Menu engineering: four boxes that tell you what to change

Menu engineering sounds academic. It is an hour of work with a spreadsheet and your own sales data, and it usually finds two dishes you should remove and one you should promote.

What you need

From your POS, for the last 90 days, per dish:

• Units sold
• Selling price
• Food cost per unit

If you cannot get food cost per dish, you need recipes costed first. That is the real work, and it is worth doing once: without it, every conclusion below is a guess.

The calculation

Contribution margin = selling price − food cost. Not a percentage — riyals. A dish at SAR 45 with SAR 15 of food cost contributes SAR 30.

Popularity = units sold ÷ total units sold across the menu.

Then compare each dish to the menu averages and you get four groups.

The four boxes

Stars — popular and high contribution

Instruction: protect them.

Do not change the recipe to save money. Do not move them on the menu. Make sure the kitchen never runs out of the ingredients. These dishes are why people come back.

Plough-horses — popular and low contribution

Instruction: fix the cost, not the price.

People love them; they earn you little. Options in order of preference:

1. Reduce portion cost through better purchasing, not smaller portions people notice
2. Reduce waste in preparation
3. Pair them with high-contribution sides and drinks
4. Raise the price by a small amount — popular dishes usually absorb a modest increase

Puzzles — unpopular and high contribution

Instruction: sell them harder before you remove them.

These earn well and nobody orders them. Usually the cause is placement, description or price anchoring, not the dish. Try:

• Moving them to a prominent position
• Rewriting the description
• Staff recommendation — genuinely the most effective lever
• A photograph

Give it four weeks. If nothing moves, they become dogs.

Dogs — unpopular and low contribution

Instruction: remove them.

Every dish on a menu costs you: ingredients that expire, prep time, kitchen complexity and menu space. A dog earns nothing and consumes all four.

The common objection is "but one customer always orders it". One customer is not a business case. If they are a regular, tell them personally — most people are flattered rather than offended.

The Saudi specifics

Family sharing changes the unit. Where dishes are ordered for the table, count by table rather than by cover, or your popularity numbers mislead.

Delivery menus need their own analysis. Commission and packaging change the contribution completely — a star in the dining room can be a dog on delivery. Seedelivery commission maths.

Seasonality is strong. Ramadan, school terms and the summer shift patterns significantly. Run this on comparable periods, not across a Ramadan boundary.

Drinks carry the margin. In most Saudi cafés and restaurants, beverages are the highest-contribution category by a wide margin. If your analysis is only about food, you are missing the part that pays the rent.

How often

Quarterly is right for most restaurants. Monthly is too noisy; annually is too slow to catch an ingredient price change.

Run it after any significant supplier price change too — a 20% increase in one ingredient can move several dishes from plough-horse to dog without anyone noticing.

What your POS needs to give you

• Units sold per dish over a date range
• Cost price per dish, from costed recipes
• The ability to separate channels — dine-in, takeaway, each delivery platform
• Export to a spreadsheet, because this analysis is easier in a spreadsheet

If your system cannot produce cost per dish, that is the gap to close first. See thereports feature.