Bengaluru's metro serves 6+ lakh commuters daily — yet most still queue for physical tokens because the app loses them mid-booking. We set out to find out why, and fix it.
Commuters were abandoning bookings at the station search step — under time pressure, on poor connectivity, often missing their train.
The Namma Metro app — provided by BMRCL — was a growing source of friction. Despite serving 800K+ daily riders, commuters were abandoning digital ticketing and reverting to token queues, defeating the purpose of going digital.
We focused on Heuristic Evaluation and moderated User Testing to assess and enhance the UX of the Namma Metro app — with the goal of improving usability, increasing adoption, and repositioning it as a valuable solution for both commuters and the Government of Karnataka.
Synthesised from interviews with 6 daily commuters — two distinct archetypes that shaped every redesign decision.
We systematically evaluated the app against each of Nielsen's 10 usability heuristics — identifying exactly where the app helped and failed commuters before touching any design.
Six critical findings — each illustrated with hand-drawn wireframes of the actual Namma Metro app screens showing exactly where the experience broke down.
Every change traces directly to a specific finding and a real user struggle. Here's the before vs. after — followed by the reasoning behind each decision.
The original home screen buried departures behind a non-standard icon. The redesign puts a live departures card directly on the home screen — the single most important piece of information for a commuter about to travel.
The Quick Book widget pre-fills the user's last origin station, reducing the booking flow from 4 steps to a single tap for repeat journeys. Card balance and active tickets are surfaced as secondary glanceable cards below.
The original station search required exact spelling of Kannada-romanised station names. 8 of 6 participants failed their first attempt. The redesign introduces fuzzy autocomplete — typing "indir" immediately surfaces Indiranagar as the top result, with phonetic matching for common misspellings.
A 3-step progress indicator at the top gives users a clear mental model of where they are in the booking flow — directly addressing the "no progress feedback" finding. Recent routes are pinned below the dropdown for instant re-selection.
A new screen that didn't exist in the original app — the redesign introduces a live train selection step showing upcoming departures for the chosen route. The selected train is highlighted in purple, making the active choice immediately clear.
Each departure card shows platform, time, stops, journey duration, and fare — giving commuters all the information needed to make a confident choice without leaving the booking flow.
The original 4-step checkout is consolidated into a single confirmation screen. The saved payment method (GPay UPI) is shown by default — no re-entry needed. An inline edit button lets users correct passenger count without going back.
The QR code is front and centre immediately after payment — no navigation required. A "Cancel Booking" button is now available post-purchase (fixing heuristic 03), and transaction details are shown for trust and transparency.
A yellow warning banner surfaces when card balance drops below the minimum fare — fixing heuristic 05 (error prevention). Preset top-up amounts (₹100/₹200/₹500) reduce cognitive load, with the most common amount pre-selected.
Second-round moderated testing with 8 of the original participants using the Figma prototype. Consistent improvements across all age groups and experience levels.
"This actually feels like something I'd use instead of standing in the queue. It just makes sense now."
— Participant P7, Software Engineer, post-redesign testing sessionThe prototype validated the direction — but three things stuck with me that I'd carry into any service design project going forward.
The prototype proved the direction with 15 users. Real confidence comes from measuring booking conversion in production — specifically whether the redesigned station search and home screen reduce drop-off at scale. I'd instrument both flows and run for 4 weeks minimum before drawing conclusions.