Pub cellar log book: what actually matters

Last updated: 26 June 2026

Most pub cellar log books are either gathering dust or filled with numbers that don’t actually prevent loss. I’ve seen licensees record gravity readings they never act on, stock figures they don’t reconcile against till data, and wastage entries that bear no relationship to what really happened in the cellar. Then they wonder why their year-end stocktake reveals a 3–5% variance they can’t explain.

The truth is that a 1% stock loss on wet sales quietly costs a typical pub £3,000–£5,000 a year—and a proper cellar log book catches it. But only if you’re recording the right numbers, checking them weekly, and acting on what you find. This article walks you through what a cellar log book should actually contain, why most pubs get it wrong, and how to build a count routine that works in real life, not just in theory.

You’ll learn why spirits hide losses differently than draught, what a weekly line check actually reveals, and how to reconcile your cellar figures against till data the same day. If you’re tired of guessing whether your stock variances are theft, measurement error, or wastage, you need to read this.

What actually goes in a cellar log book

A pub cellar log book isn’t a diary. It’s a record of measurable facts that let you reconcile what you bought against what you sold and what remains. Most pubs fail on this immediately: they record opening stock figures that were never correct to begin with, or they skip entire lines because “the rep just filled it.”

The most effective way to prevent stock loss is to record the same measurements every time: gravity readings for every cask, weight for every open spirit bottle and partial keg, cellar temperature, and the date of the last line clean. If you don’t record these things in the same order on the same day every week, you have no baseline to compare against.

At my own pub I was running stock on a tangle of spreadsheets and still losing track of partial kegs and spirit measures. I built a simple count routine around a dipstick and a set of scales, and the weekly variance went from guesswork to a number I could trust within a fortnight. Here’s what I record:

• Cask gravity: For every cask in the cellar, a gravity reading (not an estimate). A dipstick costs £8.
• Open bottle weight: For every open spirit bottle, the weight in grams. A digital scale costs £15. Weigh empty, record it once, then weigh every week. The loss is the difference.
• Partial keg volume: Either by weight (for larger kegs) or by dipping if you have a measuring tape and know the keg dimensions. Record the date received and the gravity if it’s draught.
• Cellar temperature: Record it once a week. A thermometer costs nothing. Temperature directly affects yield and shelf life.
• Line cleaning date: Record when you last cleaned each draught line. Most pubs think they clean lines weekly; they don’t. This proves it.
• Till reconciliation: At the end of the count, record what the till says you sold that week. You’ll reconcile cellar against sales the same day.

That’s it. No flowery descriptions. No notes about who worked shifts or why stock went up. Just data you can compare week on week.

Why wet GP by line is the number that matters

Most pubs chase a headline stock variance figure—a single percentage for the whole pub. This is a waste of time. A pint of lager selling at 65% GP tells you something completely different than a pint of craft IPA selling at 55%. If one line is hemorrhaging margin and another is dead solid, a headline figure hides both.

The number that actually matters is wet GP by line, not a single headline stock figure, because it tells you which specific drink is costing you money and why. At the end of each week, calculate this:

Cellar value of what left the cellar (opening stock + deliveries − closing stock) divided by till sales for that line. If it’s lower than your expected GP, something is wrong. Either you’re over-pouring, the till is not ringing it, you’re giving away samples, or there’s spillage.

I found this out the hard way. My headline variance was running at 2.8%—acceptable, but not great. When I broke it down by line, I discovered that my house lager was running at 3.8% loss (over-pouring at the bar, it turned out), my cask ales were at 1.2% (fine), and my spirits were at 4.1% (theft plus over-pouring plus the fact that I wasn’t weighing bottles). I couldn’t see any of this in the overall number.

Once I fixed the house lager pour and started weighing spirit bottles, those two lines alone clawed back 1.2 GP points. Within two months, pubs that move from a messy spreadsheet to a disciplined count routine typically claw back 1–2 GP points within a couple of months.

Building a weekly count routine that works

The biggest mistake licensees make is treating the cellar log book as a Friday afternoon job when the weekend is about to hit. You won’t get accurate numbers this way. Instead, schedule it for a quiet day—Tuesday or Wednesday morning if you can—and give yourself 45 minutes to an hour. Don’t rush.

Here’s the routine I follow:

1. Count casks and kegs first: Start with full casks (record gravity), then partial casks, then full and partial kegs. Record gravity and any visible sediment or quality issues. This takes 15 minutes if you have 10–15 lines.
2. Weigh every open spirit bottle: This is non-negotiable. A 70cl bottle open for three weeks and selling at 25ml pours should show 200ml loss. If it shows 400ml loss, you have a problem. Record the weight, not the visual estimate.
3. Record temperature and line cleaning date: If a line hasn’t been cleaned in more than two weeks, add it to your to-do list today. Temperature should be 50–52°F for ales, 38–40°F for lagers and ciders. If it’s not, you’re wasting product.
4. Reconcile against till data immediately: Don’t wait until Friday. Pull the till report for the past week, compare it against what left the cellar, and if the variance is more than 2%, investigate the same day. You want to catch errors and issues fresh, not four weeks later when you can’t remember what happened.
5. Record any delivery or adjustment: If a rep delivered a replacement cask, if you poured a pint away because it was cloudy, if you found a broken keg seal—record it. Context matters when you’re reconciling.

