Menu Engineering and Pricing
A menu is not just a list of dishes — it is the single most powerful sales and profit tool a restaurant owns. Every guest reads it, every server sells from it, and every item on it either earns money or quietly drains it. Menu engineering is the disciplined study of which items do which, so you can redesign the menu to steer guests toward the dishes that are both popular and profitable. Done well, it can lift a restaurant's bottom line by several percentage points without raising a single price or changing a single recipe.
In this lesson you will learn the classic four-box classification (stars, plowhorses, puzzles and dogs), how to calculate the contribution margin that drives it, and the main pricing methods a menu planner uses. You will also see where the whole discipline came from — a 1982 breakthrough by two hospitality academics that turned menu design from art into measurable science.
Learning Objectives
- Define menu engineering and explain why contribution margin, not food-cost percentage alone, drives profit.
- Calculate an item's contribution margin and menu mix (popularity) percentage.
- Classify menu items as stars, plowhorses, puzzles or dogs using the Kasavana-Smith matrix.
- Recommend the correct management action for each category.
- Compare the major menu pricing methods and select an appropriate one.
- Recognise common errors in interpreting menu-engineering data.
Quick Answer
Menu engineering evaluates every item on two axes: profitability (contribution margin — selling price minus food cost) and popularity (its share of items sold, or menu mix). Plotting items against the averages of these two measures creates four categories. Stars are high-profit and high-popularity — feature and protect them. Plowhorses are popular but low-profit — try to raise their margin gently. Puzzles are profitable but unpopular — reposition or promote them. Dogs are low on both — usually remove or rework them. Pricing methods (cost-plus/factor pricing, contribution-margin pricing, and demand/value-based pricing) then set prices that protect margin while remaining acceptable to guests. The goal is to shift the whole menu mix toward high-contribution items.
Where It Came From
For most of restaurant history, menu design was guided by chef's intuition and a single blunt metric: food-cost percentage. Operators obsessed over keeping food cost around 28–35% of the selling price, assuming that a low percentage automatically meant good profit. This hid a dangerous flaw. A $4 side salad with a 20% food cost looks "efficient," but it only contributes $3.20 in cash margin. A $32 steak at 40% food cost looks "expensive" to run, yet contributes $19.20 — six times more actual money to pay the rent and staff. Percentages do not pay bills; dollars do.
The breakthrough came in 1982, when Michael Kasavana and Donald Smith, both academics associated with Michigan State University's School of Hospitality Business, published Menu Engineering: A Practical Guide to Menu Analysis. They argued that menu items must be judged on two dimensions at once — how much cash each item contributes and how well each item sells — rather than on cost percentage alone. Their now-famous four-quadrant matrix gave managers a shared vocabulary (stars, plowhorses, puzzles, dogs) and a repeatable procedure. The motivation was intensely practical: thin restaurant margins meant that even a small, data-driven shift in the sales mix could be the difference between profit and closure. Their method drew on the same portfolio thinking as the Boston Consulting Group growth-share matrix (1970), which also used a "star / dog" language, but Kasavana and Smith adapted it specifically to the mechanics of a restaurant menu. It remains the industry standard more than four decades later.
The Two Measures That Drive Everything
Menu engineering rests on two numbers you must be able to calculate.
1. Contribution margin (CM). This is the cash each sale contributes toward fixed costs and profit:
Contribution margin = Selling price − Item (portion) food cost
Note it uses only the food cost of the plate, not labour or overhead. If a burger sells for $14 and its ingredients cost $4.50, its contribution margin is $9.50. Every burger sold puts $9.50 into the pot before rent, wages and utilities are paid.
2. Menu mix percentage (popularity). This is the item's share of total items sold within a category:
Menu mix % = (Number of that item sold ÷ Total items sold) × 100
If you sold 300 mains in a week and 60 were the burger, the burger's menu mix is 20%.
To classify items you compare each against a benchmark average:
- Average contribution margin = total contribution margin of all items ÷ total number of items sold.
- Popularity threshold — Kasavana and Smith set this not at a simple average but at 70% of the "equal share." If a menu has 10 items, an equal share would be 10% each; the threshold is 70% of that, i.e. 7%. Any item selling above 7% is considered "high popularity." The 70% rule prevents a single blockbuster from unfairly dragging the bar so high that every other item looks unpopular.
An item is "high" or "low" on each axis relative to these two benchmarks, and that placement decides its quadrant.
The Four Categories (Kasavana-Smith Matrix)
| Low popularity | High popularity | |
|---|---|---|
| High contribution margin | Puzzle | Star |
| Low contribution margin | Dog | Plowhorse |
Stars — high profit, high popularity. These are the menu's champions: guests love them and each sale earns strong margin. Action: protect the recipe and quality rigidly, keep them in prime "eye-magnet" menu positions, and test small price increases carefully (loyal demand can often absorb a modest rise). Never bury a star at the bottom of a list.
