Skip to content

18 November 2026Inventory, food cost & wasteBy The ServeAround team18 min read

Over pour: what a heavy hand costs in a year

Five millilitres a drink looks like nothing. Across a busy bar it is several bottles a week. What over pour costs, how to measure it and three fixes that keep service fast.

Illustration for over pour: a wine glass, a tumbler and a cocktail beside a slice of lemon, in yellow and coral on cream.

A bartender pours 30 millilitres of gin into a drink that should have 25. Nobody at the bar notices. The guest certainly does not complain. The drink goes out, the next one is poured the same way, and the extra five millilitres disappear into a busy night as if they had never existed.

They have not disappeared from the stock count. Over pour is the gap between the measure in the spec and the measure in the glass, and on a busy bar it is often the largest single cost that no report names. Five millilitres on one drink is a few pence or fils. Five millilitres on every spirit poured in a year is a figure most owners would never sign off if it arrived as an invoice.

This guide prices that figure out, product by product and across a whole bar, shows how to measure it, explains how to tell it from the other ways a bar loses stock, and sets out three fixes that bring it back without slowing service. Every bar, price and volume in it is an example, not a benchmark.

What over pour is, and what it is not

Over pour is liquid that goes into a sold drink beyond what the drink's specification says. The guest receives it and pays nothing for it. It is different from four other losses that end up in the same place on the stock count:

  • Spillage: liquid that never reaches a glass, knocked over, dropped or poured down the drip tray.
  • Giveaways: drinks handed out free and not recorded, whether a round for a regular or a staff drink.
  • Unrung drinks: drinks made and served but never put through the till, by mistake or otherwise.
  • Breakage and spoilage: a dropped bottle, a corked wine, a flat keg.

All five make the actual cost higher than the theoretical one. Only over pour reaches the guest inside a drink they paid for, which is why it is so easy to tolerate and so hard to see.

Five millilitres, priced out

Start with one drink. A house gin costs 84 for a 700 millilitre bottle, which is 0.12 a millilitre. The spec is 25 millilitres, so the measure costs 3.00. Pour 30 and it costs 3.60. The extra 0.60 is the over pour on one drink.

That drink is poured, say, 350 times a week. The heavy hand costs 210 a week on house gin alone, or 10,920 over a year. The figure comes from three ordinary numbers: a small extra measure, a modest price per millilitre and the volume a busy bar sells. None of them looks alarming on its own, and that is the point.

Our guide to pour cost shows what the same five millilitres does to one drink's percentage: a gin and tonic moves from about 14 to about 16 per cent of its price. The percentage is useful for a single drink. For the whole bar, the money is the clearer measure.

A whole bar, priced out

Now take a bar's house spirits for an illustrative week, each poured five millilitres over a 25 millilitre measure, with example bottle prices.

  • House vodka: 400 measures a week, from a 700 millilitre bottle costing 70, so 0.10 a millilitre and 0.50 over on each drink. That is 200 a week.
  • House gin: 350 measures at 0.12 a millilitre, so 0.60 over on each. That is 210 a week.
  • House rum: 150 measures from a bottle costing 77, so 0.11 a millilitre and 0.55 over. That is 82.50 a week.
  • House whisky: 120 measures from a bottle costing 91, so 0.13 a millilitre and 0.65 over. That is 78 a week.
  • House tequila: 80 measures from a bottle costing 98, so 0.14 a millilitre and 0.70 over. That is 56 a week.

Together the bar pours 1,100 house measures a week, and the over pour costs 626.50 a week, or 32,578 a year. In liquid, it is 5.5 litres a week: nearly eight full bottles poured out every week that nobody paid for.

Cost, not lost sales

It is tempting to value those eight bottles at their selling price. Five and a half litres is 220 measures, and at an example 30 a measure before tax that looks like 6,600 a week in lost sales. The figure is dramatic and mostly wrong.

The bar did not lose those sales unless it ran out of stock and turned guests away. It lost the liquid, which it had to buy again. Value over pour at cost, as the example does. It is still large enough to matter, and a figure the team can check against an invoice is a figure they will believe.

Where over pour hides

Over pour is rarely spread evenly. It gathers in a few places, and knowing them is half of finding it.

