Written by: JJ Tan, Founder, Jelly
Menu engineering classifies every dish by contribution margin and popularity, then applies clear actions to protect gross profit. Popularised by Michael Kasavana and Donald Smith at Michigan State University in 1982, it treats the menu as a financial document where each item must justify its place on the page.
Key Takeaways
- Menu engineering ranks every dish by contribution margin and popularity, which gives a clearer profit picture than food cost percentage alone.
- UK operators face rising costs from 2.2% food price inflation, 9.4% beef price increases, and a new £12.71 National Living Wage, so disciplined menu analysis now matters more than ever.
- The four-quadrant matrix (Stars, Plough Horses, Puzzles, Dogs) turns data into actions: protect high-margin favourites, fix low-margin bestsellers, promote high-margin low-sellers, and remove weak performers.
- Contribution margin per cover is the primary ranking metric, and delivery commissions of around 30% mean delivery menus need separate costing to stay profitable.
- Jelly automates invoice scanning, price alerts, recipe costing and EPOS integration to cut food costs by about 3% on average and provide real-time margin visibility.
See Jelly Working With Your Menu
Why Menu Engineering Protects UK Restaurant Margins
Food cost percentage often creeps upward without triggering an alert because month-end accounts arrive weeks after the damage. A supplier raises the price of beef mid-month, the invoice is filed, the dish keeps selling at the same price, and the GP shortfall only surfaces when the accountant calls. Delivery menus compound the problem, as a dish priced for dine-in can become loss-making on a platform charging 30% commission, and many operators never run the arithmetic to confirm it.
A disciplined menu engineering routine protects GP and speeds up pricing decisions. Some of the UK's largest restaurant groups generate profit margins of only around 1.5%. Even a single unnoticed price rise on a key ingredient can push a site into loss. The process described here replaces reactive month-end reviews with a weekly routine built on live data.
Menu Engineering Matrix and What Each Quadrant Means
The four-quadrant matrix classifies every dish by two variables: contribution margin (selling price minus food cost) and popularity (share of covers). The table below shows how each combination maps to a specific action, from protecting Stars to removing Dogs.
| Quadrant | Popularity | Profitability | Action |
|---|---|---|---|
| Stars | High | High | Protect quality, feature prominently, leave the recipe untouched |
| Plough Horses | High | Low | Re-engineer cost first, then raise price 4–7% with a visible upgrade |
| Puzzles | Low | High | Rename, re-describe, move to a high-visibility slot, add one photo |
| Dogs | Low | Low | Remove, unless ticket data shows they complete larger party orders |
Stars are the menu's engine, so never discount them. A Star at 20% off becomes a Plough Horse for the duration of that promotion. Plough Horses present the opposite problem. They drive covers but erode margin per cover, which makes cost reduction the first move. Switch supplier, adjust portion, or substitute an imperceptible ingredient, then apply a small price increase with a visible upgrade so guests feel added value rather than quiet inflation.
Puzzles look like a menu problem but usually reflect a marketing problem. Before removing one, check whether the name undersells it, whether the description feels flat, or whether servers are simply not talking it up. Dogs are the only quadrant that normally warrants removal. The exception is a dish that appears consistently on larger-party orders, where it acts as a party-completer and should be reclassified. A high-cost but high-margin dish still sits in the Star quadrant, so protect it and re-engineer the cost base rather than cutting it.
For a deeper treatment of the matrix and layout tactics, see Menu Engineering for UK Restaurants: A Complete Guide.
Contribution Margin Per Dish vs Food Cost Percentage
Most UK operators default to food cost percentage as their primary metric, yet this metric ranks dishes poorly by profitability.
Consider two dishes on the same menu. Dish A sells at £18 with a food cost of £7.20. That is a 40% food cost percentage and a £10.80 contribution margin per cover. Dish B sells at £14 with a food cost of £3.78. That is a 27% food cost percentage and a £10.22 contribution margin per cover. By food cost percentage, Dish B looks healthier. By contribution margin, Dish A earns more per cover, in this illustration £0.58 more. Across 200 covers a week, that difference becomes £116 more gross profit from the dish that looks worse on the standard metric.
The shift from food cost percentage to contribution margin per cover is the single most important change a UK operator can make to their analytical routine. A dish with a 35% food cost contributing £15 per plate will outperform a 22% food cost dish contributing £4 every service. Rank dishes by absolute contribution margin rather than by cost ratio, and the matrix will reflect commercial reality instead of accounting convention.