The entire job is 45 minutes. If you skip it because you’re busy, you might as well throw money in the bin. One missed week means you can’t reconcile the following week’s numbers accurately.

Why spirits and draught hide losses differently

Spirits and draught don’t lose money in the same way, and a cellar log book that treats them identically will miss both problems.

Spirits hide losses in over-pouring because a free-poured 25ml is often 32–35ml, and over-pouring is invisible until you weigh the bottle. A bar staff member isn’t stealing from you; they’re just pouring generously. At a pound a millilitre, a 7ml over-pour on every large spirit is £7 a pint lost. If you pour 40 spirits a week, that’s £280 a week you’re leaving on the bar. Most pubs don’t realise this is happening because they never weigh bottles.

Draught hides losses in cellar temperature, line cleaning, and poor turnover. A cask that sits for two weeks at 55°F instead of 51°F is degrading. A draught line that hasn’t been cleaned in three weeks is losing yield to sediment and bacteria. A partial keg that’s been open for six weeks is flat and unsellable. None of these show up in a till variance—they show up in your cellar count as “shrinkage” when actually they’re quality failures.

A proper cellar log book separates these. For spirits, you weigh and reconcile pours. For draught, you track temperature, line cleaning dates, and cask age. This tells you whether your loss is a people problem, a process problem, or a quality problem. The solution is different in each case.

Reconciling cellar against till data

This is where most pub log books fall apart. Licensees record the numbers but never actually compare them. Recording without reconciliation is just busywork.

At the end of each week, take your till report for wet sales and your cellar count, and calculate what should have left the cellar. Here’s the formula:

Opening stock value (at cost) + Deliveries − Closing stock (at cost) = Cost of goods that should have been used

Divide that by till sales for the week, and you get your actual cost of goods sold (COGS). If your expected COGS is 28% and your actual is 29.5%, you’re losing 1.5 points of margin. That’s real money.

Now find where it went. Was it one line, or across the board? Was it spirits, draught, or cider? Did it happen on a specific shift? Cross-reference your till data with your cellar log—if the till says you sold 50 pints of lager last Tuesday but your cask only fell two inches, something’s wrong.

Most stock loss isn’t theft. It’s measurement error and forgotten wastage. A cellar log reconciled against till data catches both within days. Then you fix the measurement (weighing bottles instead of guessing, for example) or you address the waste (cleaning lines more often, for example). Within a fortnight, you’ll see the variance tighten.

Common mistakes that wreck accuracy

After 15 years running a pub and helping other licensees build stock routines, I’ve seen the same errors again and again. Here are the ones that matter:

Estimating instead of measuring

If you’re eyeballing how much is left in a cask or estimating a spirit bottle loss, you’re not running a log book. You’re writing fiction. Get a dipstick, get scales, and record actual numbers. The upfront cost is £25. The margin you claw back is thousands.

Recording opening stock you never verified

If you inherit a “stock figure” from the previous management or from a stocktake four months ago, you’re starting from a lie. Your cellar log begins the day you first measure everything yourself. Before that, it’s a guess.

Skipping delivery days or adjustments

If a cask was replaced mid-week because it arrived damaged, you need to record that. If you poured a pint away because it was infected, you need to note it. Without context, a variance makes no sense.

Not reconciling weekly

If you only reconcile cellar against till once a month or once a quarter, you’ve lost the thread. By then, you can’t remember what happened, which shift caused the problem, or whether it was a one-off or a trend. Reconcile weekly. It takes 10 minutes and it catches issues fresh.

Recording without acting

A cellar log that isn’t acted upon is theatre. If your variance is 3.2% and you do nothing, next week it’ll be 3.4%. If you discover that your house lager is over-poured and you don’t brief the bar team about it, nothing changes. A log book is only useful if you use it to make decisions.

A cellar log book is not optional if you’re serious about margin. Most pubs that run one properly claw back 1–2 GP points within two months. That’s £2,000–£4,000 a year on a typical pub. The cost of the equipment is negligible. The cost of not running one is enormous.

If you’re managing multiple lines and want to automate the reconciliation maths instead of doing them in a spreadsheet, SmartPubTools has built cellar tracking into its system so you can record weekly counts, compare against till data, and spot variances by line in one place. But the discipline of recording accurate numbers yourself is non-negotiable either way.