Case study Ticketing Platform
A service floor that knows who is waiting, and for how long
Walk-in service environments share one operational problem: visitors arrive faster than staff can be told about them. Without a structured queue it falls back to paper tokens, and management estimates service times from memory.
Services Product engineering & real-time systems
Category Queue management system
Client Browser Automations
Stack Next.js · Supabase Realtime · Postgres · TypeScript
Ticketing Platform
The problem
Counters operating as isolated islands.
A counter agent had no idea what was waiting until someone walked over. A receptionist had no live view of how counters were performing.
✓Paper tokens or a shared spreadsheet as the queue ✓No visibility into which counters are free ✓Customers with no idea how long they will be waiting ✓No structured capture of arrival, wait, call or completion ✓Staffing decisions estimated from memory ✓Reconfiguring the floor meant touching code
What it was costing
✓Service times unknown, so staffing was guesswork ✓Visitors standing without information ✓Agents calling across the room ✓No historical picture of floor performance
Our approach
Push state changes, do not poll for them
The core runs on Postgres change subscriptions, so a ticket issued at reception appears on the counter screen and the lobby display within the same second.
✓Ticket state machine with every transition timestamped ✓Per-category numbering sequences running independent counts ✓Priority flag letting appointment and VIP tickets insert ahead ✓Row-level security scoping each role to what it may read and write ✓Unauthenticated real-time display client for unattended lobby screens ✓Counters, services, routing and hours configured through an admin layer
Core capabilities
What it does now
One connected system instead of three disconnected touchpoints.
✓Check-in in seconds with routing handled by service type ✓Live prioritised queue on every counter screen ✓Public display with audio chime on each new call ✓Wait time, service time and throughput per counter ✓No-show and abandonment rates over any period ✓Floor reconfigured by operations staff, not engineers