For a detailed walkthrough of ingredient-level cost tracking, see How to Track Ingredient Costs Affecting Menu Profitability.
How to Cost a Dish with VAT and Delivery Commission
Contribution margin only reflects reality when it starts from the correct net price. HMRC requires UK hospitality businesses to apply the standard 20% VAT rate to food and drink prepared for catering or consumed on the premises. Menu prices are VAT-inclusive, so operators cannot advertise "£10 + VAT" to diners. The VAT element inside a VAT-inclusive price is one sixth of the gross price, so contribution margin must be calculated from the VAT-exclusive net price, not the sticker price.
Mandatory service charges are treated as part of the sale and attract 20% VAT, while optional tips do not. This distinction changes the net revenue calculation on every cover where a service charge applies.
The following worked example shows a pan-fried sea bass dish. The Pan Fried Sea Bass Fillet & Prawns dish at Bluebell Bistro is priced at £15.50. For illustration, assume a VAT-inclusive menu price of £19.20.
- VAT-inclusive menu price: £19.20
- VAT-exclusive net price (÷ 1.20): £16.00
- Food cost (ingredients, garnish, oil, waste): £5.60
- Food cost percentage (on net price): 35%
- Contribution margin (dine-in): £10.40 per cover
Now apply the same dish to a delivery platform. Delivery platforms such as Deliveroo and Uber Eats charge average commissions of 30%, applied to the VAT-inclusive order value.
- VAT-inclusive delivery price: £19.20
- Platform commission at 30%: £5.76 (plus 20% VAT charged on the commission)
- Net revenue after commission: £13.44
- VAT on net revenue (÷ 1.20): £11.20 net of VAT
- Food cost: £5.60
- Contribution margin (delivery): £5.60 per cover
On these illustrative figures, the same dish at the same price earns £10.40 on a dine-in cover and £5.60 on a delivery order, a 46% reduction in contribution margin. Without a separate delivery menu priced to absorb commission, the dish loses relative profit on every delivery order. Operators who set separate target gross profits on dine-in and delivery menus see actual gross profits 2–3% higher on average.
How To Use EPOS Sales Mix Data For Menu Engineering
The menu engineering matrix needs two inputs: contribution margin per dish and popularity. Contribution margin comes from costed recipes. Popularity comes from the sales mix report in the EPOS system. UK operators commonly use integration partners such as Square, Lightspeed, EPOS Now and Toast, and each system produces a product mix (PMIX) report that shows units sold per item over a selected period.
The workflow stays simple when it runs on a schedule. Pull 90 days of EPOS sales mix data, then map each POS item to a dish in the costing system so that sales volume and margin can be read together. Without that mapping, the matrix cannot be built accurately. A POS item called "Sea Bass" and a recipe called "Pan-Fried Sea Bass with Samphire" describe the same dish, but the system will not know that until the link is made.
Jelly integrates natively with Square, Lightspeed, EPOS Now and Toast via real-time API, delivering item-level sales data the moment a transaction completes. Setup takes approximately five minutes across all four systems. POS-to-dish linking only surfaces items sold since the integration was connected, which keeps the mapping clean and free of legacy menu clutter. Connecting a POS automates 2–5 hours of weekly work and provides real-time margins and sales mix data.
See Your Live Margins in Jelly
Ingredient Cross-Utilisation and the Minimum Viable Menu
Once the matrix is built, the next lever is menu size. Fewer SKUs across more dishes reduce both waste and costing complexity. A menu built around British seasonal produce, such as spring lamb from Bidfood, heritage tomatoes from Fresh Direct and root vegetables from Brakes, naturally concentrates purchasing volume on fewer lines. That concentration strengthens negotiating position and simplifies stock management. Common menu-design guidance suggests trimming each category to roughly seven to nine dishes. That range works as a starting point, and the right number then depends on kitchen throughput and cover count.
Cross-utilisation means a single ingredient, such as a whole chicken from a named UK supplier, appears across a starter (chicken liver parfait), a main (roast half chicken) and a staff meal. Every part of the bird is costed and nothing is written off as waste. Fewer menu lines also mean fewer recipes to re-cost when a supplier changes a price, which makes the weekly GP review faster and more reliable. For the relationship between menu size and food waste, see How Menu Engineering Reduces Food Waste in Restaurants.
