How to calculate GP on drinks, without flattering yourself

The formula is short: GP% = (net selling price minus cost price) ÷ net selling price, × 100. The word doing all the work is net. Take the VAT off the till price first, by dividing it by 1.2, because 20% of every pint you sell was never your money. Skip that step and a pint that's really making 56% looks like it's making 62%, and you'll walk past a problem for months feeling quietly smug about it. That's the whole trick. The rest of this guide is worked examples for keg, cask, spirits and bottles, the difference between the GP your price list promises and the GP your stocktake proves, and how to price backwards from a target.

I run a wet-led pub in Washington, Tyne and Wear. I've had five independent liquor audits in eight months, and the education in those wasn't the counting, it was the maths that came back stapled to it. GP is the one number on that report your whole pricing lives or dies by, and I kept meeting licensees, some of them years in, who were working it out on the till price. So here it is, done properly, with sums you can re-run on your own bar tonight.

The formula, one drink at a time

Three steps. Say your standard lager is £4.89 a pint, which is the UK average this year per the Morning Advertiser's price survey.

1. Net the price. £4.89 ÷ 1.2 = £4.07. That's yours. The other 82p is VAT, and it's HMRC's on the way through.
2. Find the true cost of the serve. Not the keg price divided by 88. The keg price divided by the pints that actually reach a customer. More on yield in a minute; call it £1.79 here.
3. Do the sum. (£4.07 − £1.79) ÷ £4.07 = 0.56. That pint runs at 56% GP.

Now the wrong way, which I still see written on cellar whiteboards: (£4.89 − £1.79) ÷ £4.89 = 63%. Same pint, same cost, and the answer is seven points kinder because the VAT never left. Seven points is not rounding. Seven points is the gap between a pint that's paying its way and one that isn't, and if your "62% GP" bar has always felt skinter than the number says it should, this is usually why. I've written up the same trap on the valuation side in ex-VAT vs inc-VAT stock values, because stocktake reports mix the two bases as well.

One more habit worth building: your cost prices are already ex-VAT on the invoice, so never "add the VAT back on" to compare them with till prices. Move the selling price down to net. Everything in stock maths happens in the ex-VAT world.

Worked examples: keg, cask, spirits, bottles

Example figures so you can check the method, not benchmarks. The till prices are the Morning Advertiser's 2026 survey averages where one exists; the buy prices are round numbers picked for the arithmetic. Put your own invoice prices in.

Drink	Buying	Real yield	Cost per serve	Till price	Net price	GP%
Keg lager	£150 / 11g keg (88 pints)	~84 pints after line cleans and foam	£1.79	£4.89	£4.07	56%
Cask ale	£95 / firkin (72 pints)	~66 pints after sediment and line	£1.44	£4.91	£4.09	65%
House spirit, 25ml	£15.50 / 70cl (28 serves)	28 serves, barely any loss	£0.55	£3.50	£2.92	81%
Bottled beer	£1.10 / bottle	What you buy is what you sell	£1.10	£4.00	£3.33	67%

Two things jump out of that table, and they're both structural rather than anyone's fault.

First, yield is a real number, not a moan. An 11-gallon keg holds 88 pints, but a weekly line clean pulls product through to waste, foam gets binned, and a cask drops its last few pints as sediment before you ever get the chance to sell them. Dividing by 88 when you actually serve 84 understates your cost per pint and overstates your GP, and the error compounds quietly across every keg all year. If your cellar's calibration is off as well, the yield number is fiction before you start; that's its own guide, keg and cask calibration.

Second, this is why spirits carry pubs. Same bar, same staff, and the gin runs 25 points ahead of the lager because nothing about a sealed 70cl bottle goes down a drain. Draught sitting at 56 to 65% next to spirits at 80%+ isn't a performance gap, it's physics, and it's also the honest argument for caring about your wet mix. On the keg example above, being free-of-tie at, say, £125 a keg instead of £150 moves that same pint from 56% to 63% GP, about the size of the tied-versus-free-of-tie gap the last published survey found. What the blended number should look like for your sort of pub, tied or free, is covered with the actual survey data in what's a good GP% for a UK pub in 2026.

