Menu Engineering: Find Your Stars, Plowhorses, Puzzles and Dogs

A menu is a set of decisions about what you want guests to order. Menu engineering gives you a way to see which dishes are helping and which are not, using two numbers you may already have: how often a dish sells and how much it earns.

All figures below are illustrative examples.

The two measures

• Popularity: how many of each dish you sold, compared with the others in its category.
• Contribution margin (CM): selling price minus ingredient cost, per plate. This is gross profit per plate, before labor and other costs. See food cost percentage vs gross profit.

Menu engineering is usually credited to Michael Kasavana and Donald Smith in the early 1980s. The four labels come from that work.

The four groups

Step by step

1. Pick one category, such as mains. Compare like with like.
2. List each dish with units sold over a period of several weeks and its contribution margin.
3. Set the popularity line. A common convention is 70% of an equal share. With five dishes, an equal share is 20%, so the line is 14% of total units.
4. Set the margin line: the average contribution margin weighted by units sold.
5. Classify each dish against the two lines.

A worked example

Five mains, illustrative:

Total units: 840. Popularity line: 14% of 840, about 118 units. Weighted average CM: $10,919 of total contribution divided by 840 units, about $13.00.

What to do with each group

• Stars: protect them. Keep quality and consistency high, and feature them without discounting.
• Plowhorses: popular but low margin. Try a small price increase, a slightly lower plate cost through the recipe or garnish, or a profitable add-on. Do not shrink the portion in secret.
• Puzzles: profitable but under-ordered. Look at placement, description, the name and the price. Ask servers to recommend them. If a puzzle stays slow, it may not suit your guests.
• Dogs: weak on both. Rework the dish, replace it, or remove it, unless it serves a clear purpose such as a vegetarian option.

Limits

• It ignores labor and time. A dish that clogs the line may deserve a different decision than its numbers suggest.
• Menu design affects sales. Popularity partly reflects where a dish sits and how it is described.
• Small samples mislead. Use several weeks.
• Compare within categories, not between appetizers and mains.
• It is a prompt for decisions, not a verdict.

Common mistakes

• Removing dogs without asking why they sell badly.
• Changing several things at once, so you cannot tell what worked.
• Using outdated plate costs. See how to price a menu item.
• Treating contribution margin as net profit.

Where software can help

Menu engineering depends on current plate costs and units sold. ChefMind AI's food cost tools show food cost per dish against your target. The classification itself you can do in a spreadsheet with the steps above.

The short version

• Classify dishes by popularity and contribution margin within a category.
• Stars: protect. Plowhorses: nudge price or cost. Puzzles: promote. Dogs: rework or remove.
• Use several weeks of data and remember what the method ignores.