Written by: JJ Tan, Founder, Jelly | Last updated: 8 November 2026

Key Takeaways

  • Single-site GP calculators fail multi-site groups because each location uses different assumptions, stock timings and waste logging, which makes direct comparisons unreliable.
  • Group GP margin relies on consistent CPU transfer pricing, inter-site stock reconciliation and sales-weighted roll-ups instead of simple averages.
  • Actual versus theoretical GP gaps should be tracked weekly. Gaps above 1–2% usually indicate waste, portion drift or unrecorded sales that need investigation.
  • Spreadsheet-based GP models break at scale because of manual entry, stale prices and version drift, and they consume 10–20 hours of admin weekly across four or more sites.
  • Jelly automates invoice capture, live dish costing and multi-site roll-ups, so operators get accurate group GP visibility without the admin burden of spreadsheets.

Chat To The Team

What Multi-Site GP Margin Means and How to Calculate It

Group GP margin for a multi-site chef operation is net sales across all sites minus total cost of goods sold (COGS), including CPU transfers and waste, divided by group net sales and expressed as a percentage. The calculation happens per site first, then rolls up on a sales-weighted basis so a high-revenue site carries proportionally more weight than a low-revenue one. VAT must be stripped from sales before any calculation, and ingredient costs must also be recorded excluding VAT.

The formula stays the same at every level.

GP% = (Net sales − COGS) ÷ Net sales × 100

Five variables break single-site calculators when applied to a group. CPU or commissary transfers must be costed at a consistent transfer price rather than ignored. Inter-site stock movements and per-site waste allocation both need tracking. Delivery platform commissions and site-level discounts reduce net sales differently at each location. Clarity Stocktaking notes that the real difficulty lies in keeping the inputs, such as stock counts, purchase records and sales data, accurate. At group level, that difficulty multiplies with every site added.

How to Calculate GP Margin Across Multiple Sites

Six practical steps apply the standard GP methodology to a multi-site group.

  1. Strip VAT from each site's net sales. Use EPOS Z-reports, not bank deposits. Probal Global identifies posting takings from the bank statement as the single biggest error on inherited hospitality files, because bank deposits are net of card fees, aggregator commissions and refunds. That understatement of turnover makes GP meaningless.
  2. Cost each site's purchases from invoices, not memory. Every line item, including quantity, SKU and price, must be captured for the period.
  3. Add CPU transfers in at the agreed transfer price. The transfer price must stay consistent across all sites and all periods. A method change mid-period makes site comparisons meaningless.
  4. Take opening stock plus purchases minus closing stock for actual COGS. Adjust consistently for credits and transfers. Probal Global requires the inter-site account to net to nil across the group so nothing is double-counted.
  5. Allocate waste and staff food per site. Waste, staff meals and complimentary items should sit in dedicated categories rather than in unexplained COGS.
  6. Calculate each site's GP%, then weight by net sales for group GP%. Group GP% = sum of site GP cash margin ÷ sum of site net sales × 100.

Both GP% and cash margin need tracking side by side. A dish running 55% GP can out-earn a 70% GP dish in absolute cash terms. A steak at £28 inc VAT with £10.50 cost earns 55% GP but £12.83 cash margin, while a burger at £12 inc VAT with £3 cost earns 70% GP but only £7.00 cash margin. Group reporting needs both figures.

Costing CPU and Commissary Transfers Without Distorting Site GP

Step 3 above set the rule that CPU transfers must use a consistent transfer price, and this section explains how to set that price. MarketMan's commissary cost-control guide recommends that multi-unit operators choose one of three transfer pricing methods and apply it consistently.

  • At cost: sites are charged the ingredient cost of each commissary item. This method is simple and makes site food cost reflect true ingredient cost, but it leaves the commissary with no contribution margin, so commissary efficiency must be monitored separately.
  • Cost plus markup: sites are charged ingredient cost plus a set percentage that roughly covers commissary labour and overhead. The site's transfer price should be benchmarked against the cost of buying the equivalent product externally.
  • Fully loaded: sites are charged ingredients plus commissary labour, packaging and an overhead allocation. This method gives the truest view of the cost to produce each item, provided the allocation basis is documented and applied consistently.

