A full count every month, a 15-minute spot check on your top lines every week, and a full count every fortnight once wet takings pass about £10,000 a week. Quarterly is a number for your accountant. It is not stock control, because whatever went wrong went wrong twelve weeks ago and nobody remembers. That's the answer. The rest of this is why, what it costs to wait, and the handful of days when you count whether it's due or not.
I run a pub in Washington, Tyne and Wear, and I've been audited five times in eight months by an external stocktaker. The results swung from plus £193 to minus £755 over that stretch. The minus £755 one was minus 1.08% of sales, and it turned into a chargeback dispute. What I learned from it wasn't about the count. It was about the gap between counts, and how much you can lose inside a gap you're not looking at.
What the trade says, and what it leaves out
Ask a stocktaking firm and you'll get monthly as the sensible baseline, stretching to six-weekly or quarterly for small sites with a clean history. That's roughly right, and it's also the answer of someone who charges per visit. What it leaves out is everything between the visits. The question isn't really "how often should a stocktaker come". It's "how long am I prepared to not know". Those are different questions with different prices.
The cost of the gap
Variance is normally quoted as a percentage of sales at retail. So the pounds at risk scale with your takings and with the weeks between counts. Say a pub has a 1% wet variance it doesn't know about. Not a disaster, well inside what a lot of people would call acceptable. Here's what that 1% adds up to before the next count tells you it exists.
| Wet takings a week | Weekly check | Monthly count | Quarterly count |
|---|---|---|---|
| £3,000 | £30 | £130 | £390 |
| £5,000 | £50 | £217 | £650 |
| £8,000 | £80 | £347 | £1,040 |
| £12,000 | £120 | £520 | £1,560 |
| £20,000 | £200 | £867 | £2,600 |
Now swap 1% for 2% and double every number. The table isn't saying you'll lose that. It's saying that's how much can leave before the count you've chosen is capable of telling you. A quarterly count on £8,000 a week has a blind spot worth a thousand pounds. A weekly spot check on the same pub has a blind spot worth eighty.
There's a second cost that doesn't fit in a table. A variance you find in week two is a cause you can still find. The line clean that didn't get logged, the new starter free-pouring, the promo button set up wrong. A variance you find in week twelve is just a number. Nobody remembers what happened in week three, so you can't fix it, you can only absorb it. And if a pub company or landlord is the one finding it, you can't argue it either. My chargeback guide is mostly about evidence, and evidence is what a frequent count is.
The cadence, by pub
One size doesn't fit. What changes the answer is how much money moves through the bar each week and how many hands are on it.
| Pub | Full count | Spot check | Time a month |
|---|---|---|---|
| Micropub or small bar, under £3,000 a week wet, 40 to 80 lines | Monthly, six-weekly if the last three were clean | Weekly, 10 lines: cask and open spirits | About 2 hours |
| Community wet-led pub, £5,000 to £10,000 a week, 100 to 200 lines | Monthly | Weekly, top 20 lines by value | About 3 hours |
| Busy or food-led site, over £10,000 a week wet, or 200+ lines | Fortnightly | Twice weekly, top 20 plus a rotating 10 | About 5 hours |
| Tied or managed with an external audit | Match their cycle, and count the week before they come | Weekly, whatever your agreement asks for or more | 3 to 5 hours |
| More than one site | Same interval at every site, same week | Weekly at each, same lines | 3 hours per site |
The time column is the bit people don't believe until they've done it with a scale and a phone instead of a clipboard. A full count on a 200-line site is about 90 minutes once the product list is set up; I've put the per-line timings in the how long does a stocktake take guide. Fifteen minutes a week for the spot check. Three hours a month is the price of never being surprised. A quarterly paper marathon is three or four hours anyway, once a quarter, and tells you everything too late. Same time. Completely different value.
The tied row deserves a word. Some agreements ask for a set number of line checks every week on top of the pubco's own stocktake. Whatever you think of that, the maths is in your favour: if you've checked the lines yourself the week before their visit, their number is never news. You've already seen it, and if it's wrong you've got your own count to say so.
What the weekly spot check actually is
A spot check isn't a small stocktake. It's a different tool. You're not valuing the stock, you're testing whether the bar behaved this week.
Pick the lines that carry the money. In most wet-led pubs the top 20 lines by value are the spirits and the fast draught, and between them they're most of the cash on the shelf. Weigh the spirits, don't eyeball them. Dip the casks and kegs, with the calibration done properly or the dip is fiction. Add three or four lines at random each week so nothing is safe from a look. Then, the same day, put what the till says sold for those lines next to what moved, and get the variance in measures and in pounds.
