How to do a keg stocktake

Last updated: 29 June 2026

Most pubs lose money on draught without knowing it’s happening. A 1% variance on wet sales quietly costs a typical pub £3,000–£5,000 a year — and most licensees only notice when the monthly figures land. The problem isn’t that you’re being robbed. It’s that nobody has a system to measure what’s actually in the cellar against what the till says sold. A proper keg stocktake takes 15 minutes once you know the routine. This guide walks you through it.

Why weekly keg counts matter more than you think

The number that actually matters is wet GP by line, not a single headline stock figure. Most pubs run a monthly stocktake and find a £500 variance. They write it off as shrinkage or theft. What they don’t do is work out whether it happened in week one or week four. By then, the damage is three weeks old and the root cause is a mystery.

A weekly keg count is a snapshot. It tells you whether the draught section made the GP it promised. Spirits hide losses in over-pouring (a free-poured 25ml is often 32–35ml). Draught hides it in poor cellar temperature and bad line cleaning waste. Most stock ‘theft’ is actually measurement error and forgotten wastage. A weekly count catches all three within days.

I ran my own pub on a tangle of spreadsheets for years. Still lost track of partial kegs and couldn’t tell you whether Tuesday’s 15-pint keg was full or three-quarters gone. When I switched to a simple count routine—dipstick, scales, same day reconciliation—the weekly variance went from guesswork to a number I could trust within a fortnight. Once you trust the number, you can act on it.

Most pubs that move from a messy spreadsheet to a disciplined count routine claw back 1–2 GP points within a couple of months. That’s not magic. It’s the cellar being visible.

What you actually need to do a keg stocktake

People think you need specialist equipment. You don’t. You need five things, and four of them you already have in your cellar.

• A hydrometer or gravity dip. £12–25. Floats in the sample to show specific gravity. Tells you what percentage full a cask is. Worth every penny.
• A set of scales (digital, platform style). £30–60. Weighs partial kegs. A full 20-litre keg weighs roughly 24kg; at half gravity it’s about 12kg. Scales cut the guesswork.
• A notebook or spreadsheet. The format doesn’t matter. You need date, time, line name, brand, cask tag, dip reading, weight, and any notes (leaks, temperature issues, etc.).
• Your EPOS till data exported for the same day. Total draught sales by line. Non-negotiable. You can’t reconcile without knowing what sold.
• A quiet cellar and 15 minutes. Pick the same time each week. I do mine Tuesday morning before service. Consistency beats accuracy.

Do not wait for the brewery stocktaker to arrive. They’re counting for them, not for you. Their job is to reconcile what you owe them. Your job is to reconcile what you’re actually selling.

The step-by-step process

Step 1: Prepare the cellar

Before you start counting, make sure the cellar is cool (ideally 13–15°C for lagers, 12–14°C for ales) and stable. Temperature swings change specific gravity and make readings unreliable. Turn off the line pump. Wait two minutes. You want a still cellar so your dip reading isn’t affected by turbulence.

Write down the current date and time at the top of your sheet. Seconds matter for reconciliation. If you count at 10:45am, reconcile against till data from yesterday 10:45am to today 10:45am.

Step 2: Dip every cask

Start with the first full or near-full cask. Record the cask tag (brewery batch number), line name, brand, and date. Take a sample from the sample valve (bottom tap) into a clean glass or sample pot. Float the hydrometer in the sample. Read the specific gravity to one decimal place (e.g. 1.048). That number tells you the original strength.

The current gravity (what the hydrometer reads) compared to the original gravity tells you how much beer is left. A keg stocktake reading is meaningless without the original gravity reference—breweries print it on the cask tag or brewer’s notes. If you don’t have it, ask the brewery before your first count.

Record your dip reading. Move to the next cask. Do every full cask and every partial. It takes 90 seconds per cask if you’re steady.

Step 3: Weigh every partial keg

For kegs that are low but not ready to change, dipping alone isn’t enough. A keg that looks half-full might be three-quarters full, or it might be a quarter. Weigh it on the scales. Record the weight. If you know the full weight of that keg (usually printed on the side), you can calculate the percentage left. If not, mark it as ‘approx’ and use your dip reading as backup.

Weighting partials catches the variance that dipping misses. It’s the difference between “I think it’s half-full” and “I know it’s 11.2 litres”.

