Menu Engineering Calculator (Stars, Plowhorses, Puzzles, Dogs)

How this calculator works

Menu engineering is the analytical framework that classifies every item on a restaurant menu into one of four quadrants based on (a) how popular the item is relative to the rest of the menu and (b) how profitable the item is relative to the rest of the menu. The framework was published by Donald Smith and Michael Kasavana in 1982 and has since become the operator standard for menu analysis across full-service, fast-casual, and QSR concepts. The classification drives the per-item action plan: promote stars aggressively, re-engineer plowhorses, reposition puzzles, cut or rebuild dogs.

The calculator takes a list of menu items where each entry carries a food cost, a sell price, and a unit count sold during the analysis period. For each item it computes the contribution margin in dollars (sell price minus food cost), the total contribution margin for the period (CM times units sold), and the food-cost percentage. Across the menu it computes the mean units sold per item, the mean contribution margin per item, and the popularity benchmark (mean units multiplied by 0.70). Each item is then placed in one of four quadrants based on the comparison of its unit count against the popularity benchmark and its CM against the mean CM. The quadrants drive the recommended action per item.

This is an operating diagnostic. It is not professional advice. For consequential menu rewrites, pricing actions, or recipe-cost contract decisions, consult a restaurant-specific operating consultant or a CPA familiar with hospitality cost accounting.

The Kasavana / Smith framework

The four quadrants are defined by two binary tests: high or low popularity and high or low profitability. The popularity test compares item units sold against the popularity benchmark, defined as 70% of the mean units sold across the menu. The profitability test compares item contribution margin against the mean contribution margin across the menu.

STAR — high popularity, high contribution margin. The cornerstone of the menu. Stars carry their weight in both volume and per-unit dollars. The action is to promote stars aggressively through menu placement (prime real estate in the section), server scripting (lead with stars when guests ask what is good), and supplier-relationship protection (the stars depend on the recipe and the ingredients staying stable). A typical full-service menu has 15-25% of items in the star quadrant.

PLOWHORSE — high popularity, low contribution margin. The most actionable quadrant in the framework. Plowhorses move volume but leave money on the table. The volume is already there; the operator does not need to grow demand, only lift the CM dollars per unit. Standard fixes: portion reduction (the entrée moves to a slightly smaller plate at the same price), accompaniment substitution (swap a high-cost side for a moderate-cost side), modest price increase (3-5% rarely depresses demand on a popular item), or a partial recipe rebuild that preserves the visible plate while compressing the cost line. A typical full-service menu has 20-30% of items in the plowhorse quadrant; this is where most menu-engineering quick wins live.

PUZZLE — low popularity, high contribution margin. Puzzles have the margin to be valuable but guests are not ordering them. The fix is a merchandising and placement problem, not a pricing problem. Standard moves, in order: improve menu placement, train servers to suggest the item, rewrite the name and description, feature the item as a weekly special. If none of these moves the item out of the puzzle quadrant after one or two analysis cycles, the item is structurally a dog with a misleading CM and should be cut.

DOG — low popularity, low contribution margin. Most dogs should be removed at the next menu rewrite. They occupy menu real estate, kitchen inventory, prep time, and recipe-card maintenance without paying for any of it. A small number should be retained for strategic reasons: a vegetarian entrée for category coverage, a kids' menu item for table-filling, a signature legacy item that anchors brand identity. Strategic dogs should be a small share of the menu, typically 5-10% of items.

Inputs explained

Item name. Display label only; does not enter the math. Use the operator's internal menu item name for consistency with POS reports and recipe cards.

Food cost per unit (dollars). The fully-loaded food cost of one serving of the item, computed from the recipe card. Include every ingredient on the plate (the protein, the starch, the vegetable, the sauce, the garnish, the bread), the cooking oil, the salt-and-pepper, and the disposables (the takeout container, the napkin, the cutlery) that the operator pays for and the guest consumes. Do not include labor or overhead — those are not part of the menu-engineering CM calculation. Recipe cards in most POS or back-office platforms compute this automatically once the ingredient list and yield percentages are entered.

Sell price per unit (dollars). The menu price the guest pays before tax and tip, net of any promotional discount the operator is currently running. Use the regular menu price for the analysis even if the item is being promoted, because the analysis is meant to drive the menu rewrite, not the promotional calendar.

