Bar inventory management without closing the bar

Bar inventory management has a reputation for being slow, fiddly and disruptive. Counting hundreds of open bottles, estimating what is left in each one, reconciling kegs and wine by the glass, all after a long night, feels like a job that needs the bar closed for a morning. So it happens rarely, roughly, or not at all, and the bar becomes the one part of the business where nobody really knows what is going on.

That is expensive, because a bar holds a lot of value in a small space and loses it in small, constant amounts. A heavy pour here, a spill there, a drink for a friend, a bottle broken during a rush, a round rung through as the wrong item. None of these is dramatic on its own. Over a month they can add up to a meaningful share of drinks revenue, and without counting, there is no way to tell which of them is happening.

This guide sets out a bar count that fits into the hour between last orders and lock up, once a week. It covers how to count open bottles quickly and accurately, how to work out pour cost by category, how to set bar par levels, a bar stock checklist, and how to find the handful of lines that usually explain most of the variance. It assumes no special equipment beyond a set of scales.

Bar losses come from a short list of causes, and it helps to know them before designing the count, because the count is how you tell them apart.

Most bars experience all of these to some degree. The point of bar inventory management is not to eliminate them, which is impossible, but to measure them so that the large ones can be found and reduced.

A working bar system has four parts:

The comparison in step three is the heart of it. If the till shows forty measures of a gin sold and the stock shows sixty measures used, twenty measures went somewhere. The count tells you that the gap exists and how big it is; working out why is the next job.

The best time for a bar count is after the last drink is served and before the bar is locked, on the same night each week. Nothing is moving, the deliveries for the week have arrived, and the people who ran the bar that night are still there to answer questions. Counting the next morning risks a delivery arriving or a bottle being used for prep before the count is done.

Pick the quietest closing night of the week if you can, so the count does not fall at the end of an exhausting shift. Keep it the same night every week. The count defines the boundary of the week, and sales need to be cut off at exactly the same moment.

The route matters as much as the method. Count in the order the bar is laid out: the speed rail first, then the back bar from left to right, then fridges, then the cellar or store room. A count sheet that follows that route, rather than an alphabetical list, cuts the time in half. Our guide on how to run a restaurant stock count properly covers the same principle for the kitchen.

Unopened bottles are easy: count them in whole units. Open bottles are where most of the time and most of the error sit. There are two ways to count them.

The first is by eye, in tenths. The counter looks at each bottle and estimates how full it is: seven tenths, four tenths, and so on. It is quick and, with practice, reasonably consistent, but it drifts, and it is easy to round in the bar's favour.

The second is by weight. Weigh each open bottle, subtract the weight of the empty bottle for that product, and convert the remaining weight to volume. It takes a little setting up, because you need the empty weight of each product, recorded once. After that, it is fast and much more accurate, and it removes the temptation to round. For spirits, where a few millilitres per pour matter, weighing is worth the effort.

Draught is harder to count by eye. If your system has meters, read them. If not, weigh kegs: a full keg and an empty keg of the same product give you the scale, and the current weight tells you roughly what is left. Record line cleaning and the beer pulled through during it, because that is a real, regular loss that should not be mistaken for something worse.

Count unopened wine by the bottle. Count open bottles of wine sold by the glass in tenths, like spirits, and record any bottles that were opened and spoiled. Wine by the glass is one of the most common sources of loss in a bar, because open bottles oxidise and are poured away, and generous pours are easy.

Pour cost is the cost of the drinks used, as a share of drinks revenue. It is the bar's equivalent of food cost percentage, and it is calculated the same way:

Calculate it by category, not just for the bar as a whole. Spirits, wine, beer and soft drinks have very different natural costs, and a healthy overall figure can hide a wine programme leaking money behind a strong beer trade. If your selling prices include tax, remove it before dividing, or the figure will look better than it is.

As with food, there is no universal target that suits every bar. A cocktail bar, a pub and a restaurant with a serious wine list will all run different pour costs for good reasons. The useful number is your own, measured the same way every week, and its movement over time. The idea of gross margin is the other side of the same coin: what each sale leaves over after the cost of the drink itself.

If each drink has a recipe, with a standard measure and garnish, you can calculate a theoretical pour cost from sales: the cost each drink should have had, multiplied by the number sold. The gap between theoretical and actual is the bar's variance, and it is where over-pouring, spillage, comps and errors live.

In most bars, a small number of products account for most of the variance. They tend to be the products that sell in the highest volumes, are poured by eye, or sit open for long periods. Typical candidates:

Start your variance work with your own top six by revenue and by volume. If the variance on those six is under control, the rest of the bar usually follows. If it is not, those six are where the money is.

Bar par levels work the same way as kitchen pars: the quantity of each product you need to get through to the next delivery, plus a small buffer. Work them out from usage per week or per cover, multiply by the days each delivery must cover, and round to the case size.

The difference at the bar is shelf life. Most bar stock lasts a long time, so the risk of over-stocking is tied-up cash rather than spoilage. A cellar holding three months of a slow-moving spirit is money that could be doing something else. Keep buffers modest on expensive products and generous only where running out would be serious. Our guide to par levels covers the method in detail.

Remember that the bar itself holds a working stock that is topped up from the cellar. Set a par for the bar, which the bartender restocks to from the cellar at close, and a separate par for the cellar, which drives orders from suppliers.

