Key things to evaluate:

1. Pre-ordering and demand forecasting
   

   If employees pre-order, the kitchen cooks to actual demand instead of estimating. Ask whether forecasting uses historical consumption, attendance data, and day-of-week patterns, and whether it plugs into your desk booking or attendance data to know who’s actually in.

2. Menu and vendor management
   Multiple counters, multiple vendors, rotating menus, dietary tags. Can vendors update their own menus, or does everything route through your admin?

3. Entitlement and subsidy rules
   This is where subsidy leaks. Rules should support per-employee daily limits, grade-based subsidies, meal-type restrictions, guest meals and shift-specific entitlements. Ask whether a limit can be set as an amount, a count, or both.

4. Payments
   Wallet, payroll deduction, UPI, card, or corporate account. Payroll deduction matters most for subsidised setups because it removes cash handling entirely.

5. Queue reduction
   Contactless checkout via QR or employee ID, order-ahead with pickup slots, and live queue or wait-time visibility. Ask how long the average transaction takes at the counter.

6. Reconciliation
   The finance question nobody asks in the demo. Can you reconcile vendor invoices against actual verified transactions, line by line? Manual reconciliation of cafeteria bills is where disputes and overpayment live.

7. Waste and consumption reporting
   Prepared versus consumed, by item, by day. Without this, forecasting improvements can’t be proven.

8. Multi-site and multi-currency
   If you run offices across cities or countries, one platform with site-level configuration beats separate systems per location.

Questions that expose weak products

• Show me a subsidy rule being created live, not a screenshot.
• What percentage of orders at an existing client are pre-ordered?
• How does the system know how many employees are in the office tomorrow?
• Show me a vendor invoice reconciled against transactions in the product.

Common mistakes to avoid

• Treating canteen management as a payments project. Digitising payment without pre-ordering might solve the queue but leaves the waste untouched. The forecasting is where the savings are.
• Buying without attendance data. A cafeteria system that doesn’t know how many people are coming in tomorrow is forecasting blind. If you already run desk booking, that data should feed the kitchen.
• Ignoring shift workers. Night-shift and rotating-shift employees have different meal windows and entitlements. Systems built for a single lunch peak break here.