Step 4: Record temperature, pressure, and notes

Once you’ve counted all kegs, write down cellar temperature and any issues: is a line running slow? Is a cask leaking? Is a keg sitting at an odd angle (affects dip readings)? These notes are gold when you reconcile. A line that’s supposed to sell 60 pints but only sold 40 isn’t loss—it’s a problem (temperature, line blockage, tap issue). Your notes spot the difference between loss and a problem.

Step 5: Reconcile against till data the same day

Export your EPOS data for the count period (e.g. yesterday 10:45am to today 10:45am). Total draught sales by line. For each line, work out expected stock movement: opening stock + deliveries minus sales should roughly equal closing stock. If your count shows 8 pints but the till says 50 pints sold, something’s wrong. Your notes will tell you what.

A variance of 2–3% is normal (line waste, over-pours, sampling). Anything above 5% needs investigation. Anything above 10% needs immediate action.

How to reconcile with till data

This is where the count stops being a number and becomes a decision.

Formula: Opening Stock + Deliveries − Sales = Expected Closing Stock. Your physical count is the actual closing stock. If they match, you’re clean. If they don’t, the gap is loss (waste, theft, spillage, measurement error, or till error).

Example: Guinness line on Tuesday.

• Opening stock: 240 pints (from last week’s count)
• Deliveries: 180 pints (one full keg, one partial)
• Till data: 310 pints sold
• Expected closing: 240 + 180 − 310 = 110 pints
• Your count: 95 pints
• Variance: 15 pints (12%)

Twelve percent is high. Check your notes. Did you record a leak? Is the line running slow? Did you miss a delivery? Most of the time, the variance is measurement error—you miscounted, or the till picked up a void or adjustment you forgot about. Ring the manager. Ask what happened Tuesday. Often the answer is “we had a line blockage and pulled 20 pints trying to clear it” or “the till went down and I manually keyed in three rounds”. Once you know, you adjust your expected closing stock and the variance disappears.

If the variance is real and unexplained, investigate: check the cellar for leaks (a slow drip under a cask adds up), ask bar staff whether they vented a line, look at till for voids (staff comping drinks). Most of the time, the root cause is not malice—it’s waste you didn’t know was happening. A weekly count is your only way to see it.

Common stocktake problems and how to fix them

Problem: Cask tags are missing or illegible

Breweries label casks with batch numbers and OG (original gravity). If the tag is gone, you can’t dip accurately. Solution: label every new cask yourself when it arrives. Use a permanent marker. Write the brand, delivery date, and OG on the cask. Takes 30 seconds per keg. Saves your entire count.

Problem: Temperature keeps changing

If your cellar swings from 10°C to 16°C, your gravity readings will vary. Specific gravity changes with temperature. A keg at 10°C reads different than at 16°C. Solution: fit a cellar thermometer (£8) and check it before every count. If temperature is unstable, your variance won’t be reliable. Fix the cellar temperature first, then trust the count.

Problem: Partial kegs are hard to read

A cask that’s one-quarter full gives a less reliable dip reading than a full cask. Solution: combine dipping with weighing. Dip to get the gravity, weigh to confirm the volume. Two data points beat one. And mark the side of the keg with a marker at key points (full, three-quarter, half, quarter) so you can eyeball it next week and spot problems fast.

Problem: The till doesn’t match the count and you can’t work out why

This happens. You counted right, the till data is right, but something sold that didn’t ring through. Solution: don’t panic. Add a ‘variance adjustment’ line to your spreadsheet and investigate weekly. If the variance is consistent (always 5 pints over), you’ve found a pattern. It might be free pours you forgot to account for, or a till issue. Once you spot the pattern, you can fix it. A weekly count makes patterns visible.

Spreadsheet versus a proper system

I built my count routine on a spreadsheet for years. It worked. But spreadsheets have blind spots. You’re typing numbers at 10:45am. By the time you export till data and reconcile, you’ve lost context. Did I remember to note that Tuesday delivery? Did I mark down the temperature? Is this week’s Guinness variance the same as last week’s?

A proper pub stocktaking system handles the repetition for you. The StockTap pub stock app is built for this exact problem. You dip and weigh on your phone, it records the time and cellar data automatically, and it reconciles against your till the same day. No manual export. No copying numbers between sheets. No wondering whether you’ve counted this cask yet.

Most importantly, a system that’s built for pub counting (not generic inventory software) understands the math: specific gravity, keg weights, partial cask logic, and the difference between loss and a problem. SmartPubTools was built by a working licensee. It knows what you’re counting and why.

The choice is simple: if you’re doing this once a week, a spreadsheet works. If you want to know your numbers in real time and catch problems before they cost you money, you need a system that’s built for pubs.