Most menus are designed by feel. A dish goes on because the chef loves it, because a competitor has one, or because it sounded good at the time. Menu engineering replaces that guesswork with a simple, repeatable method: score every item by two numbers you already have — how many you sell, and how much profit each sale leaves behind — then sort the menu into four groups that each demand a different action. Done well, menu engineering routinely lifts profit by ten to twenty percent without raising a single base price. This guide walks through the full method and how to run it with your own data.
The Two Numbers That Matter
Every dish has a popularity and a profitability. Popularity is just unit sales over a period — how many plates left the kitchen. Profitability is the contribution margin: the menu price minus the food cost to make it. Note that this is margin in currency, not percentage. A dish that earns $9 profit on 40 sales contributes far more to your rent than a dish earning 80% margin on 5 sales.
You need both numbers per item. Sales counts come from your order history. Food cost comes from costing each recipe — the per-unit cost of every ingredient that goes into the plate. If you have entered a food cost for each item, your analytics can compute margin automatically; if not, a one-time costing exercise on your top 30 sellers covers the vast majority of your volume.
The Four Quadrants
Plot every dish on a grid: popularity on one axis, profit margin on the other. Draw a line at the average of each. Four groups appear.
**Stars** — high popularity, high profit. These are your winners. Guests love them and they make you money. Your job is to protect them: keep quality flawless, give them prime menu real estate (top-right of the page, where eyes land first), and never let a supplier change quietly erode their margin.
**Plowhorses** — high popularity, low profit. Guests order these constantly, but each sale barely pays. They drive traffic but starve your margin. The move is to nudge profitability without killing demand: trim portion or plating cost, raise the price modestly, or pair them with a high-margin side or drink. A small price increase on a beloved plowhorse often passes unnoticed and drops straight to the bottom line.
**Puzzles** — low popularity, high profit. These earn well but few people order them. The problem is usually visibility or description, not the dish itself. Rename them, write a more appetizing description, add a photo, move them higher on the page, or have staff recommend them. A puzzle that becomes popular turns into a star — the highest-leverage move on the whole menu.
**Dogs** — low popularity, low profit. Few order them and they make little when they do. They clutter the menu, complicate prep, and tie up inventory. Most should be cut. The exception is a strategic dog — a vegan option, a kids' dish — that you keep for menu completeness even though it does not earn its keep on its own.
How to Run Your First Analysis
Start with one menu section or your full menu over a representative period — a month is ideal, long enough to smooth out daily noise. Pull unit sales per item and calculate margin per item. Many restaurant tools, including the Menu Engineering view in GetFreeMenu, do this automatically once you have entered a per-item food cost, classifying each dish into the four quadrants for you and updating as new orders come in.
Sort your list and look at the extremes first. Your top three stars and your three worst dogs are where the fastest wins live. Resist the urge to act on everything at once. Pick three to five changes — promote two puzzles, reprice one plowhorse, cut one dog — implement them, and measure the result over the next month.
Common Mistakes
The biggest mistake is confusing margin percentage with margin dollars. A 75% margin sounds great until you realize the dish sells twice a week. Always rank by total contribution.
The second mistake is acting on too little data. A dish that sold poorly during one slow week is not necessarily a dog. Use enough history that the numbers are stable, and watch for seasonality — a hot soup is a dog in July and a star in January.
The third is forgetting that the menu is a layout, not just a list. Where a dish sits on the page changes how often it sells. Once you know your stars and puzzles, design the menu to push eyes toward them. The order of items, boxes, photos, and "chef's pick" tags are all levers.
Make It a Habit, Not a Project
Menu engineering is not a one-time cleanup. Costs drift, tastes change, and a star can slide into plowhorse territory when a key ingredient gets more expensive. The restaurants that win treat it as a quarterly ritual: pull the numbers, re-plot the grid, make a handful of changes, and measure. Over a year those small adjustments compound into a materially more profitable menu — built not on guesswork, but on what your guests actually order and what each plate actually earns.
