A New York hotel GM can quote occupancy to a tenth of a point and still run the lobby bar on a blended pour-cost number nobody has rebuilt since opening. Restaurant groups do the same: one monthly percentage, three outlets, no specs attached. Pour cost is the number that tells you whether the bar is a business. It is also the number most hospitality teams track loosely, calculate inconsistently, and “fix” by raising prices. Here is the formula done properly, the difference between pour cost and variance, and seven levers that recover margin before the list reprints.
The calculation, done right
Pour cost formula
Pour cost % = (Cost of goods used ÷ Beverage revenue) × 100
Used, not purchased: Opening inventory + Purchases − Closing inventory.
Run it monthly, minimum. Weekly if the room is a lobby bar, a banquet-heavy hotel, or anything doing serious volume.
Two mistakes show up in almost every NYC property audit. First: using purchases instead of usage. A big opening order or a banquet pre-pull makes the month look terrible; a month you live off the cellar makes it look brilliant. Neither number is the program. Second: blending the hotel. The lobby bar, the restaurant, in-room dining, and banquets are four businesses. One blended percentage hides a 31% banquet pour inside a respectable lobby number.
Healthy benchmarks: cocktail programs typically target 18–24%; wine runs higher, beer lower. A blended beverage pour cost in the low twenties is a well-run bar. Anything drifting past 26–28% is leaking money — unless the list is deliberately spirit-heavy and priced for it. Print the target on the spec sheet, next to the recipe. A number that lives only in the controller’s workbook does not change a Saturday night pour.
What four points are worth
Hotel lobby bar, $80,000 a month in beverage revenue
At 26% pour cost: $20,800 in product
At 22%: $17,600
Difference: $3,200 a month — about $38,000 a year
That is the fee conversation. The menu, press, and covers it draws sit on top.
Variance vs pour cost — two numbers, two jobs
Pour cost is a ratio: what the product you used cost against what you sold. Variance is a count: what left the shelf against what the POS says was sold. A bar can hold a respectable 22% while a case of gin walks, because pour cost only sees what was rung. If a drink was over-poured, spilled, comped off the record, or never entered, pour cost does not flag it. It drifts half a point and nobody can say why.
The healthy variance range is 1 to 2% of inventory. Industry reporting puts many bars at 5 to 25% without knowing it. I wrote the weekly routine in The Variance Report. The short version for a GM: if pour cost is drifting and the team swears they are on spec, run variance on the top twenty SKUs before you rewrite the menu. You may have a paper problem, a free-pour problem, or a theft problem. Those are different meetings.
Seven levers that work before you raise a single price
Price is a lever. It is not the first one. The rooms that recover four points usually do it in this order.
1 · Specs — cost every recipe, to the dash
Not the top sellers. Every drink. A hotel list of 24 cocktails plus a seasonal and an NA set is forty-odd recipes if you count the banquet builds. If three of them were never costed, you do not have a pour-cost target. You have a hope. Put ingredient cost, contribution dollars, and target pour cost on the spec sheet the bartender actually uses. Re-cost when invoices move — quarterly at minimum in New York, where distributor pricing is not a still life.
This is the cheapest work in bar consulting. It is also the work most opening teams skip because the tasting was more fun than the spreadsheet.
2 · Batching — kill free-pour variance on the house drinks
Free-pour is the biggest silent killer on a busy stick. A bartender pouring 60 ml into a 45 ml spec gives away a bottle every 30 drinks. Batch the stirred list and the multi-ingredient bases. Ticket time drops. The hundredth Negroni tastes like the first. Variance on those SKUs collapses because the bottle is no longer the unit of improvisation. The labor argument is in Your Labor Problem Is a Prep Problem; the draft version of the same idea is in Cocktails on Draft.
3 · Prep lab — stop doing commissary in a kitchen corner
Citrus husks become oleo and cordials. Trim becomes syrups. A hotel that juices to order on four outlets is buying the same lime three times and dumping half of it. Centralized prep is waste reduction with a flavor upside — and it is the only way banquet volume and à la carte stay on the same costing. When throughput no longer fits behind one stick, that is a prep-lab design engagement, not a new Vitamix.
