Menu Engineering Basics: Stars, Plow Horses, Puzzles, Dogs
Menu engineering is a two-question sort applied to every item: does it sell? does it earn? Cross the answers and every dish lands in one of four boxes — each with a different playbook.
The four quadrants
Stars — popular and profitable.Protect them. Never cut the portion, never let quality drift, feature them in the menu's prime real estate (top-right of a panel, boxed, photographed). Test gentle price rises — stars usually have headroom.
Plow horses — popular, low margin.Your burger, your wings. Guests come for them, but they don't earn. Fix from the cost side first: re-spec the build, tighten the portion by half an ounce, negotiate the protein. Then nudge price in small steps — popularity buys pricing power in $0.50 increments.
Puzzles — profitable, slow.The dish that earns beautifully and nobody orders. Rename it, move it, have servers speak to it, put it on a feature card. If three months of promotion can't move it, it wasn't a puzzle — it was a dog with good margins.
Dogs — slow and unprofitable.Cut them. Every dog costs you inventory slots, prep time, and menu attention. The only reprieve: a dish that exists for a reason the matrix can't see (the owner's mother's recipe, the one vegan anchor).
Doing the sort
- Pull a month of sales counts per item from your POS.
- Cost every item honestly — yield-adjusted, current prices (this is the step that breaks in spreadsheets; how to cost a recipe).
- Item margin = price − plate cost. Popularity = its share of category sales.
- Above/below the category average on each axis → its quadrant.
Redo the sort quarterly. Dishes migrate — a star whose ingredient costs crept becomes a plow horse without changing a thing on the plate, which is exactly why live costing matters.