Liquor inventory: counting to the tenth of a bottle

A liquor inventory is a count of every bottle behind the bar and in the store, full and open, turned into a figure you can set against what the bar sold. The full bottles are quick. The open ones are where the time goes and where the errors hide, because a bottle that is somewhere between half and two thirds full has to be written down as a number.

There are three ways to put a number on an open bottle. You can judge it by eye in tenths, you can weigh it, or you can scan it with an app that identifies the bottle and does the weighing maths for you. Each gives a different accuracy, and each costs a different amount of time on a night when the team would rather be going home.

This guide compares the three honestly. It shows the arithmetic behind weighing spirits, how to build a count sheet that is quick to fill in, how to turn a count into usage and shrinkage, and how spot checks keep the figures honest between full counts. Every number in it is an example, not a benchmark.

A count on its own tells you what is on the shelf. It becomes useful when it is joined to two other figures: what came in, and what was sold. Opening stock plus deliveries, minus closing stock, gives what the bar used. Sales, multiplied by the measure in each drink's spec, give what it should have used. The difference between the two is the variance, and the variance is what bar inventory management is really about.

That is why accuracy matters more than it first appears. The variance on a well-run bar is a small number, often a fraction of a bottle per product over a week. If the count itself can be wrong by more than that, the variance stops meaning anything, and the team learns to ignore it.

The same count feeds the pour cost you report for the bar. Our guide to pour cost shows how the actual figure is built from usage, and our guide to actual vs theoretical food cost explains the same gap from the kitchen side. A liquor inventory that is close enough for the accounts can still be too loose to manage a bar with.

Sealed bottles are counted as whole units, wherever they are: on the back bar, in the speed rail, in the cellar, in the function bar that only opens for events. The common mistakes with full bottles are not counting errors but location errors, a case left by the kitchen door or a bottle on a shelf nobody thought of as stock.

Open bottles are where the method matters. The traditional convention is to record each one in tenths, from 0.1 for a bottle that is nearly empty to 0.9 for one that has had a measure or two taken out. Tenths are popular because they are quick to judge and quick to write, and because they turn into money with simple arithmetic.

It helps to know what a tenth is worth before deciding whether it is precise enough. On a 70 centilitre bottle, one tenth is 70 millilitres, a little under three 25 millilitre measures. On a one litre bottle, a tenth is 100 millilitres, four measures of 25. A count recorded in tenths can therefore never be more precise than two or three drinks per bottle, and in practice it is usually less precise than that.

Write the bottle size on every line of the sheet. A 70 centilitre and a one litre bottle of the same spirit are different products for counting, and mixing them up is one of the most common reasons a liquor inventory looks wrong.

The quickest method is to look. Hold the bottle upright at eye level, judge the liquid against the shoulder and the base, and write the tenth. A practised counter can move along a back bar at a steady pace, and nothing is needed but the sheet and a pen.

The accuracy is the problem. For this guide's examples, assume that a careful counter is usually within half a tenth either way. On a 70 centilitre bottle, half a tenth is 35 millilitres, more than one measure. On a bar with 120 open bottles, those small errors add up quickly, and they are not always random.

Three things make the eye less reliable than it feels:

Random errors partly cancel across many bottles. Habitual rounding does not. That is the real case against relying only on the eye: not that any single judgement is far out, but that the errors lean in one direction and quietly move the result.

Weighing replaces the judgement with a scale. A bottle of spirit weighs the glass plus the liquid, and the liquid can be worked out from the difference between what the bottle weighs now and what it weighs when it is empty.

You need a scale that reads to one gram and holds at least three kilograms, and two figures for every product on the bar: the weight of the bottle when it is full and sealed, and its weight when it is empty. The empty weight is often called the tare. Weigh empties as they come off the bar, and weigh a full bottle when the next delivery arrives.

With those two weights, the share of the bottle that is left needs no knowledge of the liquid at all:

This is the safest way to work, because it uses each product's own figures. Spirits at 40 per cent alcohol weigh about 0.95 grams per millilitre, a little less than water, because alcohol is lighter than water; the density tables for mixtures of the two are on Wikipedia's ethanol data page. Cream and sugar-based liqueurs can be heavier than water. Using one density for the whole bar would introduce an error the full and empty weights avoid.