  • The speed rail. The house spirits are poured most often, fastest and most often by eye.
  • Doubles and long drinks. A double poured as "a good splash" is easier to overdo than a single.
  • Cocktails with several spirits. Three spirits each poured a little heavy add up within one glass.
  • The last hour. Tired hands and a full bar drift towards generosity.
  • New starters. Someone still learning the bar pours to look confident, not to the line.
  • Regulars. The guest the bartender knows by name gets the generous hand, every time.

None of these is a character flaw. They are what happens when pouring is done by feel under pressure. The fixes later in this guide work by changing the conditions, not by blaming the person.

Wine by the glass

Wine has its own version. A 175 millilitre glass poured to 190 is 15 millilitres over. From a 750 millilitre bottle costing 60, which is 0.08 a millilitre, that is 1.20 a glass. At 300 glasses a week, it costs 360 a week, or 18,720 a year.

The yield tells the same story another way. A bottle holds just over four 175 millilitre glasses. Poured at 190, it gives fewer than four, so every bottle sold by the glass loses part of a sale's worth of wine. Large glasses make this worse, because a generous pour looks modest in a big bowl. A line on the glass or a measured pour is the only reliable defence.

Draught beer

Draught over pour is mostly beer that goes over the top of the glass and down the drip tray: the overflow while topping up a pint, or a glass held at the wrong angle. Say 30 millilitres goes to the tray on each 568 millilitre pint. From a 50 litre keg costing 180, that is about 0.11 a pint. At 1,000 pints a week, it is 108 a week, or 5,616 a year.

The keg's yield shows it too. A 50 litre keg holds about 88 pints. Lose 30 millilitres on each and it gives closer to 84. Line cleaning and the first pulls of a new keg are separate, legitimate losses. Record them as such, so they do not hide inside the over pour or get blamed on it.

Why the guest never notices

Five millilitres in a long drink full of ice and tonic is invisible. Nobody holds a gin and tonic up to the light and checks the measure. In a short drink, a slightly fuller glass reads as generosity, and generosity is pleasant. There is no complaint, no returned drink and no moment where anyone has to decide anything.

That is why over pour survives in well run bars. Every other loss creates a moment: a broken bottle is swept up, a comp is asked for, a missing drink is queried. Over pour creates none. The only place it shows up is the count.

Why the stock count does notice

The count compares what the bar should have used with what it did use. If the till says 350 gin and tonics were sold at 25 millilitres each, the bar should have used 8.75 litres of gin. If the count shows 10.5 litres gone, 1.75 litres is unaccounted for. That is the variance, and at five millilitres a measure, 1.75 litres is exactly what over pour on 350 drinks would leave behind.

So the count is how over pour becomes visible. Our guide to bar inventory management sets out a weekly count that fits between last orders and lock up. Without a count, over pour is a suspicion. With one, it is a figure per product.

Three ways to measure over pour

There are three ways to put a number on over pour, and a bar that takes it seriously uses all three.

  1. The pour test: each bartender pours measures of water into a measuring cylinder, and you read the result.
  2. Variance by product: the count against what the recipes say was used, product by product.
  3. Watching: a manager who stands at the end of the bar for an hour on a busy night learns more than a spreadsheet.

The pour test names the hand. The variance names the product. Watching tells you the conditions, such as the rush, the rail or the glassware, that the other two cannot.

The pour test, step by step

A pour test takes ten minutes per person and needs a bottle of water with the bar's usual pourer, a jigger and a measuring cylinder.

  1. Ask each bartender to pour ten single measures exactly as they would on a busy night, into a glass, one after another.
  2. Tip each pour into the cylinder and write down the volume.
  3. Work out the average and the spread. An average of 29 millilitres for a 25 millilitre spec is four millilitres over on every drink.
  4. Repeat with a double and with whatever long drink the bar sells most.
  5. Show each bartender their own numbers privately, then repeat the test monthly.

The spread matters as much as the average. A bartender who averages 25 but ranges from 19 to 32 is giving some guests short measures and others free liquid. Consistency is the aim, and the test shows it plainly.

Reading variance by product

The count's variance, read product by product, tells you where the liquid went. For each product, compare three figures: what the recipes say the sold drinks used, what the count says was used, and the difference at cost.

