Your menu is your single most powerful revenue tool — and most restaurants are leaving 10-15% of income on the table by not engineering it deliberately. Menu engineering is not about changing recipes or raising prices. It is about understanding which dishes drive profit, which drive volume, and how to position them so guests naturally make choices that are good for both them and the restaurant.
The research foundation comes from two institutions: Kasavana and Smith at Michigan State University (1982), who developed the BCG matrix for restaurant menus, and Cornell's Center for Hospitality Research, whose menu psychology studies have been validated across thousands of restaurants over four decades.
Every item on your menu falls into one of four quadrants, defined by two axes: popularity (how often it is ordered relative to your menu average) and profit margin (contribution margin per dish, not just price).
Your best performers. Feature prominently on the menu, train staff to recommend them, prioritize in AI recommendations. Protect these at all costs — do not change recipes or raise prices aggressively.
Guests love them but they don't make you money. Strategy: suggest premium add-ons ("add truffle for €4"), slightly increase price, or reduce portion cost through supplier negotiation.
Hidden gems. Great margin but guests don't order them. Strategy: reposition on the menu, improve descriptions, have staff recommend them, use AI to suggest them to adventurous guest profiles.
Neither popular nor profitable. Strategy: remove from the menu, reformulate to improve margin, or keep only if essential for dietary coverage. Dogs create kitchen complexity with no return.
Cornell's research goes beyond item classification into the psychology of how guests read and respond to menus. These findings are consistent across decades of research and directly applicable to any restaurant:
Applying BCG classification manually requires pulling sales data, calculating contribution margins, plotting items, and repeating every time the menu changes. FlashBook's Chef Agent does this automatically by combining three data sources: order frequency from booking history, menu item cost data, and real-time availability.
The AI then uses this classification in every guest recommendation: Stars are always suggested first, Puzzles are recommended to guests with adventurous dining history, Plowhorses get upsell suggestions attached, and Dogs simply don't appear in recommendations. The result is that every guest interaction subtly steers toward higher-margin choices — without any feeling of being sold to.
"The 26% of restaurant operators already using AI for menu optimization report it as one of the highest-ROI technology investments they've made — because it works on every single ticket, every day, without additional staff effort." — Toast 2025 Survey, 712 operators
Automatic BCG classification, descriptive name generation, and personalized recommendations that steer guests toward your Stars.
See FlashBook IA