Quick answer
Menu engineering is how you design a menu that sells: you classify every dish by crossing popularity with margin, then use that map to push the right items, rework the weak ones and cut the dead weight. With smart placement and pricing you can lift your average check by 10-15% without changing prices. You don't need software, just your POS data and a spreadsheet.
What "a menu that sells" really means
A menu isn't a list of dishes, it's the most important sales tool in the room. A guest looks at it for 90-120 seconds, and in that window they decide how much they'll spend. Designing it means steering that decision toward the dishes that work for you, without the guest feeling pushed.
Menu engineering was born in the United States in the 1980s and is now the backbone of organized foodservice. The logic is simple: not every dish contributes equally to profit. Some sell a lot but earn little, others earn a lot but nobody orders them. Your job is to make those two things meet.
The two numbers that matter
For each dish you need only two figures:
- Popularity: how often it's ordered relative to other dishes in the same category (starters, mains, desserts).
- Contribution margin: selling price minus ingredient cost, in currency, not as a percentage.
That last point is critical: menu engineering runs on margin in money, not food-cost percentage. A dish at 35% food cost that leaves $12 per plate beats one at 22% that leaves $4. Absolute margin is what pays rent, payroll and profit.
You'll need the per-portion margin for every dish. If your recipe cards don't already show it, build it line by line with the dish margin calculator; to model what a menu change does to your whole cover, use the menu margin projection.
The four-quadrant matrix
Crossing popularity and margin gives you four groups. Each dish is compared with the average of its own category, not the whole menu.
| Quadrant | Popularity | Margin | What to do | |---|---|---|---| | Star | High | High | Protect and feature | | Plowhorse | High | Low | Recover margin without losing guests | | Puzzle | Low | High | Push with placement and copy | | Dog | Low | Low | Rework or remove |
A main ordered above the mains average and with margin above average is a Star. If it sells like crazy but margin is below average, it's a Plowhorse: it brings people in but eats your profit.
A worked example on five mains
| Main | Sales/month | Cost | Price | Margin | Quadrant | |---|---|---|---|---|---| | Carbonara | 180 | $2.80 | $12.00 | $9.20 | Star | | Cacio e pepe | 160 | $1.90 | $11.00 | $9.10 | Star | | Amatriciana | 140 | $3.20 | $12.00 | $8.80 | Plowhorse | | Truffle risotto | 40 | $5.50 | $18.00 | $12.50 | Puzzle | | Veg pasta | 30 | $2.40 | $10.00 | $7.60 | Dog |
Category sales average: 110 portions. Margin average: $9.44. Dishes above both thresholds are Stars, dishes below both are Dogs, the rest split across the two mixed quadrants.
The carbonara is the engine of the menu: leave it alone. The truffle risotto is a Puzzle, great margin but low sales, worth pushing. The veg pasta is a Dog: either rebuild it or drop it.
What to do, quadrant by quadrant
Star — Don't touch the recipe, give it the best position, don't lower the price. Watch costs: if an ingredient jumps, a Star can slide into Plowhorse territory.
Plowhorse — The trickiest group. Cutting them upsets guests, keeping them as-is erodes profit. Trim the portion slightly (120 g to 110 g of pasta), swap a costly ingredient for an equivalent cheaper one, raise the price 0.50-1.00 if the market holds, pair it with a high-margin side or wine.
Puzzle — The margin is there, the pull isn't. Rename it with a more evocative name, rewrite the description telling the ingredient and origin story, move it to a visible spot, have servers recommend it, run it as the special.
Dog — Takes up space with no return. Remove it or rebuild it completely. If you keep it for completeness (a vegetarian option, say), at least raise its margin.
Visual placement: where guests actually look
Eye-tracking shows the gaze follows predictable paths.
- Two-page open menu: the eye goes top-right first, then top-left, then center. The top-right corner is the "sweet spot", put a Star or the Puzzle you want to push there.
- Single-page menu: read top to bottom. The first two dishes in each category get the most attention.
Practical rules:
- Highlight only high-margin dishes with boxes, icons or frames.
- Never highlight more than 2-3 items per page, too many cues cancel out.
- Keep 7-10 dishes per category. Long lists paralyze choice and push guests back to the usual suspects.
Pricing psychology
How you write the price changes how much a guest spends.
- Drop the currency symbol: "12" reads lighter than "$12.00". Common in mid and upper tiers.
- Don't right-align prices in a column: columns make guests compare numbers and pick the lowest. Put the price at the end of the description.
- Decoy effect: one very expensive dish in the same category makes the others feel accessible. The $18 risotto makes the $12 carbonara look like a deal.
- Anchoring: the first price a guest sees becomes the reference. Open with a $16 dish and everything at $11-13 after it feels reasonable.
Common mistakes
- Thinking in percentages instead of money. Chasing low food cost pushes dishes that leave few dollars per plate.
- Changing the menu on a hunch. Without POS data you decide on a server's impression, not facts.
- Highlighting everything. Feature ten dishes and you feature none.
- Endless menus. More items mean more waste in the kitchen and more indecisive guests.
- Raising prices as the first move. It's the last lever, not the first: work portions, pairings and placement first.
- Running the analysis once. Costs move: revisit the matrix every 3-4 months.
Related resources
- Dish margin calculator — work out the contribution margin of each dish from its ingredient cost.
- Menu margin projection — model how a menu change affects your average check and total margin.