Menu engineering in restaurants: definition and practical formula

Menu engineering: the complete guide, from the matrix to the menu

The menu engineering matrix: stars, plowhorses, puzzles and dogs

The menu engineering matrix classifies every dish on two axes: popularity, how many units it sells compared with what you would expect, and contribution margin, how much each unit leaves compared with the menu average. The classic method, developed by Kasavana and Smith, produces four groups. Stars sell a lot and earn a lot. Plowhorses sell a lot but earn little. Puzzles earn a lot but sell little. Dogs sell little and earn little. The matrix does not tell you which dish is good or bad in the kitchen: it tells you which menu decision makes sense for each one.

How to do menu engineering step by step?

Menu engineering takes five steps. First, pick a representative period and pull the units sold for each dish from your POS. Second, take the standard recipe cost and the menu price before tax of each dish and calculate its unit contribution margin:

price minus cost. Third, multiply that margin by units sold to get what each dish contributes. Fourth, compare each dish's popularity and margin with the rest of its category and place it in the matrix. Fifth, decide one action per dish and measure again after you apply it.

What to do with each dish in the menu engineering matrix?

Each quadrant of the matrix calls for a different move. Stars stay as they are, protected from recipe changes and highlighted on the menu.

Plowhorses are worked on cost: review the recipe, the portion or the supplier, or raise the price carefully, because guests already want them. Puzzles are worked on visibility: better placement, a better name and description, a server recommendation or a more attractive price. Dogs are replaced or dropped, unless they serve a purpose, like a kids' or vegetarian option that brings a whole table in.

Menu engineering and menu design

Menu engineering decides which dishes to push; menu design is how you push them. Your best-margin dishes go where the eye lands first in each section, with a concrete description of ingredients and preparation, and sometimes inside a box that sets them apart. Prices lined up in a column invite guests to choose by price, so it helps to place them at the end of each description. Fewer dishes per category make choosing easier and the kitchen steadier. Measure every design change with the same matrix in the next period: if the dish did not move quadrants, the change did not work.

Menu engineering in a spreadsheet: the columns you need

A spreadsheet is enough to do menu engineering. Each row is a dish, and the columns are: category, units sold, share of category sales, recipe cost, menu price before tax, unit contribution margin, total margin, and two comparison columns, one against expected popularity and one against the category's average margin. With those two comparisons, a formula assigns each dish its quadrant. What the spreadsheet cannot fix is recipe cost: if your recipe cards are not updated with real purchase prices, the whole matrix is wrong.

What is menu engineering?

Menu engineering measures how much real money each dish puts in the cash box—not printed margin, but margin multiplied by volume sold—to optimize prices, eliminate products without marginal profitability, and strengthen dishes that drive the most cash flow.

According to research by James Love at Cornell in the 1980s, every menu decision without cash data is guesswork: a high-margin dish sold rarely can generate less pure margin than a lower-margin dish sold several times more often. Diego F. Parra has audited over 8,400 restaurants and sees the identical mistake: owners confuse profit percentage with actual profit, leaving dead weight in the kitchen. Menu engineering measures it, names it, and removes it from the menu.

Unit margin versus cash flow

Here lives the first major misunderstanding: margin is what remains on each portion sold; marginal profitability is how much money that dish puts in the cash box, shift after shift, year after year. For example, an appetizer priced above its cost with a handful of daily sales generates a meaningful margin on its own. The second dish, despite its lower percentage margin, brings several times more money to the cash box because it sells far more often. Masterestaurant flags this difference as the #1 axis for menu redesign: the cash register does not read percentages, it reads dollars.

Volume versus optimal mix

Selling high volume of a low-margin product burns your cash opportunity. Picture two mixes on the same shift: one leaning on a low-margin, high-volume item, the other on a mid-margin item paired with a higher-margin shared board. The second mix can end the shift with meaningfully more total margin than the first, even though neither dish alone looks like the star. High volume with low margin is a trap: you strain the kitchen, burn portions, push the cash box backward. Menu engineering reorders the menu, prices, and visibility to flow toward profitability.

How marginal profitability is calculated?

You take the gross margin of each dish (price minus portion cost), multiply by volume sold over a period (shift, week, month), and compare the result against kitchen space, prep complexity, and time consumed.

