Bar Beverage Yield and Variance: Kegs, Spirits and Wine

Bar yield control compares how much beverage should have been used for recorded sales with how much stock was actually used. The difference—variance—can come from standard-serve settings, over-pouring, foam, line cleaning, spillage, tastings, comps, transfers, incorrect POS buttons, counting errors, supplier shortages, waste or theft.

The objective is not to accuse staff from one number. It is to measure consistently, record legitimate movement and investigate material patterns while maintaining responsible service, food safety, fair workplace practice and licence compliance.

Define the units first

Choose one base unit for each product, usually millilitres or litres.

• A keg is recorded in litres and converted to millilitres if serves use millilitres.
• A spirit bottle is recorded by bottle size and millilitres.
• Wine is recorded by bottle size and approved serve size.
• Packaged beer, cider or premix may be counted as units.

Use the product’s actual labelled or supplier specification. Do not assume every wine bottle, spirit bottle or keg uses the same nominal size.

Record standard serves from approved recipes and POS products. A licence or responsible-service rule may affect service, but this article does not prescribe one universal pour size.

Calculate theoretical yield

For a full package:

Theoretical serves = package volume ÷ standard serve volume

Examples:

• 700 mL spirit bottle ÷ 30 mL standard recipe measure = 23.33 theoretical serves.
• 750 mL wine bottle ÷ 150 mL serve = 5 theoretical serves.
• 50 L keg = 50,000 mL. At 425 mL per recorded serve, nominal theoretical yield is 117.65 serves.

These are mathematical yields, not guaranteed saleable yields. Keg line loss, foam, ullage, wine sampling, bottle residue, measured recipes and legitimate waste can reduce actual saleable output. Define expected operating allowances from reliable venue evidence, equipment and supplier information rather than hiding all loss inside a generous target.

Calculate actual usage

For a stock period:

Actual usage = opening stock + purchases + transfers in − closing stock − transfers out

Use the same units and valuation basis.

For example:

• opening vodka: 4,200 mL;
• purchases: 2,100 mL;
• transfers in: 0;
• closing stock: 2,900 mL;
• transfers out: 0.

Actual usage is 3,400 mL.

If recorded sales and approved movements imply theoretical usage of 3,180 mL, the usage variance is:

3,400 mL actual − 3,180 mL theoretical = 220 mL unfavourable variance

Variance percentage can be stated against theoretical usage:

220 ÷ 3,180 × 100 = 6.9% unfavourable

State the formula used. Some systems reverse the sign or use sales value; inconsistent definitions create false arguments.

Build theoretical usage from sales

Map every POS item and modifier to its recipe quantity.

For spirits:

• single and double pours;
• cocktails;
• tasting flights;
• staff training or approved sampling;
• bottle service;
• complimentary drinks; and
• voids and remakes.

For wine:

• glass sizes;
• bottle sales;
• tasting pours;
• cooking transfers where used;
• functions and packages;
• corked or oxidised product; and
• staff or customer samples.

For kegs:

• each glass size;
• jugs;
• tasting paddles;
• function packages;
• line cleaning and changeover;
• foam and spill records; and
• tap or product changes.

Use the recipe version active during the period. If the POS says 150 mL but staff were trained to pour 180 mL, the variance report exposes a system mismatch rather than proving misconduct.

Count stock consistently

Set a count time when movement is controlled—often after close or before opening. Freeze or record deliveries, transfers, staff drinks, functions and wastage during the count.

Sealed bottles and packaged stock

Count units by product and size. Separate damaged, held or customer-owned stock.

Open spirit bottles

Use a consistent approved method:

• calibrated scale and tare data;
• reliable measuring device where safe and suitable;
• system-specific bottle gauge; or
• another method validated for the venue.

Visual “tenths of a bottle” estimates can be quick but produce large errors across many bottles. If used, document the limitation and keep the same trained counters.

Open wine

Measure or weigh consistently, considering container weight and product. Record wines held under preservation systems and bottles allocated to functions.

Kegs

Use suitable cellar or dispensing-system information, a calibrated keg scale, or another validated method. A tap counter may measure dispensed flow but still needs reconciliation to keg changes, cleaning and stock.

Do not lift heavy kegs unsafely to estimate content. Follow manual-handling and cellar procedures.

Record every legitimate movement

Create simple reason codes:

• spill;
• foam or line issue;
• line cleaning;
• incorrect pour remade;
• customer complaint replacement;
• tasting or quality check;
• authorised complimentary drink;
• staff training;
• kitchen transfer;
• function transfer;
• broken package;
• supplier shortage or credit;
• theft or unexplained loss under investigation; and
• stock-count correction.

Record product, quantity, reason, time, person and manager approval where appropriate. “Wastage” without a reason is not useful.

Keep responsible-service decisions separate from pressure to protect yield. Staff must be able to stop service, discard an unsafe or contaminated drink and correct an error without being told to hide it.

Reconcile by product, not only by category

Total bar cost can look acceptable while one product has repeated variance.

Review:

• volume variance;
• cost variance;
• sales-value exposure;
• frequency;
• shift, station or dispenser;
• serve type;
• new staff or training;
• maintenance and line condition;
• promotion or function;
• supplier delivery; and
• count method.

Use priority based on value, safety, pattern and controllability. A one-off 50 mL count difference is not the same as repeated full-bottle loss.

Set tolerances from measured operations

A tolerance is an investigation trigger, not permission to waste stock. Build it from:

• accuracy of the count method;
• package and recipe units;
• documented line cleaning, foam and quality checks;
• equipment performance;
• sales volume;
• product value;
• normal timing differences; and
• comparable historical periods after known errors are corrected.