Units sold during the analysis period. Total units moved during the analysis period. The standard analysis period for full-service is one month or one quarter; for QSR it is one week or one month. The same period must apply to every item in the analysis — mixing a month of entrée sales with a week of dessert sales produces benchmarks that misclassify both.

Industry benchmarks and typical results

A well-run, mature full-service menu typically produces a quadrant distribution in the range of 20% STARS, 25% PLOWHORSES, 25% PUZZLES, 30% DOGS at the first analysis. After two or three quarterly menu rewrites driven by the framework, a healthy distribution shifts toward 30% STARS, 30% PLOWHORSES, 25% PUZZLES, 15% DOGS — fewer dogs, more stars, and a stable plowhorse bench that the operator continues to optimize. A distribution heavier than 40% DOGS typically signals a menu that has been built by accretion (add an item, never remove one) and is overdue for a rewrite; a distribution lighter than 10% PUZZLES typically signals a menu where the operator has been afraid to take a margin risk on creative items.

Overall food-cost percentage at the menu level (total food cost across all items divided by total revenue across all items) should land in the 28-32% range for full-service, 25-30% for QSR, and 32-38% for fine dining. The component analysis surfaces which items are driving the food-cost result — a 30% overall food cost masking a few 50% items the operator never noticed, for example. The total contribution margin in dollars is the figure that ultimately pays the labor, rent, and overhead lines, and it is the figure to track quarter over quarter as the menu evolves.

What this calculator does NOT model

Several material parts of menu and operating economics fall outside the menu-engineering quadrant analysis:

Labor cost. The framework deliberately excludes labor because labor is fixed against the schedule, not the menu mix. The operator does not avoid labor cost by selling a different item, only by changing the schedule. Labor analysis belongs in the prime-cost diagnostic and the labor-scheduling model.

Overhead allocation. Rent, utilities, marketing, insurance, debt service, and depreciation are below the CM line. A complete per-item profitability analysis would allocate overhead per unit; the menu-engineering framework intentionally stops at CM because per-unit overhead allocation requires assumptions (per-cover allocation, per-revenue-dollar allocation, per-square-foot-of-kitchen allocation) that vary by concept and add complexity without changing the operator action.

Plate visibility and check-average effect. A signature appetizer that is photographed by every diner and drives the check average through a beverage attach is a menu fixture even if its quadrant classification is mediocre. The framework does not capture this category of strategic value; the operator should overlay strategic judgment on the mechanical quadrant output before cutting items.

Substitution dynamics. Cutting a dog may push some of its (limited) demand to a star or a plowhorse and lift overall menu economics, or it may push that demand off the menu entirely and the guest walks. The framework treats each item in isolation; substitution effects require a more detailed model and operator judgment.

Sales velocity by daypart. The framework analyzes total period units. A breakfast-only item with strong morning velocity may look like a dog against a menu of all-day items even when it is a star in its daypart. For multi-daypart concepts, run the analysis per daypart.

For any of the above, and for consequential menu pricing decisions on premium items or supplier contracts, consult a restaurant-specific operating consultant or a hospitality-trained CPA.

Sources

• Kasavana, M. and Smith, D., "Menu Engineering: A Practical Guide to Menu Analysis," Hospitality Publications, 1982. The original formulation of the four-quadrant framework with the 70% popularity benchmark and the mean-CM profitability benchmark.
• Cornell School of Hotel Administration — Hospitality Quarterly. Periodic peer-reviewed revisits of the menu-analysis methodology, including alternative threshold studies that have generally affirmed the 70% baseline.
• Pavesic, D., "The Psychology of Menu Pricing." Cornell Hospitality Quarterly. Operator-facing analysis of menu placement, pricing psychology, and the merchandising fixes appropriate to each quadrant.
• National Restaurant Association — annual Restaurant Industry Forecast, for the concept-level food-cost and CM benchmarks used in the industry-benchmarks section.

Last reviewed: 2026-05-16 against the Kasavana and Smith 1982 framework, Cornell Hospitality Quarterly menu-analysis methodology, and National Restaurant Association industry benchmarks current as of 2026.