For example, if a dish carries a fixed margin and sells several times per service, that volume can add up more than a high-margin, low-turnover dish; if it also takes several minutes of kitchen time per order, check how much time it costs your line. Another dish with a smaller margin but sold far more often per service can outearn it in less kitchen time overall. This is where mix analysis enters: which combination of dishes fills your kitchen with maximum profit. Restaurants with consolidated menu engineering reallocate a meaningful share of their mix toward higher-margin dishes without losing total volume.

Alcohol: the highest-margin category

Alcoholic beverages is where menu engineering plays hardest. For example, a glass of wine priced well above its cost carries a wide margin, with no limit per customer. For example, an entrée priced above its marginal cost still leaves you a real margin even if the customer orders only one. When you raise price (if positioning allows), when you lower it in cross-promotions for volume, when you highlight it on the POS screen: all of that is a menu engineering decision. The mistake is letting the bartender set the wine list; the cash register must speak first.

Common misinterpretations of the concept

Menu engineering is not «raise prices until it hurts»; it is knowing exactly which price on each dish delivers maximum cash flow while accounting for real elasticity of demand (not theoretical). Nor is it «sell only premium items»: it means mixing. It is not killing the signature dish that defines your brand because it scores low in margin; it is repricing it or adjusting if it is irreplaceable. Some confuse «menu engineering» with «maximize percentage margin»: they are opposites. The second mistake is ignoring cooking cost (how much it costs to produce beyond ingredient expense) and attributing the difference to margin alone. A made-to-order dish costs kitchen time; a pre-assembled item costs less. Engineering includes both, not just ingredient cost per unit.

Elasticity: price and volume in tension

You raise price and volume drops more than the per-unit margin gain: you made less than before. You raise price and volume drops less than the margin gain: you made more despite selling less. That is elasticity: the actual relationship between price and demand in your restaurant, in your neighborhood, with your clientele. No «correct» universal price exists; there exists the price where margin multiplied by volume is highest in your context. Masterestaurant measures elasticity in each category by auditing positioning, location, customer type, and frequency. A quick-service restaurant in a high-traffic area has different elasticity than fine dining with reservations. Menu engineering without elasticity is guessing in the dark.

Concrete application: from audit to decision

You walk into a restaurant. You audit 30 days of sales and portion cost. You calculate unit margin on each dish; you rank by total marginal profitability (margin × volume). Three groups emerge: stars (high margin × high volume), workhorses (medium margin × high volume, funding the business), and dead weight (low margin × low volume). The typical recommendation: eliminate dead weight or reprice it higher; promote workhorses on the POS screen with beverage bundling; highlight stars on the physical menu. If a star fades, promote it on social channels. If the dead weight is iconic (brand identity), raise price 15% and cut advertising: preserves brand, raises margin. In Diego F. Parra's experience advising restaurants, well-executed menu engineering improves EBITDA on a sustained basis without losing customers.

The most common confusions

Unit margin vs. contribution margin: margin is % profit per plate sold; contribution margin is how much REAL money each dish brings to cash. A dish with a high margin but almost no sales per shift loses against one with a lower margin and several sales per shift, because contribution per shift is what pays the bills. The first looks profitable on paper, the second actually is. Volume vs. mix: selling a lot of a low-margin product compresses total cash. If you sell 100 tacos at $2.5 margin each and 5 ribs at $18 margin, total margin is $340; if you invest that capacity in 45 ribs, you hit $810. Menu engineering measures which mix leaves more money in cash, not how many portions went out. Fixed price vs.

The most common confusions — in practice

actual elasticity: the myth is «all customers react the same to price». Reality: alcoholic beverages have low elasticity (raise price, volume drops little); salads have moderate elasticity (raise price, volume drops close to proportionally); water and bread have no competition, so elasticity is minimal. Diego F. Parra's experience advising restaurants confirms a clear pattern: the menu has three price-psychology zones, not one rule. Recipe cost vs. service cost: they confuse costing the dish (ingredients + waste) with costing the experience (server, kitchen energy, rent per square meter). For example, a dish with a high ingredient cost doesn't cost the restaurant that same amount if it takes a few minutes on a range that cooks several dishes simultaneously. The correct formula is: recipe cost + (cook time × (energy + depreciation + labor) / units cooked at once).