Two restaurants can serve the same food, at the same prices, to the same number of guests — and one makes double the profit. The difference is rarely the chef. It's the menu, and specifically whether anyone has done the math on it.
The four-box matrix
Menu engineering classifies every dish by two numbers — popularity and profit margin — into four boxes:
- Stars — popular and profitable. Protect them, feature them, never let quality slip.
- Puzzles — profitable but unpopular. These need better placement, better photos, or a better name.
- Plowhorses — popular but low-margin. Small price adjustments or portion re-costing here move real money.
- Dogs — unpopular and unprofitable. Cut them; every dog steals kitchen time from a star.
The theory is decades old. The reason most restaurants never apply it: it requires knowing the true cost of every dish — every gram of every ingredient — and keeping that current as supplier prices move. That's a spreadsheet nobody maintains past week two.
What happens when AI does the spreadsheet
DineCore automates the whole loop. Recipes are costed ingredient by ingredient, supplier invoices are scanned and extracted with AI, and every sale updates the matrix. The platform then recommends prices, flags margin leaks, and even generates fully-costed menu ideas. Because it works on top of your existing POS, you don't have to replace anything to start.
Profit is bigger than the menu
The same platform runs QR dine-in ordering per table, delivery with your own drivers and live tracking, loyalty and customer CRM, two-level inventory that deducts automatically with every sale, and a live P&L with payroll and cash-box reconciliation. Ask "William", the built-in AI assistant, how yesterday went — it knows your numbers.
Start with one report
Run the matrix once and you'll find at least one Dog to cut and one Puzzle worth promoting. Most operators find the analysis pays for the software in the first month — the rest is compounding.