Transit agencies across the US were launching on-demand microtransit services to help riders connect to buses and trains, but nobody had focused on timing the connection. Ford, in partnership with King County Metro and Cornell University, designed a feature called "Transit Connect" allowing users to book trips around the real-time transit schedule.
My work focused on the human side of that question: who are the users, do they trust it, and and how do we need to shape the service in response to their behavior?
I led the Ford core research team on a five-year, $2.5M DOE-funded project. My contributions spanned the full research lifecycle: co-designing the national user survey and value-of-time study used to calibrate service parameters, leading service design collaboration with King County Metro, and contributing to qualitative analysis of pilot user surveys and interviews. I also represented Ford in multi-party planning across Ford, Cornell, King County Metro, and the City of Kent.
The project had two main research threads running in parallel: modeling and simulation (led by Cornell), and user behavior research (shared between Ford and Cornell). My work sat primarily in the latter.
Before any pilot launched, we needed to understand how riders actually valued microtransit service attributes. We designed and deployed a survey of 2,400 commuters across four US metro areas: Washington D.C., Miami, Seattle, and Minneapolis. All respondents lived within five miles of a transit station and commuted regularly. The survey included a discrete choice experiment asking respondents to choose between hypothetical service configurations at different fare, wait time, and travel time levels.
The goal was to answer two questions: what is riders' value of time for microtransit, and who is most likely to use it?
Once the service launched in Kent, WA (branded "Ride Pingo to Transit"), two surveys were fielded to understand who was actually using it and why. The first, conducted by King County Metro in December 2021, gathered early feedback from 37 riders. The second, a larger-scale effort by Cornell and the University of Oregon in late 2022, reached 169 residents, including both users and non-users who had downloaded the app. Fifteen riders were also recruited for follow-up interviews to understand how the service was affecting their lives.
Running alongside the surveys, I worked closely with King County Metro through an iterative service design process spanning multiple years. This involved translating simulation findings into policy decisions around waiting time limits, detour parameters, transit hub placement, and service hours.
Riders value wait time far more than travel time. The national survey found that commuters valued their access time (waiting and walking) at $75.38 per hour, compared to $18.63 per hour for in-vehicle time. Riders were nearly four times more sensitive to wait time than to how fast the shuttle drove. This was the most consequential behavioral finding for service design.
The service reached the people it was designed to reach. Early pilot survey data showed that roughly 50% of riders came from low-income households and nearly 60% lived in car-free households. Most (62%) were using the service to get to work. Riders chose Ride Pingo primarily because it was convenient (51%), safe (41%), and cheaper than alternatives (27%).
The Transit Connect feature had a committed core but a usability problem. Transit Connect allowed riders to request a guaranteed connection to a specific bus or train departure. About 40% of riders used it at least once, and among that group, 60% became regular users. However, a significant portion tried it once and stopped. Post-pilot analysis revealed a likely cause: the app allowed riders to book Transit Connect trips far in advance, so many arrived at the station long before their connection, undermining the on-demand value of the feature. This was a product design problem, not a demand problem.
Operational performance was strong but revealed a design gap. The shuttle successfully made 97% of bus transfer connections and 99% of train connections. However, shuttle arrival times showed standard deviations of 6 to 7 minutes, and the absence of a maximum advance booking constraint in the app meant guaranteed connections often came with unexpectedly long station wait times.
Service parameters were grounded in behavioral data. The value-of-time survey directly shaped key operational constraints. Given riders' high sensitivity to wait time, the team set the maximum wait time at 35 minutes at launch, later calibrated to 45 minutes on weekends based on real demand patterns. The 8-minute minimum buffer for Transit Connect connections was validated by the high transfer success rates observed in the field.
Transit hub configuration was shaped by simulation informed by user research. Simulation, calibrated with survey data, showed that adding a second transit hub reduced average travel time but increased wait times at both locations. King County Metro accepted that tradeoff, and the two-hub design was adopted for the final service.
Marketing investment was validated as the primary lever. A coordinated summer 2022 campaign including social media, radio, outdoor signage, and street teams doubled weekly unique riders in a single week (from 60 to 122), providing concrete evidence that awareness was the primary adoption barrier, not demand.
Transit Connect needs a booking window cap. The finding that many riders were arriving at stations far too early pointed directly to a product fix: implementing a maximum advance booking constraint in the app. This was identified as the primary improvement needed before scaling the feature to other cities.
The equity case for permanent microtransit was strengthened. Survey findings showing disproportionate use by low-income and car-free residents gave King County Metro a strong rationale for continuing first/last mile microtransit as a permanent service rather than a time-limited pilot.
Ford Motor Company / U.S. Department of Energy. "Integrating Microtransit with Public Transit for Coordinated Multi-Modal Movement of People." Final Technical Report, Award DE-EE0008464, December 2023. https://www.osti.gov/biblio/2335891
“Commuter Preferences for a First-Mile/Last-Mile Microtransit Service in the United States.” Transportation Research Part A: Policy and Practice 167 (January 1, 2023): 103549. https://doi.org/10.1016/j.tra.2022.11.009