Written by: JJ Tan, Founder, Jelly
Key Takeaways
- Stocktake variance analysis compares physical counts with theoretical stock to reveal waste, over-portioning, unrecorded comps or data errors that erode gross margin.
- Even a 5% variance on a £500,000 annual turnover site can cost £25,000 in lost margin every year, so early detection matters.
- Standard formulas calculate actual usage (opening stock + purchases − closing stock) and variance percentage against theoretical usage derived from recipes and EPOS sales.
- Common causes include portioning inconsistency, unlogged waste, POS mapping errors and supplier discrepancies; theft is rarely the primary driver when investigated properly.
- See how Jelly automates variance tracking to protect your margins from day one.
The Problem: Margin Erosion from Unexplained Stock Losses
Small stock variances of a few percent of theoretical COGS sit within normal noise for many UK hospitality venues. Once variance climbs above 5%, the financial damage becomes material. On a site turning over £500,000 annually, a 5% food cost variance translates to around £25,000 in lost margin every year, which disappears without a clear line on the P&L.
A small variance on wet sales can cost a typical UK pub several thousand pounds annually, equating to roughly one lost gross profit point. A bar running £20,000 in weekly wet sales at a recurring 2% liquor variance loses £20,800 per year through small, repeating issues such as over-pours or unclaimed credits.
The manual process of investigating these losses consumes considerable admin time, with hours spent reconciling spreadsheets rather than running the business. That burden compounds as sites and supplier relationships multiply, and each extra venue adds another layer of stock, invoices and counts to manage.
See how Jelly eliminates that admin burden from day one.
How to Calculate Stocktake Variance
UK hospitality operators use a simple set of formulas to calculate stocktake variance. Actual consumption equals opening stock plus purchases minus closing stock, and theoretical consumption comes from recipe specifications multiplied by portions sold. Variance is the difference between the two, expressed as a percentage of theoretical usage.
- Actual Usage = Opening Stock + Purchases − Closing Stock
- Variance = Actual Usage − Theoretical Usage
- Variance % = (Variance ÷ Theoretical Usage) × 100
To see how these formulas work in practice, here is a worked example using beef mince over one month.
Worked £ example for beef mince over one month:
- Opening stock: 20 kg @ £8/kg = £160
- Purchases: 50 kg @ £8/kg = £400
- Closing stock: 18 kg @ £8/kg = £144
- Actual usage: £416
- Theoretical usage (recipe × portions sold): £380
- Variance: £36 | Variance %: 9.5%
Even smaller percentages compound quickly: a 2% variance on £50,000 monthly spend costs £12,000 per year. The table below shows how a monthly variance report should be structured.
| Item | Theoretical Cost (£) | Actual Cost (£) | Variance % / Root-Cause Category / Recommended Action |
|---|---|---|---|
| Beef mince | £380 | £416 | 9.5% / Over-portioning / Retrain kitchen on portion weights, introduce scales |
| House spirits | £210 | £224 | 6.7% / Over-pouring / Implement measured optics, spot-check pours |
| Draught lager | £540 | £556 | 3.0% / Line loss / Check cellar temperature, review line-clean frequency |
| Chicken breast | £290 | £293 | 1.0% / Acceptable noise / Monitor, no action required |
Why Stock Variance Happens in Bars and Kitchens
Kitchen and bar operational causes
Unlogged waste and write-offs can account for a significant share of stocktake variance in venues that properly investigate root causes. Equally common is portioning inconsistency, which creates variance through a different mechanism. A recipe calling for 180g of protein but consistently served at 200g generates 10% ingredient overconsumption per cover without any waste recorded.
Beyond these measurement and execution issues, recipe drift, comps and staff drinks, and theft also contribute to variance, though typically at lower levels. On the bar side, over-pouring on spirits is a significant source of bar variance in many UK pubs, especially when staff pour by eye. Draught beer line loss can represent a notable proportion of volume in UK pubs due to cellar temperature, line-cleaning frequency and pour technique.
System and administrative causes
- POS and recipe mapping errors are a common source of variance.
