If you were an Uber driver or a dasher for Doordash, which neighborhood in the above picture do you want to go to?
Scale matters greatly if we want to keep the price of goods or services low. It determines the economics of the business.
In transportation, scale is measured by the population density, or the area size and population size, something very intrinsic to the region and hard to change. Scale matters because it directly relates to the sales of goods or services. It underlies the availability, accessibility and affordability of transportation service. If a service can be rendered in one place and used in another place, such as legal services or accounting services, and the cost of delivering the service is very low, then the usage can be scaled by reaching out to more regions. However that’s not the case for local transportation. S, services available in one area may be accessible in the adjacent areas, but can’t be used in farther areas without significant transferring cost, i.e., paying drivers to drive from one area to another. This means the scale of usage (# of rides) that local transportation service can achieve is largely determined by the population density of its service area.
However the overall cost of providing the same bus service (i.e., vehicle and driver costs) may not vary much from region to region in the US, which means the regional cost per ride or per rider can vary greatly with the population density. For example, the cost per ride for the bus services is about $25 in San Jose with an urban density of 5700 per square mile, and $35 in San Ramon with 4000 per square mile. Both fall within the nation’s range of $20 to $35. Therefore each city or municipality has to plan their transit services based on their budget and funding if they want to keep the fare affordable for the masses. A threshold, i.e. $35 per ride, is typically used to determine whether a regular service can be provided. For the areas with higher than this threshold, then dial-a-ride service is provided instead of regular service to strike a balance between the budget limit and service availability.
The biggest challenge for providing transportation services in the US is that the country predominantly consists of suburban and rural markets. About 40% of the US population live in one of the suburbs with a population between 10,000 and 1,000,000, and 45% live in rural villages of less than 10,000 residents. Only 5% of Americans live in one of the nine cities with one million or more residents, compared with more than 100 cities in China.
Another challenge for reining in transportation cost is sprawling. Suburban Americans live in single-family houses, far apart from each other. This makes designing routes for fixed-route bus services extremely difficult. In an ideal market, i.e., densely populated area, fixed route bus service is the most effective way to maximize ridership. Riders can easily walk to and cluster at the bus stop, and buses can collect clusters along the defined route. However in suburbs, it becomes very ineffective as potential riders are scattered and away from bus stops and the last-mile hurdle deters them from using the service. Hence the ridership in American suburbs is very poor. For example, 5 riders per hour is considered an excellent ridership, and a dream for many suburban transit agencies.
Obviously, fixed-route bus services don’t work well in the largely suburban US market. After seeing on-demand service’s growth, many companies then wanted to apply dynamic route services to the suburban market and leverage the on-demand technologies similar to Uber or Lyft’s which uses real-time dispatching and real-time rider matching for rider pooling. However we do not believe that this is the right approach. Yes, on-demand dynamic route service can eliminate the last-mile problem, but that’s not enough to increase ridership in the sprawling suburban areas because the probability of finding two matched rides on-demand decreases rapidly with the drop of population density.
We believe dynamic route service on its own is not enough to improve ridership in the suburban market. The ability of pooling multiple riders in one trip is critical for ridership improvement. Since it’s ineffective to pool riders on-demand in the suburban market, we created a new method of booking rides which can be adjusted based on the population density to optimize pooling rate in all markets. In addition, since riders like the convenience of on-demand, we added the on-demand experience to our new booking design which relies on real-time data collection and processing capability of our system. We invented a new service model to optimize ridership in different urban density. We call our proprietary service model, Duet ride, which is a dynamic route, on-demand like service. Moreover, we’ve demonstrated 20 to 40% utilization improvement, i.e., rider per vehicle hour, in two of our initial pilots.