Convert the difference into measures. In the gin example, 1.75 litres over on 350 drinks is five millilitres a drink. A difference that matches a steady extra measure on every drink of that product looks like over pour. A difference that does not fit any pattern of drinks looks like something else, and the next section covers how to tell.

Keep the results by product, week by week. One bad week can be a miscount. Three in a row on the same product is a habit.

Telling over pour from other losses

The variance mixes every loss together, so separating them needs some detective work. The patterns help.

  • Over pour grows with sales. A busy week shows more of it, and products poured often show the most.
  • Breakage arrives in steps. One bottle's worth gone in a week, nothing the next.
  • Unrecorded giveaways cluster around people and times: closing time, a manager's shift, a regulars' night.
  • Unrung drinks show up as variance on products that are easy to pour without a ticket, often beer and house spirits, on nights when the takings look light for how busy the bar was.

Record what you can, so it leaves the unexplained gap. Breakage, comps and staff drinks each have a place on the record. Whatever remains, and grows with sales, is the over pour.

Fix one: the right tools

The quickest fix is to make the right measure the easy one. A jigger, a measured pourer, an optic on the back bar or a line on the glass makes the spec the default rather than an act of willpower.

Jiggers suit cocktail bars where every drink is built by hand. Measured pourers and optics suit high volume bars where speed matters most. Lined glassware suits wine by the glass and spirits served neat. Choose by the drink and the pace, not by habit. The tools cost a fraction of a single week's over pour in the example.

Fix two: the spec and the habit

Tools only work when everyone pours to the same spec. Write each drink's measure on a spec card, keep the cards behind the bar, and train every new starter on them before their first busy shift.

Then build the habit. Repeat the pour test monthly. Recheck anyone whose numbers drift. Make measuring a point of pride, the way a kitchen is proud of consistent plates, rather than a sign that someone is not trusted. Our guide to staff training that fits around service covers how to do this without pulling people off the floor.

Fix three: the count and the conversation

The third fix is to close the loop. Share each week's variance by product with the bar team, without blame and without names. "We were 1.75 litres over on house gin, about five millilitres a drink" is a fact everyone can act on. "Someone is over pouring" is an accusation nobody can.

When the team can see the numbers, the numbers improve. They know which products to watch, they see the effect of the new pourers, and they notice when a busy weekend pushed things back up. The count stops being something done to the bar and becomes something the bar uses.

Free pour vs jigger, by cost

The free pour vs jigger argument is mostly argued on style and speed. On cost, it is simpler. A skilled free pourer can be accurate, but accuracy has to be proven by testing and kept by retesting. A jigger is accurate by default, as long as it is filled to the line and not overfilled.

The practical answer for most bars is to measure the house spirits, which carry the volume, and to allow free pouring where it is part of the bar's style only for staff who pass the pour test regularly. Whatever the policy, the variance by product will show whether it is working.

Does measuring slow service?

The fear is that measuring every drink slows the bar down on the busiest nights. In practice, a well set up bar measures quickly: the jigger lives in the hand, the pourers are set to the house measure, and the most popular drinks are prepared in ways that cut steps.

Batched cocktails, prepared and measured before service, are both faster and more accurate than building each drink to order. Measured pourers are faster than counting. A bar that is slow when it measures is usually slow because of its layout, not because of the jigger. Our guide to bar operations covers speed and stock in the last busy hour.

Generosity without liquid

Part of over pour is a wish to look generous. Guests like a full looking glass, and bartenders like guests who feel well treated. There are cheaper ways to give that feeling than five millilitres of spirit.

A glass sized to the drink, so the measure fills it properly. Generous ice. A considered garnish. A water glass topped up without asking. A bartender who remembers a name. Each of these reads as hospitality and costs a fraction of the liquid it replaces.

When a generous pour is a decision

Some venues decide to pour more than the standard measure as part of their offer, such as a large house measure that is part of the brand. That is a pricing decision, not over pour, and it is perfectly sound if it is costed. Write the measure into the spec, cost it, price it and count against it.

The problem is never generosity. It is generosity that nobody decided, costed or recorded, which turns up only as a gap in the count.