Target GP and actual GP are two different numbers

Everything above is target GP, sometimes called theoretical GP: what your price list promises if every millilitre you bought ended up in a paying glass. It never does. The number that tells the truth is actual GP, and it only comes from a stocktake.

Actual GP for a period works like this:

1. Opening stock at cost, plus purchases at cost, minus closing stock at cost = what you actually used. Say £6,200 + £8,400 − £6,000 = £8,600 consumed.
2. Net your wet sales for the same period. Till says £28,800; ÷ 1.2 = £24,000 net.
3. GP = (£24,000 − £8,600) ÷ £24,000 = 64.2%.

Now put that against your target. If the price list says this bar should run 67% and the stocktake says 64.2%, the missing 2.8 points is your variance: line cleaning, over-pours, comps nobody logged, breakages, and anything walking out the back door, all bundled into one figure. A gap of a couple of points on a draught-heavy bar is normal life. A gap that was two points in spring and five points now is a leak with a start date. What counts as normal is real-audit territory and I've put my own five sets of numbers on it in acceptable stocktake variance, and on why the headline net figure can hide bigger movements underneath, see net vs gross stocktake results.

The mistake to avoid is comparing one period's actual GP against a target you calculated years ago. Costs move under you. Draught duty alone is £19.45 per litre of pure alcohol since February this year, which is about 50p of the cost of every 4.5% pint before the brewery has paid for anything else, and this February's rise repriced every keg in the country. If your supplier passes through a rise and your board price doesn't move, your target GP just fell and nobody told you. Recalculate targets every time an invoice price changes, not annually.

Pricing backwards from a target GP

The same formula flipped is how you set a price instead of discovering one. Net selling price = cost per serve ÷ (1 − target GP). Then multiply by 1.2 to get back to a till price.

Take the keg lager example: £1.79 a pint cost, and say you want 62%. £1.79 ÷ 0.38 = £4.71 net. × 1.2 = £5.65 on the till, so £5.70 on the board. If £5.70 is unsayable in your room, and in plenty of rooms it is, then you're choosing a lower GP on that line with your eyes open, and making it up on spirits and softs. That's fine. That's running a pub. The point of the formula isn't that maths sets your prices, it's that you know the cost of the kindness.

Three habits that keep the number honest:

• Recalculate cost per serve from the invoice every delivery, using real yield, not container size.
• Re-run target GP on any line whose invoice price moved, the week it moves.
• Compare target to actual monthly via a stocktake, because target GP with no stocktake behind it is a hope, not a margin. If you're doing that count yourself, the method is in the DIY pub stocktake guide.

If you'd rather not do any of this arithmetic by hand, I built a free GP and profit leak calculator that runs these sums, and stocktaking software keeps cost prices against every line so target and actual GP fall out of the count on their own. StockTap stores selling prices ex-VAT for exactly the reason this guide exists: it's the only basis the maths is true on.

Common questions

Is 70% GP good on drinks?

On spirits it's ordinary; on draught it's exceptional and worth double-checking your yield assumptions before you celebrate. Blended across a wet-led bar, 70% net would put you well ahead of the last published segment averages. Just be sure it's a net figure. A "70%" calculated on till prices is really about 64%.

Does GP include wages, rent or duty?

Wages and rent, no. GP is sales minus cost of goods only; everything else comes out of it afterwards, which is why a healthy GP% can still coexist with a skint bank account. Duty, yes, in effect: it arrives inside your invoice price from the brewery, so it's already in your cost per serve whether you think about it or not.

Should I use 25ml or 35ml maths for spirits?

Whichever you actually pour, and check the optic or jigger matches the till button. A bar pricing on 25ml serves while pouring 35ml is handing out 40% more product per sale, and it shows up as a spirits GP that refuses to hit target. Free pours are the same problem without the honesty; I've costed that in what bad pours actually cost.