no revenue attached. A persistent one-point shrinkage figure is a persistent one-point food cost problem.
Where should I start if the number comes back high?
Verify the inputs first: recipe prices, yields, date alignment and missing sales channels. If the figure survives that, weigh portions on your three highest-volume dishes for a week. Portioning and unlogged waste explain most genuine shrinkage before theft ever enters the picture.