A worked example, with invented weights. A 70 centilitre bottle of gin weighs 1,180 grams full and 515 grams empty, so the liquid in a full bottle weighs 665 grams. Tonight the open bottle weighs 860 grams:

By eye, the same bottle might have been called 0.5 or 0.6, a difference of 70 millilitres. Weighing spirits brings the uncertainty down to a few millilitres per bottle, as long as the empty and full weights are right.

Two practical points decide whether weighing works in your bar. First, decide whether bottles are weighed with their pourers in, and keep to it; a pourer can weigh as much as a measure of spirit. Second, re-weigh full and empty bottles when a supplier changes the bottle design, because a new bottle with the same label can be heavier or lighter by enough to matter.

Scanning tools combine the two earlier methods with a phone or tablet. The counter scans a bottle's barcode, the app finds the product and its bottle weights in its own database, and a connected scale passes the weight straight to the app, which works out what is left. Some tools estimate the level from a photograph instead of a scale.

The gains are real. Nobody looks up a tare weight, nothing is copied from paper into a spreadsheet, and the count is in a system the moment it is finished. On a large bar with several counters, removing the transcription step can save as much time as the weighing itself.

The costs and limits are real too. Scanning tools carry a subscription and usually need their own scale. Their bottle databases are only as good as their data, and local spirits, own-label bottles and decanted products may not be in them. A photograph-based estimate is still an estimate. And a wet bar at the end of the night is not kind to phones, so the routine needs a dry place to work and a charged device.

Whatever the tool, check its numbers when you start. Weigh ten bottles by hand, count them with the tool, and compare. If the two disagree, find out why before trusting either.

The right method is the one whose accuracy is worth the time it costs in your bar. To compare them, take an example back bar with 120 open bottles and assume one experienced counter. The timings below are illustrations; time your own team before deciding.

Now put a value on the difference. Suppose that half of the 120 open bottles are judged half a tenth too high, which is easy to do with habitual rounding. That is 60 bottles times 35 millilitres, or 2,100 millilitres: three bottles of stock that exist only on paper. At an example cost of 18 per bottle, the count has added 54 of stock that is not there, and the week's variance will look 54 better than it is. If your usual variance is smaller than that, eyeballing cannot tell you whether the bar is getting better or worse.

There is no single answer, but the bar's mix usually decides it.

A cocktail bar or a bar with a serious spirits list carries a lot of value in open bottles, and the gap between a careful and a careless count is worth real money. Weighing, or scanning with a scale, pays for itself there.

A pub that sells mostly draught and bottled beer has fewer open spirit bottles and more value in kegs. Tenths may be good enough for the spirits, as long as the counter is consistent, and the effort is better spent on weighing kegs and reconciling deliveries.

A restaurant with a small bar of 30 or 40 bottles can use a hybrid. Count most bottles in tenths, and weigh the ten that carry the most value: the house spirits that pour all night and the expensive bottles where a small error is a large sum. Ranking stock by value and spending effort on the top of the list is the same thinking as setting par levels by value rather than treating every line the same.

Whichever method you choose, keep it. A bar that switches between tenths and weights from one count to the next creates variance out of the change of method alone.

A good sheet makes the count faster and more accurate before anybody picks up a bottle.

Our guide to running a restaurant stock count properly covers the same discipline for the kitchen, and the same sheet design works for both.

Most bars cannot close for an hour to count, and they do not need to. The bar inventory management guide sets out the full routine; the essentials for a liquor inventory are these.

Once the count is in, the arithmetic is simple and the same for every product:

An example for one gin, in bottles of 70 centilitres. The week opened with 6.4 bottles, 12 were delivered, and the closing count is 7.9. Usage is 6.4 plus 12 minus 7.9, which is 10.5 bottles, or 7,350 millilitres.

Sales for the same week show 270 gin and tonics at 25 millilitres each, which is 6,750 millilitres, about 9.6 bottles. The difference is 600 millilitres, or 24 measures. Whether that is acceptable depends on what sits behind it, but it is now a question with a size, not a feeling.

Value the usage at cost and the same calculation gives the cost of what the bar used, which is the starting point for the actual pour cost. Do it product by product. A total for the whole bar can look fine while one product loses more than everything else put together.