4 · Menu engineering — sell the margin, not the showpiece
Every list has a workhorse at 15% pour cost and a showpiece at 30%. Placement, naming, and server language decide which one sells. Plot contribution dollars, not pour-cost percentage: a $22 martini at 30% can still beat a $16 daiquiri at 20%. The method is in Your POS Knows Which Cocktails to Cut. Cut the dogs. Protect the stars. Rework the plowhorses before you reprint a higher price.
5 · Theft, waste, and the paper you do not run
Four buckets, in the order you should check them: counting errors, over-pours and spills, invoices and POS mapping, then unrecorded product — comps, staff drinks, theft. Everyone jumps to the last bucket. Most of the money is in the first three. Pars, a named owner for the count, and a weekly variance on the premium SKUs routinely find 1–2 points on their own. A hotel banquet bar that does not date and FIFO its bottled builds is not “hospitality.” It is shrinkage with a logo.
6 · Pricing — last, and never as a substitute for costing
If the specs are costed, the mix is engineered, and variance is inside 2%, then price is fair game. Tier the list: entry, core, statement. Do not take the Manhattan up two dollars because the controller wants a point — guests have a reference price for classics and they will punish you. Take margin on drinks without a street price. The pricing logic is in Guests Are Drinking Less. Price for It. Separate conversation: supplier pricing set at opening rarely gets revisited. Volume changed. Your deal should have too.
7 · Retrain the pour, then leave a book
Jiggers, spec adherence, and a culture where precision is craft rather than surveillance. The team that measures is the team that hits margin. Training without a spec book dies at the next resignation. Training with a book, a par sheet, and a named owner is how a hotel survives turnover. That is the handoff in a beverage consulting engagement: the house owns the documents.
When it is a program problem, not a discipline problem
If the specs were never costed, the prep load does not match the ambition, and the menu was designed for Instagram instead of the P&L, no amount of jigger discipline saves it. Hotel F&B that inherited an opening list and a head bartender who left is the usual case. That is a program refresh: audit, re-costed menu, retrained team — six to ten weeks. Properties still inside a year of doors should read Opening a Hotel Bar and the hotel bar consultant landing. Groups that already have a competent internal team usually want the advisory retainer, not another rebuild.
What that work costs, in honest 2026 ranges, is in How Much Does a Bar Consultant Cost? The ROI math is the four-point example above. The audit that starts it is this article, run against your numbers, across three shifts of live service.
Working with Drinks By Neat
Pour-cost recovery is bar consulting and beverage consulting — a refresh when the list is the problem, a prep-lab when the volume is, a retainer when the team is good and the numbers still drift. Hotel properties: hotel bar consultant. If the percentage has been moving and nobody can say why, fifteen minutes usually locates it. Or write.
Questions GMs actually ask
What is a healthy pour cost for a hotel or restaurant cocktail program?
Cocktail programs typically target 18–24%. Wine runs higher, beer lower. A blended beverage pour cost in the low twenties is a well-run room. Anything drifting past 26–28% is leaking money — unless the list is deliberately spirit-heavy and priced for it.
What is the difference between pour cost and variance?
Pour cost is a ratio: cost of goods used against beverage revenue. Variance is a count: what left the shelf against what the POS sold. Pour cost can look healthy while product walks. Run both. The weekly routine is in The Variance Report.
How often should a hotel bar calculate pour cost?
Monthly is the minimum. Weekly if volume is high — lobby bars, banquet-heavy properties, and any room doing six figures a month in beverage. A monthly number hides a bad Saturday inside 30 days of noise.
Can you reduce pour cost without raising menu prices?
Usually yes. Specs, batching, prep-lab waste recovery, menu mix, and theft/waste controls move the number before price does. Price is a lever, not the first one.
When is pour cost a program problem instead of a discipline problem?
When specs were never costed, the prep load does not match the ambition, and the menu was designed for Instagram instead of the P&L. Jigger discipline cannot save an uncosted list. That is a beverage consulting engagement — refresh or retainer — not a memo.
Consistency at volume isn’t a flex. It’s a margin strategy.