Classic menu engineering plots popularity against contribution margin. The framework still works — what changed is how fast you can refresh the matrix. Weekly, not quarterly.
Stars (popular + profitable) belong in the prime visual real estate of your digital menu and on server suggestions. Protect their quality; do not discount them casually.
Try this in QuickMenuQR
QR menu, POS, kitchen display & inventory — free for 14 days.
Start free trialPuzzles (profitable but unpopular) need storytelling: better photos, sharper names, or a prix-fixe pairing. AI-written descriptions can help, but a human should approve tone so it still sounds like your restaurant.
Plowhorses (popular but thin margin) are candidates for a small price test, a portion tweak, or a forced side that lifts ticket without angering regulars.
Dogs (unpopular + weak margin) should earn their keep as brand statements or leave the menu. Demand forecasting helps you see seasonal dogs before you prep a case of product that will not sell.
Start with one category this month — mains only. Act on two items. Measure. Menu engineering fails when it becomes a 40-item science project instead of a weekly habit.