How to Cost a Cocktail: Pour Cost, Explained

A cocktail is cheaper to cost than a plate of food. There's no bone-in yield loss, no overnight braise, no trim. It's a handful of liquids measured in millilitres and a garnish. And yet bars run wider margin leaks than most kitchens — because the thing you're pouring is expensive per litre, easy to over-pour, and invisible once it's in the glass.

This is how to work out pour cost — the drinks-world name for food cost percentage — from the bottle price up. We'll cost per millilitre, add the parts everyone forgets, build a full margarita from scratch, and then talk about the leak that no spreadsheet catches: what actually leaves the bottle versus what the recipe says should.

Pour cost is just food cost with a different unit

If you've costed a dish, you already know the maths. The formula doesn't change; only the ingredients do.

If a drink costs you $2.40 to make and sells for $12, your pour cost is 20%. That's a healthy number. Most bars target a drink cost percentage of 18–24% — lower than a kitchen's 28–32% because bar labour, glassware breakage, and spillage aren't in that figure and have to be earned back elsewhere. Spirit-forward drinks (a Negroni, an Old Fashioned) run at the higher end because they're almost all liquor. Anything long and mixer-heavy — a highball, a spritz — should sit near the bottom, and if it doesn't, you're either over-pouring or under-charging.

The reason bars still bleed is not the formula. It's that "cost of ingredients" for a drink is deceptively fiddly to pin down honestly.

Costing per millilitre from the bottle price

You buy spirits by the bottle. You pour them by the millilitre. The first job is turning one into the other, and it's a single division:

A 700 ml bottle of a mid-shelf tequila at $34 works out to $34 ÷ 700 = $0.0486 per ml. A standard 45 ml pour of it costs you $2.19. That's it — that's the core of every cocktail cost. Do it once per bottle on your shelf and you have a price-per-ml table you can build any drink from.

Two things trip people up here. The first is bottle size: a 750 ml US fifth, a 700 ml EU bottle and a 1 L bottle are three different divisions, and buying the litre changes your per-ml cost even when the brand is identical. Always divide by the actual volume on the label you bought, not a round number. The second is the pour standard itself. A "shot" is not a universal measure. Costing assumes a defined pour — 45 ml, 30 ml, a specific jigger — and it only holds if the bar actually pours that. Free-pouring quietly detaches your real cost from your costed cost, which is the whole problem we come back to at the end.

For reference, the US NIAAA (standard drink) defines a standard serving of distilled spirits as 1.5 fl oz — about 44 ml — of 40% ABV liquor, which is why 44–45 ml is the pour most cost models are built around. It's a sensible default to cost against even if your house pour differs.

The parts everyone forgets to cost

The spirit is the expensive bit, so it gets all the attention. The margin usually leaks somewhere else.

Garnish

A wheel of lime, a sprig of mint, a dehydrated orange slice, a Luxardo cherry, a salt rim. Individually trivial — a few cents each. Across a busy Friday, a bar going through 300 drinks with a 15-cent garnish each is spending $45 a night, $16,000 a year, on garnish nobody costed. And the fancy stuff isn't cheap: a single quality cocktail cherry can run 12–20 cents, an edible flower more. Cost the garnish. It's a line, not a rounding error.

Mixers, juice and syrups

Fresh lime juice has a yield problem exactly like a kitchen ingredient: a lime costs you 30 cents and gives roughly 25–30 ml of juice, so your juice is running around $1.10 per 100 ml — far more than the bottled stuff, and worth every cent, but only if you've costed it. House syrups (a 1:1 simple, an oleo-saccharum, a spiced honey) need to be costed as their own sub-recipe: batch the syrup, work out its cost per ml, then draw from it. Tonic and soda from a gun are near-free; premium bottled tonic at 50–70 cents a bottle is not, and in a gin and tonic the tonic can cost more than the gin.

Ice and dilution

Ice reads as free and behaves as expensive. Not the water — the water is nothing — but a good clear cube or a hand-cut block has real labour and machine cost behind it, and more importantly ice is where dilution lives. A stirred drink picks up 20–25% of its volume in water off the ice; a shaken one more. That dilution doesn't change your ingredient cost, but it changes your finished volume, which changes glassware, and it means a drink built to 90 ml of spirits leaves the shaker at 115–120 ml. If you're batching (more on that next), you have to account for the water you're not adding so the batch tastes like the single. Ice is a cost line and a recipe variable at the same time.