Plowhorses — low profit, high popularity. Workhorses that sell in volume but contribute little each. Think of the crowd-pleasing pasta or the value burger. Action: raise contribution margin without killing demand — a small price nudge, a modest portion reduction, cheaper garnish, cross-selling a high-margin side or drink, or reducing the item's visual prominence so it does not cannibalise better items. Be cautious: because these are price-sensitive favourites, aggressive price rises can drive guests away.
Puzzles — high profit, low popularity. They make good money but few people order them. Action: the puzzle is a selling and positioning problem. Rename or rewrite the description to sound more appealing, move it to a highlighted box or "chef's recommendation," train servers to suggest it, or lower the price slightly to widen appeal. If it stubbornly refuses to sell despite effort, it may need retirement.
Dogs — low profit, low popularity. Unloved and unprofitable. Action: usually remove them; they clutter the menu, complicate the kitchen and tie up inventory. Exceptions exist — a dog may stay if it is a necessary option (a plain kids' dish, a vegetarian staple), a signature that supports brand image, or uses ingredients shared with a star so removing it wastes nothing.
Worked example
Suppose a small bistro sells four mains in a week:
| Item | Sold | Price | Food cost | CM/unit | Total CM |
|---|---|---|---|---|---|
| Ribeye steak | 40 | $34 \vert $14 | $20 \vert $800 | ||
| House burger | 120 | $15 \vert $5 | $10 \vert $1,200 | ||
| Seafood risotto | 20 | $28 \vert $9 | $19 \vert $380 | ||
| Veg curry | 20 | $16 \vert $8 | $8 \vert $160 |
Total sold = 200. Total CM = $2,540. Average CM = $2,540 ÷ 200 = $12.70.
Popularity threshold with 4 items: equal share = 25%; 70% of that = 17.5%. So an item is popular if its mix is above 17.5% (i.e. more than 35 sold out of 200).
- Ribeye: CM $20 (above $12.70 = high); mix 20% (above 17.5% = high) → Star.
- Burger: CM $10 (below average = low); mix 60% (high) → Plowhorse.
- Seafood risotto: CM $19 (high); mix 10% (low) → Puzzle.
- Veg curry: CM $8 (low); mix 10% (low) → Dog.
The action plan writes itself: protect the ribeye, nudge the burger's margin (perhaps a $1 rise or a cheaper bun contributes an extra $120/week on 120 covers), promote the risotto with a better description and server push, and consider replacing or reworking the curry.
Menu Pricing Methods
Classification tells you what to do; pricing methods tell you how to set the number.
1. Factor (mark-up) pricing. Set a target food-cost percentage, then multiply the item's food cost by a pricing factor = 100 ÷ target%. For a 30% target, factor = 3.33. A dish costing $5 → $5 × 3.33 = $16.65. Simple and fast, but it blindly protects percentage rather than cash margin, so it can under-price expensive proteins and over-price cheap items.
2. Contribution-margin pricing. Add a target dollar contribution to the food cost. If the operation needs an average $11 contribution per cover to cover overhead and profit, a dish costing $5 is priced at $16. This aligns pricing directly with the metric menu engineering cares about, and is excellent for reworking plowhorses.
3. Prime-cost pricing. Includes both food cost and the direct labour cost of preparing an item before applying a mark-up — useful for labour-intensive dishes (hand-made pasta, elaborate desserts) that factor pricing would under-price.
4. Demand-based / value pricing. Sets price by what guests will pay and how they perceive value, not by cost. Fine-dining tasting menus, "market price" seafood and psychological pricing ($14.95 vs $15, or dropping the currency symbol entirely) live here. Powerful for stars and puzzles where perceived value is high.
Most real menus blend these: cost methods set a floor, demand methods set the ceiling, and menu engineering data decides which items to push toward the ceiling.
Real-World Applications
- Menu redesign / layout: engineering data drives the "golden triangle" and eye-flow placement — stars and puzzles go in high-attention zones, plowhorses and dogs are de-emphasised.
- Costing discipline: the exercise forces accurate, up-to-date recipe costing, which in turn tightens purchasing and portion control (links to
../../7._Hotel_Accounting/index.md). - Seasonal engineering: running the matrix each quarter catches items that drift categories as costs and tastes change.
- Upselling and training: servers can be scripted to convert plowhorse orders into higher-margin combinations, and to recommend puzzles.
- Multi-outlet chains: standardised menu-engineering reports let head office compare outlets on mix quality, not just revenue.
Common Mistakes
-
Judging items by food-cost percentage alone. Why wrong: a low percentage can still mean tiny cash contribution. Correction: always compute contribution margin in dollars — a "high food-cost" steak often out-earns a "cheap" salad.