Use this as the weekly routine. Each step is short.

The bar stock checklist includes recording comps, staff drinks and breakages, and this step is what separates explained variance from unexplained. A drink given to a guest as an apology, a drink for the team at the end of the night, and a bottle broken during service are all legitimate. When they are recorded, they come out of the variance and into their own lines. When they are not, they sit in the gap and look like theft.

Make recording easy. A single sheet or a dedicated button on the till for comps and staff drinks, used every time, is enough. Agree a policy for staff drinks so that everyone knows what is allowed, and record breakages with the product and the reason.

Here is an illustrative week for one house gin, to show how the comparison works. The figures are examples rather than benchmarks.

Recorded usage is 457.5 centilitres against 469 used, an unexplained variance of 11.5 centilitres, roughly four or five measures across a week. That is small, and probably within the normal noise of free pouring. If the same calculation showed sixty centilitres missing, the conversation would be very different, and it would be about specific nights and specific shifts.

When a product shows a larger variance than usual, work through the causes in order. Check the count first: a misweighed bottle or a missed case explains many surprises. Then check the paperwork: deliveries, transfers and credits in the right week. Then check the till: is the drink being rung as the right item, and are cocktails linked to their spirits correctly? Only then look at pouring, spillage and loss, ideally by looking at which shifts the variance is concentrated in.

Most of the variance a bar can control comes down to how drinks are poured. Measured pours, using jiggers or measured pourers, reduce over-pouring sharply. Free pouring is faster and suits some bars, but it needs regular checks: a timed pour test with water, where each bartender pours a set measure into a measuring cylinder, keeps everyone honest and shows who needs practice. Post drink recipes, with measures and garnishes, where the team can see them.

Events and holiday weekends distort bar numbers more than kitchen ones, because drinks volumes swing harder and packages blur what was actually poured. A wedding with an open bar, a corporate party on a drinks package or a long bank holiday weekend can each move a week's pour cost by several points without anything going wrong.

Treat these as their own periods where you can. Count before and after a large event, so its usage can be separated from normal trade. Record what the package included and what was actually served, and compare the two, because a package priced on an assumed number of drinks per guest is only profitable if the assumption holds. After a busy weekend, expect the weekly variance to be a little wider and look for patterns rather than single numbers. Over several events, you will learn how much your guests really drink on a package, which is worth knowing the next time you price one.

A bar count is only as consistent as the people doing it. When a new bartender joins, show them the count route, the sheet, how to read tenths and how to weigh. Count alongside them for their first two or three weeks. The same goes for pouring: new staff should learn the house measures and the drink recipes before their first busy shift, not during it. A short pour test with water, done at the start of a shift every few weeks, keeps measures honest without turning it into a disciplinary matter.

Give the bar count a clear owner, usually the bar manager or head bartender, and a second person who counts with them. The owner is responsible for the count, the comparison and the weekly list of largest variances. The general manager reviews the pour cost and the variances weekly, briefly, and agrees what will change. Rotate the second counter so that the count stays independent.

The count only works if it happens every week, so treat it as part of closing on the chosen night rather than an extra job. Put it on the rota, give it the time it needs, and keep the sheet in the same order every week. After a few weeks, a well-organised bar count genuinely fits between last orders and lock up, and the bar stops being the part of the business nobody understands.

The count happens behind the bar, with scales and a sheet. Where servearound helps is with the sales side of the comparison and with the records that explain variance.

Drinks are items on a menu like any other, and the bar menu can run on its own hours. Once a drink's cost is set, the menu report in Analytics shows its units sold, revenue and margin from what was actually rung through, which is what you compare with usage. The payments report breaks voids and comps down by member of staff, which is where unrecorded giveaways start to show. And if the bar runs out of a spirit, marking it sold out takes it off every menu and terminal at once, so nobody keeps selling a cocktail the bar cannot make. If you would like to see it, get in touch.

Weekly is the rhythm most bars can sustain and the one that catches problems while they are still small. Monthly is the minimum. Daily counts of a few high-value lines can help when chasing a specific problem.

For spirits, usually yes. It removes the estimate from the most valuable part of the count, and once the empty weights are recorded, it is quick.

That is a policy decision for each venue. Whatever the policy, record staff drinks every time, so they are a known cost rather than an unexplained one.

Measure your own, by category, for several weeks, and aim to stabilise and then reduce the gap between actual and theoretical. A target borrowed from a different kind of bar is rarely meaningful.

Give each cocktail a recipe with measured quantities of each ingredient, so that sales of the cocktail can be converted into usage of each spirit, liqueur and mixer.

They can be counted on the same night, which keeps the cut-off consistent, but keep them on separate sheets and separate figures. Food cost and pour cost behave differently and need different conversations. Ingredients used by both, such as citrus, herbs or cream for cocktails, should be counted where they are stored and transferred to the bar on paper when the bar takes them, so neither figure is distorted.

Record them as their own line, like staff drinks, so they are explained rather than lost in the variance.

Bar inventory management does not need the bar closed. Count once a week between last orders and lock up, in route order, weighing open spirits and recording kegs, wine by the glass, comps, staff drinks and breakage. Compare usage with sales by product, calculate pour cost by category, set bar par levels for the bar and the cellar, and start with the handful of lines that carry most of the variance.

Done every week by a clear owner, the count turns the bar from a mystery into the best-understood part of the business.