Salem's experiment with on-demand bus service didn't work out, but the innovative service yielded valuable lessons for future transit projects.
More people might ride buses if they worked like Uber. That was the thinking behind PSU planning student Matt Berggren’s flexible on-demand bus plan for a 2.4-square-mile affluent neighborhood in West Salem.
Cherriots, Salem’s transit agency, asked PSU’s planning master’s program to solve its West Salem transit woes. Over five months in 2014, Berggren and several classmates designed the Connector as their second-year workshop project.
The transit agency later hired Berggren as a transit planner, and in June 2015 implemented his plan as a pilot program. It was an elegant idea: Cherriots overlaid a “battleship grid” of stops in the neighborhood. Users requested rides through a smartphone app, website or by phone. Once an hour, a driver in a 14-seater cutaway bus collected ride requests. Using an iPad equipped with Demand Trans software, the driver charted a course.
A Connector driver uses Demand Trans software to plot a route
If running an on-demand bus network sounds like a headache, that’s because it is. As a Salem-Keizer planner, Berggren juggled competing bookings, anemic ridership and other challenges until the Cherriots board voted to discontinue his system. Cherriots will replace the Connector in January 2018 with fixed-route buses.
The project wasn’t a loss, however. Berggren has shared two years of insights with transit agencies around the country: LA Metro, ODOT and the City of Sacramento. The Los Angeles transit agency recently requested proposals for a “microtransit” Connector-like system.
“People call all the time to ask about it,” Berggren said. “I’ve talked to 30 or 40 agencies over the past few years.”
On a Connector ride through the hills of West Salem, Berggren and I talked about the Connector experiment, and the lessons learned.
The conundrum: On-demand bus systems are an attractive option for suburban residents, but they also need density to pencil out.
We took the 17, a fixed-route bus, to the Glen Creek “Transit Center”: two glass booths outside a strip mall. This was the gateway to West Salem.
Before the Connector, some West Salem residents waited nearly two hours for a bus. They walked uphill for a half-mile or so through a neighborhood built for automobiles: large, two-car garages flanking affluent homes perched on wide lots. The last time one of these neighborhoods, Eola, had bus service was in 2009.
The Cherriots board thought the Connector fit West Salem, board President Robert Krebs said, because of the area’s low density, narrow streets, and non-grid-based street pattern.
But with only one bus, the Connector booked up too quickly.
Theoretically, riders could book a trip up to 30 minutes in advance. In reality, if one person a half hour away from the Transit Center requested a ride two weeks in advance, subsequent riders faced a trip of at least an hour.
“People who book two weeks in advance or are subscribers end up creating this custom route for two or three people,” Berggren said.
The Connector is smaller than Cherriots' fixed-route buses
On-demand transit works best in dense urban cores, with faster turnover between trips. That’s why Uber started in San Francisco.
While a Connector would work well in downtown San Francisco, rural areas are in dire need of flexible transit systems. Several Oregon public officials called Berggren asking how to adapt his system for rural routes. He said it would suffer from the same booking challenges.
Krebs said an on-demand network could work in South Salem, a neighborhood that has seen increased development but still lacks good transit coverage.
“I think it has a purpose,” Krebs said. “But it’s for areas with very light traffic.”
Map from Connector evaluation showing the most-used stops
A few people depend on it
On average, about four people board the Connector every hour. During our 20-minute ride, we picked up and dropped off two riders.
At least 13,275 people have ridden the connector since it began. The flexible bus service sees about 52 boardings daily, compared to estimates of around 86 boardings for the fixed-route buses that served the same area.
Do West Salem residents who can afford cars opt for the Connector? Berggren answered with a resounding “No.”
“Essentially everyone wants a car,” he said. “We had nowhere near the level of service to compete with the private automobile.”
A Connector driver
I spoke with a housekeeper who schedules rides between job sites. Like her, 54% of Connector users ride to and from work. Another third are high school students, whose only public transit option is the Connector.
“It’s a low ridership, but these people need the service,” Krebs said.
Others are non transit-dependent senior citizens going to the library or store.
“A lot of the riders I talk to they ride once every three months,” Berggren said. “Like if they have a doctor’s appointment or their car broke.”
A flexible route provokes anxiety in some riders…
At one stop, we spotted a thin pole with a Connector sign. Cherriots had to add the poles because riders got stressed out just standing on a curb with no signage, Berggren said.
Cherriots also started taking walk-ons every hour at the Glen Creek Transit Center and sending texts five minutes before rides.
“Flexed route service has a little more angst to it than fixed route,” Krebs said.
Picture a continuum with “freedom” plotted on one side, and “certainty” on the other. The Connector experiment proved that most people hug the “certainty” end of this spectrum. Though it’s nice to request a ride whenever you want, it’s better to know the bus will come no matter what every 15 minutes.
“Wherever we could we put certainty into this thing,” Berggren said.
…But it also feels cool to ride
Berggren hails from Chicago, where he said people judged you for taking the bus instead of the train. At school in Portland, he found a different culture.
“In Portland, people are respected for taking the bus and biking,” Berggren said. “Salem hasn’t really gotten to that point. There’s still a stigma for some people.”
In some ways, with its iPad software and mobile booking options, the Connector overcomes that stigma. In other ways, users feel less valued than they did on fixed-route. They have to wait longer for service. They can’t travel on a whim.
“Is this a premium service or a last mile service for people who really need it?” Berggren said. “We struggled trying to define that.”
Cherriots downtown transit center, where the fixed-route 17 departs for Glen Creek
Rebranding the bus might not require iPads and apps. It could be as easy as providing better fixed-route service.
Around $7 million in funding from the recently passed $5.3 billion state transportation package will “change everything,” for Cherriots, Berggren said. The agency will put the money toward Saturday, Sunday and evening service; 15-minute service in more areas; and computer aided dispatch for real-time tracking and electronic fare options.
Flexible transit is not yet cost-effective
The Connector costs $234,352 to operate, compared to $365,277 for the planned fixed-route replacement. Cherriots saved money by using smaller buses and contracting labor to a private company.
Taking ridership into account, however, a fixed-route system proves more efficient.
Connector Costs and Efficiency
Total operating cost
Boardings per revenue hour
Cost to Cherriots per ride
As with most U.S. bus systems, about 10% of the Connector’s funding comes from fares, the rest from taxpayer dollars paid into a general transit fund.
Adding another bus would solve many of the Connector’s problems, but that would pull too much money away from Cherriots buses that serve more riders per hour.
“If we kept adding service, since we’re subsidizing it so heavily with tax dollars, we would just blow our budget immediately,” Berggren said.
“Being innovative is not a reason to use a service,” Berggren. "You have to think about the needs of your riders. If you can get what you want out of Fixed-route, or just a smaller bus, that’s way easier to manage.”
With adaptations, bus on demand may work.
Berggren hopes the transit agencies that receive his report will find a way to make on-demand bus service work. As the LA Metro RFP suggest, the experiment has already gained traction.
“The thought is spread that knowledge around and eventually it will come back to us,” he said.
Follow me next week as I learn about expansion of the Columbia Gorge Express and other transit routes.