Over pour and the rest of the bar's waste

Over pour is one part of a wider picture. To reduce bar waste as a whole, look at spillage, breakage, flat kegs, out of date mixers, garnishes cut and thrown away, and wine left open too long. Each has its own fix, and over pour is usually the largest. Our waste audit for a single service works for the bar as well as the kitchen.

A weekly routine

  1. Count the bar at the same point each week.
  2. Compare each product's counted use with what the recipes say was used.
  3. Convert the biggest differences into millilitres per drink.
  4. Check the records for breakage, comps and staff drinks.
  5. Share the top three products with the team.
  6. Pour test anyone whose station shows a pattern, and recheck the following week.

Common mistakes

  • Valuing over pour at selling price, which makes the figure easy to dismiss.
  • Treating every variance as over pour, when part of it is breakage or giveaways nobody recorded.
  • Blaming individuals before measuring.
  • Buying jiggers and never pour testing again.
  • Ignoring wine by the glass and draught because spirits are easier to measure.
  • Forgetting that a short pour is no fix: it is a different problem with guests and, in many places, with the law.

Where ServeAround fits

Over pour shows up as a difference between what the drinks should have used and what the shelf says they did, product by product. ServeAround gives you both halves once the bar's drinks have recipes and the bar is counted.

With inventory, part of Order Fulfillment, each drink's recipe holds its measure, such as 25 millilitres of the house gin taken from a bottle counted in litres, so an open bottle is counted as part of a litre. Every drink rung through the point of sale takes its recipe off the expected stock. A count of the bar, as one storage area or with the rest of the stock, is set against what was expected, with the variance item by item and at cost.

The inventory report in Analytics has a panel for exactly this question, Usage against the recipes: item by item, what the counts say was used against what the recipes say the orders used, with the difference at cost, and anything used faster than the recipes account for marked in red. That names the product, and the pour test names the hand. Waste can be recorded with a reason, such as dropped or broken, so a smashed bottle does not pass for a heavy pour.

It does not weigh bottles or connect to measured pourers, and the difference is by product and by count, not by bartender. If you would like to see it on your own bar, get in touch.

Questions operators ask

How much over pour is normal?

There is no standard figure. Some variance between theoretical and counted use is normal in any bar, because nobody pours to the millilitre. Measure your own by product for several weeks, then work to make it smaller and steadier. A gap that grows with sales and sits on the house spirits is the one to tackle first.

Is over pour the same as pour cost?

No. Pour cost is the cost of what is in a drink as a share of its price. Over pour is extra liquid beyond the spec. Over pour raises the actual pour cost above the theoretical one, which is how it shows up in the numbers.

Should we use jiggers or measured pourers?

Choose by the drink and the pace. Jiggers suit drinks built by hand, measured pourers and optics suit high volume house spirits, and lines on the glass suit wine by the glass. Most bars use more than one.

Can a free pourer be accurate?

Yes, with practice and regular testing. The risk is drift: accuracy fades under pressure and over time. If free pouring is part of your style, pour test every bartender monthly and watch the variance on the house spirits.

How do we measure over pour on wine by the glass?

Pour test with water into the wine glasses you use, and compare the glasses each bottle actually yields with the number it should. Lined glassware or a measure removes most of the problem.

Is under pouring the answer?

No. A short measure shortchanges the guest, and in many places it breaks the law. In the UK, for example, gin, rum, vodka and whisky must be sold in set measures under the Weights and Measures (Intoxicating Liquor) Order 1988. The aim is the spec, poured consistently.

How often should we pour test?

Monthly for everyone, and again the week after any change to the team, the pourers or the glassware. Retest anyone whose station shows a pattern in the variance.

Does over pour matter in a restaurant that sells little alcohol?

Less in money, but the habits are the same. Wine by the glass is where a restaurant usually over pours most, because glasses are large and pours are made by eye. A line on the glass costs very little and settles it.

The short version

Over pour is liquid that goes into a sold drink beyond its spec. Five millilitres looks like nothing on one drink and becomes several bottles a week across a busy bar, valued at cost. The guest never notices. The count always does.

Measure it three ways: the pour test for each hand, the variance for each product and an hour of watching for the conditions. Fix it with the right tools, a written spec and a weekly conversation with the team. Keep generosity as a decision that is costed and written down, not a habit that turns up only as a gap in the count.

Written byThe ServeAround team

Keep reading