You're a PM at Lyft and want to deploy a ride-sharing service using a self-driving car in Palo Alto. How would you choose around 10 pickup and drop-off locations, assuming routes are safe and pre-programmed?

To select 10 key pickup and drop-off locations for a self-driving car (SDC) pilot in Palo Alto, I would focus on areas that maximize learning and rider adoption while minimizing risk. My approach would involve:

1. Identifying High-Density Demand Zones: Pinpoint neighborhoods, business districts, and popular transit hubs (like Caltrain stations or university areas) where ride requests are frequent. This ensures we have sufficient rider volume to test the service.
2. Selecting Diverse Travel Patterns: Choose locations that represent a variety of trip lengths and purposes (e.g., commuting, errands, entertainment). This helps assess SDC performance across different use cases.
3. Prioritizing Areas with Predictable Traffic: While routes are pre-programmed, selecting areas with generally stable traffic patterns will allow for more reliable ETAs and a smoother initial rider experience, reducing frustration and improving trust in SDCs.
4. Considering SDC Operational Constraints: Factor in locations that are easily accessible for SDCs, avoiding complex intersections or areas prone to unexpected road closures that could disrupt the pre-programmed routes.
5. Balancing Pickup and Drop-off Points: Ensure a good mix of locations that serve as both origins and destinations, creating a network effect and encouraging repeat usage within the pilot zone.

By focusing on these criteria, we can strategically choose locations that provide the most valuable data on rider acceptance, SDC performance, and operational efficiency within the Palo Alto pilot.