Counting Stock in a Café: How Often, and How to Do It Fast

Most café owners know they should count stock. Fewer do it, and the reason is almost always the same: the last time they tried, it took three hours, half the numbers were guesses, and the result didn't tell them anything they could act on. So it becomes a quarterly ordeal, or an annual one, or something that happens the month the accountant asks. A count that takes that long isn't a discipline problem. It's a method problem, and the method is fixable in an afternoon.

What the count is actually for

Counting stock answers one question: does what you physically have match what the system says you should have? Everything useful comes out of the gap between those two numbers. Without a count, your stock figure is a theory — it assumes every delivery arrived complete, every recipe was made to spec, nothing was dropped, over-poured, given away or written off unrecorded. All of those things happen in every café every week. The count is what turns the theory back into a fact.

This is different from deciding what to order, which is a forward-looking job driven by par levels and lead times, and different again from investigating a gap once you've found one, which is a shrinkage question. The count sits between them. It's the measurement that makes both of the others possible, and on its own it accuses nobody of anything.

How often should you count?

Not everything needs counting at the same rhythm, and trying to count everything at once is what makes the job unbearable. Split the list by what the item is worth and how fast it moves. A tiered approach means the expensive, fast-moving things — the ones where a problem costs you real money and shows up quickly — get counted often, while the slow, cheap, stable things get counted when there's time.

The day matters more than the frequency. Count at the same point in the week, before opening or after close, with the till shut and no deliveries mid-count. A count taken while service is running and stock is moving isn't a count, it's an estimate, and you will not be able to tell later which of the two you were looking at.

Building a count sheet that follows your shelves

This is the single change that turns a three-hour count into a twenty-minute one. Most count sheets are ordered alphabetically, or in whatever order the system exported them. That forces the person counting to walk the store room over and over, hunting for the next line. Instead, order the sheet the way your shelves are physically arranged: start at the left of the dry store, work along and down, then the under-counter fridge, then the walk-in, then the bar. The sheet becomes a route.

Once the sheet follows the room, the rest is mechanics:

• One unit per line, and write it on the sheet. Decide once whether milk is counted in litres or in cartons, and never let two people answer that differently. Most variance nonsense traces back to a unit mismatch, not a missing item.
• Two people is faster than one. One counts and calls, one writes. It also removes the temptation to write down what you expect to see, which is the commonest error in a solo count.
• Count sealed and open separately. Full boxes first, then the opened one. Guessing 'about half a bag' consistently is more useful than guessing precisely once.
• Never count what you're mid-way through receiving. Finish the delivery, book it in, then count — or count before it lands. A delivery counted on the floor before it's recorded is the classic phantom variance.
• Write zeros. A line left blank is ambiguous forever; a zero is information.

Do that consistently and the count stops being an event. It becomes a routine that one trained person can run on a Monday morning before the first customer arrives, which is the only version that survives past the first month.

Reading the variance without jumping to conclusions

A variance is the difference between counted and expected. Nearly every line will have one. The instinct is to treat any gap as loss, and that instinct causes more damage in small teams than the gap itself, because it puts the owner in an accusation before the arithmetic is finished.

Work through the boring explanations first, in order. Was the delivery booked in at the quantity that actually arrived, or at the quantity on the order? Is the recipe in the system what the barista actually makes — a 20-gram dose against an 18-gram recipe is a five-percent-plus gap on your biggest line, and it is not theft. Were staff drinks and comps recorded? Was waste written off, or just binned? Did somebody take a bag of beans to the second site and not record the transfer? Is the unit right? Most first-time variances are explained entirely by these questions, and each one you close makes the next count more meaningful.

What matters isn't the size of any single variance — it's the direction over time. One count tells you almost nothing. Four counts of the same line, all short in the same direction, is a finding. Set your own baseline from your own counts rather than comparing yourself to a number from somewhere else, then manage against movement from that baseline. The same principle applies to food cost percentage, and for the same reason: your café's normal is the only benchmark that means anything.

Where MidaOne fits

A count only works if there's a credible expected figure to count against, and that's the part MidaOne handles. Stock moves as you sell — recipes deduct their ingredients at the till, so the system's figure is built continuously rather than reconstructed at month end from a pile of invoices. Waste written off is recorded as waste rather than quietly vanishing, and the waste-by-item report will often show you where to look before the count does. Because sales, stock and accounting sit in one system, the count you take on Monday lands against the same records your cost of goods is built from, instead of in a spreadsheet that has to be reconciled with everything else later. If you're still setting the underlying stock system up, managing café inventory in the UAE is the place to start.

Frequently asked questions

The café that counts fifteen lines every Monday knows more about itself than the one that counts four hundred lines every three months — and it finds out in time to do something. Pick the day, write the sheet in shelf order, and let the first few counts be messy. The information starts the week you do it twice.