Menu engineering: how to classify every menu item
Menu engineering ranks each dish on two axes — how well it sells and how much margin it keeps — and sorts the whole menu into four groups you act on differently. Here's the matrix, and how methodus builds it for you from live sales and real cost.
14-day trial · No credit card · Cancel anytime
Definition
Menu engineering is the practice of classifying every menu item by two measures — popularity (how much it sells) and profitability (the margin it keeps) — to decide what to promote, reprice, rework, or remove. Items fall into four groups: stars, plowhorses, puzzles and dogs.
Popularity line = (1 ÷ number of items) × 70% · Profitability line = average contribution margin (price − cost)
The four groups on the matrix
Plot popularity on one axis and profitability on the other, and every dish lands in one of four quadrants. Each one has a different move.
Stars
High popularity · High margin
Your winners: loved and profitable. Protect them — keep the recipe exact, feature them on the menu, and never let their cost quietly drift.
Plowhorses
High popularity · Low margin
Popular but thin. They pull people in but keep little. Re-engineer the cost (portion, yield, supplier) or reprice carefully — a small change moves a lot of covers.
Puzzles
Low popularity · High margin
Profitable but overlooked. The margin is there; the sales aren't. Rename, reposition, describe better, or have the team recommend them — small nudges, real upside.
Dogs
Low popularity · Low margin
Neither loved nor profitable. Rework them into something better or take them off — they cost prep, stock and menu space for little return.
How to do menu engineering, step by step
You don't need software to run the analysis once — you need two numbers per dish and two threshold lines. Here's the classic method (Kasavana & Smith), the one every hospitality-school version comes from.
- 1
List each item with units sold and contribution margin
For a menu section (say, mains), write down how many of each dish sold over the period and its contribution margin — ex-tax price minus real ingredient cost. Cost the edible portion, not the invoice price, or every number that follows is wrong.
- 2
Draw the popularity line
If all items sold equally, each would take 1 ÷ (number of items) of the mix. The convention sets the bar at 70% of that equal share: popularity line = (1 ÷ number of items) × 70%. A section of 10 dishes gives 7% — sell more than 7% of the section's covers and the dish is 'high popularity'.
- 3
Draw the profitability line
Average the contribution margin across the section. Any dish above the average is 'high margin', any below is 'low'. Note: menu engineering uses contribution margin in currency, not food cost percentage — a low-food-cost dish that nobody orders contributes little.
- 4
Place each dish and act on the quadrant
High popularity + high margin = star (protect). High popularity + low margin = plowhorse (re-cost or nudge the price). Low popularity + high margin = puzzle (reposition, rename, recommend). Low + low = dog (rework or cut).
Worked example
A 4-main section sells 400 covers. Popularity line = (1 ÷ 4) × 70% = 17.5% (70 covers). The burger keeps $9.20 margin on 160 covers — above average margin, above the line: a star. The risotto keeps $11 on 30 covers — high margin, below the line: a puzzle to push. The fish-and-chips keeps $5 on 150 covers — popular but thin: a plowhorse to re-cost. The salad keeps $4 on 60 covers — a dog to rework or drop.
methodus builds the matrix for you
Running it once by hand is fine. Running it every week, on current costs, across the whole menu, is where it breaks down — because popularity means pulling POS numbers and profitability means a real, current cost per dish. methodus already holds both, so the matrix builds itself — no spreadsheet, no month-old costs.
- 1
Popularity from your POS
Daily sales sync from your till, matched to each fiche — so popularity is what actually sold, not a guess.
- 2
Profitability from real cost
Every dish is costed from your live purchase prices, updated as invoices come in. Set the price against a target with the selling-price calculator →
- 3
The margin, watched
When a supplier price moves, a dish can slip from star to plowhorse overnight. Check the margin any time with the gross-margin calculator →
Menu engineering only works on real costs. If your ingredient costs are stale, start with how to calculate food cost percentage → — the number the whole matrix is built on.
Menu engineering FAQ
Menu engineering classifies each menu item by popularity (how much it sells) and profitability (the margin it keeps), sorting the menu into stars, plowhorses, puzzles and dogs. You then promote, reprice, rework or remove each item based on where it lands.
List each dish with its units sold and contribution margin (ex-tax price minus real cost). Set the popularity line at (1 ÷ number of items) × 70%, and the profitability line at the menu's average contribution margin. Place each dish above or below each line — that gives the star, plowhorse, puzzle or dog quadrant — then act accordingly.
There are two threshold lines, not one formula. Popularity: an item is a high-seller if its share of the menu mix beats (1 ÷ number of items) × 70%. Profitability: an item is high-margin if its contribution margin (price minus cost) beats the menu's average. Both are needed to place a dish on the matrix.
Stars are high-popularity, high-margin winners. Plowhorses sell well but keep little margin. Puzzles are profitable but under-sold. Dogs are neither popular nor profitable. Each quadrant calls for a different action — protect, re-engineer, reposition, or remove.
Profitability is the gross margin each item keeps: its selling price (ex-tax) minus its real ingredient cost. That cost must include yield loss and current purchase prices — a dish looks profitable until the supplier price moves and nobody updates the recipe.
Review it whenever the menu changes and at least quarterly — high-volume operators run it monthly. The classification drifts as supplier prices move and seasons change, so a star can quietly become a plowhorse between reviews if no one is watching the cost side.
No — you can run it once by hand with a spreadsheet. The problem is keeping it current: popularity needs live POS data and profitability needs up-to-date costs. methodus syncs both, so the star/plowhorse/puzzle/dog matrix stays current instead of being a spreadsheet you rebuild each month.
Who builds methodusmethodus is built by an operator. Nathaniel Gilliand, École hôtelière de Lausanne graduate, runs restaurants, bars and beach clubs. This is the tool he built to find the margin that was evaporating in his own P&L.
14-day trial, no credit card.
See which dishes are stars
methodus reads your sales and your real costs and shows you the matrix — which plates to protect, fix, or drop. Start free.
14-day trial · No credit card · Cancel anytime
Cookies & measurement. Essential cookies run methodus.bar. Analytics & ads only with your OK. Details in the privacy policy.