Cut your pub's GP losses in 30 days: the plan, week by week

The GP% you price for and the GP% you bank are two different numbers, and the gap between them is the most fixable money in your pub. 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. Thirty days is enough to find your gap, close the leaks you control, and prove it with a second count. Here's the plan I'd run, week by week.

First, what the gap is worth

Take a pub doing £6,600 a week ex VAT on wet sales, which is about £8,000 through the till. One GP point on that is £66 a week, call it £3,400 a year. Most pubs I've seen numbers for are running two to four points under their theoretical GP, and they've normalised it, because the deficit arrives in dribs and drabs rather than as one invoice. Three points is £10,000 a year. That's the prize for a month of paying attention.

One rule before you start: you fix nothing in week one. The first job is measurement, because a fix you can't measure is a guess with better PR.

Days 1 and 2: the baseline count

Do a full count, properly. Every line, every part bottle, every keg and cask. Weigh the open spirits rather than squinting at them, because an eyeballed tenth on a 70cl bottle is £10 of revenue either way, and the errors don't cancel out, they pile up. Part bottles are where spirit counts go wrong covers the method, and the DIY stocktake guide covers the whole job. If you're on paper, the free template has the formulas built.

Set real container yields while you're down there. An 11-gallon keg holds 88 pints on paper and a tidy pub sells 84 to 86 of them. A firkin says 72 and gives you mid-60s once sediment and the three-day window take their share. Count against brochure yields and you'll chase a phantom deficit all month. Keg and cask calibration is the five-minute fix.

Write down three numbers: stock value, the till's wet sales for the same period, and the variance between what the stock says you used and what the till says you sold. That variance is the number the next 28 days exist to shrink. What's acceptable is already written up, but the short version is that a tidy wet-led pub runs 4 to 6% on draught and much less on spirits.

Week one: make the numbers honest

Most of a scary variance isn't loss at all. It's things that happened and never got written down, and you can kill the biggest ones in a week without spending a pound.

Map the till buttons. If a guest ale is rung through the button for the old guest ale, your stock report is fiction for both lines. Walk every pump and every fridge shelf against the till layout. Till buttons are a variance machine and fixing them costs nothing.

Sort the "one in the pump" pints. Prepaid pints for later are lovely hospitality and a stocktake nightmare, because the till says sold and the stock says still here, or the other way round. Pick a system, poker chips behind the bar, a tab, a dedicated till button, anything, as long as the sale and the pour land in the same counting period.

Put a wastage log on the wall. Line cleans, fobbing, a blown keg, breakages, out-of-dates. Logged, they're housekeeping. Unlogged, they look like theft, and you'll waste week four suspecting people who did nothing wrong. What belongs in the log is its own guide, including which losses get you supplier credit back.

Check deliveries before you sign. A short delivery you signed for is stock loss you agreed to in writing. Count cases and casks at the door, every time.

Week two: the cellar week

The cellar is where good stock quietly becomes waste. Clean every line on a seven-day cycle and log the beer that goes with it, roughly 1.25 pints sitting in every 10 metres of 3/8 inch python, plus the two or three pints you pull through per line to clear the cleaner. Vianet's data says 27% of beer is served through lines outside the seven-day window, and tired lines don't show up in the variance, they show up in regulars drinking somewhere else. The full arithmetic is in the line cleaning guide.

Run the cask timetable off the cellar card: tap within two days of venting, empty within three days on sale, and order pins for any line that can't drink a firkin in three days. Check the gas connections with soapy water. Set the cooler to hold 11 to 13 degrees and check it with a thermometer, not the dial. The cellar routine, with prices on the shortcuts, is the checklist. None of this is new advice. The point is that this week it gets done on schedule and everything it costs in beer gets logged, so the next count can tell routine from problem.

Week three: the pricing week

Now the other side of GP, what you charge against what you pay. Pull your last month of delivery notes and compare line prices against the oldest invoice you can find. Suppliers move prices mid-year and most of us find out at the year end. Heineken's February 2026 rise averaged 2.7% before duty on draught, and duty itself went up 3.66% in the same week. The invoice-checking habit takes fifteen minutes a week.

Then do the sum that catches nearly everyone: to hold your GP after a cost rise, the shelf price has to rise by the cost rise divided by (1 minus your GP). At 65% GP, a 5p cost rise on a pint needs about 14p ex VAT, which is 17p on the shelf. Put 5p on because the cost went up 5p and your margin just quietly shrank. Reprice the laggards this week, and while you're at it check what basis your GP report is even using, because three calculators will give you three answers for the same pint.

Week four: count again, and read the answer

Days 24 to 30, run the second full count, same method, same yields, same person if you can. Now the variance means something. The recording fixes from week one will have taken their bite, the cellar log explains its share, and what's left is real loss: over-pouring, mismeasures, or someone helping themselves.

Don't jump to the last one. Pouring, waste or measuring error shows how to tell the fingerprints apart, and free-poured spirits alone can eat a full GP point without a single dishonest person in the building. Fix pour discipline with measures and training before you fix anything with an accusation.

Then keep the rhythm. A monthly count found this money; a weekly one keeps it, because a leak caught at day 7 costs a quarter of a leak caught at day 30.

The 30 days on one page

The plan, with the number each stage should produce. Figures sourced in the text.

Days	The job	The number you write down
1–2	Full baseline count, real yields, weighed spirits	Stock value, wet sales, variance %
3–9	Till buttons, prepaid pints, wastage log, delivery checks	Waste logged this week, in pints and pounds
10–16	Line cleaning on schedule, cask timetable, gas, temperature	Cleaning waste per week, cellar temp
17–23	Invoice check, hold-your-GP repricing, GP report basis	Lines repriced, new theoretical GP
24–30	Second count, compare, diagnose what's left	New variance %, explained vs not

What 30 days won't fix

Be straight with yourself about the limits. This plan does nothing about rent, energy, duty or the card machine's cut, and if the pub's problem is footfall, stock discipline won't invent customers. What it fixes is the gap between the margin you planned and the margin you got, and in most pubs that gap is worth more than any supplier switch you could make this month. Halving your unexplained variance in 30 days is a realistic result. Zero isn't, and anyone selling you zero is selling.

The other honest limit: a monthly paper count on 200 lines is three or four hours, twice, plus the discipline in between. Software shortens the counts and does the yield and GP arithmetic for you, but the plan works on paper too. The plan is the point. The tool just makes it stick.