Case study ·
2024
Atlas Rail
A seat booking flow for night and day trains across Europe, with a seat map that loads before the price does.
Role
Booking flow
Client
Atlas Rail
Stack
React Native · TypeScript · GraphQL
Result
Search to booked seat in 48 s

01 — The problem
Six screens to pick a seat.
Choosing a seat took six screens, and the seat map appeared only after payment details, so people paid for seats they did not want.
Desktop · 1440 px

01 —
Pick a train in seconds
02 — What I built
Seat map, filters and price together.
A single screen where the seat map, filters and trip summary sit together. The map is cached per train, so it opens instantly while the price loads in the background. The ticket works offline on the phone.
Phone · 390 px

02 —
48 seconds to a seat
Component · 5 states

03 —
One seat, five states
03 — What changed
A booked seat in 48 seconds.
Median time from search to a booked seat fell to 48 seconds, and seat change requests dropped by a third.
48 s
median from search to a booked seat
−33%
seat change requests
