Why your till and your stocktake never agree: the eight causes, in the order to check them

The till says you sold 41 pints of Madri. The stocktake says 46 left the keg. Five pints of gap, every single week, and nobody can tell you where it went. Before anyone mutters about staff, here's the thing: the till counts money and the stocktake counts liquid, and there are at least eight ordinary, boring reasons the two disagree. I've put them in the order I'd actually check, starting with the one that catches out more pubs than anything else and never gets written down anywhere: the pint that's already been paid for.

I'm a licensee in Washington, Tyne and Wear, and I've sat on the wrong side of enough audit results to have worked through this list for real. Every cause below leaves a slightly different shape in the numbers, so you can find yours instead of shrugging at the total.

What you're actually comparing

A till-versus-stock check is two estimates shaking hands. The till gives you expected usage: everything rung through, converted into pints and measures. The stocktake gives you actual usage: opening count, plus deliveries, minus closing count. When people say "the till doesn't match the stocktake", they mean expected and actual usage disagree, and the difference, the variance, gets valued in pounds at selling price.

Both sides carry error. The till only knows what was pressed, not what was poured. The count is only as good as the counting. So the gap is never zero, and chasing zero will drive you daft. What you're looking for is a gap that's small, steady and explainable. My acceptable variance guide covers how big is normal; this one is about where the gap comes from.

1. Pints already paid for, the "one in the pump"

Someone buys their mate a pint to collect later. Someone leaves at half time and pays for one for after the match. The charity bloke buys a round for the darts team who haven't arrived. Every one of those is money through the till tonight and beer through the line next Tuesday.

Look what that does to the reconciliation. This week the till records sales for beer that never left the keg, so the till side reads high and the count side hasn't moved: a positive gap, you look weirdly good. Next week the beer pours with nothing rung in, and the same pints show up as a loss. One in the pump doesn't lose you a drop, it smears sales and stock across different weeks so both look wrong, and in a proper local at Christmas there can be twenty or thirty pints floating like this at once. Thirty pints is about £147 of retail sitting in the gap at 2026 prices, in one direction one week and the other direction the next.

The fix is dead simple and older than the till: make the promise physical and countable. A jar of poker chips behind the bar, a tally on a whiteboard, a page in the diary, whatever survives a busy Friday. When the pint is poured, it gets rung through a zero-priced "prepaid" button so the till knows the liquid moved. And when you count, the jar gets counted too, because those chips are pints the till has sold that your cellar still owes. Do that and this whole cause disappears from your variance overnight.

2. The till buttons themselves

Doubles rung as singles because the double button is buried two menus deep. The 2-4-1 cocktail rung as one. The "house spirit" button that assumes 25ml when the optic pours 35ml. The guest ale sold through the button of the ale it replaced, so one keg looks like a hosepipe and the other looks immortal. The till can only convert button presses into expected pints, and if the mapping is wrong the reconciliation is wrong before anyone pours anything. I wrote a whole guide on till button variance because it's probably the most common cause after prepaid pints, and the cheapest to fix.

3. Line cleaning nobody logged

Every line clean pulls the beer standing in the python through to the drain. About 1.25 pints per 10 metres of 3/8 inch line, per tap, plus the pints you pull through to clear the water and get the beer back on. A tidy cellar routine on a 6-tap bar bins somewhere around 20 pints a week, which is beer the till will never see. If it isn't logged as waste, it lands in your variance looking like loss. It isn't loss, it's hygiene. The line cleaning guide has the full arithmetic per tap and metre, and the logging habit that keeps it out of your gap.

4. Waste and comps that never get written down

The drip trays. The first cloudy pint of the session. The lemonade the kid knocked over that you replaced with a smile. The birthday shot you put on the house. The bottle that came up smelling of wet cardboard. None of it is sinister and all of it is stock leaving without a sale, which is exactly what a variance is. A waste log only works if it's effortless, a pad by the taps or a button in an app, filled in at the moment, not reconstructed on Sunday. And spoilt kegs are their own subject: logged and reported properly some of that money comes back as supplier credit, which is the ullage guide's job to explain.

