Booking a train in three taps

Research

We started with the booking data: 71% of trips were a repeat of one of the customer's last five journeys. I ran a diary study with twelve commuters for two weeks and asked them to record every booking they made, in the app or elsewhere. The website won whenever they already knew the train they wanted.

We also learned when people book. Most repeat bookings happened on the platform or on the walk to the station, with one hand free and a train about to leave, so every decision was tested standing up, outdoors, with gloves on.

Decisions

The home screen became a list of recent journeys with today's next departure and price on each. One tap opens the trip with the last fare type and railcard already chosen, a second confirms seats, a third pays with the saved card. Search is still there, one swipe away.

The fastest booking form is the one you have already filled in.

We removed the seat map from the default path and showed it only when a customer asked for a specific seat. That one change accounted for most of the drop in abandonment.

Shipping

We tested the new flow against the old one for six weeks with half of the iOS users, then on Android. Checkout completion rose 14%, the median booking took 41 seconds less, and the store rating climbed from 3.9 to 4.7 over the following quarter.

Accessibility was part of the brief from the start. Every step works with VoiceOver and TalkBack, departure times are read out in full, and the fare list keeps its order when text is enlarged to the largest size the phone allows.