Menu engineering in an afternoon, with our worksheet

You can engineer a menu in one afternoon. Export 30 to 90 days of item sales from your POS, add each dish's plate cost, and sort every item into Stars, Plowhorses, Puzzles or Dogs by popularity and profit. Our free menu engineering worksheet does the maths, so you spend the afternoon deciding what to change.

What is menu engineering, and where does it come from?

It is a way of rating every dish on two things at once: how often it sells and how much it earns per plate. Michael Kasavana and Donald Smith developed it at Michigan State University's School of Hospitality Business around 1982 (AHLEI, Kasavana, 2025). Their book, Menu Engineering: A Practical Guide to Menu Analysis, was published that year by Hospitality Publications (WorldCat).

The four labels come from a two-by-two grid:

Kasavana's recommended move for each is short: Stars are retained, Plowhorses repriced, Puzzles repositioned on the menu, and Dogs removed (AHLEI, Part Two).

What data do I need before I start?

Three numbers per dish, all for the same period:

1. Units sold. From your POS item-sales report. Use 30 to 90 days so one festival week or one bulk order does not skew it.
2. Menu price. Before GST. If you sell on delivery apps at a different price, run delivery as a separate analysis.
3. Plate cost. The cost of everything on the plate: ingredients, garnish, and packaging for takeaway items. If you have not costed a dish recently, do it first with our food cost calculator.

Work one menu section at a time. Kasavana and Smith's method compares each item with others in its "competing menu group" (LeBruto et al., 1997). Starters compete with starters and mains with mains. Put a cup of chai in the same list as a curry and both averages stop meaning anything.

How do I decide what counts as popular?

Use the 70% rule. Divide 100% by the number of dishes in the section to get an even share. A dish is popular if its share of sales is at least 70% of that even share (LeBruto et al., 1997; AHLEI, Part Two).

In a section of ten dishes, the even share is 10% and the popularity line is 7% (LeBruto et al.). The rule is set below the even share on purpose. Not every dish can be above average, and a dish that sells steadily but a little under the average is usually still pulling its weight.

How do I decide what counts as profitable?

Compare each dish's contribution margin with the section's weighted average. Contribution margin is price minus plate cost. The weighted average is total contribution margin for the section divided by total units sold. A dish at or above that average is "high"; below it is "low" (LeBruto et al., 1997).

This is where many kitchens go wrong. They rank dishes by food-cost percentage. In the worked example below, the fish curry has the higher food-cost percentage of the two, yet it leaves far more money per plate than the chole bhature, which looks efficient on percentage. Rent and wages are paid in rupees, not percentages.

What should I do with each group?

Each group gets a different kind of change. Make one change per dish, not five, so you can tell what worked.

• Stars (popular, high margin). Keep them and protect them. Give them the best spot on the menu, keep the recipe and portion consistent, and be careful with price rises.
• Plowhorses (popular, low margin). People love them, so do not remove them. Try a small price rise, a slightly smaller or cheaper garnish, or pairing them with a high-margin side or drink.
• Puzzles (unpopular, high margin). They earn well when they sell. Move them to a better spot, rename or describe them better, have staff recommend them, or add a photo on delivery apps.
• Dogs (unpopular, low margin). Remove them, or rework the recipe for a lower cost and test again. Keep one only if it serves a real purpose, such as the only Jain or vegan main.

How do I run it with the worksheet?

The worksheet comes as an Excel file with the formulas built in and a PDF that walks through the method.

1. Download the worksheet and open the Excel file.
2. Paste in one menu section: dish name, units sold, menu price and plate cost.
3. The sheet works out contribution margin, each dish's share of sales and its label.
4. Circle the Plowhorses and Puzzles first. They are where small changes pay back fastest.
5. Write one action beside each dish, then make the changes on the menu, POS and delivery apps together.
6. Run the same section again after about 30 days of sales at the new settings.

Worked example: ten mains at a casual dining restaurant

These numbers are illustrative, chosen to show the maths. They are not benchmarks. Ten mains sold 1,360 plates in 30 days. Total contribution margin was ₹2,74,775, so the weighted average is ₹2,74,775 ÷ 1,360 = about ₹202 a plate. With ten dishes, the popularity line is 70% of 10% = 7% of sales.

What we would do with this menu:

• Chicken biryani is the best seller but earns ₹2 below average. A ₹20 price rise would take its margin to ₹220 and make it a Star if volume holds.
• Chole bhature earns ₹130, well below the average. Pair it with a lassi as a combo rather than raising the price of a comfort dish.
• The three Puzzles earn well but are hidden. Move the risotto and fish curry into the first box customers read, and have servers recommend the rogan josh.
• Veg biryani is a Dog, but it may be the main vegetarian rice option. Rework its cost or fold it into a thali instead of cutting it outright. Kadai veg can go.
• Dal makhani is a Star by only ₹3. A small rise in dairy costs would make it a Plowhorse, so watch it.

What are the limits of the method?

It is a sorting tool, not a verdict. It measures food cost only, not labour. LeBruto and colleagues added a third test for labour effort in 1995, turning the grid into a three-by-two model, because a high-margin dish that ties up your best cook may not be worth it (LeBruto et al., 1997).

It also looks backwards. A new dish needs a few weeks of sales before its label means anything. And it cannot see dishes that bring people in, such as a signature item that appears in every review, even when that item sells less than others. Use the labels to start the conversation with your chef, then decide.