Design
Menu engineering for American restaurants
Menu engineering in dollars: stars, plowhorses, puzzles, dogs, and why the wine list is still a catalog. For F&B managers and owners in the US.
Written by Marco Azzari on
Menu engineering is the practice of reading a food menu as a grid of popularity and contribution, then putting the plates that pay where guests actually look. The term came out of American hospitality research in the early 1980s (Kasavana and Smith) and it stuck. The work underneath is older than the buzzword: count what sells, know what each plate leaves after true cost, stop treating every line as equal.
This page is for F&B managers and owners who already cost the kitchen and have never run the same logic on the wine list.
The grid (use it, then get back to service)
The usual four boxes:
- Stars: high popularity, high contribution. Keep them visible. Do not hide them in a corner to look modest.
- Plowhorses (or workhorses): high popularity, low contribution. These feed covers and starve the month. Rework the recipe, the garnish, or the price. Do not celebrate volume that does not pay.
- Puzzles: low popularity, high contribution. The plate is fine; the page is not. Rename, move, have the floor say it, or pair it. If it still sits, it is not a puzzle. It is a leftover.
- Dogs: low popularity, low contribution. Cut them unless they are a regular's identity dish and you have decided to subsidize that identity.
Run the grid in dollars, on the pre-tax menu price. Sales tax is added at checkout and is not your revenue, and neither is the tip. If your costing sheet pulls tax-inclusive totals from the POS, you are flattering every plate. See food cost versus wine cost.
A full engineering pass also needs a period long enough to be dull: not a holiday weekend, not the week the sous chef was out. Four weeks of POS data beats a feeling.
Where the eye goes
Engineering fails when the grid lives in a spreadsheet and the menu still leads with a photograph of a dish you cannot make money on. Guests scan. They land on corners, boxes, the first line of a section, the second price in a cluster. Put stars there. Put dogs anywhere you like, including off the menu.
You do not need dark-pattern boxes. You need honesty about attention. A boxed "signature" that is a plowhorse trains the room to sell the wrong plate.
The wine list is the unengineered half
Most operators who search menu engineering have already done some version of this to food. The wine list is still a distributor catalog: long, famous, priced by a flat markup, ordered by color and country because that is how the book came.
Wine can be a high-contribution line. In many full-service restaurants, beverage alcohol is a fifth or more of sales, while typical restaurant profit sits in the low single digits, often 3 to 5 cents on the dollar. Those two facts together are why a messy list is not a lifestyle issue. It is a P&L issue that never gets a grid.
Apply the same four boxes to bottles and glasses. Your "house red" may be a plowhorse: everyone drinks it, nobody costed it since 2019. Your second-cheapest bottle may be a dog that sells anyway, because guests use that slot to avoid looking cheap. That sale is not a star. It is a design accident.
Wine-list design is the page build. Pricing the list is true cost. By the glass is usually the highest-margin engine and the one that never gets monitored once busy season starts.
If you want this done on the list you already have, rather than as a method on a wiki page, Wine CFO runs a free Wine List Checkup. Independent. No cellar to empty. Capacity is limited because it is done by a person.
About the author
Marco Azzari
Founder of Wine CFO · Wine List Engineer for restaurants
I help restaurants, hotels and bars earn more from the wine they already sell. I price lists wine by wine, cost by-the-glass programs on their real yield, rebuild selections around the food menu, and rewrite lines so guests understand what they're ordering. Every Wine Profit Checkup goes across my desk personally.
More about Wine CFO