Rumbo: Designing a 0→1 public transport app for Mexico City's 9 million daily commuters
500k+
Downloads in Mexico City at launch
1M+
Users across Mexico City, Lima, and Bangkok
3
Emerging-market cities launched on a single design system
Role
Lead Product Designer
Duration
1 year
Scope
Research · 0→1 Design · IA · UI · Launch
Platform
Android (emerging markets)
The problem
WhereIsMyTransport had built something genuinely rare: comprehensive, real-time mobility data for cities where 80% of commuters rely on informal and semi-formal transport — minibuses, shared taxis, and routes that don't exist on any official map. The data was extraordinary. The challenge was turning it into a product that everyday commuters would actually use.
I was brought in alongside a PM to design Rumbo from scratch — from blank canvas to App Store launch in Mexico City, a megacity where the average commute spans multiple modes, costs are unpredictable, and disruptions are a daily reality.
There was no playbook for this. No comparable app existed in these markets. We were designing for a user who had never had reliable transit information, on a device with limited data, in a city that never stops moving.
What I found in research
Before any wireframes, I ran research with commuters in Mexico City to understand how people actually navigated the city. A few things shaped every design decision that followed.
Most users relied on a combination of word-of-mouth, WhatsApp groups, and memorised routes built up over years. Disruptions — which happen around 60 times a day in Mexico City — were handled by asking other passengers or calling contacts near the destination. There was no single source of truth, and people had learned not to expect one.
Data connectivity was intermittent. Many users were on prepaid plans with limited MB. Any solution that required constant data to be useful would fail in the real world.
Trust was the deepest problem. Users were sceptical that any app could cover informal routes — because none ever had. The product had to prove its data was complete before people would rely on it.
Key decisions I made
Map-first architecture
I designed the entire experience around the map rather than a search-first flow. In a city with flexible, informal routes, search assumes users know the name of their stop or line — they often don't. Building from the map let users orient themselves spatially first, then drill into details. This was a deliberate departure from how apps like Google Maps or Citymapper approach transit.
Offline-first by default
I pushed hard for static offline maps as a core feature, not a nice-to-have. If the app failed on a degraded connection, it would lose the trust of exactly the users who needed it most. This required close collaboration with engineering to cache route data intelligently without bloating the app size — a genuine constraint on lower-end Android devices.
Real-time alerts as the hero feature
WhereIsMyTransport's data advantage was its real-time disruption coverage — every mode, not just formal metro lines. I made alerts the feature we led with in onboarding and on the home screen. It was the proof point that would convert sceptics: "we know about the route you take, and we'll tell you when something changes." This became Rumbo's most-used feature post-launch.
Designed to scale to three cities
From day one, I designed the component library and IA with Lima and Bangkok in scope. Different cities have different transport modes, naming conventions, and UI density needs. Building a flexible design system upfront meant we didn't have to redesign from scratch for each market — we adapted.
The outcome
Rumbo launched in Mexico City and reached 500,000 downloads — without paid advertising, primarily through word of mouth and partnerships with transit card issuers. The product subsequently expanded to Lima and Bangkok, reaching over 1 million users across three markets on the design system we built for the original launch.
The alert system delivered more than 8 million real-time disruption notifications since launch, which became the metric WhereIsMyTransport used most in external communications — a signal that the feature hierarchy I'd prioritised was the right call.
What I'd do differently
I'd invest more in community contribution features earlier. The biggest gap at launch was that users trusted the data but had no way to enrich it. Commuters knew things our field teams didn't — hyper-local shortcuts, stops that had moved, routes that only ran at certain times. We shipped community reporting in a later version, but designing that feedback loop into the 0→1 product would have accelerated trust and data quality from day one.