Menu Engineering

What menu engineering actually means

Menu engineering is the discipline of treating your menu like a portfolio. Every item gets scored on two things: how profitable it is per sale, and how often it sells. The goal is to push customers toward the items that make the most money while quietly removing or repositioning the ones that do not.

It is part math, part psychology, and it is the highest-impact marketing work a bar manager does.

How it is used on the floor

Operators rank every menu item into one of four categories (stars, puzzles, plow horses, dogs) based on popularity and profit margin. Then they make layout, pricing, and descriptive decisions based on where each item lands.

Stars go in the top right of the menu (highest visibility). Puzzles get more prominent placement or better descriptions to boost sales. Plow horses get quiet price increases or margin work. Dogs get cut, repriced, or repositioned.

The four categories

The classic menu engineering matrix, developed by Michael Kasavana and Donald Smith at Michigan State in 1982:

• Stars: High popularity, high margin. Protect these.
• Puzzles: High margin, low popularity. Promote these.
• Plow Horses: Low margin, high popularity. Reprice or rework these.
• Dogs: Low margin, low popularity. Cut or rework these.

Most bar menus have 15 to 30 cocktails. A healthy distribution is roughly 20 percent stars, 20 percent puzzles, 40 percent plow horses, and 20 percent dogs. Anything further skewed is a sign the menu needs work.

A real example

A bar has 20 cocktails. The classic Old Fashioned sells 280 a month at 20 percent pour cost, so it is a star. A $16 mezcal cocktail sells only 40 a month but at 18 percent pour cost, so it is a puzzle. A vodka soda sells 450 a month at 28 percent pour cost because of over-pouring, so it is a plow horse. An amaro-forward drink sells 18 a month at 32 percent pour cost. Dog.

Actions: feature the Old Fashioned on the menu, add a story line to the mezcal cocktail with a better description, jigger the vodka soda to fix pour cost, cut the amaro drink or rework it with different ingredients.

Common mistakes

Classifying without real data (gut-feel ranking). Only running the analysis once a year. Ignoring seasonality. Assuming dogs should always be cut (some are staff favorites or cultural signatures). Not testing price changes before rolling them out.

How PourIQ handles it

PourIQ automatically classifies every menu item using live POS sales data and current recipe cost data. Stars, puzzles, plow horses, and dogs are labeled on the dashboard, updated continuously as sales and costs shift. No spreadsheet exports, no quarterly analysis sessions.