Menu engineering is the practice of using sales data plus item-level cost data to redesign a menu so it sells more of the high-profit items and less of the low-profit ones. It is part finance, part design, part psychology.

The classic framework groups every menu item into one of four quadrants based on popularity (units sold) and contribution margin (profit per unit): Stars (high popularity, high margin — promote them), Plowhorses (high popularity, low margin — re-engineer recipe or raise price), Puzzles (low popularity, high margin — redesign placement or description), and Dogs (low popularity, low margin — remove).

The redesign work follows from the categorization. Stars get top-right placement on the menu, larger photos, and named callouts. Plowhorses get tested with a price increase or a recipe rework to lift margin. Puzzles get repositioned with better names and descriptions. Dogs come off the menu, freeing prep complexity.

Menu engineering is cyclical. Run the analysis quarterly. The data tells you what changed — a Star can drift into Plowhorse territory as costs rise, a Puzzle can become a Star with the right description. The menu is a living document, not a static asset.

Most restaurants set menu prices once and then never touch them. Even small menu engineering work — moving an item, raising one price by 50 cents, rewriting one description — can lift contribution margin by 3-8%. On a $1.5M restaurant that is real money, and the work is mostly free.