Bar inventory and pour cost: counting, variance, and where the ounces go

The drawer balances every night. The card batches settle. And the month-end liquor invoice still says you bought noticeably more whiskey than you sold. That gap is what bar inventory management exists to find, and in most bars nobody is looking for it — counts happen “when things slow down”, the spreadsheet lives on a laptop nobody opens, and pour cost is a number the owner heard once at a trade show. Liquor is the highest-margin thing you sell and the easiest thing in the building to lose an ounce at a time. Here is how to count it, cost it, and trace the missing ounces to a cause, without turning Sunday night into a second full shift.

What pour cost actually tells you

Pour cost is the cost of the alcohol you used divided by the alcohol sales it generated, as a percentage. Use twenty dollars of vodka to ring a hundred dollars of vodka sales and your vodka pour cost is 20%. Most full-service bars land somewhere in the high teens to mid-twenties overall, with liquor typically running leaner than beer and wine — but the target matters less than the trend. A bar that runs 22% every week is healthy; a bar that drifts from 20% to 24% over six weeks has a leak, and the drift is the alarm. Pour cost is the bar-side sibling of food cost percentage: one number that compresses purchasing, pricing, portioning, and shrinkage into something you can watch weekly.

The formula needs real usage, not purchases. Usage for a period is opening inventory plus what you received minus closing inventory. That is why the count matters: without a real opening and closing number, “pour cost” is just your invoice total wearing a costume.

How to take bar inventory without losing your night

The count is the unglamorous half of bar inventory, and the way you structure it decides whether it takes ninety minutes or four hours.

• Count when the bar is closed — after close or before open. A count taken during service is a guess with a clipboard.
• Arrange the count sheet in shelf order, not alphabetical order. You walk the room once, top shelf to well to walk-in, and never double back.
• Count full bottles as ones and open bottles in tenths, by eye against the bottle shoulder. Tenthing feels crude and is accurate enough — consistency beats precision here.
• Two people, one counting aloud, one writing. It is close to twice as fast and catches transposed numbers on the spot.
• Same day, same time, every week. Weekly counts make variance a seven-day puzzle; monthly counts make it an unsolvable thirty-day one.
• Log breakage, spills, comps, and kitchen transfers the moment they happen, not from memory at count time.

How to calculate pour cost, category by category

Work it per category — liquor, beer, wine — and then per bottle for anything that looks off. An overall number hides too much: a lean liquor cost can mask a wine program quietly bleeding out. The math per line is the same: usage in bottles times cost per bottle gives usage cost; divide by the sales that category rang in your POS.

The table above is an illustration, not a benchmark — your numbers depend on your market, your pricing, and your pour sizes. What it shows is the shape of the work: the total looks fine at 21%, and the wine line is still worth a conversation. Maybe wine pricing has not kept up with invoice creep, maybe by-the-glass pours grew, maybe a case walked. The category view tells you where to look.

Variance: the number that finds the missing ounces

This is the step most bar inventory guides stop short of, and it is the one that actually finds money. Your count tells you what you really used. Your POS tells you what you sold — and if your drink recipes are mapped, it can tell you how many ounces of each spirit those sales should have consumed. Actual usage minus POS-expected usage is your variance, bottle by bottle. A bottle of well tequila that shows a liter gone beyond what sales explain is not a rounding error; it is overpouring, unrung drinks, an untracked comp, or a pour that went home in a coffee cup.

• Overpouring: free-pouring heavy by a quarter ounce is invisible per drink and adds up to entire bottles per week on a busy well.
• Unrung and “forgotten” drinks: rung after the rush, or never. Tight comp and void controls with named reasons keep honest habits honest.
• Kitchen transfers: the wine that went into the short ribs counts as kitchen use, not bar shrinkage — but only if someone logs the transfer.
• Breakage and spills that never hit the log, so the count eats them as mystery loss.
• Theft — rarer than managers fear and realer than they hope. Variance narrows it to a bottle and a week, which is usually enough to change behavior on its own.

Variance is also the honest way to judge whether fixes work. Retrain the free pour, then watch the well tequila line for two weeks. If the variance closes, the training held. If it does not, the problem was never technique.

Par levels: order from the count, not from vibes

Once the count is weekly, ordering gets simple. A par level is the amount of each product you want on hand at the start of the week — enough to cover expected sales plus a cushion. The order is par minus what the count says you have. No more shelf-squinting in the walk-in at 4pm, no more emergency runs for the vodka you somehow ran out of on a Friday. Pars also make prep-list logic work behind the bar: batched cocktails, juices, and syrups get built to par from the same count discipline.

Where your POS carries the load

A spreadsheet finds the leak; the POS keeps it closed. Recipe-mapped drinks deplete inventory as they ring, so expected usage updates itself. Comps and voids carry a name and a reason. Transfers to the kitchen get logged in seconds instead of remembered never. And if you run more than one room or more than one location, multi-location inventory means the same count discipline and the same pars everywhere, visible from one screen. On Novaryq, that sits in the same platform as your restaurant inventory, your sales, and your labor — one platform with per-location pricing, and the separate costs disclosed up front.

The weekly rhythm that makes it stick

Pick the night, protect the ninety minutes, and run the same loop every week: count in shelf order, enter the numbers, read three lines — overall pour cost, worst category, worst bottle — and pick one fix. Jigger the well for two weeks. Reprice the two wines that crept. Chase the missing case of pilsner. One fix a week, verified by next week’s count, beats a quarterly deep-dive that everyone dreads and nobody acts on. Six weeks in, the count stops feeling like homework and starts feeling like the scoreboard.