Bar Stock Control: The Tightest Ship in the House
The bar is where hospitality margins go to be tested — pour discipline, nightly bottle counts, empties and deposits, and the variance math that makes the tightest ship in the house.
Every hospitality operator eventually learns the same lesson: the kitchen loses money in percentages, the bar loses it in bottles. Liquor is dense value in portable form, sold in pours nobody weighs, in an environment designed to be convivial rather than supervised. That combination makes the bar the single most leak-prone point in the house — and, paradoxically, the easiest to lock down, because unlike a kitchen, a bar's stock is countable to the shot.
The arithmetic that runs a bar
A bottle of spirits is a fixed number of portions: a 750ml bottle at 25ml pours is 30 shots — 28.5 with an honest spillage allowance. Every bottle issued to the bar should therefore produce a knowable revenue, and the nightly reconciliation is three numbers:
- Consumption by count: opening bottle levels + issues − closing levels, in tenths of a bottle (or by weight — a bar scale reads a bottle's remaining pours in seconds).
- Sales by POS: what the till says was sold, item by item — tots, cocktails (via their recipe cards), bottles served.
- The variance: consumption valued at selling price minus recorded sales. That gap, per night, per bar, per bartender shift, is the number the bar answers to.
Percentage points are bottles
A busy bar turning KES 1.5M a month at a 5% variance is losing KES 75,000 — roughly two bottles of spirits a night walking, over-pouring, or selling off-book. Bars that post the nightly variance where staff see it routinely land under 2% within a month. The math being visible is the control.
Where bars leak, specifically
| Leak | How it works | The counter |
|---|---|---|
| Over-pouring | Generous free-pours, friends' doubles | Measured pourers or jiggers as house standard; variance by shift |
| Own-bottle sales | Bartender's bottle, house's customers, pocketed revenue | Nightly counts + sealed-stock rotation; consumption that exceeds issues is the tell |
| Unrecorded comps | "On the house" without a void trail | Comp button with reason and approver on the POS |
| Cocktail short-pours | Recipe says 50ml, glass gets 35ml, difference accumulates | Cocktail recipes on the POS; periodic test rounds |
| Empties fraud | Refilled premium bottles, recycled seals | Empties broken/marked at count; premium stock sealed and serialized |
| Store-to-bar shrinkage | Bottles vanish between issue and shelf | Issues signed at handover; bar holds its own location stock |
Par levels, issues, and the beverage supply chain
- Each bar carries a par per SKU — enough for the busiest expected night plus margin — and is topped back to par by daily issue, signed at handover.
- Kegs and draught lines get their own math: a 50L keg is a fixed number of glasses; line cleaning and foam loss are recorded allowances, not explanations.
- Empties, crates, and deposits are stock too — beverage suppliers' deposit systems leak real money when nobody counts returns against deliveries.
- Wine cellars run on bin cards with bottle-level counts monthly — low turnover, high value, and the easiest place for a bottle to be "breakage".
The nightly close, bar edition
Fifteen minutes at last call
- Bottle levels counted (or weighed) and entered — tenths matter.
- POS sales pulled by category; comps and voids reviewed with reasons.
- Variance computed and posted — by bar, by shift.
- Cash and M-Pesa reconciled with the same discipline as any till.
- Tomorrow's par top-up requisition raised from tonight's counts.
None of this needs exotic hardware — a bar scale, measured pourers, and a system where issues, recipes, and the POS meet cover a full-service operation. What it needs is the nightly ritual, unskipped, including the nights the owner is buying.
Run the tightest ship in the house
Nightly bottle math, pour-cost by outlet, variance by shift, and par-driven issues — bar control that pays for itself in bottles.
See bar control in AWRAFrequently asked questions
Free-pour or measured pour — does it really matter?
A skilled free-pourer is accurate to ±10%; a jigger is accurate to ±2%. On a bar's volume, that difference is the whole variance target. Measured pouring also protects good bartenders — their shift variance proves their honesty in a way style never can.
How do we handle happy hour and promotional pricing in the variance math?
Value consumption at the price actually charged in each period — the POS knows which price band every sale used. What breaks variance math is unrecorded discounting: promos through the till are fine; "special prices" negotiated at the counter are leakage wearing a smile.
Should bartenders see their own variance numbers?
Yes — per shift, posted where the team sees them. Named, visible variance changes behavior faster than cameras or searches, and it converts the honest majority into the control system: nobody wants to inherit the shift after a bad number.
What about stock for events and outside catering?
Issue event stock as its own mini-location — provisioned against the function sheet, returns counted back, variance closed per event. Event bars run the same math as fixed bars, just with a defined start and end; the [banqueting pattern](/blog/hotel-restaurant-inventory-kenya) applies bottle-for-bottle.
