Line checks: the five-minute count that settles stock arguments

A line check is a five-minute count of a handful of products, checked against what the till says should be there. It's the fastest early-warning system a pub has, it's the evidence that settles arguments when an audit goes against you, and most licensees who are asked to do them quietly don't. My own agreement asks for fifty a week. Here's what they're actually for, how to do one that means something, and why fifty is less work than it sounds.

First, which "line check" we're talking about

The trade uses the phrase for two different jobs. Cleaning the beer lines is one, and it has its own guide and its own arithmetic. This guide is about the other one: a line check as in a line on the stock sheet. You count a few products, the system compares your count with what's been sold and delivered since the last count, and you get a variance per product, today, not at the end of the month.

That definition isn't mine. Memsec, an EPOS supplier, describes it the same way in their stocktaking manual: enter a count for selected lines, and the system instantly shows sales, deliveries and other movements since the last count, plus the variance. The point they make, and it's the right one, is that line checks are for watching specific products between full counts, often daily.

Why bother, when you already do a stocktake

A monthly stocktake tells you what you lost. A line check tells you when. If your best lager goes four pints light, the monthly count hands you a number and thirty days of possible explanations. A daily or twice-weekly line check on that same product narrows it to a day or two, and suddenly you can line the variance up against what actually happened: a delivery, a busy Saturday, a line clean, a function with pints left in the pump.

The scale of what goes missing between counts is not small. Vianet's dispense data from 11,000 monitored sites says 61% of pubs miss their target pouring yield, and 2% of drinks poured never reach the till at all. Four weeks is a long time to not know which of those is happening to you.

The fifty-a-week problem

Some tied agreements ask for a set number of line checks a week. Mine asks for fifty. And I'll be honest about what most licensees do with that: nothing. Not because they're lazy, but because the pubco's own system produces numbers that rarely match what's on the shelf, so the whole exercise feels like homework that marks itself wrong. You count 43 pints, the screen insists on 51, and after three weeks of that you stop counting.

Here's the reframe that made me start again: do the checks on your own numbers, for your own benefit, and treat the agreement's requirement as a happy side effect. A line check log you built yourself, from your own counts against your own till, is worth more to you than any figure the pubco's system generates. It's your record. Nobody can quietly recalculate it.

How to do a line check that means something

Pick the lines that move. Your top five draught products are most of your wet turnover, so start there, and add two or three spirits that walk, usually the house vodka and whatever the current shot of the month is. Ten lines is plenty for one day.

Then, same time every day, count them. Draught needs a keg-weight or dip reading rather than a guess, which is why calibration matters; spirits are quickest weighed as part bottles. Write down the count, the till sales for that product since the last check, any delivery, and the difference. Date it, initial it. Two minutes a line once you've done it a few times.

What you're looking for is not pennies. A pint out on a busy lager is measurement noise. What you're looking for is pattern: the same product light on the same shift pattern, a variance that appears the day after every line clean because the flushed beer never got logged, a Sunday shortfall that turns out to be a till button ringing the wrong product. Run line checks for a month and they mostly convict the paperwork, not people: unlogged waste, wrong buttons, prepaid pints, deliveries signed short. That's good news, because paperwork is cheap to fix.

The arithmetic on one line

Worked example, illustrative numbers. Monday's check on your best lager: 45 pints in the cellar. During the week a keg goes on (another 88 pints on paper, call it 84 sellable once normal waste is out). The till says 92 pints sold. Friday's check finds 33. So 45 plus 84, minus 92, says the shelf should hold 37, and it holds 33. Four pints adrift in five days, on one product. At £5.34 a pint that's £21 a week, £1,100 a year, from a single tap, and without the two checks you'd have met it a month later as one anonymous lump in the stocktake variance.

Line checks as evidence

Now the part that made me take them seriously. When a stock audit goes against you and a chargeback lands, the difference between grumbling and disputing is documentation. An auditor's report is a set of dated, product-level figures. "My cellar was fine" is an opinion. A line check log is the same class of thing as the audit: dated counts, product by product, written down at the time. If their count says you lost 200 pints over a period, and your log shows twice-weekly counts across that period with variances a tenth of that size, you're no longer arguing feelings against figures. You're arguing figures against figures, alongside your wastage log and delivery notes. The audit guide covers the whole file to build; the line check log is its spine.

I won't promise a log reverses a chargeback, because that depends on what the numbers show and who's reading them. I will say that disputing one without records is close to hopeless, and every route through a dispute starts with "show us your documentation".

Making fifty a week not a nightmare

Fifty sounds like a shift's work. It's ten lines a day, five days a week, and with a routine it's five to ten minutes a day: weigh the open spirits while the coffee machine warms up, dip or weigh the busy kegs before doors. The trick is that the counting has to be faster than the arguing it replaces.

This is also, in fairness, exactly the job StockTap's Spot Check was built for: count a few lines, it pulls the till figures for the same window, shows the variance per product and keeps the dated log. That's the product plug, and the honest version of it is that a ruled notebook does the same job if you'll keep it up. The tool just removes the part where you fetch till reports and do the subtraction, which in practice is the part that kills the habit. Where line checks fit alongside full counts is its own question, but the short version is: they don't replace the stocktake, they make it boring, and boring is what you want a stocktake to be.