Written by: JJ Tan, Founder, Jelly

Key Takeaways

  • Menu engineering quadrants classify every dish by popularity and contribution margin into Stars, Plowhorses, Puzzles, and Dogs, giving UK operators a repeatable way to protect gross profit.
  • Accurate calculations require stripping 20% VAT from menu prices, using recipe yields for portion costs, and running separate analyses for delivery menus to reflect 25–35% platform commissions.
  • Apply the 70% rule to set popularity thresholds within each category, then use weighted average contribution margin to plot dishes and assign them to the correct quadrant.
  • Follow a quadrant-specific action playbook: protect Stars, lift margin on Plowhorses, increase visibility for Puzzles, and remove or rework Dogs that lack a clear strategic role.
  • Jelly automates sales mix tracking, live dish costing, and price alerts so quadrants update in real time and operators can react before margin erosion compounds.

Start tracking your menu quadrants in real time

Before You Begin: Data And Setup You Need

Run menu engineering only when four core inputs are in place.

  • Item-level POS sales data covering at least 4–12 weeks of normal trading (exclude weeks with closures, holidays, or one-off events)
  • Current supplier invoices or food cost data per dish, updated for recent price changes
  • Recipe cards with yields, so portion cost reflects what is actually plated rather than raw purchase weight
  • Clarity on VAT treatment for each menu category, since HMRC applies standard-rate 20% VAT to hot food and drink consumed on the premises, while some cold takeaway items may be zero-rated

Ownership usually sits with the owner, operations manager, or head chef, with finance support. Multi-site operators should centralise the method and run the analysis per site, because performance varies by location.

Treat menu engineering as a recurring process. Initial setup takes a few hours. Ongoing tracking works best when automated. Delivery menus always need separate analysis because platform commissions change the effective selling price of every item.

See how Jelly automates your menu analysis

Why Menu Engineering Quadrants Matter For UK Restaurants

Quadrant analysis replaces guesswork about which dishes to promote, re-price, or remove. Guesswork usually erodes margin, especially when supplier prices move and nobody has refreshed recipe costs since the last menu print.

A correctly run process delivers specific outcomes: clearer menu decisions, faster reaction to ingredient cost changes, stronger evidence for supplier negotiations, and improved gross profit. Menu engineering quadrant analysis degrades quietly without continuously refreshed data, because a Star becomes a Plowhorse the moment a key ingredient price rises and nobody recalculates the recipe cost.

The maths must stay accurate. Using VAT-inclusive prices or ignoring delivery commissions misclassifies dishes and drives poor decisions. A dish that looks profitable for dine-in can become loss-making on a delivery platform once commission (typically 15–35% of order value in the UK) is deducted.

Step 1: Calculate Contribution Margin Per Dish (UK Worked Example)

Contribution margin (CM) is the selling price excluding VAT minus the food cost per portion. It is an absolute pound figure, which makes it the correct profitability axis for menu engineering.

UK menu prices are VAT-inclusive, so the 20% VAT already sits inside the price the customer pays, unlike US sales tax. Strip VAT out before calculating margin.

The formula is:

  • Ex-VAT selling price = menu price ÷ 1.20
  • Contribution margin = ex-VAT selling price − food cost per portion

Worked UK example: a dish priced at £15.00 on the menu has an ex-VAT selling price of £12.50. If the food cost is £5.00, the contribution margin is £7.50. Using the £15.00 figure directly would overstate margin by £2.50, which belongs to HMRC.

For delivery menus, subtract the platform commission from the ex-VAT selling price before calculating margin. Delivery platforms such as Deliveroo and Uber Eats charge average commissions of 30%, which squeeze restaurant margins. A dish sold at £15.00 on a delivery platform at 30% commission yields £10.50 after commission. Once VAT is removed, that is approximately £8.75 ex-VAT, before food cost is deducted.

