Menu Engineering for Restaurants: Design a More Profitable Menu

Menu engineering is the practice of sorting every dish by two things — how often it sells and how much profit it makes per plate — and then redesigning the menu to steer guests toward the items that make you the most money. It is one of the highest-return moves an operator can make, because it needs no new customers and no ad budget: you are simply getting more profit out of the guests already walking in. This guide covers what menu engineering is, the four-box matrix at its heart, how to actually run the analysis, and the design levers that change what people order — with an honest look at which of those levers the research supports.

We sell nothing here — no POS, no menu software, no analytics tool. Menu engineering is a way of thinking you can do with your sales report and a spreadsheet, and that is exactly how we cover it.

What menu engineering actually is

Every dish on your menu has two numbers that matter:

1. Popularity — how many you sell in a given period, relative to your other dishes. This is your "menu mix."
2. Contribution margin — the profit each plate makes, in dollars: the menu price minus the plate cost (what the ingredients cost you).

Menu engineering plots those two numbers against each other and asks a simple question of every item: is this dish popular, profitable, both, or neither? The answer tells you what to do with it — feature it, fix it, promote it, or cut it.

The single most important idea here is to judge profit in dollars, not in food-cost percentage. Operators are trained to watch food-cost percentage, but a great percentage can hide a poor dollar margin. A soft drink might have a 90 percent margin as a percentage yet only put a dollar or two in the till, while a steak with a "worse" percentage contributes far more real money per plate. Menu engineering follows the dollars.

The menu engineering matrix

The framework was introduced by hospitality researchers Michael Kasavana and Donald Smith in 1982, and it still holds up. Sort every dish into one of four boxes based on whether its popularity and its contribution margin are above or below your menu average:

Each box has a different job:

• Stars are what you build the menu around. Don't hide them, don't discount them into oblivion, and never let their quality slip — a star that gets worse quietly becomes your biggest problem. Give them the best real estate and your staff's best recommendation.
• Plowhorses are the dishes everyone orders that barely make money — often the burger, the wings, the house pasta. You rarely want to remove them (they may be why people come), so the move is to lift the margin gently: a modest price increase, a slightly trimmed plate cost, a high-margin side or drink attached, or a small portion adjustment guests won't notice.
• Puzzles make good money but few people order them. Before you cut one, try to sell it: move it higher in its section, give it a mouth-watering description, add a photo, and train servers to suggest it. Many puzzles are just badly positioned, not genuinely unwanted.
• Dogs earn their name — low sales and low profit. Most should be reworked or removed to simplify the menu. The one exception is a dish that plays a role beyond its numbers: a vegetarian option, a kids' item, or a signature that defines you. Keep those on purpose, not by accident.

An elegant printed dinner menu with cutlery on a table

How to run a menu-engineering analysis

You don't need software. You need your sales data and your recipe costs. Here is the whole process:

1. Pull your sales counts. Export how many of each dish sold over a meaningful period — a month is a good start, or a full quarter if your sales swing by season. This is your menu mix.
2. Cost each plate. Work out the ingredient cost of each dish from your recipes and current supplier prices. This is the number that drifts most, so use current costs, not what you paid last year.
3. Calculate contribution margin. For each dish: menu price minus plate cost. That is the profit that dish makes every time it sells.
4. Find your two averages. Work out the average popularity (a dish's share of sales) and the average contribution margin across the menu. These two lines are what "high" and "low" are measured against.
5. Place each dish in a box. Above-average on both is a star; above on popularity but below on margin is a plowhorse; below on popularity but above on margin is a puzzle; below on both is a dog.
6. Act on each category using the moves above — then re-check next quarter.

A quick worked example shows why both numbers matter. Say a pasta sells for $22 with $6 of ingredients — a $16 contribution margin — while a steak sells for $38 with $18 of ingredients, a $20 margin. Per plate, the steak wins. But if the pasta sells three times as often, it contributes far more total profit to the restaurant. Popularity and margin only mean something together, which is exactly what the matrix captures.

A restaurant owner reviewing sales figures on a laptop with paperwork at a cafe table

Menu design levers that change what guests order

Once you know which dishes you want to sell more of (your stars and puzzles), menu design is how you nudge guests toward them. These levers genuinely work — but each has an honest limit, and some popular "menu psychology" advice is more myth than evidence.

• Placement. Put the dishes you most want to sell near the top of their section. A popular belief holds that diners' eyes jump to a top-right "sweet spot," but eye-tracking research has challenged that, finding people read menus more sequentially — roughly top to bottom within each section. So the reliable move is simply: high-value items near the top of the category, not buried in the middle of a long list.
• Descriptions. A short, sensory, specific description — where an ingredient is from, how it's cooked, why it's good — consistently outperforms a bare dish name. Write them for the dishes you want to sell; you don't need one on every line.
• Photos, used sparingly. A single photo draws the eye to whatever it sits beside, which makes it a good tool for a puzzle you want to promote. Photograph too many dishes and the effect disappears (and the menu starts to look like a diner placemat). Shoot them well — see our guide to food photography for restaurants.
• How you show prices. Avoid a neat right-aligned column of prices, which invites guests to scan down the cheapest options and order by price. Tuck the price in at the end of the description instead. There is also some evidence here: in one often-cited Cornell study, guests given menus that listed prices without dollar signs spent more than those given menus with them. Treat effects like that as small nudges, not magic.
• Fewer, better choices. Behavioral research on "choice overload" suggests that too many options can make people slower to decide and less happy with what they pick. A tighter menu is usually easier to run, cheaper to stock, and easier for guests to order from — and it makes your stars stand out.
• A light highlight. A box, a subtle "chef's pick," or a small section for signatures can lift a dish — as long as you highlight only a few. Highlight everything and you highlight nothing.

The honest through-line: menu design steers attention toward dishes people are already willing to order. It does not create demand for a dish nobody wants, and it should never oversell — a description or photo that promises more than the plate delivers just buys you a disappointed guest.

A chef plating a gourmet dish in a professional kitchen

Common menu-engineering mistakes

• Chasing food-cost percentage instead of dollar margin. The percentage is a useful guardrail, but the dollars are what pay rent. Follow contribution margin.
• Too many items. A sprawling menu buries your stars, slows the kitchen, raises food cost through waste, and overwhelms guests. Menu engineering usually argues for fewer dishes.
• Cutting a puzzle before trying to sell it. Low sales often mean poor positioning, not a bad dish. Reposition and describe it first; cut it only if it still won't move.
• Discounting your stars. Your most popular, profitable dishes don't need a discount — that just gives away margin on items people would happily pay full price for.
• Ignoring prep time and complexity. A dish can look profitable on ingredients alone yet tie up your kitchen and slow every other ticket. Factor in labor and complexity, not just plate cost.
• Changing everything at once. Re-price, re-describe, and re-lay-out all in one go and you'll never know what worked. Change deliberately and watch the sales mix respond.

A waiter serving plates of food to a group of diners in a busy restaurant

Where menu engineering fits

Menu engineering sits at the join between your restaurant's costs and your marketing. It works on the profit side of the same equation the 30/30/30 rule describes: your food-cost bucket is decided dish by dish, and the menu is where you control it. It also overlaps with marketing, because a menu is a selling document — the levers above are the same craft as a good Google Business Profile photo or an appetizing social post, just aimed at the guest already holding the menu.

Do this analysis once and you'll never look at your menu the same way again. For how the menu fits into the bigger picture of filling tables and bringing guests back, see our complete restaurant marketing guide.