Whichever method you choose, transfer prices must be calculated from actual batch inputs divided by actual usable yield, because the theoretical recipe yield overstates what a batch really produces. Onions cook down. A batch that the recipe card claims yields 20 lb may actually yield 15 lb, and transferring at the recipe-card rate gives every site a hidden discount that only shows up in the commissary's numbers. Transfer prices should be refreshed whenever ingredient costs move. Every transfer should be recorded as an internal invoice stating item, quantity, unit cost and total value, so the inter-site account nets to nil at group level and nothing is double-counted.

Site-by-Site Variance: Theoretical vs Actual GP

With transfer prices set and recorded, the next question is whether each site is actually hitting the margin its recipes promise. Theoretical GP is the margin a site should achieve based on costed recipes multiplied by the POS sales mix. Actual GP is what the site achieves once stock variances, waste, overpouring and unrecorded sales are accounted for. Clarity Stocktaking describes the gap between the two as "where your missing profit is going."

Actual COGS = opening stock + purchases − closing stock, adjusted for credits and transfers. The gap between theoretical and actual GP should be reconciled to waste, shrinkage and portion drift per site. Clarity Stocktaking notes that a gap of 1–2% may be acceptable. Anything above that level warrants investigation into the operational cause. A protein portion running 20g over spec on every plate across 200 covers a day, five days a week, adds up to significant food cost overrun quickly.

The scale of the problem is substantial. WRAP stated on 4 November 2024 that the UK hospitality and food service sector throws away 1.1 million tonnes of food every year, 75% of which could have been eaten, costing businesses and organisations £3.2 billion. That waste is one reason actual GP drifts below theoretical. For a sense-check on where a group should sit, the UKHospitality × Christie & Co Benchmarking Report 2022 puts whole-survey food GP at 67% across 4,791 licensed managed outlets. Food GP across the eight segments ranged from 65.9% to 69%, with accommodation-led venues at 69% and high-street pubs also reported at 69%, so operators should treat 67% as a reference point rather than a fixed target.

Worked Example: Group GP Across Four UK Sites

This fictional four-site UK group example shows how one site's GP% shift affects the weighted group figure. The group includes two high-street restaurants (Sites A and B), one suburban pub-restaurant (Site C) and one site supplied by a CPU (Site D). All figures are illustrative and are not drawn from any published benchmark.

  • Site A (high-street): Net sales £85,000 | COGS £28,050 | GP% 67.0% | Cash margin £56,950
  • Site B (high-street): Net sales £72,000 | COGS £25,200 | GP% 65.0% | Cash margin £46,800
  • Site C (suburban): Net sales £60,000 | COGS £19,800 | GP% 67.0% | Cash margin £40,200
  • Site D (CPU-supplied): Net sales £43,000 | COGS £14,190 | GP% 67.0% | Cash margin £28,810

Group totals: Net sales £260,000 | Total COGS £87,240 | Weighted group GP% 66.4% | Total cash margin £172,760. Site B already trails the others on GP% at 65%, so it starts as the weakest performer.

Now suppose the CPU refreshes its transfer prices mid-quarter because a key protein rises in cost. Site D's COGS increases to £15,480, which drops its GP% to 64.0% and its cash margin to £27,520. At group level, total COGS rises to £88,530 and weighted group GP% falls to 65.9%. That 0.5-point movement is visible and explainable. Without the internal invoice trail and consistent transfer pricing, that movement would appear as unexplained variance at Site D and the group figure would be unreliable.

Building a GP Model in Excel or Google Sheets and Where It Breaks

A workable multi-site GP spreadsheet needs at minimum an invoice input tab per site, a recipe costing tab, a stock tab, a per-site GP tab and a group roll-up. That structure is sound in principle. The failure points emerge at scale.