How to Keep Food Cost Down with a Supplier Price-Alert Routine
Most operators discover a supplier price rise at month-end, weeks after it appeared on an invoice. By that point, the dish has been sold at the wrong price dozens or hundreds of times, and the GP shortfall is already locked in.
A weekly price-alert routine triggered by invoices fixes this lag. Every invoice that arrives, whether by email or photographed into the system, is scanned line by line. Any ingredient whose price has moved since the previous invoice generates an alert, with the supplier name, the ingredient, the old price and the new price shown side by side. The chef or operations manager can then decide within the same week whether to challenge the supplier, switch to an alternative, or adjust the dish price.
Jelly's Price Alert feature flags every price increase or decrease by ingredient and supplier. Jelly users cut food costs by about 3% on average in the first three months. At Amber, a Mediterranean restaurant in East London, the combination of invoice automation, price change alerts and real-time costing saves £3,000–£4,000 per month, which delivers approximately 68× ROI on the platform cost.
Seven-Step Menu Engineering Routine for UK Restaurants
The following seven-step sequence can be handed to a head chef on Monday morning and completed within a standard working week.
- Pull 90 days of EPOS sales mix and map every POS item to a dish.
- Cost every dish from received invoices, including garnishes, oils and waste.
- Rank dishes by contribution margin per cover rather than food cost percentage.
- Classify each dish into the four quadrants.
- Re-engineer or re-price Plough Horses and Puzzles, and remove Dogs.
- Build the minimum viable menu and cost the delivery menu separately.
- Set the supplier price-alert routine and review GP weekly.
Recipe costing must use received invoice prices, not supplier price lists, and must include every variable cost such as garnishes, cooking oils and portioning waste. Get this step wrong and the matrix rests on false numbers, which makes the costing step the one most worth automating. What used to take 28 minutes per dish in a spreadsheet takes approximately 3 minutes in Jelly's Kitchen section, where chefs build recipes by clicking on ingredients already populated from scanned invoices.
For a full treatment of best practices and tooling, see Menu Engineering Best Practices & Tools for UK Restaurants.
Frequently Asked Questions
How Often Should Menu Engineering Be Re-Run?
A full matrix review, which means re-costing every dish, pulling fresh sales mix data and reassigning quadrants, should run every 90 days. A spot review is warranted any time a key ingredient rises in cost by more than 5%, so the supplier price-alert routine should trigger a targeted re-cost rather than waiting for the quarterly cycle. A full menu redesign is appropriate every 12 to 18 months.
What Should an Operator Do When a Supplier Changes a Price Mid-Month?
Act within the same week. When an invoice arrives with a higher line-item price, the price-alert routine flags it immediately. The options are:
- Challenge the supplier with the invoice data and request a credit note.
- Switch to an alternative supplier for that ingredient.
- Adjust the dish price on the menu.
- Substitute a lower-cost component in the recipe.
Waiting until month-end means the dish has already been sold at the wrong margin for weeks.
Should Delivery Menus Be Costed Separately from Dine-In Menus?
Yes, always. Delivery platform commissions of around 30% applied to the VAT-inclusive order value reduce contribution margin substantially on a typical dish priced for dine-in. An £8 dish that leaves around £5.60 of gross margin on a dine-in cover earns only £2.27 on a delivery order at the same price, after 30% platform commission, VAT on that commission and £0.45 packaging. That separate-costing discipline is the same one that delivers the 2–3% gross profit uplift noted earlier. The delivery menu should be built as a separate document, with each item re-costed to confirm it is profitable after commission.
Who Should Own the Menu Engineering Process Across Multiple Sites?
The operations manager or finance manager owns the weekly GP review and the quarterly matrix cycle. The head chef or executive chef owns recipe costing and supplier negotiation. Management should have direct access to the same live data the kitchen uses, not a summary report produced by the chef, so that decisions are made on a shared, trusted source of truth. At multi-site operators, the routine must be standardised across sites because menu mix varies significantly between locations even when the format is identical.
Conclusion
The menu functions as a financial document, and the arithmetic that determines which dishes earn their place, such as contribution margin per cover, VAT-exclusive net price and delivery commission, is the arithmetic many UK operators never run in full. A repeatable weekly routine built on live invoice data, real-time dish costing and EPOS sales mix replaces month-end guesswork with decisions made in the same week the numbers change. As the routine matures and scales across sites, it becomes the operational standard that protects GP regardless of which site the owner is physically present at.
Talk Through Your Menu With Jelly
