Neihu is one of Taipei’s most congested employment districts. Every weekday, tens of thousands of office workers leave at unpredictable times. Taiwan’s workplace culture operates on a “stay until the job is done” basis, and leaving time is rarely fixed. With ongoing MRT construction making public transit even more overwhelmed, carpooling has become one of the most common ways for commuters to get home. But finding someone to share that ride has never been easy.

In Taipei, commuters rely on two ways to find a carpool home.

The first is digital. People join LINE groups to post their route, departure time, and fare — then wait for someone to reply. But LINE is fundamentally a chat room. To find a match, commuters have to manually scroll through a mix of posts and unrelated replies, filtering for origin, destination, and time. And because leaving time is rarely predictable, by the time someone posts or finds a match, the ride is often already gone or no longer available.

The second is physical. At designated spots near Neihu’s office buildings, commuters queue up and ask passing drivers where they’re heading. But the window is extremely tight. Police regularly clear the area when too many cars gather, which means drivers can’t afford to wait. And if they leave without a passenger, coming back means another 15 to 20 minutes of traffic. One missed connection, and the opportunity is gone.

Both approaches share the same core problem: commuters have no efficient, reliable way to find a match. Whether it’s scrolling through a chat room or standing at a street corner, getting home still comes down to timing and luck.

Metropia already had the technical infrastructure for real-time matching. The challenge wasn’t building the system. It was designing an experience that commuters would actually choose over a LINE group or a street corner queue.

I worked with the PM to map out what each role actually needed to accomplish. Four user stories drove every major design decision.

Setting up trip preferences

As a driver / rider, I want to set up my trip preferences so that I only get matched with people who fit my route and conditions, without having to filter through everyone myself.

Based on what both drivers and riders needed, the MVP came into focus quickly. The core hypothesis was simple: both sides want to carpool, but have no efficient way to find each other. If we could solve that, giving both sides a way to find a match without having to do everything manually, everything else would follow.

The MVP focused on the daily Taipei commute, the highest-frequency, highest-friction use case. Intercity carpooling was a deliberate second phase, not an afterthought. The existing Facebook communities showed the demand was real, and the same infrastructure could extend to cover it. But solving the hardest problem first, unpredictable, real-time, high-pressure matching, would prove the core experience before expanding scope.