Shrinkage is the stock that left the bar without being sold. Some of it is unavoidable: a dropped bottle, a measure poured for a drink sent back, the last drops in a bottle that never reach a glass. The rest has causes worth finding, and the order in which you look matters.

Read shrinkage by product and by week. One week's spike is often a count error. A gap that repeats on the same products is a pattern.

A full liquor inventory once a week, or once a month in a quieter bar, leaves long gaps in which nobody knows how the bar is doing. Spot checks fill them. Pick 10 to 20 lines, weigh them at the start and the end of a shift, and set the difference against the sales of those drinks for the same shift.

Choose the lines with care. The house spirits pour all night and show pouring habits quickly. The expensive bottles show the cost of mistakes and the temptation of free drinks. Rotate a few others in so that nobody can predict which bottles will be checked.

Spot checks are quick because the list is short, and they are useful because they are close to the event. A variance found on the night it happened can be explained by the people who were there. One found three weeks later rarely can.

Kegs are best weighed too. Record the empty weight of each keg size from your brewery, weigh the open keg at each count, and convert the difference to litres. Beer weighs a little more than water, so use full and empty keg weights from your own kegs rather than assuming one kilogram per litre.

Wine by the glass is counted like spirits: in tenths, or by weight using the full and empty bottle. Open bottles of wine move quickly and spoil, so count them every night if wine by the glass is a large part of sales, and record the end of a bottle that is poured away.

Some products always need weighing. Liqueurs in opaque or oddly shaped bottles, cream liqueurs, syrups made in house, batched cocktails in their own containers and bag-in-box products are all close to impossible to judge by eye. Give each one its own full and empty weight, and count it the same way every time.

Each of these moves the result without anything happening at the bar, which is exactly why a figure built on them stops being trusted.

A liquor inventory has two halves: the stock that was used and the drinks that were sold. servearound covers the sales half, and the records that explain a gap. The count itself stays on your sheet or in your counting tool.

Every drink rung through the point of sale appears in the menu report in Analytics, with units sold, revenue and share of sales for each drink and each category over any range of dates. That is the sales column for working out what the bar should have used. The payments report shows discounts, voids and comps by member of staff, which is where unrecorded giveaways start to show. When a bottle runs out, 86 the drink on the till and it shows as sold out on every table's menu at once.

servearound does not track stock, bottle weights or recipe costs today, so the count and the costing stay in your own sheet. Inventory tracking, for the stock behind the menu, is coming. If you would like to see the sales side for yourself, get in touch.

Often enough that a variance can still be explained by the people who were working. Many busy bars count spirits weekly and everything else monthly, with spot checks in between. A quieter bar may count monthly. Our guide to what a stocktake is and how often to do one sets out the trade-off in more detail.

Often, for part of the bar. Weigh the bottles that carry the most value and count the rest in tenths. A scale costs less than the stock a habitually rounded count can hide in a single week.

It is the traditional unit for open bottles: 0.1 is almost empty, 0.9 has had a little taken out. On a 70 centilitre bottle a tenth is 70 millilitres; on a one litre bottle it is 100.

No. Record each product's full and empty weight and work out the share that is left from those two figures. That avoids density altogether, and it handles liqueurs and syrups correctly.

Decide once and keep to it. Either weigh every bottle with its pourer and include the pourer in the empty weight, or take every pourer off before weighing. Mixing the two moves the result by up to a measure a bottle.

There is no figure that holds across bars. Measure your own by product for several weeks, then work on two things: making the count trustworthy, and narrowing the gap on the products where it repeats.

They know the bar best, so involve them, but pair them with someone else. Two people counting is faster and more accurate, and it protects the team as much as the stock.

Yes, if you weigh them. Weighing gives a figure to the nearest few millilitres, and it is better to record that figure than to round it back to tenths.

A liquor inventory is only as good as its open bottles. Tenths are quick but rarely closer than half a tenth, and habitual rounding pushes the error one way. Weighing spirits against each product's full and empty weight brings the error down to a few millilitres, and scanning tools add speed by removing the copying. Choose by the value in your open bottles, and keep the same method every time.

Turn the count into usage with opening stock plus deliveries minus closing stock, set it against sales product by product, and investigate shrinkage in order: the count, the deliveries, the rings, the comps, the pour, and only then the people. Use spot checks on 10 to 20 lines between full counts, so a gap is found while it can still be explained.