What is an acceptable stocktake variance in a UK pub? The working ranges, what 1% costs you a week, and why direction beats size

There's no published UK standard for an acceptable stocktake variance, and anyone quoting one as fact has borrowed it or made it up. In practice most pub companies treat a liquor result inside plus or minus 0.5% of sales as normal trading and start asking questions around 1%. On a £5,000 week, 1% is £50 a week, £2,600 a year, about a third of what a typical pub keeps. But the number that matters more than any single result is the direction over several stocktakes: three small deficits in a row is a leak, one big one is often a mistake.

I've had five independent liquor audits, and one of them got charged back to me, so I've spent more time staring at this number than I'd like. Here's what's real, what isn't, and how to read your own.

How the variance is worked out

A UK pub stocktake usually works at selling prices. Opening stock, plus deliveries, minus closing stock, gives you what went out of the building, valued at what it should have taken. Compare that with what the till actually took.

A worked stocktake period at retail value, illustrative figures, my arithmetic

Line	At selling price
Opening stock	£8,000
Plus deliveries	£10,000
Minus closing stock	£7,950
Should have taken	£10,050
Actual takings for the period	£10,000
Variance	−£50, or −0.5% of takings

Two things to check on your own report before you read anything into the percentage. First, what it's a percentage of. Most UK reports use takings or retail sales, but some software works against cost of goods, opening stock or purchases, and the same trading gives wildly different percentages. Second, whether the takings figure is clean. Food, cashback or room money rung on wet buttons will make the drinks look better than they are.

The number everyone quotes, and where it comes from

Search this and you'll get 1 to 2% of sales on site after site. The clearest version is on Backbar's inventory guide, which says an acceptable liquor variance should fall within 1 to 2% of sales, with no source given. Backbar is American, writing for spirit-led American bars, and their variance formula works off cost of goods.

Carry that into a wet-led British local and it falls apart. It's measured against a different base, and draught behaves completely differently from a speed rail of spirits. A 2% allowance across the whole bar would hide a real spirits problem behind normal draught movement.

Working ranges, for what they're worth

A working reading of liquor variance against takings. A convention, not a standard.

Result	On a £5,000 week	What it usually means	What I'd do
Inside ±0.5%	up to £25 a week	Normal trading noise	Log it, move on, don't go looking for a culprit
±0.5% to ±1%	£25 to £50	Worth a look if the direction repeats	Split it by category, compare with the last two periods
Beyond ±1%	over £50	Where an operator usually picks up the phone	Work the list below before you accept it
Same direction three periods running	any size	A systematic error, not noise	Find it. This is the one that costs you over a year

Why £50 a week matters: on a £5,000 week the average pub keeps about £150, which I worked through in how much profit a pub makes. A steady 1% deficit takes a third of it.

And a surplus isn't automatically good news. A steady small surplus on draught is normal. A big one usually means a missing delivery, a closing count that's too high or the wrong prices, and it often comes back as next period's deficit. More in what a stocktake surplus means.

My own five audits

Five consecutive independent liquor audits, one site, December 2025 to July 2026

Audit	Result	Direction
1	+£193	Surplus
2	+£65	Surplus
3	+£51	Surplus
4	−£756	Deficit, −1.08%
5	+£13	Surplus

Four of those landed inside three tenths of one per cent. The fourth was −£755.71, which the report put at −1.08%, and that's the one that got charged back. It's one site, one stocktaker, seven months, so it's not a benchmark. What it shows is the shape: four results tight around zero and one four times further out than the rest. That shape is the signal. It's also the only result anyone ever asked me about, which tells you roughly where the line sits in practice. Not a published threshold. Just where somebody picked up the phone.

Draught: what's real loss, and what the head does

Draught is where most licensees panic about a number that's fine, and where the old version of this guide got one thing wrong, so let me be precise.

Real draught loss is physical: the beer that goes down the drain on line-cleaning day, fob at keg changeover, the bit left in the keg, the drip tray. Built up from those, a well-run pub loses about 4 to 6% of its draught volume before anything has gone wrong, which I worked through in what draught wastage is normal. Line cleaning alone is a few pints a line every clean, costed in line cleaning and your variance.

