Wayfind

The problem

Ride diaries and intercept interviews with 14 commuters showed the same pattern: riders opened the app 6–10 times per trip, always for the same two facts — the next departure and whether their connection was safe. Every open cost them 20+ seconds of navigating to that answer.

The team had been shipping features (trip planning, saved places, alerts) while the core loop stayed slow. Success metrics rewarded feature adoption, not time-to-answer.

Process

I reframed the product around a single measure: seconds from open to answer. We mapped every rider question against frequency and urgency, which made the priority order obvious — departures first, disruptions second, everything else behind a tap.

Prototypes went from paper to code in weekly cycles. The breakthrough was a “commute card” that learns your two or three routine trips and pins live departures for them on the first screen — no search, no map, no taps.

The solution

The redesigned home screen answers the top question in under two seconds: your next sensible departure, with a confidence indicator when live data is stale. Disruptions rewrite the card in place instead of firing notifications riders had learned to ignore.

A restrained visual system — one accent color, big numerals, generous spacing — keeps the glanceable hierarchy honest even when service is falling apart.

Outcome

Median time-to-answer dropped from 22 seconds to under 3. Weekly retention rose 38% over two quarters, and support tickets about “where is my bus” fell by half. The commute card is now the pattern the whole product is being rebuilt around.