Inputs required at this stage are POS sales data, supplier invoices, and recipe costing with yield adjustments. The output is a table of every dish with an accurate contribution margin figure.

Step 2: Calculate Popularity For Each Dish

With contribution margin calculated for every dish, the next input is popularity. Popularity measures how often each dish is ordered relative to total items sold within the same menu category. Menu engineering works best within comparable categories, such as starters against starters and mains against mains, because mixing very different price structures distorts the analysis.

The formula is:

  • Popularity (sales mix %) = (units sold of dish ÷ total units sold in category) × 100

Use at least four weeks of data to smooth anomalies. Exclude comps and voids, because a comped plate adds apparent popularity with no margin behind it. The output is a popularity percentage for every dish in each category.

Step 3: Set Your Popularity Threshold With The 70% Rule

The standard popularity threshold in menu engineering is the 70% rule: (100 ÷ number of items in the category) × 0.70. This rule recognises that menus never sell evenly, so a dish slightly below its equal share can still contribute well.

Worked example: a mains section with 8 dishes has an equal share of 12.5% per dish. The popularity threshold is 12.5% × 0.70 = 8.75%. Any dish with a sales mix at or above 8.75% counts as high popularity. Any dish below that threshold counts as low popularity.

Apply this threshold consistently within each category and across every analysis period. Consistent rules keep results comparable over time.

Step 4: Plot Dishes On The Menu Engineering Matrix

Once contribution margins and popularity percentages are ready, assign each dish to one of the four quadrants. The profitability threshold is the category’s weighted average contribution margin, which equals total contribution margin of all dishes sold in the category divided by total units sold.

Using a simple unweighted average instead of the weighted average can incorrectly classify borderline dishes.

The four quadrants are:

  • Stars: high popularity, high contribution margin
  • Plowhorses: high popularity, low contribution margin
  • Puzzles: low popularity, high contribution margin
  • Dogs: low popularity, low contribution margin

Plot each dish on a 2×2 matrix with popularity on one axis and contribution margin on the other. The popularity threshold and weighted average CM divide the matrix into four quadrants, so every dish lands in exactly one quadrant.

Step 5: Use The Quadrant-By-Quadrant Action Playbook

Each quadrant comes with a standard set of actions. Apply these per category and per site, and document every change so you can measure its effect.

Stars deserve protection. Protect Stars by keeping the recipe and portion size unchanged and giving them the best menu positions. A small price rise of 3–5% is usually safe, but never discount them.

Plowhorses need margin improvement. Start by reducing portion size by 8–12% or swapping an expensive component for a cheaper alternative, because both options lower cost without changing the price customers see. If that still leaves margin too low, raise the price in small steps or pair the dish with a high-margin add-on. A one-dollar price increase on a Plowhorse selling 505 units a month generates $505 a month, or about $6,000 a year, from a single line on the menu.

Puzzles need visibility. Rewrite the description with sensory and origin language, move the dish to the top of its section, brief front-of-house staff to recommend it by name, or run it as a daily special. If visibility improvements do not shift sales after 4–8 weeks, review the price.

Dogs usually need removal or replacement. They consume menu space, kitchen prep time, and stock without contributing meaningfully to margin. Before removing a Dog, check whether it serves a strategic purpose, such as the only vegan main, a dietary anchor, or a dish with a small but loyal following. Rework its cost if it plays that role.

Multi-site operators should apply the action playbook per site and then look for consistent patterns across locations before locking in permanent menu changes.

Step 6: Adjust For Delivery Commissions And VAT

A dish classified as a Star on dine-in can become a Plowhorse or Dog on delivery once platform commission is deducted. Sushi Revolution uses Jelly to set separate target gross profits on dine-in and delivery menus, accounting for 30% delivery commissions, and achieves actual gross profits 2–3% higher on average.

