Menu Engineering: Make Your Menu Push Customers Toward Your Most Profitable Items

You've got 40 items on the menu, and you know they don't all bring in the same money, but you can't say exactly which 5 items carry your profit and which ones eat your time and stock for nothing. Menu engineering is the tool that answers that with numbers, not gut feel: you take each item, look at how often it sells and how much it makes, then decide what to do with it. This guide walks you through it step by step: we explain the matrix, run a worked example on a small cafe, and tell you exactly what to do with each type of item.

What is menu engineering?

Menu engineering is simple at its core: you look at your menu on two axes and put every item where it belongs. The first axis is popularity: how often does the item sell? An item that sells 500 times a month is popular, one that sells 40 times is not. The second axis is profitability: how much does the item leave in your pocket after you subtract the cost of its ingredients? That's called the unit margin, and you get it from your recipe cost. If you haven't costed your items yet, that's a prerequisite before any analysis, so go back to our guide on costing a dish and food cost percentage and do that first.

When you put the two axes together, you get a square with four quadrants, and every item you sell lands in one of them:

• Stars: popular and profitable. These are your restaurant's stars, the items that sell a lot and make good money at the same time.
• Plowhorses: popular but thin. People love them and they sell all day, but they leave you very little.
• Puzzles: profitable but rarely ordered. The margin is great, but hardly anyone orders them, so they're hidden from the customer's eye.
• Dogs: neither popular nor profitable. They don't sell and they don't earn, they just take up space on your menu and in your stock.

Example: a 6-item cafe menu

Let's run a simple example. We have a cafe that sells 6 items, and for each one we pulled how many times it sold last month and its unit margin after ingredient cost. The table below shows each item and its classification:

How did we classify them? We drew two lines: the average units sold and the average margin. Total units are 1920 divided by 6 items = 320 units per item, so that's the popularity line: above it is popular, below it is not. The average margin (11 + 10 + 5 + 14 + 13 + 4 divided by 6) = 9.5, so that's the profitability line. Cappuccino and latte are above both lines, so they're stars. The croissant is popular (480 above 320) but its margin of 5 is below average, so it's a plowhorse. Cheesecake and smoothie have high margins but sell below 320, so they're puzzles. Fresh orange juice is below both lines, so it's a dog.

What to do with each quadrant

Once you know which quadrant each item is in, the real work begins. Each quadrant calls for a different move:

Stars: lock them in and put them front and center

Your stars are the goose that lays golden eggs, so don't mess with them. Lock down the recipe and quality so you never let a customer down, and place them at the top of the menu in a prominent spot so they sell even more. And if you run a deal or a combo, build it around your stars, not around items nobody asks for.

Plowhorses: lift the margin carefully

Plowhorses are loved but thin, and your goal is to lift their margin without killing their popularity. Three ways: reprice carefully (a small bump on a beloved item usually goes unnoticed), tweak the recipe (a cheaper ingredient at the same quality lowers your cost), or shrink the portion a little, as long as it still looks like a fair serving. Try one of them and watch the sales before you try the next.

Puzzles: market them

A puzzle is profitable but hidden, and your job is to make it seen. Move it to a spot the eye goes to, give it a name and description that make it tempting, add a nice photo if your menu has images, and have the cashier suggest it ('want a slice of cheesecake with that?'). Sometimes marketing alone moves an item from puzzle to star.

Dogs: swap them or cut them, boldly

A dog neither sells nor earns, and the longer it stays, the more it complicates your stock and kitchen. Give it one last chance (price, recipe, or placement), and if it doesn't move, cut it from the menu without hesitation. A shorter, clearer menu sells better than a long one full of dead items.

Where the numbers come from

The whole analysis rests on two numbers per item: how many it sold, and its margin. You get the first from the sales-by-item report in your POS, and the second from the recipe cost. Sales by item are ready in Loqma's reports at a tap, and you can export them for any period you want, so you never have to count invoices by hand or piece your numbers together off paper.

Your menu layout sells for you, quietly

Menu design isn't decoration, it's a sales tool. The moment a customer opens the menu, their eye goes to familiar spots: the top of the list, and the first and last item in each section. Put your stars and puzzles in those spots, don't bury them in the middle. And don't line up all the prices in one column on the side, because that makes the customer compare on price and pick the cheapest; keep the price next to the item name and description so they compare on value, not on the number. The same principles apply whether you run a cafe or a full restaurant with a longer menu.

Run the analysis in five steps

Ready to run it on your own menu? Walk through these five steps:

1. 1
   Export a full quarter of sales

   Pull the sales-by-item report for the last 3 months by unit count, not just by amount. A full quarter gives you a truer picture than a week or two.

2. 2
   Attach each item's margin

   Next to each item, put its unit margin (price minus ingredient cost from the recipe). This ties sales to real profit, not just revenue.

3. 3
   Draw the two lines

   Work out the average units sold and the average margin, and draw them as two lines. Each item lands above or below each line.

4. 4
   Classify each item

   Put each item in its quadrant: star, plowhorse, puzzle, or dog. A simple Excel table is enough, no special software needed.

5. 5
   Work on one quadrant a month

   Don't try to fix everything at once. Pick one quadrant (say plowhorses this month), apply the changes, and watch the result before moving to the next.