Menu engineering: which dishes actually make money

In most restaurants the menu is just a list of dishes. It is actually a revenue tool. Menu engineering means scoring every item on two numbers and rebuilding the menu around the result.

The two numbers

1. Contribution margin

Not the food cost percentage — the margin in money:

Margin = selling price − ingredient cost

Why not the percentage? Because the till collects money, not percentages. Tea runs an 11% food cost but leaves 7,000 UZS per pot. Kebab runs a 50% food cost but leaves 16,000 UZS per skewer. Judge by percentage alone and you will drop the wrong dish.

2. Units sold

Portions sold over a period, usually a month. This is the guest voting.

ABC analysis

Start by sorting dishes by revenue and calculating the cumulative share:

• Group A — the dishes producing roughly 80% of revenue, usually 20% of the menu. Work here pays off fastest.
• Group B — the next ~15%. Keep them under observation.
• Group C — the final ~5%. They bloat the menu and create stock and waste.

Group C items are often kept alive for one ingredient, and that ingredient is the one that keeps expiring in the walk-in. That is the first benefit of ABC analysis: it shows what can safely leave the menu.

The matrix: star, plowhorse, puzzle, dog

Now draw two lines: average margin and average units sold. For the sales threshold, 70% of the average is the usual cut-off. That gives four quadrants:

• Star — high margin, high volume
• Plowhorse — low margin, high volume
• Puzzle — high margin, low volume
• Dog — low margin, low volume

A worked example with seven items

One month of illustrative data:

• Plov — 38,000 / cost 13,000 / margin 25,000 / 1,200 portions
• Lagman — 55,000 / 18,000 / margin 37,000 / 480 portions
• Kebab (skewer) — 32,000 / 16,000 / margin 16,000 / 1,500 units
• Samsa — 12,000 / 4,500 / margin 7,500 / 900 units
• Caesar salad — 45,000 / 20,000 / margin 25,000 / 180 portions
• Pot of tea — 8,000 / 900 / margin 7,100 / 1,400 units
• Dessert — 25,000 / 14,000 / margin 11,000 / 90 portions

Average margin ≈ 18,400 UZS. Average volume ≈ 821, so the threshold at 70% ≈ 575. The result:

• Star: plov (25,000 / 1,200)
• Plowhorses: kebab, samsa, tea — high volume, thin margin
• Puzzles: lagman (37,000 / 480), Caesar (25,000 / 180)
• Dog: dessert (11,000 / 90)

What to do in each quadrant

Stars — protect them

Do not touch the quality or the portion size. Keep the item in the most visible position on the page. Raise the price carefully, in steps of no more than 5-7% — demand absorbs it.

Plowhorses — raise the margin

They bring guests in but leave little behind. Options: cut the cost (supplier, portion discipline), attach add-ons (bread, sauce, a drink), or raise the price quietly. On kebab, 1,500 units × 2,000 UZS is an extra 3M UZS of profit a month.

Puzzles — raise the volume

The margin is good, but guests are not ordering. The cause is usually one of three: the item is buried on the page, the description is weak, or servers never suggest it. Move lagman to the top of the page, write a proper description, give servers a line to say. Cutting the price is the last resort.

Dogs — remove them

Low volume, low margin, but they still consume purchasing and attention. The exception is a baseline item guests expect to see, such as a children's dish.

Putting the matrix onto the menu page

So the analysis does not stay on paper, the result has to move onto the menu itself:

• Move stars and puzzles to the top. Guests do not read a page end to end — the first two or three items and the last one get seen most.
• Keep 5-7 items per category. The wider the choice, the more likely a guest defaults to the most familiar dish, which is not necessarily your most profitable one.
• Do not line prices up in a column. A column of prices invites comparison and the eye lands on the cheapest row. Put the price at the end of the description instead.
• Give photographs only to stars. Every image on the page pulls attention; spend it on a dish with margin, not on a dog.
• Write a line for the servers. One sentence about a puzzle item often works faster than reprinting the menu.

Where to start

1. Export the last 30 days of sales: item name and quantity.
2. Attach ingredient cost from the recipe cards.
3. Calculate the margin column, then the average margin and sales threshold.
4. Sort the dishes into the four quadrants.
5. Make only five changes — not everything at once.
6. Recalculate after 30 days and see what moved.

One condition matters: the analysis has to be repeated. A one-off menu engineering exercise is stale within three months because purchase prices move. Keeping sales volume and ingredient cost in the same place is normally the POS system's job — it calculates margin from the recipe card and ranks the dishes for you.