Menu engineering, with a worked example

The idea, in one paragraph

Every dish on your menu occupies the same expensive things: a line on the card, space in your fridge, prep time, and a slot in your guest's attention. Very few dishes earn that equally.

Menu engineering sorts them by two measures you can get from your own sales data, and tells you what to do with each group. It is not a theory exercise — it is an afternoon with two columns of numbers that usually finds five figures a month.

The two axes

Popularity — units sold over a defined period. One month minimum; a quarter is better.

Contribution — menu price minus true plate cost, in rupees. Not percentage. The full costing method, including the accompaniments and wastage most owners forget, is in how to price a dish properly.

Plot every dish against the two averages for your menu, and four groups fall out:

A worked example

A casual dining restaurant, one month. Average units sold across the menu: 142. Average contribution: ₹168.

Two things are already visible that a food-cost report would never show you. The papad — the single best-selling item on the menu — is contributing less per plate than almost anything else. And the lamb shank, which the owner has been considering cutting, is the highest-contribution dish in the building.

What to do with each group

⭐ Stars — protect and promote

Popular and profitable. These are the dishes people describe to a friend.

• Do not touch the recipe or the portion. The temptation to shave cost off a star is the most expensive mistake in this whole exercise.
• Raise the price cautiously — 3–5%, and only if it is not one of your anchor dishes. A star that is also a price anchor should be left alone entirely.
• Give them the best placement on the menu: top of category, top-right of a printed page.
• Make sure your staff know which they are. A captain recommending a star instead of a dog is worth real money and costs nothing.

🐴 Plough-horses — the biggest opportunity

Popular, low contribution. Guests love them, you barely profit from them. Because the volume is high, small changes compound fast.

Options in order of preference:

1. Raise the price slightly. On the masala papad above, +₹15 across 340 units is ₹5,100 a month — from a change most guests will not register.
2. Reduce the plate cost without touching perceived value. Better supplier terms, tighter portioning of the expensive component, less waste.
3. Bundle it. Pair a plough-horse with a star at a set price so the combined contribution improves.
4. Reposition it as an accompaniment rather than a standalone order.

What not to do: shrink it visibly. A plough-horse is popular precisely because it feels like good value, and the moment it stops feeling that way you lose the volume that made it worth fixing.

🧩 Puzzles — usually a visibility problem

High contribution, low sales. Before concluding there is no demand, check the three cheap explanations:

• Placement. Is it at the bottom of a long category where nobody reads?
• Description. "Lamb shank — ₹520" sells far worse than a sentence explaining what it is and why it takes four hours.
• Name. Unfamiliar names need a translation; familiar names need a reason to choose yours.

Then try, in order: rewrite the description, move it up the menu, brief the staff to recommend it, photograph it. In the example above, the lamb shank at ₹315 contribution needs only a handful more orders a week to be worth the effort — 20 extra units is ₹6,300 a month.

If it stays unpopular after a fair run with a good description and good placement, it becomes a dog.

🐕 Dogs — usually remove, but check first

Neither popular nor profitable. Two checks before cutting:

Is it somebody's reason for coming? A dish ordered by twelve regulars every single week is worth more than its line in this table suggests. Losing those twelve guests costs more than the dish earns.

Is it doing a menu job? A vegetarian option in a meat-heavy list, a mild dish on a spicy menu, a salad for the one person in a group of six who wants one. A dish that unlocks a table of six is not really a dog.

If neither applies, remove it. You get back fridge space, prep time, an ingredient to stop holding, and a line on the menu for something better.

What removing dogs is actually worth

Owners underrate this because the dish's own contribution is small. The saving is elsewhere:

• One or two ingredients you no longer buy, hold or throw away
• Prep time returned to the kitchen during service
• One fewer recipe to train every new joiner on
• A shorter menu, which is easier to choose from and easier to execute consistently

Cutting the bottom five dishes from a 55-item menu typically has no measurable effect on revenue and a visible effect on kitchen speed.

Doing it without a spreadsheet you dread

The reason most restaurants try this once is that assembling units-sold per dish by hand from a month of bills takes an evening. Then it never happens again.

Two ways to make it repeatable:

• Sample rather than census. Two representative weeks tells you almost as much as a full month for this purpose.
• Get the numbers from wherever your orders are recorded, rather than counting slips. Top items by quantity and by revenue for a date range is a standard report; if your system produces it, the exercise drops to under an hour.

Then do it quarterly. Ingredient costs move, a price rise reshuffles the groups, and seasons change what sells. A dish can spend nine months in the wrong group if you only look once a year.

Common mistakes

Using food cost percentage instead of rupee contribution. The prawn curry above has the worst percentage on the menu and the second-best contribution. Percentage would have you cut your best earner.

Shrinking a star. The one dish where guests are paying attention.

Cutting puzzles before fixing their description. You are removing profit because of a copywriting problem.

Cutting every dog mechanically. Some are holding a regular or unlocking a table.

Doing it once. The whole value is in the repetition.

Ignoring drinks. Beverages usually have the highest contribution on the menu and are almost never included in this exercise. Include them.

What to do this month

1. Pull units sold per dish for the last full month or quarter.
2. Cost your top twenty dishes properly — whole plate, accompaniments, 5–10% wastage.
3. Calculate contribution in rupees and the two menu averages.
4. Sort every dish into the four groups. An afternoon, on one sheet.
5. Act on one group first — plough-horses. Highest volume means fastest return.
6. Diarise the next pass for three months' time.

The short version

• Two axes: units sold, and contribution in rupees. Never food cost percentage.
• Stars get protected and promoted. Never shrink them.
• Plough-horses are the money: high volume means a ₹15 rise can be ₹5,000 a month.
• Puzzles are usually a description or placement problem, not a demand problem.
• Dogs go — unless they hold a regular or unlock a table of six.
• Quarterly, including drinks. Doing it once is how the wrong dish stays on for a year.