- Supplier shorts and delivery discrepancies can contribute to variance.
- Stocktake counting errors are also frequent.
- System data errors, such as a supplier renaming an SKU, cause recipe components to become unmatched and create phantom variance.
Theft is often assumed first by operators but is usually the smallest of the seven common causes in restaurants, pubs and bars that investigate properly.
How to Build a Monthly Variance Report That Drives Action
A reliable monthly variance report follows a consistent process from opening stock through to accounting. Tracking variance separately by category, such as spirits, wine, draught and food, rather than using a single blended figure, reveals exactly where margin leakage occurs.
- Lock opening stock from the previous period’s closing count. Prevent post-submission edits to maintain data integrity.
- Record all purchases by scanning invoices line by line. Jelly’s automated invoice scanning captures every SKU, quantity and price via photo or email, with no manual entry required.
- Pull theoretical usage from your POS. Jelly integrates natively with its integration partners Square, Lightspeed, EPOS Now and Toast, pulling item-level sales data in real time so theoretical consumption updates automatically with every transaction.
- Conduct the physical count using a standardised method at the same point in the delivery cycle each period.
- Calculate and categorise variance using the template above. Flag any line exceeding your threshold for investigation.
- Push to Xero via Jelly’s one-click accounting integration, which removes a separate bookkeeping step and can reduce reconciliation time by up to 90%.
Investigation Workflow and Decision Tree for Unexplained Variance
Steady variance of 2–3% over consecutive months should be treated as acceptable noise, while variance exceeding 5%, changing suddenly, concentrating in one category or correlating with operational changes requires immediate investigation.
The five-step workflow:
- Isolate the largest single variance line item rather than the total figure. A 24-unit pizza chain found that a single ingredient had created €91,000 in uncontrolled losses, and ingredient-level reporting shifted the conversation from a general problem to a targeted fix.
- Test the seven causes in order of likelihood: over-portioning, unlogged waste, comps or voids, POS mapping errors, supplier shorts, counting error, theft.
- Apply the decision tree threshold: variance steady at 2–3% equals noise to monitor, above 5% or a sudden change equals an issue to investigate this period, and above 10% equals a process failure requiring immediate structural review.
- Implement a targeted fix such as retraining on portion weights, adding a waste log, correcting a POS mapping or raising a credit note with the supplier.
- Re-measure the same line item in the next period to confirm the fix has closed the gap before you move to the next largest variance.
Turning Variance Data into a Live KPI
Well-run multi-location restaurant groups target a food cost variance of 2–3% of theoretical, as outlined earlier. The recommended site-level targets for UK operators are:
- 2–3%: Excellent, maintain current processes.
- 3–5%: Watch, investigate root cause before the next period.
- Above 5%: Action, treat as a systemic issue requiring immediate intervention.
Operations that commit to weekly inventory tracking see a 3–6% improvement in food cost within a single quarter. Jelly’s daily Flash report, generated automatically from POS sales and scanned invoices, surfaces gross profit margin every day so variance trends are visible in real time rather than discovered at month-end. Daily and weekly exception reports surface issues such as poor inventory variance for quick action, while monthly reviews assess trends and structural issues.
See Jelly’s Flash report and live KPI dashboard in action.
How Real-Time Invoice and Recipe Automation Removes the Manual Burden
When supplier invoices, recipe specs and menu pricing connect in real time, costs update automatically as invoices arrive, which closes the 30-day visibility gap between price changes and month-end discovery. Without this connection, a dish that was profitable last week can be losing money today with no visible signal until the next stocktake.
Jelly replaces that manual cycle with a connected workflow. Every invoice is scanned automatically by photo or email, and every line item is matched to the relevant recipe ingredient. Dish costs and GP margins update the moment a new invoice lands. Jelly’s Cookbook lets chefs build recipes by clicking on ingredients already populated from scanned invoices, with all unit conversions and maths handled automatically. What previously took 28 minutes per dish in a spreadsheet now takes three minutes.