Use tighter review for high-value or high-risk products where the measurement supports it. Avoid one percentage across kegs, wine, spirits and packaged stock. A small-volume product can show a large percentage from one count error, while a high-volume product can hide a material dollar loss inside a small percentage.

Set two levels where useful:

• a routine review trigger for checking counts and records; and
• a material escalation trigger for manager investigation.

Document who approves tolerances and when they are reviewed. Do not quietly widen them after an unfavourable period.

Separate quantity, cost and sales variance

Quantity variance asks how much product is unexplained. Cost variance applies a consistent product cost to that quantity. Potential sales value estimates what recorded sales might have been at the relevant mix and prices, but it is not proven lost revenue.

Report the three separately. Calling every unexplained millilitre “lost retail sales” overstates the evidence and can lead to unfair employment decisions.

Worked keg example

A venue opens the week with 22 L of a beer across connected and spare stock, receives a 50 L keg, and closes with 18 L. Actual usage is 54 L.

Recorded sales and standard recipes equal 51.5 L. Approved line cleaning and documented foam from a coupler fault equal 1.2 L. Adjusted expected usage is 52.7 L.

Unexplained unfavourable variance is 1.3 L.

The manager checks the keg count method, POS mapping, glass size, changeover record and fault period. They find one jug button mapped to the old 1,000 mL recipe even though the current jug is 1,140 mL. Correcting the recipe explains part of the gap. The remaining difference is monitored after the coupler repair.

The investigation fixes data and equipment before considering individual performance.

Worked wine example

For a 750 mL wine with a 150 mL standard serve, the POS records 40 glasses and six bottles sold.

Theoretical usage:

• glasses: 40 × 150 mL = 6,000 mL;
• bottles: 6 × 750 mL = 4,500 mL;
• total = 10,500 mL.

Actual usage from stock movement is 11,250 mL. The waste log records one corked 750 mL bottle approved for disposal. After that legitimate movement, the unexplained variance is zero.

Without the waste record, the report would wrongly suggest five missing glasses.

Investigate in a fair sequence

1. Confirm opening, purchases, transfers and closing count.
2. Check product size and unit conversions.
3. Check POS recipe mapping and modifiers.
4. Verify voids, comps, refunds and function packages.
5. Review waste, line cleaning and quality records.
6. Check equipment calibration, leaks and maintenance.
7. Observe actual pouring and glassware.
8. Compare shifts or periods.
9. Speak privately with relevant staff and invite information.
10. Escalate unresolved material patterns through the appropriate operational, employment or security process.

Do not publicly post individual variance rankings or deduct unexplained stock from wages without qualified legal and payroll advice. Workplace surveillance and searches may also be regulated.

Improve the control system

Possible actions:

• calibrate or replace measures;
• standardise glassware;
• update POS recipes;
• train and observe pouring;
• repair couplers, taps, lines or refrigeration;
• adjust keg-change and line-clean records;
• separate function and kitchen transfers;
• require authorisation for comps and tastings;
• improve receiving checks;
• increase count frequency for selected products; and
• review supplier pack or delivery discrepancies.

Verify the action over comparable periods. Do not declare success after one quiet shift.

Connect variance to costing and ordering

Actual usable yield affects product cost and par levels. Feed reviewed data into How to Calculate Food Cost Percentage, How to Reduce Stock Waste and the supplier review plan.

Do not automatically increase recipe cost using an unexplained variance. First determine whether the issue is data, equipment, process, legitimate loss or another cause.

Common mistakes

• Using different units for stock and recipes.
• Assuming every bottle or keg is the same size.
• Treating mathematical yield as guaranteed saleable yield.
• Counting while stock is moving.
• Estimating open bottles inconsistently.
• Omitting comps, tastings, line cleaning and transfers.
• Using outdated POS recipes.
• Reviewing only total beverage cost.
• Blaming staff before checking data and equipment.
• Hiding responsible-service or quality disposals to protect a target.
• Making wage deductions without advice.
• Changing the allowance to make variance disappear.

Beverage yield and variance checklist

• ☐ Define base units and approved standard serves.
• ☐ Map every POS item and modifier to a recipe.
• ☐ Confirm package sizes and supplier specifications.
• ☐ Set a controlled stock-count time and method.
• ☐ Record opening, purchases, transfers and closing stock.
• ☐ Record spills, foam, cleaning, comps, tastings and breakage by reason.
• ☐ Calculate theoretical and actual usage consistently.
• ☐ State variance formula and sign convention.
• ☐ Review by product, shift, dispenser and serve type.
• ☐ Check count, unit, POS and record errors first.
• ☐ Inspect equipment and observe actual service.
• ☐ Investigate people issues privately and fairly.
• ☐ Verify corrective actions across comparable periods.
• ☐ Confirm liquor, WHS, employment, tax and record requirements.

Take the next useful action

Choose one high-volume keg, spirit and wine. Confirm package size, every POS serve mapping, current stock-count method and legitimate movement codes. Run one controlled weekly reconciliation before expanding the system.

Use the Food Cost Percentage Calculator for cost checks. The Bar & Pub Operations Pack provides editable bar controls.

General-information limitation: General operational information only. Confirm responsible-service, liquor-licence, workplace, surveillance, wage-deduction, tax, food-safety, manual-handling and record requirements with the relevant state or territory authority and qualified advisers. Yield calculations do not prove theft or compliance.

References

• Record keeping for business, Australian Taxation Office. Accessed 25 August 2026.
• Model Code of Practice: How to manage work health and safety risks, Safe Work Australia. Accessed 25 August 2026.
• Accommodation and food services industry, Australian Government. Accessed 25 August 2026.

Source review date: 25 August 2026