To build a delivery-specific matrix, follow this sequence:

  • Start with the ex-VAT selling price (menu price ÷ 1.20)
  • Subtract the platform commission (for example, 30% of the ex-VAT price for Deliveroo or Uber Eats)
  • Subtract packaging costs
  • Subtract food cost per portion
  • Treat the result as the delivery contribution margin

Commissions on major UK delivery platforms typically run between 25% and 35% of order value, with the exact rate varying by platform, contract, plan, and delivery model. Use your actual contracted rate for accuracy.

Create a separate delivery menu engineering matrix and apply the same quadrant action playbook independently of your dine-in matrix. Dishes that are marginal on delivery may be worth removing from platform menus or repricing upward to recover margin.

Step 7: Automate And Repeat Your Menu Review

This process delivers value only when it runs consistently. Aim for a monthly or quarterly cycle and trigger extra reviews whenever a key ingredient price changes significantly.

Manual spreadsheets usually break down over time. Supplier prices move, POS data sits in a separate system, and nobody has capacity to rebuild the model every month.

Jelly’s Sales Mix feature pulls item-level sales from your POS and combines them with live dish costing from scanned invoices, so quadrant tracking updates without manual data entry. When an ingredient price changes, Jelly’s Price Alert flags it immediately and gives you time to act before the shift moves a dish into a different quadrant.

Jelly onboards in the first week and adds on average 2 percentage points to gross margins in the first 3 months. Connecting a POS automates 2–5 hours of weekly work and delivers real-time margins and sales mix data.

See how Jelly automates your menu analysis

Common Mistakes To Avoid In Menu Engineering

How To Measure Success From Menu Engineering

A complete menu engineering cycle ends with every dish classified into a quadrant and a documented action plan for each. Success shows up in clear financial and operational signals.

  • Improved gross profit margin, as noted earlier, with Jelly customers typically seeing this improvement within the first quarter
  • Reduced food cost, with Jelly users cutting food costs by 3% on average in the first 3 months
  • Time saved on admin, with Jelly saving 10–20 hours of admin every month according to customer evidence
  • Faster reaction to price changes, such as catching a supplier increase in the same week rather than the same quarter

Amber, a Mediterranean restaurant in East London, saves £3,000–£4,000 per month using Jelly’s invoice automation, price alerts, and real-time costing. Populu lifted gross profit from 68% to 72% across 16 locations after connecting their POS through Jelly.

Advanced Tips For Multi-Site Operators

Multi-site operators face a compounding version of the same challenge. Ingredient prices move, POS data lives in separate systems, and by the time a manual analysis is complete the data already feels stale. Real-time menu engineering insights replace the 45 to 60-day lag between analysis and action that manual spreadsheet processes create.

Use these practical steps to keep multi-site menu engineering under control.

  • Use Jelly’s Price Alert to support supplier negotiations, because a sudden increase in a Plowhorse’s key ingredient price becomes visible immediately and gives you grounds to challenge the supplier and claim credit notes.
  • Create separate menus for dine-in and delivery, each with its own quadrant analysis and margin targets.
  • Standardise the method and recipe cards across all sites, then run the analysis per site so local performance differences stay visible.
  • Connect menu engineering to stock management so high-margin Stars never run out of stock.
  • Treat group averages with caution, because a dish that looks strong on a group report can underperform at a specific location.
  • Consider menu design and psychological pricing to nudge customers toward Puzzles and Stars. A 2012 eye-tracking study by Sybil Yang found that diners read menus top to bottom and left to right, so the first items in each section get the most attention.

Frequently Asked Questions

What Is A Menu Engineering Quadrant?

A menu engineering quadrant is one of four categories in the Kasavana–Smith menu engineering matrix. The matrix classifies every dish by two dimensions: popularity (sales mix within its category) and profitability (contribution margin, which equals selling price excluding VAT minus food cost per portion). The four quadrants are Stars (high popularity, high margin), Plowhorses (high popularity, low margin), Puzzles (low popularity, high margin), and Dogs (low popularity, low margin). Each quadrant has a standard set of recommended actions.