The time saving on stocktakes is equally significant. Sushi Revolution’s monthly stocktake using Jelly takes 5–20 minutes instead of 2–3 hours, and their actual gross profits are 2–3% higher on average as a result of tighter cost control.
AI-powered invoice scanning reads supplier invoices via OCR, matches line items to purchase orders and flags price variances before approval, preventing undetected overcharges that previously diverged budget versus actual by 8–12%. Jelly’s Price Alert feature surfaces every ingredient price movement, up or down, so operators can negotiate credits or switch suppliers before the variance appears in the next stocktake.
Jelly customers see gross margins increase by an average of 2 percentage points within the first three months, and food costs fall by 3% on average over the same period. At £500,000 revenue, that 2-point GP lift is worth £10,000 annually, recovered from losses that previously had no name on the P&L.
Frequently Asked Questions
How do you calculate stocktake variance?
Stocktake variance is calculated in three steps. First, calculate actual usage as opening stock plus purchases minus closing stock. Second, calculate theoretical usage as the sum of each recipe ingredient multiplied by the number of portions sold, drawn from your POS data. Third, subtract theoretical usage from actual usage to get the variance in units or £ value, then divide by theoretical usage and multiply by 100 to express it as a percentage. A positive result means you used more stock than your recipes predict, which indicates waste, over-portioning, unrecorded comps or data errors.
What does 20% variance mean?
A 20% stocktake variance means your kitchen or bar consumed 20% more of an ingredient than your recipes and sales data predict. This sits well above the 5% threshold that warrants investigation and indicates a structural problem rather than counting noise. Common causes at this level include a systematic portioning error, such as consistently serving double the recipe weight, a POS mapping fault that has been misreporting sales for several periods, a supplier SKU change that has broken recipe links or a combination of unlogged waste and unrecorded comps accumulating over time.
A 20% variance on a high-cost protein or spirit line can represent thousands of pounds in annual losses and requires immediate root-cause investigation using the five-step workflow above.
Is variance analysis a KPI?
Stocktake variance expressed as a percentage of theoretical COGS is a direct operational KPI that measures how accurately your kitchen or bar converts purchased stock into sold dishes. It sits alongside food cost percentage, gross profit margin and waste percentage as a core performance measure for hospitality operators. Well-run multi-site groups set site-level variance targets, typically 2–3% for excellent performance, 3–5% as a watch zone and above 5% as an action trigger, and review them on a weekly or monthly cadence alongside daily Flash reports. Tracking variance as a KPI turns a reactive monthly count into a proactive margin-protection tool.
How can automation reduce stocktake admin time?
Automation reduces stocktake admin time by removing the three most time-consuming manual steps: invoice data entry, theoretical usage calculation and variance reporting. When invoices are scanned automatically and matched to recipe ingredients, and when POS sales data flows directly into the system in real time, theoretical stock updates continuously rather than requiring manual calculation at period end.
The physical count becomes the only remaining manual step, and with pre-populated count sheets and mobile counting tools, that process shrinks from two to three hours to as little as five to twenty minutes per month. Jelly automates the full flow from invoice scanning through to dish costing and Xero integration, which saves operators a significant amount of admin time every month.
Conclusion: Protect Profit with a Repeatable Variance System
Unexplained stocktake variance is not an accounting problem, it is an operational one. The margin it erodes is recoverable, but only with a repeatable system that combines the right formula applied consistently, a structured investigation workflow, category-level KPI tracking and automation that keeps theoretical costs aligned with real supplier prices every day.
Moving from manual spreadsheets to automated invoice scanning, live recipe costing and POS integration removes the considerable admin time that currently stands between operators and the data they need. It also closes the 30-day visibility gap that allows supplier price increases and portioning drift to compound undetected. The result is tighter variance, higher gross profit and a business that reacts to cost changes in hours rather than weeks.
Jelly delivers all of this at a flat rate of £129 per location per month, with onboarding measured in days rather than months and value visible from the first scanned invoice.
Turn your next stocktake into a live margin-protection system, and book a demo to see how.