Same day matters. A spot check you reconcile on Thursday for a count you did on Monday has three days of sales in it that you can't separate. Count, reconcile, done, before the doors open.
What you're looking for isn't a single bad week. It's the shape across weeks. One line down two measures once is nothing. The same line down two measures four weeks running is a habit, a pour, a button or a leak, and it's cheap to find because it's still happening. I wrote up how to tell those causes apart from the numbers in the stock loss diagnosis guide.
The counts you do whether they're due or not
Cadence is the rhythm. These are the exceptions, and skipping any of them is how a clean pub gets a dirty audit.
When responsibility changes. A new manager, a relief coming in, a long holiday. Count the night before and the night after. It protects both people. Taking over a pub is the extreme version, where the count sets the price you pay.
When prices change. A new price list means a new theoretical GP on every line. A count on the change date splits the old margin from the new, otherwise the next stocktake blends two price lists into one number that means nothing.
After a big function or a bank holiday weekend. Volume goes up, extra staff come in, and comps and spillage go up with them. A count the morning after tells you what the weekend cost while the bar staff still remember it.
After a short delivery or a line-clean weekend nobody logged. Both show up on the next count as a loss. Both have an innocent explanation that only exists if someone wrote it down within a day or two.
The week before an external audit. Whether that's a pubco stocktaker, an independent, or your accountant's year-end valuation. Your count first. Then theirs is a check on you, not a verdict.
Why more often isn't the same as more accurate
The temptation once you've got a scale and an app is to count everything weekly. Don't, at least not to start. A count is only as good as its consistency. Same dead hour, same route through the building, same method on each line every time. Weigh the spirits every time, not weigh them this week and tenth them next week. Two counts done the same way give you a variance you can act on. Two counts done differently give you a number that's mostly the difference in method, and you'll chase ghosts.
Get the monthly full count boringly repeatable first. Then add the weekly spot check on the lines that matter. Only then think about tightening the interval, and only if the money says so. A £4,000-a-week pub counting weekly in full is spending time it should spend on the bar.
The short version
Monthly full count. Weekly spot check on the top 20 lines, reconciled against the till the same day. Fortnightly full count past about £10,000 a week wet. Extra counts on handovers, price changes, big weekends, and the week before anyone else comes to count you. Quarterly for the accounts and nothing else. It's three hours a month, and it's the difference between a variance you can fix and one you can only pay for.
Common questions
How often should a pub do a full stocktake?
Monthly is the floor for a pub that wants to control stock rather than just value it. Fortnightly once wet takings pass roughly £10,000 a week, because the pounds at risk between counts double. Quarterly is fine for the accounts and useless for control, because by the time a problem shows it is three months old.
How often should a pub do spot checks?
Weekly, on your top 20 lines by value plus a few rotating lines, logged against the till the same day. It takes about 15 minutes with a scale for the spirits. That is where most of the money sits, and a weekly check catches a drift in the week it starts rather than the month after.
Is a quarterly stocktake enough for a pub?
For year-end accounts, yes. For running the bar, no. A 1% wet variance on £8,000 a week is £80 a week, so a quarter between counts means over £1,000 has gone before you know anything is wrong, and nobody can remember what happened in week three. Count quarterly for the books and monthly for the bar.
When should a pub do an extra stocktake?
Any time responsibility or prices change: a manager or relief handover, a big function, a price list change, a line-clean weekend that nobody logged, a short delivery, and the day before an external stocktaker or pub company audit so you already know the number they are about to tell you.
Does counting more often make the count more accurate?
No. Consistency does. Two counts done the same way, on the same route, at the same dead hour, with the same method per line, give you a variance you can trust. A weekly count done differently every week gives you noise. Fix the method first, then raise the frequency.
Sources
- Author’s own five independent liquor audits on a wet-led site in Washington, Tyne and Wear, December 2025 to July 2026: results of +£193, +£65, +£51, −£755.71 (−1.08% of sales) and +£13. Count timings from the author’s own stocktakes.
- Sterling Stock Auditors, How often should I stocktake? — the trade view: monthly as baseline, six-weekly to quarterly for smaller sites with a clean record.
- Cost-of-the-gap table is the author’s arithmetic: weekly wet takings × 1% × weeks between counts (4.33 weeks per month, 13 per quarter).