50 bar metrics defined, with formulas: the operator's glossary

By Jack Hawkins · Updated 19 October 2026

A bar metric is a repeated measurement that turns trade behaviour into a number you can act on, and this page defines the fifty that matter for UK venues, with calculation conventions and the decisions they support; where a band is used it is an illustrative review policy. Bookmark-level reference: every guide on this site links back here, and if your software uses a term that isn't here, it's probably repackaging one that is.

Conventions used throughout, so the definitions stay comparable: net means ex-VAT (menu ÷ 1.20); loaded cost includes employer NI, pension and holiday accrual; period-fixed means the same window every time, because a metric with a drifting window measures nothing. Bands marked "trade-estimate" are indicative practitioner ranges, not audited statistics.

Money & margin

1. Net revenue — gross takings ÷ 1.20. The only denominator worth using in the UK.
2. Gross profit (GP) — net revenue − cost of goods. The ingredient gross-profit definition before labour and overheads.
3. Gross margin % — GP ÷ net revenue. Illustrative wet-led planning band: 55–65%; calculate the actual sales-weighted result.
4. Pour cost % — cost of goods ÷ net revenue; the mirror of margin. A 55–65% GP scenario implies 35–45% pour cost on the same basis.
5. Contribution per serve — (net price − ingredient cost) per drink. The menu's true ranking column.
6. Total ingredient contribution — contribution per serve × units; contribution margin % is contribution ÷ net revenue × 100 after the variable costs included in the definition. Ranks drinks by money, not popularity.
7. Markup — profit ÷ cost. Different from margin; the ×4 folklore is a markup statement.
8. Break-even revenue — fixed costs ÷ contribution margin after all variable costs. The number under every lease argument.
9. Break-even coverage — net revenue ÷ break-even revenue. Above 110% trading; below 90% deciding.
10. Prime cost — COGS + loaded labour. Express it as a share of net revenue and model ingredients and labour together; no universal comfort line is assumed.

Drinks performance

1. Units per product (28-day) — POS units, item level, fixed window. The PMIX spine.
2. Product mix % — units ÷ total category units. The popularity axis of menu engineering.
3. Popularity hurdle — 70% ÷ number of items. Clears the star/plowhorse line.
4. Rate of sale (ROS) — units per outlet per period (week is standard). See distribution metrics.
5. Dead-list count — count of available items below a stated units-per-open-day threshold over a stated window. The cut list writes itself.
6. Category mix — net revenue share by spirit/beer/wine/soft. Drift here is the market arriving.
7. Average unit price — net revenue ÷ units. Your realised price after every discount.
8. Premium share — net revenue share from the top price tier. The premiumisation dial.
9. Zero-ABV share — net revenue share of no-alcohol serves; watch venue-level, not national noise.
10. Price ladder depth — count of distinct current menu price points within a stated category. Flat ladders cap spend per head.

Guests & covers

1. Covers — guests served, same definition all year. The dirty prerequisite of everything below.
2. Spend per head — net revenue ÷ covers. The diagnostic that revenue hides.
3. Drinks per head — units ÷ covers. Service speed's shadow metric.
4. Repeat rate — identified returning guests ÷ all identified guests × 100, where usable consented guest data exists. Honest note: most bars can't measure this cleanly.
5. Dwell time — sum of measured visit durations ÷ visits measured, with the start/end event defined. Rises with food, falls with a hurried room.
6. Table turn — seatings per table per service. Pubs: mostly a myth; restaurants: real.
7. Party-size mix — bookings in each party-size band ÷ all bookings; average party size is covers ÷ bookings. Two-tops and six-tops are different businesses wearing one table.

Stock & loss

1. Actual usage — opening + purchases − closing, at cost. What physically left the shelf.
2. Theoretical usage — POS units × spec volumes. What the recipes say left the shelf.
3. Stock variance % — (actual depletion − theoretical usage) ÷ theoretical usage × 100; unexplained variance also subtracts logged explained depletion. Under 5% disciplined (trade-estimate).
4. Shrinkage £ — sum of product quantity differences × consistent unit cost; label total versus unexplained loss. The number, not the adjective. See shrinkage guide.
5. Keg yield — saleable pints ÷ nominal pints. Measure per product and dispense setup; a chosen investigation target is not a national observed average.
6. Stock cover (days) — closing stock value ÷ daily usage cost. Interpret cover against supplier lead time, minimum order, seasonality and desired availability.
7. Waste rate — logged waste ÷ actual usage × 100, in matching units or values by category. The clipboard number, not the till number.
8. Delivery variance — invoiced quantity minus physically received quantity, in consistent product units. The paperwork leak with a name.