5. The pour itself

A pint with a proper head should still be at least 95% liquid, that's the industry standard, and a bar that habitually fills to the glass line is giving away the difference on every pull. Free-poured spirits drift the same way, a generous thumb turns 25ml into 32ml and nobody notices except the stocktake. This is a training and glassware subject, not a discipline one, lined glasses and thimble measures fix more than any team talk. The numbers are in the pour cost guide.

6. The delivery you signed for but didn't get

Eleven kegs on the note, ten on the dray. A split case signed as full. The right invoice against the wrong gyle. If a delivery goes into your stock system at more than actually arrived, your book stock starts high and the count will find a "loss" that was never in the building. Count the cellar drop before you sign, every time, and check quantities against the note line by line. Boring, two minutes, and it removes a whole cause. Deliveries also interact with timing, which is next.

7. Timing: when the count happens versus when the till cuts off

Count on Monday morning against a till report that runs to Sunday close and you're fine. Count Monday lunchtime while the bar's open and every pint sold during the count sits on one side of the line but not the other. Same if a delivery arrives between the till cutoff and the count, or if your card machine batches sales into the next day. The gap this creates is pure paperwork, and it reverses next period, which is the tell. Fixed cutoffs, counted at the same point in the week, before opening. Make it a rule and this one vanishes.

8. The count itself

Last because you should rule the others out first, but not least. Eyeballed tenths on open spirits can be 100ml out per bottle without anyone squinting unusually. A keg judged by rocking it is a guess with a confident face on. Two different people counting two months running will disagree with each other's eyes even when nothing was poured. If your variance flip-flops, loss one month, magic surplus the next, suspect the measuring before anything else: it's the classic fingerprint, and the stock loss diagnosis guide shows how to read it. Weighing open bottles and calibrating your kegs and casks takes the eyes out of it.

Work the list, then look at the shape

Run the eight in order and most till-versus-stock gaps die at one, two or three. What's left, if anything, is your genuine loss to manage, and the shape of it across two or three counts tells you whether it's pouring, unlogged waste or measuring. Only after all of that, with numbers in hand and every ordinary cause excluded, does anything harder deserve a thought, and in my experience it almost never gets that far. Chasing your own team on a gap that turns out to be the till buttons is how you lose good staff over a mapping error.

The practical cadence: a monthly full count, and a two-minute spot check on your five biggest lines each week so a gap gets caught while it's small and the cause is yesterday, not "sometime since the 4th". That's the counting rhythm guide's subject, and it's exactly what I built StockTap's Spot Check for: pick a handful of products, count them, and it shows the variance against the till there and then. Fourteen days free, no card, and the first thing most people find is a till button.

Common questions

Why doesn't my till match my stocktake?

Because they measure different things. The till records what was rung in, the stocktake measures what physically left. Prepaid pints, wrong till buttons, unlogged line cleaning and waste, over-pouring, short deliveries, timing cutoffs and counting error all open a gap between the two before anything sinister does.

What is a pint 'in the pump' and why does it affect the stocktake?

It's a pint paid for now and poured later. The till records the sale this week while the beer leaves the keg next week, so one period looks artificially good and the next shows a loss. Track them physically, chips in a jar or a tally, and ring the pour through a zero-priced prepaid button so the till knows when the liquid moved.

How close should the till and the stocktake be?

Never exactly level, both sides carry error. A well-run wet-led bar keeps the gap within about 1% of wet sales. Steady and small is fine; growing, or flip-flopping between loss and surplus, is a signal worth chasing.

Does a gap between till and stock mean staff are stealing?

Almost never as the first explanation. The ordinary causes, buttons, prepaid pints, unlogged waste and measuring error, account for the overwhelming share of variance. Work through those with numbers before any conversation about people, because accusing a team over a till-mapping error does damage you can't undo.

How do I stop the gap coming back?

Fixed count cutoffs, a waste log that's effortless enough to be used, a physical tally for prepaid pints, and a weekly spot check on your biggest five lines so a new gap gets caught inside a week while the cause is still findable.