On average it takes 28 minutes of work in a spreadsheet to cost a single menu item. Across a group with four sites and rotating menus, that time compounds into days of admin every quarter. Beyond costing time, five structural failures compound at scale. Manual line-item entry from invoices is slow and error-prone. Supplier prices go stale because nobody updates them when invoices change. There is no live link to POS sales data. Each chef maintains a slightly different spreadsheet, so versions drift between sites. Together these consume 10–20 hours a week of admin that should be spent on operations. Probal Global estimates that a two-point GP swing on a £1.2m site equates to £24,000. That error hides comfortably in a file where stock is only counted twice a year.

The Group KPIs to Report Every Week

A multi-site group needs a consistent weekly reporting set to catch margin movement before it compounds. GP% sits at the centre of that set because it is the one figure that compares cleanly across sites of different sizes.

  • Food GP% per site and weighted group GP%
  • Actual vs theoretical food cost per site
  • GP variance by site, week-on-week and versus budget
  • Waste % per site
  • GP per site versus group average

GP% is the right comparison metric across sites because it sits on the same scale regardless of site size or sales mix. Comparing food cost percentages across a kitchen and a bar tells you far less, because the two categories have fundamentally different cost structures. HOPS makes this point directly. A multi-site operator comparing Site A at 69% blended GP with Site B at 63% has a comparison that is immediately legible and shows which site has a margin problem. A single blended figure still only tells you something is wrong, not where, which needs departmental margin visibility. Comparing food cost percentages across different venue types, by contrast, produces a figure that reflects neither category accurately.

Spreadsheet vs Automated Software: Choosing a GP Margin Calculator for Multi-Site Chef Operations

Those spreadsheet failure points are what push multi-site groups towards dedicated software. In the category of UK hospitality GP calculators and restaurant GP software, Jelly is the primary tool for multi-site chef operations. DataChef is a general data and AI implementation firm that builds custom platforms for enterprise clients in chemicals, logistics and telecoms, so any hospitality GP model would be bespoke consultancy work rather than an out-of-the-box calculator. Onecall Hospitality Consultants and Integra (McVey) Limited offer templates and consultancy instead of live invoice-to-dish automation. Access Hospitality's Access Evo is an end-to-end hospitality software platform connecting reservations, EPoS, scheduling, procurement and guest data, while Tenzo is a hospitality analytics and PerformanceOps platform that pulls POS, labour, inventory, reservations and review data into one view rather than replacing existing systems. Lightspeed is a POS system, and Jelly integrates with it natively. Google Sheets is the manual method described in the section above.

Jelly wins for this specific job because it is purpose-built for invoice-level costing, POS-linked sales mix and one live group view. That combination delivers initial value in the first week rather than after months of bespoke configuration.

The specific capabilities relevant to multi-site GP reporting are clear.

  • Automated invoice scanning via photo or email, with every line item captured without manual entry
  • Live dish costing that updates the moment a new invoice arrives, with a red or green margin indicator per dish
  • Price Alert, which flags supplier price changes as they come in, highlighting changes in supplier pricing and giving a clearer view of spend trends, so CPU transfer prices can be refreshed promptly
  • Flash Report, which gives a daily, weekly or monthly view of GP margin calculated from invoice costs and POS sales
  • Sales Mix, integrating with Square, Lightspeed, EPOS Now and Toast via real-time API, with setup taking approximately five minutes across all four systems
  • Xero accounting integration for a seamless payables process

Evidence from operators using Jelly shows the impact. Dish costing takes 3 minutes rather than 28. Teams save 10–20 hours of admin every month. Gross margins increase on average by 2 percentage points in the first three months, and food costs fall by 3% on average. Populu lifted GP from 68% to 72% across 16 locations. One operator moved from 65% to 72% within 12 weeks on approximately £500,000 in revenue. Amber restaurant in East London saves £3,000–£4,000 a month, roughly 68× ROI, with chef-owner Murat Kilic stating: "Jelly keeps my business alive."

For existing single-site GP and food cost analysis, see Gp Margin Calculator For Restaurant Back Of House Management and Gp Margin Calculator For Restaurant Food Cost Analysis Uk.

