Stars, plowhorses, puzzles and dogs. A worked example, what to change in each quadrant, and menu design that moves the mix.
By Rob Bruni, founder of OpsEdge Hospitality, twenty years running restaurants, bars, clubs and resorts. Updated November 2026.
Menu engineering is the practice of deciding what stays on the menu, what gets repriced and where things sit on the page, based on two numbers per item: how much margin it makes and how often it sells. It is the highest-return food cost exercise there is, because it changes what guests order rather than just what you pay for it. This guide covers the data you need, the matrix, a worked example, the actions for each quadrant, and how to run it every quarter.
Most menus are built once and then edited by habit: a dish the chef likes stays, a dish that a regular asks for stays, prices go up a dollar across the board every year or two. Menu engineering replaces habit with a simple analysis. For every item, what is the gross profit in dollars when one sells, and how many sell? Put those two numbers on a grid and the decisions make themselves.
It matters because food cost percentage is an average of every plate you sell. If guests order more of your high-margin items and fewer of your low-margin ones, food cost falls and profit rises with no change to purchasing, portions or prices.
Plate cost is every ingredient at current invoice prices, including garnish, sides, oil and the bread basket. Use dollars, not percent. A $34 steak at a 40 percent food cost contributes $20.40. A $12 pasta at 22 percent contributes $9.36. The steak has the worse percentage and the better margin, and margin is what pays the rent.
Pull it from the POS product mix report for a period long enough to be representative, usually four to eight weeks. Compare items within their own category (entrees against entrees, appetizers against appetizers), because a side will never outsell an entree and should not be judged against one.
For each category, work out the average contribution margin and the average mix percentage. Every item is then high or low on each axis, which gives four quadrants.
| Quadrant | Margin | Popularity | What it is |
|---|---|---|---|
| Stars | High | High | Your best items. Protect them. |
| Plowhorses | Low | High | Guests love them, they do not make enough. Re-engineer. |
| Puzzles | High | Low | Profitable but nobody orders them. Promote or reposition. |
| Dogs | Low | Low | Neither. Remove or replace. |
The popularity threshold is usually set at 70 to 80 percent of the category average rather than the average itself, so that a category with one runaway seller does not push everything else into "low." If you have 10 entrees, the average mix is 10 percent, and a reasonable threshold is 7 to 8 percent.
Six entrees, four weeks, 2,400 entrees sold
| Item | Price | Plate cost | Margin | Sold | Mix | Quadrant |
|---|---|---|---|---|---|---|
| Ribeye | $38 | $15.20 | $22.80 | 420 | 17.5% | Star |
| Chicken parm | $22 | $6.60 | $15.40 | 560 | 23.3% | Star |
| Burger | $16 | $5.90 | $10.10 | 640 | 26.7% | Plowhorse |
| Salmon | $29 | $10.20 | $18.80 | 250 | 10.4% | Puzzle |
| Pasta primavera | $19 | $4.40 | $14.60 | 330 | 13.8% | Star (borderline) |
| Pork chop | $27 | $11.80 | $15.20 | 200 | 8.3% | Dog |
Average margin: $16.15. Average mix: 16.7%, threshold at 75% of average: 12.5%.
The burger is the most popular item on the menu and the lowest margin. Moving it from $16 to $17 adds $640 a month at the current volume, and almost no guest changes their order over a dollar on a burger. The salmon makes $18.80 every time it sells but only one guest in ten orders it, so it needs a better spot on the page and a server mention. The pork chop sells 50 a week at a middling margin and is tying up a protein that spoils; replace it with something that uses an ingredient you already stock.
Once the numbers say what you want to sell, the menu's job is to sell it. A few things that move the mix measurably:
Operations doing this for the first time usually see a one to two point drop in food cost and a 3 to 5 percent lift in average check within one cycle, with no change to purchasing. The Prime Cost System page covers the help available if you want the first pass done with you.
The free prime cost calculator shows what each point of food, beverage or labor cost is worth in dollars over a year for your operation. No signup.
Open the calculator All free toolsFree tool: the menu price calculator turns a plate cost into a menu price at your target food cost, with the margin on every plate. No signup.
Menu engineering is analysing every menu item by two numbers, its contribution margin in dollars and its popularity in the sales mix, then deciding what to keep, reprice, promote or remove and where to place each item on the menu. It raises profit by changing what guests order, not just what you pay.
They are the four quadrants of the menu engineering matrix. Stars are high margin and popular. Plowhorses are popular but low margin. Puzzles are high margin but unpopular. Dogs are neither. Each quadrant has a different action: protect, re-engineer, promote, remove.
Contribution margin, in dollars. A $34 steak at a 40 percent food cost makes more money per sale than a $12 pasta at 22 percent. Percentages are useful for tracking the whole operation, but item decisions should be made on dollars.
Quarterly. Re-cost every item against current invoices, pull eight weeks of product mix, rebuild the matrix and make no more than five changes, then measure four weeks later.
Seven to ten items per category is where most operations find the balance between choice and decision fatigue. Fewer items usually raise average check, reduce inventory and lower waste.
Yes, measurably. Position on the page, boxes, removing dollar signs and price alignment, two-line descriptions and a premium anchor item all shift the mix toward the items you place and describe well, typically by several points of mix for the items involved.