Labour & productivity

1. Labour cost % — loaded hours cost ÷ net revenue. 25-35% wet-led (trade-estimate); see labour guide.
2. Sales per labour hour — net revenue ÷ paid hours. The productivity mirror of the ratio.
3. Bar covers per bartender-hour — covers ÷ bar-hours. Venue-specific; compare to yourself.
4. Coverage mismatch £ — excess staffed hours × marginal loaded hourly cost; report estimated lost peak contribution separately rather than adding hours and money. The rota's true error.
5. Overtime share — overtime hours ÷ total. Rising quietly is the first warning the rota is fiction.

Sales quality & leakage

1. Discount value % — discounts ÷ gross sales, with reason codes. Above ~4% needs a policy defence.
2. Void rate — voided units ÷ sold, per person per hour. A recording check that is not proof of theft.
3. Comp rate — comps ÷ sales, by reason. Generosity with a paper trail is fine; the other kind is variance.
4. Round time — sum of order-to-delivery durations ÷ completed rounds measured, per section. The second-drink killer, measured.
5. Card mix — receipts by payment type ÷ all receipts × 100, using a consistent value or transaction-count basis. Drives the acquirer line on the overheads page.
6. Processing cost % — merchant fees ÷ net revenue. Use actual merchant charges; a rate quoted on gross payment value is not the same percentage of net revenue.

Trading shape

1. Daypart curve — net revenue by hour, weekdays overlaid. The rota's source document.
2. Weekend share — Fri-Sun net ÷ week total. Above ~60% means three nights drive the year.
3. Same-weekday trend — this Monday vs last four Mondays. Still affected by weather, events, opening hours and availability.
4. Forecast error % — (actual − forecast) ÷ forecast, logged weekly. The meta-metric; see forecasting.

How to use this list

Nobody instruments fifty metrics. Pick a small working set, such as five related metrics, into the weekly review workspace, and let the other forty-five be lookup terms when a supplier, consultant or dashboard waves them at you. The instrument is the weekly five with fixed definitions; this page is the dictionary the instrument's language is checked against.

Where a metric has a full guide, the row links to it; where a number has a band, it's labelled trade-estimate, per our data standards. A glossary that quietly passes estimates off as standards is how the industry argues with spreadsheets forever.

Use the glossary as a data dictionary

Percentages here are ratios multiplied by 100; a percentage is undefined when its denominator is zero. Use unavailable rather than zero for missing inputs, because a zero cost creates an apparent high margin. Document whether contribution deducts ingredients only or all variable costs. Keep forecast error against forecast distinct from percentage error against actual: both are useful but they answer different questions. A source export, product mapping, recipe version and period cutoff should travel with the report. That small dictionary lets a venue compare weeks, a group compare sites and a brand compare accounts without accidentally changing the measurement halfway through the story.

FAQ

What are the most important KPIs for a bar? Five carry most of the load: net revenue with covers and spend per head, blended gross margin, stock variance, contribution top/bottom movers, and discount/comp value. Everything else on this page is either those five, or context for them.

What is the difference between margin and pour cost? Mirror images: gross margin + pour cost = 100% when both are computed on net revenue. A 22% pour cost is a 78% margin. The bar trade speaks pour cost; accountants speak margin; both must use ex-VAT numbers in the UK.

What is rate of sale (ROS) in a bar? Units of a product sold per outlet per period, usually weekly: 40 bottles across 20 outlets is ROS 2.0. It's the core brand-side metric in distribution analysis and a useful venue-side lens on your own dead list.

How do I benchmark my bar's metrics? Against your own history first (same-weekday, rolling windows), then format-matched peer ranges where honestly available. Published spend-per-head and labour averages need format, sample and definition checks; treat any page quoting one without a method as marketing.

Method and assumptions

Bands labelled trade-estimate are indicative practitioner ranges, not audited statistics; formulas follow the Kasavana lineage for mix analysis and standard UK stock-control conventions. VAT conventions per HMRC rules reviewed October 2026. Full standards: data standards.

Last reviewed 19 October 2026.