Batching and what it does to cost

Batching — pre-mixing a cocktail in bulk for service or an event — barely touches your ingredient cost. The same spirits, juice and syrup go in whether you make one drink or a hundred. What batching changes is everything around the ingredient: speed, consistency, and waste.

On the good side, a batched drink pours in five seconds instead of ninety, which is pure labour saving, and every glass is identical, which kills the over-pour variance that inflates cost on a made-to-order bar. On the risky side, a batch is a commitment: mix 5 litres of a spritz base for a wedding and serve 3, and you've thrown 2 litres of costed product down the sink. Batch waste is real waste, and it shows up as a pour cost spike after the event, not during it.

The dilution point matters most here. A batch stored and poured over ice still picks up dilution in the glass, so batch specs usually add water deliberately (often 20–30% of the spirit volume) to land at the right strength — and that added water has to be in your yield calculation or you'll think you have fewer servings than you do. Cost the batch as a sub-recipe with its true finished volume, divide by servings, and you get an honest per-drink number that's usually a shade lower than the à la carte version because there's less spillage.

A worked margarita

Let's cost a full drink the way you'd cost a plate — every line, current prices, nothing hand-waved. Classic spec: 50 ml tequila, 25 ml triple sec, 20 ml fresh lime, a bar-spoon of agave, a salt rim, a lime wheel. Prices are illustrative — plug in yours.

Sell that margarita at $14 and your pour cost is $3.44 ÷ $14 = 24.6%. That's at the top of the healthy band — fine for a spirit-forward signature, a touch high for a house pour. Notice what moved it: the tequila is 71% of the cost. If you dropped to a $26 bottle you'd shave the drink to about $3.05 and a 21.8% pour cost — a real decision with a real quality trade-off, which is exactly the kind of choice a cost card exists to make visible. Notice also that the garnish, salt and ice — the "free" stuff — added 9 cents, which is small here but is the line most bars simply leave off.

Flip it around and the card becomes a pricing tool. At a 22% target pour cost, a $3.44 drink wants a menu price of $3.44 ÷ 0.22 = $15.60, rounded to a clean $15 or $16. Cost first, then price to the target — never the other way round.

Overpour and shrinkage: the silent leak

Everything above gives you the theoretical pour cost — what the drink costs if the bartender pours exactly 50 ml and never spills, comps, or breaks a glass. Your actual pour cost, the one your stock count reveals at the end of the month, is always worse. The gap is the whole game.

Free-pouring is the biggest culprit. A bartender who pours "about a shot" by feel is high far more often than low — the generous hand is a real, measurable bias. An extra 10 ml on a 45 ml pour is a 22% over-pour on that ingredient. Do that on half your drinks and your tequila runs out a fifth faster than your sales say it should, and the difference is money that walked out in glasses you were paid for at the 45 ml rate.

You measure the leak the same way a kitchen measures food variance: theoretical usage (drinks sold × recipe pour) against actual usage (opening stock + deliveries − closing stock). If your POS says you sold enough margaritas to use 4 litres of tequila and your stock count says 5 litres left the shelf, that extra litre is your leak — about $48 at cost, every count, if it's consistent. Once you can see it, you close it: jiggers or measured pourers instead of free-pouring, a comp log that's actually filled in, and a weekly count on your top ten bottles rather than a quarterly panic.

What to do this week

You don't need to cost the whole back bar on day one. Start where the money is:

1. Build a price-per-ml table for your ten most-poured spirits. One division each: bottle price ÷ bottle ml.
2. Cost your five best-selling cocktails in full — spirits, juice, syrup, garnish, ice allowance. Not just the booze.
3. Work out the pour cost % on each. Anything over 25% is either priced too low or poured too heavy. Decide which.
4. Put a jigger on the well for a week and watch whether anyone notices. Your variance will.
5. Count your top ten bottles at the start and end of that week and compare theoretical to actual. That number is your leak.

Pour cost isn't a report you file. It's the difference between a bar that looks busy and a bar that's actually profitable — and on a product priced by the bottle and served by the millilitre, that difference lives in a few dozen invisible millilitres per drink. Count them, cost them, and the margin stops disappearing.

"We were selling more cocktails than ever and making less. Turned out the well was pouring 60 ml for a 45 ml drink. Nobody was stealing — everybody was just being generous."

That's the most common bar leak there is, and it never shows up on a menu or a recipe card. It shows up in the count. Cost the drink, then go and check what actually left the bottle.