Menu Engineering: Which Café Items Actually Make You Money

14 August 2026 · MidaOne

Every café menu has a handful of items that pay the rent, a larger group that keeps the lights on, and one or two that quietly cost you money every time somebody orders them. Most owners have a feeling about which is which, and the feeling is usually built on what they enjoy making or what customers compliment. Menu engineering replaces the feeling with two numbers you already have in your till. It is not a redesign and it is not a rebrand — it is a sorting exercise you can do in an afternoon, and it tends to move your margin faster than a price rise.

The two numbers every menu item has

Menu engineering rests on exactly two figures per item, and almost everyone gets the second one wrong.

The first is popularity — how many of that item you sold in a period, as a share of everything sold in its category. Compare cold drinks to cold drinks and pastries to pastries; a croissant will never out-sell a flat white, and putting them in the same league table tells you nothing.

The second is contribution margin — the dirhams left after ingredients, not the percentage. This is the part owners get wrong, because food-cost percentage is the number everyone quotes. A drink at 15% food cost sounds better than a sandwich at 40%, but if the drink leaves AED 12 and the sandwich leaves AED 22, the sandwich contributes almost twice as much towards your rent every time it sells. Percentages tell you whether an item is priced sensibly; dirhams tell you what it actually pays you. Work from a proper recipe cost per item, which is what our guide to pricing a café menu sets up.

Strip VAT out before you calculate anything. UAE menu prices are shown VAT-inclusive, so the 5% inside the number on your board was never yours to count as margin.

The four boxes: stars, plough horses, puzzles and dogs

Once every item has those two figures, split each into high and low against the average for its category. That gives four groups, and the whole method is deciding what to do with each one.

Category	What it looks like	What to do with it
Star — sells well, high margin	Your signature drink, the one regulars order without reading	Protect it. Keep quality and portion identical, put it where the eye lands first, never discount it
Plough horse — sells well, low margin	The everyday item people come for but that barely pays	Fix the cost, not the popularity. Tighten the recipe, renegotiate the ingredient, or raise the price in small steps
Puzzle — sells poorly, high margin	The item you are proud of that nobody orders	It is a visibility problem. Rename it, move it up the menu, let staff recommend it before you give up on it
Dog — sells poorly, low margin	The item that survives because nobody has looked at it in two years	Remove it. It costs you stock, prep, menu space and attention

The categories come from a well-known restaurant costing framework, and the labels matter less than the sorting. What matters is that the four boxes each have a different fix, and applying the wrong fix is how menu work goes wrong — a puzzle does not need a discount, and a plough horse does not need to be deleted.

Where the money usually is

In most small cafés the biggest single win is not the dogs. Removing four items nobody orders feels satisfying and changes very little, because by definition they were barely selling. The money is almost always in the plough horses — the high-volume, low-margin items — because a small improvement multiplied by a large number of covers is a real amount.

Take an item selling 40 a day. Finding two dirhams of margin on it, through portion control, a cheaper packaging line or a modest price move, is roughly AED 2,400 a month before anything else changes. The same effort spent rescuing an item that sells twice a day is worth almost nothing. Sort by total contribution — margin multiplied by volume — and the priority order writes itself.

The second place money hides is in items whose costs have crept while their price stood still. Dairy, coffee and packaging prices move, and a menu priced eighteen months ago is priced against eighteen-month-old costs. That drift is invisible on the menu board and obvious the moment you re-cost the recipes.

How to actually run it

Pull 60 to 90 days of item-level sales from your POS. Shorter periods get distorted by a single busy weekend or a quiet week.
Cost every recipe properly, including the parts nobody counts — the cup, the lid, the sleeve, the napkin, the syrup pump. On a drink these are a bigger share than owners expect.
Work out contribution margin in dirhams for each item, on the VAT-exclusive price.
Split each category into above and below average on both popularity and contribution margin. Four boxes.
Pick three items to act on, not thirty. One plough horse to fix, one puzzle to promote, one dog to cut.
Re-run it a quarter later and see whether the three moved. This is the step almost everyone skips, and it is the one that tells you whether the exercise worked.

One caution on cutting: check what a dog is doing for the people who order it before deleting it. An item that sells rarely but is the only reason a particular group comes in — the one vegan option, the one thing a regular's child will eat — earns its place through the covers it brings, not its own line in the report. Menu engineering is a tool for thinking, not a rule that overrides what you know about your room.

Menu design does half the work

Once you know which items you want to sell, where you put them matters. People do not read a menu evenly — they scan it, and the first two items in a section and the last one get looked at far more than the middle of a long list. Keeping sections short is the simplest lever there is: a section with six items gets read, and a section with twenty gets skimmed until something familiar appears.

Two habits worth breaking. Listing prices in a neat right-hand column invites customers to read down the column and pick by price rather than by what they want, so put the price at the end of the description instead. And a long list is not a generous menu — it is more prep, more stock lines, more waste and a slower kitchen, which is the same problem our guide to reducing food waste approaches from the other end. A shorter menu of items you have actually costed usually earns more than a longer one you have not.

Where MidaOne fits

The reason most cafés never do this is not that the method is hard — it is that assembling the data takes a weekend of exporting spreadsheets and matching them up by hand. MidaOne ties recipe costs to live sales in one system, so item-level volume and per-item margin come from the same record instead of two that disagree, and stock is deducted by recipe as you sell. That makes the sorting exercise something you can repeat every quarter instead of once, heroically, and then never again. How that flows through to the number at the bottom of the year is covered in our guide to café profit margins.

See the real margin on every item you sell. Free for 14 days, no card needed.

Start your free trial

Frequently asked questions

What is menu engineering?

It is sorting every menu item by two figures — how much it sells and how many dirhams of margin it leaves after ingredients — and then treating each group differently. Items that sell well and pay well get protected, items that sell well but pay poorly get their costs fixed, and items that do neither get removed.

Should I use food-cost percentage or contribution margin?

Contribution margin in dirhams, for this exercise. A percentage tells you whether an item is priced sensibly relative to its cost, but it hides size — a low-percentage drink can leave you less actual money than a higher-percentage sandwich. Use the percentage to check pricing and the dirham figure to decide what to promote.

How often should a café review its menu this way?

Roughly quarterly for the full exercise, and any time a major supplier price moves. Ingredient costs drift quietly, so a menu priced a year ago is priced against last year's costs even if nothing on it has changed.

How many items should a café menu have?

Fewer than most cafés carry. There is no correct number, but every extra item adds a stock line, prep time and potential waste, and long sections get skimmed rather than read. If an item is not selling and not costed, it is usually easier to defend removing it than keeping it.

Do I need a POS to do menu engineering?

You need item-level sales data and accurate recipe costs. You can assemble both by hand from receipts and supplier invoices, but it takes long enough that most owners do it once. A system that links recipe cost to sales produces both automatically, which is what turns this into a routine rather than a project.

The useful thing about this exercise is not the four boxes — it is that it forces you to look at your menu as a set of individual businesses rather than one list. Some of those businesses are excellent and some are not, and you cannot tell which from behind the counter. Pull the numbers once and you will find at least one item you were certain about that turns out to be wrong.