The head works the other way. BBPA guidance says a pint served with a head must be at least 95% liquid, and a top-up should never be refused. A pint with a decent head uses a little less beer than a full liquid pint, so generous heads push draught towards a surplus, not a loss. Refused top-ups push it further. If your draught is running well over, that's worth knowing before you congratulate anyone.

Then there's the allowance. Most stocktakers apply a standing draught allowance before they call a result. Ask what yours uses, line by line. A blanket allowance can excuse an honest pub or hide a real problem, and you can't argue with a number when you don't know what was taken off it.

Net and gross are not the same number

On one of my reports the bottom line came back £12.84 up. I barely read the next page. Underneath it, £1,165 of stock had moved, categories up and down, cancelling out. On another period, £11,598 of gross category movement netted down to minus £434. A bottom line near zero can sit on top of two real problems cancelling each other out. Read the categories, not the total. I wrote that up properly in net versus gross variance.

Before you accept a bad number, work this list

1. Till buttons. A button pointing at the wrong product, or the right product at an old price, makes a steady, fictional variance every week until someone spots it. The most common cause and the easiest to prove. See the till button problem.
2. Delivery timing. A delivery signed for one side of the count date and counted the other throws a period and corrects the next. A bad result followed by an equally good one is almost always this.
3. Price changes mid-period. If sales are valued at one price and you traded part of the period at another, the gap is arithmetic, not stock.
4. Line cleaning and wastage not recorded. If it isn't written down, it's counted as loss. A wastage log fixes that.
5. Cellar temperature. Warm cellar, fobbing, beer down the drain. Check the temperature log before anything else.
6. Spirits measures. A small, steady over-pour compounds fastest on spirits because the unit is small and the price is high.
7. The count itself. A part bottle guessed by eye isn't a count. Weigh it. How, in counting part bottles.

Most of the time it's on that list. Weekly line checks catch it between stocktakes, which is the whole point of doing line checks properly.

What gets a chargeback reversed

If the number sticks and comes back to you as money, what matters is what you wrote down on the day, not how convincing you are three weeks later. I've had two reversed, and both worked because there was a second piece of evidence in the room. Without one, the stocktaker has a spreadsheet and you have a story. The full write-up is in how to dispute a stocktake chargeback, and if you're tied, what to expect on the day is in the pubco stock audit.

What I still don't know

Nobody publishes UK pub variance benchmarks. Not the BBPA, not the stocktaking firms, not the pub companies. The people with the biggest datasets sell stocktakes. So the ranges above are a working convention from one site's results and from where an operator actually rang, not research. If your operator works to a different threshold, theirs is the one that costs you money, so ask for it in writing.

What I'm sure of is the method. Split draught from packaged from spirits. Read the movement, not the bottom line. Watch the direction over several periods rather than the size of one. Write things down on the day.

Common questions

What is an acceptable stocktake variance percentage for a pub?

There's no published UK standard. In practice most pub companies treat a liquor variance inside plus or minus 0.5% of sales as normal and start asking questions around 1%. That's a working convention, and it varies by operator, by site and by how the last few periods have looked.

How do you calculate stock variance in a pub?

Opening stock plus deliveries minus closing stock, valued at selling prices, gives what should have been taken. Compare that with the actual takings for the period. The difference is the variance, usually shown as a percentage of takings.

Is the 1-2% figure I keep reading online correct?

It comes from American bar inventory guidance with no source given, and is measured against a different base. UK pub stocktake reports are usually expressed against takings, and UK draught behaves differently from spirit-led American bars, so applying 1 to 2% to a wet-led UK pub can hide a real problem.

How much does a 1% stock variance cost a pub?

On a £5,000 week, 1% is £50 a week or about £2,600 a year. With a typical pub keeping about £150 a week from £5,000 of sales, a steady 1% deficit takes about a third of the profit.

Does a bad variance mean someone is stealing?

Usually not. The common causes are till buttons pointing at the wrong product or price, deliveries counted in the wrong period, price changes mid-period, wastage and line cleaning not recorded, cellar temperature, spirits over-pouring and part bottles guessed rather than weighed.