Where Café Stock Actually Disappears — and How to Find It

You bought 60 litres of milk this week. Your sales say you should have used about 44. You count the fridge and find 9. Sixteen litres went somewhere and nobody can tell you where. That gap has a name — shrinkage — and in most cafés the first instinct is to suspect somebody. That instinct is usually wrong, and acting on it before you have measured anything is how good owners lose good staff. Shrinkage is a measurement first. Only once you have measured it for a few weeks does it start telling you what kind of problem you have.

Shrinkage is a number, not an accusation

Shrinkage is simply the difference between the stock you should have — opening stock, plus what you bought, minus what your sales say you used — and the stock you actually count. It is normal for it to be something other than zero. Coffee is lost to grinder purges and dialling in. Milk is lost to steaming more than the drink needed. A shot gets pulled badly and binned. None of that is theft; it is the ordinary friction of making things by hand.

What matters is the size and the shape. A small, steady variance across everything is friction. A large variance concentrated on a few specific items, or a variance that appears on the same shift week after week, is a process problem — and only sometimes a people problem. Get the number first, then read its pattern. Publishing an accusation before you have four weeks of data is a mistake you cannot take back.

The five places café stock actually goes

Work through these in order, because between them they explain the large majority of what you will ever find:

• Over-pouring and free-hand portions. The single biggest cause in most cafés, and entirely invisible. A barista steaming a jug for a single flat white and tipping half of it away does that forty times a shift. Nobody is doing anything wrong on purpose, and the loss is larger than almost anything else on this list.
• Staff drinks and comps that never get rung in. Ordinary and usually authorised — but if they are not entered at zero or at cost, your stock and your sales permanently disagree, and every future variance calculation is built on a number that was already broken.
• Voids, refunds and remakes. A drink made twice was paid for once. If a remake is not recorded, the ingredients leave your stock with no sale attached, and it looks identical to theft in the report.
• Supplier short-delivery and substitution. You were invoiced for 24 and 22 arrived, or the case size changed and nobody updated the unit. This one is worth checking early because it is the easiest to fix and the cheapest to recover.
• Waste and spoilage that was never logged. Stock that expired, spilled or was dropped is a real loss, but if it goes in the bin unrecorded it shows up as an unexplained gap instead of a known cost. Our guide to reducing food waste covers the logging routine that turns this from a mystery into a line item.

Notice that four of the five are recording failures rather than losses. That is the usual shape of it: most cafés do not have a stock problem, they have a stock-recording problem, and the recording problem is what makes the real losses impossible to see.

How to measure it without turning the week upside down

You do not need a full stocktake to find shrinkage. You need a small number of items counted often, which is far more useful than everything counted rarely.

1. Pick five to eight items that are high-value, high-volume or both — milk, house beans, the syrup that goes in everything, the pastry you sell most of. These few will carry most of your loss.
2. Count them at the same point in the cycle, ideally the same morning each week before any delivery arrives. Consistency matters more than precision.
3. Record purchases against the same period, using delivery notes rather than invoices where the two differ.
4. Calculate theoretical usage from your sales: what each item sold should have consumed, based on the recipe. This is the step that needs recipe-linked stock to be practical, and it is why recipe-based inventory is the foundation of everything here.
5. Write the variance down every week, even when it is small. One week's figure is noise. Six weeks is a pattern, and the pattern is the whole point.
6. Only then look at who and when. By this stage the data will usually have pointed at a process rather than a person.

Do this for a month before you change anything. It is tempting to fix the first thing you find, but you will not know whether a fix worked without a baseline to compare it to, and the baseline takes four counts to establish.

The controls that close the gap

Almost all of these are process changes rather than surveillance, and process changes are what actually hold. The ones worth doing first:

• Weigh the pour, at least for a week. Scales on the bar for a few shifts recalibrate everyone's sense of a portion, and the effect outlasts the scales. Milk jugs sized to the drink do the same job permanently.
• Give staff drinks a button. A zero-price or cost-price item on the till takes two seconds, keeps your stock honest, and removes any ambiguity about what is allowed.
• Require approval for voids and refunds. Not because you expect the worst, but because an approved void is attached to a person and a time, which means it can be explained. An unapproved one is just a hole.
• Check deliveries against the note before signing. Ten seconds at the door, and it is the only moment you have any leverage with a supplier.
• Keep the waste log at the bin, not in the office. A log that requires walking somewhere does not get filled in.

One more, and it is the one owners resist: separate the person who counts the stock from the person who orders it, if you have enough people to do so. This is not about suspicion — it is the same principle that says the person counting the drawer at close should not be the only person responsible for the shift's takings, which our guide to the daily cash-up sets out. Good controls protect honest staff from being suspected as much as they deter anything else.

When the stock gap and the cash gap are the same gap

Stock variance and till variance are usually treated as two separate investigations, and they are often the same event seen from two sides. A sale rung in under the wrong item takes the wrong ingredients out of stock and puts the right money in the drawer. A drink handed over and never entered leaves the stock short and the cash correct. A refund given from the drawer without a record leaves the cash short and the stock correct.

If you are running two investigations and neither resolves, put the two records side by side for the same shift. The direction the two numbers move tells you more than either figure alone, and it is usually the fastest route to an answer.

Where MidaOne fits

The whole exercise depends on knowing what your sales *should* have consumed, and that number only exists if stock is tied to recipes rather than counted separately. MidaOne deducts ingredients by recipe as each item sells, so theoretical usage is already calculated when you count, and the variance is a subtraction rather than a reconstruction. Voids, refunds and staff comps are scoped by role and attached to the person who made them, and because stock, sales and accounting sit in one system, the stock story and the cash story come from the same record instead of two you have to reconcile by hand. See the café POS guide for how the pieces connect.

Frequently asked questions

The cafés that lose the least stock are rarely the ones watching hardest. They are the ones where the ordinary events — a remake, a staff coffee, a dropped jug, a short delivery — all have somewhere to be recorded, so the number at the end of the week means something. Start counting five items every Monday. In a month you will know more about your café than a camera would have told you.