How Do I Calculate Popularity In Menu Engineering?

Calculate popularity as the number of times a dish is sold divided by the total number of items sold in the same menu category, multiplied by 100. For example, if a main course sells 150 times out of 1,000 total mains sold, its popularity is 15%. Use at least four weeks of normal trading data and exclude comps, voids, and weeks distorted by closures or events. Analyse each menu category separately, because mixing categories produces meaningless thresholds.

What Is The 70% Rule In Menu Engineering?

The 70% rule sets the popularity threshold for classifying dishes as high or low popularity. The formula is: (100 ÷ number of dishes in the category) × 0.70. For a mains section with 8 dishes, the equal share is 12.5% and the threshold is 8.75%. Any dish with a sales mix at or above 8.75% is high popularity, and any dish below that level is low popularity. The 70% discount exists because menus never sell evenly, so a dish slightly below its equal share can still contribute meaningfully to the category.

How Does UK VAT Affect Menu Engineering Calculations?

UK menu prices are VAT-inclusive, so the 20% standard rate already sits inside the price the customer pays. For menu engineering, strip VAT out before calculating contribution margin. The formula is: ex-VAT selling price = menu price ÷ 1.20. As shown in Step 1, a £15.00 menu price becomes £12.50 ex-VAT. Using the sticker price directly overstates margin by the £2.50 VAT element. VAT treatment can vary by product and channel, so check HMRC guidance for your specific menu items.

Can I Use Menu Engineering For Delivery Menus?

Menu engineering works well for delivery menus when you run a separate matrix. As noted in Step 6, platform commissions of 25–35% reduce the effective selling price of every dish. To calculate delivery contribution margin, take the ex-VAT selling price, subtract the platform commission and packaging costs, then subtract food cost per portion. Run a separate delivery matrix and apply the quadrant action playbook independently of your dine-in analysis.

How Often Should I Review Menu Engineering Quadrants?

The recommended cadence is quarterly, plus an additional review whenever a key ingredient price changes by more than about 10%, when new dishes launch, or at each seasonal menu change. After making changes, allow 4–8 weeks before re-running the analysis so you can measure the effect of each change. Avoid changing multiple variables at once, because repricing, renaming, and repositioning a dish in the same period makes it impossible to see which change worked.

What If I Have Multiple Sites?

Apply the method per site using site-specific POS data and ingredient costs. Standardise the method, recipe cards, and thresholds across all sites, and avoid averaging results across locations, because performance can differ sharply between sites. Look for consistent patterns across sites before making permanent menu changes, and use Jelly’s multi-site dashboard to track quadrant positions across all locations in real time.

How Does Jelly Help With Menu Engineering Quadrants?

Jelly’s Sales Mix feature integrates with your POS to pull item-level sales data and combine it with live dish costing from scanned invoices. Quadrant classifications then update in real time as sales accumulate and ingredient prices change, without manual spreadsheet work. Jelly’s Price Alert flags every ingredient price increase or decrease, giving you the data to act before a Star slides into Plowhorse territory. Jelly also supports separate delivery menu costing, so you can maintain distinct dine-in and delivery matrices from a single platform.

Conclusion: Turn Menu Engineering Into A Monthly Discipline

The value of this process comes from running all seven steps consistently. Operators who do this protect gross margin, make faster menu decisions, and build a system that scales with the business.

Regular quadrant reviews create clearer menu decisions and a repeatable rhythm for pricing, promotion, and dish development. The operators who gain the most treat menu engineering quadrants as a monthly discipline that sits alongside stock checks and labour reviews.

Jelly connects your invoices, recipes, and POS data in a single platform. It automates invoice scanning, live dish costing, and POS-integrated sales mix, so your quadrants update in real time and your team can act on accurate data every week rather than every quarter.

See Jelly in action for your menu