Book A Demo

Conclusion: One Live Group GP Figure, Not Five Spreadsheets

Jelly is a strong GP margin calculator for multi-site chef operations in the UK for three reasons. It captures all supplier invoices automatically using a Human + AI system that extracts full line-item detail, whether invoices are photographed, emailed or received directly from suppliers. It combines that data with live POS sales to give real-time gross profit visibility comparing actual against theoretical margins. It also offers a dashboard that rolls sites up for multi-site groups. The multi-site methodology, including CPU transfer costing, actual-versus-theoretical reconciliation and weighted group roll-up, is sound and repeatable. The spreadsheet is the bottleneck described above, and it means the group GP figure is always behind, always approximate and never fully trusted. Jelly removes those bottlenecks and delivers a defensible, comparable group GP figure without the 10–20 hours a week of admin.

See Your Group GP in One Dashboard

Frequently Asked Questions

Is 75% Gross Profit Margin Good for a UK Hospitality Operation?

For food, 75% is above the UKHospitality × Christie & Co Benchmarking Report 2022 whole-survey figure of 67% and above every individual segment reported in that survey, where the highest was 69% for accommodation-led venues. A 75% food GP is therefore strong by UK managed-outlet benchmarks, and it warrants a check for under-portioning or unusually aggressive pricing. A food cost below 28% can indicate either. For drinks, 75–80% is achievable on spirits and wine, where structural losses from line cleaning, foam and spoilage are absent. Strapture's June 2026 guidance recommends aiming for 75–80% on drinks as a realistic target for well-run UK operations. In all cases, the right GP target is the level that covers labour, rent and other overheads and leaves a net profit, rather than a fixed industry figure.

Is 30% Markup the Same as 30% Margin?

No. Markup is calculated as a percentage of cost, while gross profit margin is calculated as a percentage of the selling price. A 30% markup means selling at 130% of cost, which produces a GP of 23.1%, calculated as 30 ÷ 130. A 100% markup, selling at double cost, equals a 50% GP. Confusion between the two is a common cause of underpricing. An operator targeting 70% GP who applies a 70% markup instead will price dishes significantly below their intended margin. Always confirm which basis is being used before setting or comparing targets.

What Is a Healthy GP Margin for a UK Restaurant Group?

The UKHospitality × Christie & Co Benchmarking Report 2022 puts whole-survey food GP at 67% across nearly 4,800 managed outlets, with all eight segments falling between 65.9% and 69%. Strapture's June 2026 guidance recommends aiming for 65–72% on food and 75–80% on drinks, with a blended target of around 70%. For a multi-site group, the most useful benchmark is a weighted group GP% calculated from each site's own theoretical and actual figures, validated against the group's own accounts. A group GP consistently below 65% on food warrants investigation into waste, portion drift, CPU transfer pricing or stale recipe costs. Group GP should always be weighted by net sales so that larger sites carry proportionally more weight in the overall figure.

How Do I Calculate GP Margin for Multiple Sites in Excel?

Follow the six steps set out in this article: strip VAT from each site's net sales, cost purchases from invoices, add CPU transfers at the agreed transfer price, calculate actual COGS from opening stock plus purchases minus closing stock, allocate waste and staff food per site, then calculate each site's GP% and weight by net sales for the group figure. The arithmetic is straightforward. The practical problems are the same four failure points set out earlier, manual entry, stale prices, no POS link and version drift, and they still cost 10–20 hours a week of admin across four or more sites. The group GP figure also lags reality by days or weeks at any given moment.

Can Jelly Calculate GP Across All My Sites?

Yes. The invoice capture, live dish costing and POS integrations described above all roll up to group level, so you get one live group GP figure rather than a per-site spreadsheet. Jelly's Flash Reports give a daily, weekly or monthly view of gross profit margins broken down by item, connecting POS sales to recipe costs. Its price alert system flags supplier price changes as they come in, highlighting changes in supplier pricing and giving a clearer view of spend trends, so margins can be protected promptly rather than discovered at month-end. The result is one live group GP figure, updated continuously, without manual consolidation. Operators typically see initial value within the first week of onboarding.