Most café owners stocktake the same way they did when they opened: a clipboard, a Tuesday morning, and a sinking feeling about waste.

There's a faster way. And it catches more.

Why most café stocktakes are a waste of time

A stocktake on its own tells you nothing. You wrote down that you have 14kg of coffee beans. Lovely. Now what?

A stocktake is only useful when you have a number to compare it to. Without that, you've just done two hours of admin and learned that yes, there is coffee in the cupboard.

The number you're comparing to is your theoretical COGS - what your POS sales say you should have used. The gap between theoretical and actual is the whole point of the exercise. That gap is where waste, theft, over-portioning, and miscoded sales live.

If your stocktake doesn't end with a comparison, you're not stocktaking. You're counting jars.

What to actually count

Here's where most people go wrong. They try to count everything. Every teabag, every sachet of sugar, every napkin.

Don't.

The top 20 ingredients in a typical café drive around 80% of your COGS. Count those properly. Eyeball the rest.

For most UK cafés, the high-value list looks something like this:

Coffee beans (almost always your single biggest line)
Milk - whole, semi-skimmed, oat, almond
Alcohol if you sell it
Meat and fish for hot food
Cheese
Bread and bakery items you buy in
Butter
Eggs
Premium pastries and cakes
Speciality syrups
Anything you've had a recent price hike on

If you've got CostingBrik set up, your ingredient list is already ranked by total cost - start with the top 20 by spend and ignore the long tail. The teabags will not make or break you. Cheese will.

The walkthrough

Here's the stocktake itself. Six steps. About 30 minutes if you've prepped properly.

1. Prep your count sheet

Print a sheet with your top ingredients listed, grouped by storage location. Walk-in fridge in one block, dry store in another, bar fridge, cake display.

Counting in location order means you don't loop back. You walk the room once.

Leave a column for quantity and a column for unit (kg, litres, units). Don't try to convert in your head while counting - you'll get it wrong.

2. Count at the same time each week

Same day, same time, before deliveries arrive. The whole point is to capture a clean closing snapshot. If a milk delivery rolls in mid-count, your numbers are dead.

Most operators do it first thing on a quiet morning. Monday before open is common. Sunday night also works if you don't trade Mondays.

3. Use two people if you can

One counts, one writes. It's twice as fast, and the second pair of eyes catches the obvious mistakes ("you've written 14, that's clearly 1.4kg").

If you're solo, that's fine - just budget a bit more time and double-check the high-value lines.

4. Value at the latest invoice price

This is the bit people get wrong.

Multiply each quantity by your current per-unit cost - the price on your most recent invoice, not the price baked into your recipe cost.

Why does it matter? Because suppliers move prices weekly. If your recipe was built when butter was £4.20/kg and it's now £6.10/kg, valuing your stock at the recipe price understates your closing stock by about 45%. Your COGS calculation will be nonsense.

5. Calculate actual COGS

The formula every operator should know by heart:

Opening stock + Purchases - Closing stock = COGS

So if you opened the week with £2,400 of stock, bought £3,800 of stock during the week, and closed with £2,100, your actual COGS for the week is £4,100.

That's what you actually used. In pounds.

6. Compare to theoretical COGS

Now the bit that matters. Pull your sales for the same period and the recipe cost of what you sold. That's your theoretical COGS - what you should have used if every recipe was portioned correctly and nothing was wasted.

If theoretical says £3,950 and actual says £4,100, you're £150 over. That's a 3.8% variance.

Now you have something to investigate.

Reading the variance

Variance is where the gold is. Some variance is normal. A lot of variance is a problem.

When you're in the 3%+ zone, the usual suspects are:

Waste. Burnt shots, dropped pastries, returned drinks, end-of-day milk dumps. Get staff logging it for a week and you'll find most of it.

Portioning. The classic. A 7g shot has become a 9g shot because the grinder drifted and no one re-dosed it. Across 200 coffees a day, that's 400g of beans gone walkabout. Multiply by a week and it's noticeable.

Theft. Less common than people think, but it happens. Look at staff coffees, friend-of-staff discounts, voided sales, comped items.

Miscoding on the POS. Someone's been ringing up flat whites as Americanos because the button is closer. Your theoretical COGS for flat whites is too low, your Americano COGS is too high, and the variance hides between them. This is where POS data and recipe costing have to agree - if they don't, the variance lies.

The trick is to investigate one category at a time. Don't try to solve the whole variance. Pick milk this week, coffee next week. Narrow it down. There's a fuller guide to investigating variance here if you want the deep dive.

From monthly chore to weekly habit

Here's the case I want to make: stop doing big quarterly stocktakes.

A quarterly stocktake takes four hours, finds variance that's three months old, and gives you no chance of fixing the root cause. By the time you spot a problem, the staff member who did it has left, the supplier price has changed, and the recipe has been tweaked twice.

A weekly stocktake takes 30 minutes. The variance is fresh. You can walk out onto the floor and watch the next shift to see if the portioning has drifted.

The maths:

Quarterly: 4 hours × 4 = 16 hours a year, with 3-month-old data
Weekly: 30 minutes × 52 = 26 hours a year, with 7-day-old data

Yes, weekly is more hours overall. But the data is actionable. Quarterly data isn't. You're paying for admin you can't use.

Most operators who switch to weekly find their variance drops within two months, simply because the feedback loop is short enough to catch problems while they're still happening. There's a simple weekly stocktake system you can run yourself - same principles as above, just tightened up for speed.

The takeaway

A stocktake without a comparison number is just admin.

Tie the count to your costing data and the same 30 minutes becomes a diagnostic. You'll know whether your coffee margin is what you think it is. You'll know whether the milk waste is normal or a problem. You'll know whether the new pastry supplier is actually cheaper once you account for shrinkage.

That's the whole game. Count the right things, value them at current prices, compare to what your POS says you sold, and chase the gap.

CostingBrik does the maths for you - it holds your latest invoice prices, links your recipes to your POS sales, and tells you what your theoretical COGS should be for any period. The counting is still on you. The thinking is, mostly, not.

Ed O'Brien has run Hunters Cake Company for 17 years across cafés in Witney, Burford, and a bakery in Carterton, Oxfordshire. He's building Brikly - modular tools that give independent café owners the same data the big chains have, without the big chain price tag.