What menu engineering is
Menu engineering is a method for deciding what to do with each dish on the menu using two numbers you already have: how many times it sold, and how much money each sale left after the cost of the food on the plate. The method was developed by Michael L. Kasavana and Donald I. Smith at the Michigan State University School of Hospitality Business, circa 1982, according to the American Hotel and Lodging Educational Institute.
The idea is simple. A dish that sells a lot is not necessarily a good dish for the business, and a dish that earns a lot per plate is not much use if nobody orders it. Putting both numbers side by side tells you which lever to pull.
The menu engineering matrix
Draw two lines: one for popularity, one for contribution margin. Every dish lands in one of four boxes.
| Group | Popularity | Margin | What it means |
|---|---|---|---|
| Star | High | High | Sells and earns. Protect it. |
| Plowhorse | High | Low | Guests love it; it does not earn its keep per plate. |
| Puzzle | Low | High | Earns well when it sells, which is not often. |
| Dog | Low | Low | Takes space, prep, and inventory for little return. |
Contribution margin is the menu price minus the plate cost: the money each sale leaves to cover labor, rent, and everything else. It is a dollar figure, not a percentage, and that is the point. A steak with a high food cost percentage can still leave more dollars per plate than a pasta with a low one.
How to compute it from POS sales and plate cost
- Pick a period and a section. Pull item sales from your POS for a few weeks that did not change menus or seasons. Compare dishes that compete with each other: mains against mains, not mains against desserts.
- Cost every plate. Recipe by recipe at current invoice prices, including garnish and sauce. The menu item price calculator helps you check price against cost one dish at a time.
- Work out each dish’s contribution margin. Price minus plate cost, then multiply by the number sold for the dish’s total contribution.
- Set the margin line. We use the average contribution per plate across the section: total contribution divided by total plates sold. Some versions use a simple average of the dishes’ margins instead. Pick one and stay with it.
- Set the popularity line. A common threshold is 70 percent of an even share. With five mains, an even share is 20 percent of sales, so the line is 14 percent. Anything above it counts as popular.
Here is a worked example. The dishes and numbers are made up for illustration; use your own.
| Dish | Sold | Price | Plate cost | Margin | Group |
|---|---|---|---|---|---|
| Chicken piccata | 120 | $26 | $7 | $19 | Star |
| Burger | 180 | $18 | $6 | $12 | Plowhorse |
| Braised short rib | 60 | $34 | $12 | $22 | Puzzle |
| Grilled halibut | 40 | $36 | $16 | $20 | Puzzle |
| Mushroom risotto | 30 | $19 | $5 | $14 | Dog |
That is 430 plates and $6,980 in total contribution, so the margin line is about $16.23 a plate. The popularity line is 14 percent of 430, about 60 plates. The short rib sold exactly 60, just under the line, which is a good reminder that the lines are a convention, not a verdict.
What to do with each quadrant
- Stars. Keep the recipe and portion steady. Give them good placement near the top of their section, and do not bury them under a longer description than they need.
- Plowhorses. Guests want them, so move carefully. Take a small price increase, trim a costly side, or pair the dish with a higher-margin add-on. Do not move them to the best spot on the menu; they sell anyway.
- Puzzles. They earn well but are not being chosen. Move them higher in the section, rewrite the description so a guest knows what arrives, rename them if the name confuses, and have servers mention them. Sometimes the price is simply too high for the room.
- Dogs. Cut them, replace them, or rework them into something that shares prep with a star. If a small group of regulars orders one, it can live on as an off-menu request.
Where the numbers mislead
The matrix is only as good as the costing. A plate cost that is three months old will put dishes in the wrong box. It also ignores labor: a dish that takes twelve minutes of a cook’s attention can look like a star and still be a problem on a full night. And it treats each dish alone, when a starter that nobody orders may be the thing that brings a certain table back.
Change a few dishes at a time, then run the numbers again over a similar period. If you change everything at once, you will not know which change did the work. For deciding the overall size of the menu, see how many items a menu should have.
Where Gaslamp fits, and where it does not
Gaslamp does not do the analysis. It has no POS connection and no sales data, and it does not store plate costs, so it cannot tell you which dish is a star. Pull those numbers from your POS and your costing sheets.
It is where you act on the answer. In the editor, which runs in your phone’s browser, you can drag dishes up or down within a section and reorder the sections themselves, rewrite a puzzle’s description, change a plowhorse’s price with the number keypad, and switch a dog off. Tap Save & update live menu and the QR menu shows the change about thirty seconds after you save. Version history holds the last 50 saves, so a change that does not work can be rolled back. When you are ready for new paper, download a print-ready PDF on a paid plan. More in menu management software and how to write menu descriptions.