-
Deleting every dog immediately. Why wrong: some dogs anchor the brand, serve a needed dietary segment, or share ingredients with stars, so removing them costs sales elsewhere. Correction: investigate the dog's strategic role before cutting; try repositioning first.
-
Raising a plowhorse's price aggressively. Why wrong: plowhorses are popular because they are perceived as good value; a big rise can collapse demand and turn a plowhorse into a dog. Correction: raise margin gently — small increments, portion tweaks, cheaper components, or bundled high-margin sides.
-
Using a plain average for popularity. Why wrong: one runaway best-seller skews a simple average so most items look "unpopular." Correction: use the Kasavana-Smith 70% rule (70% of equal menu share) as the threshold.
-
Running the analysis once and forgetting it. Why wrong: ingredient costs, seasons and trends shift categories over time. Correction: re-engineer the menu at least each season.
Comparison and Connections
| Concept | Measures | Best use | Limitation |
|---|---|---|---|
| Food-cost % | Cost as share of price | Quick benchmarking | Ignores cash contribution |
| Contribution margin | Cash earned per sale | Profit decisions | Ignores labour/overhead |
| Menu engineering (K-S matrix) | CM + popularity | Menu redesign | Needs accurate sales & cost data |
| Prime-cost analysis | Food + direct labour | Labour-heavy dishes | Harder to allocate labour |
Menu engineering pairs naturally with menu psychology and design (how placement, descriptions and typography influence choice) and with standard recipe costing. The BCG matrix is its business-strategy cousin — same quadrant logic, different axes (market growth vs share).
Practice Questions
Recall
Q: Name the four menu-engineering categories and the two axes that define them. A: Stars, plowhorses, puzzles, dogs — defined by contribution margin (profitability) and menu mix percentage (popularity).
Understanding
Q: Why do Kasavana and Smith use 70% of the equal share instead of a simple average for popularity? A: A simple average is distorted by one or two blockbuster items, which raises the bar so high that most items falsely appear unpopular. The 70%-of-equal-share threshold sets a fairer, more stable cut-off so genuinely popular items are recognised.
Application
Q: A dish sells for $18 with a food cost of $6 and accounts for 5% of items sold on a 12-item menu. Average CM is $9.50. Classify it. A: CM = $18 − $6 = $12 (above $9.50 → high). Popularity threshold = 70% of (100 ÷ 12 = 8.33%) = 5.83%. Its 5% mix is below 5.83% → low. High CM + low popularity = Puzzle — promote and reposition it.
Analysis
Q: Your top plowhorse contributes $8 and sells 200 units/week; a puzzle contributes $18 but sells 15/week. You can only invest effort in one. Which, and why? A: There is no single answer, but the analysis matters. Converting the plowhorse: a $1 margin gain across 200 units = $200/week extra with low risk. Boosting the puzzle from 15 to, say, 45 units = $540/week extra but requires successful promotion and carries demand risk. Weigh guaranteed small gain versus larger uncertain gain — many operators secure the plowhorse improvement first, then invest in the puzzle.
FAQ
Does contribution margin include labour and overhead? No. Standard menu-engineering CM is selling price minus plate food cost only. Labour and overhead are covered collectively by the total contribution pooled across all sales. Prime-cost pricing adds direct labour when you need finer accuracy on labour-heavy dishes.
Isn't a lower food-cost percentage always better? No — this is the classic trap. A low percentage on a cheap item can yield very few dollars. Always check the actual cash contribution; high-percentage items frequently generate the most profit.
How often should I re-run menu engineering? At minimum every season, and whenever key ingredient prices move sharply or you launch new items. Categories drift as costs and preferences change.
What if an item is right on the borderline between two categories? Treat it as a candidate for the more favourable action and watch it. Borderline items are ideal test subjects for small price or description changes; re-measure after a few weeks.
Can menu engineering work for a small café or a bar?
Yes. The maths scales to any menu with sales counts and recipe costs. A bar, for instance, can engineer its cocktail list (see ../../21._Bar_and_Beverage_Management/index.md) exactly the same way.
Quick Revision
- Two axes: contribution margin (price − food cost) and popularity (menu mix %).
- Star = high CM + high popularity → feature and protect.
- Plowhorse = low CM + high popularity → raise margin gently.
- Puzzle = high CM + low popularity → promote and reposition.
- Dog = low both → usually remove or rework.
- Popularity threshold = 70% of equal menu share, not a simple average.
- Cash dollars beat food-cost percentages for profit decisions.
- Origin: Kasavana and Smith, 1982, Michigan State — the four-box matrix.
- Pricing methods: factor, contribution-margin, prime-cost, demand/value.
- Re-engineer the menu at least each season.