Basics: Slower Speed is a Service Cut

Does your transit agency have a recent history of operating speed that looks like this?

Source: Portland Bureau of Transportation, Enhanced Transit Corridors Plan, Feb. 2018, page 9

This is the “boiling frog” problem of bus operating speed. In a dense and growing city, it’s not unusual to see speeds falling by about 1% a year, as in this data for Portland’s busiest lines.

If you’re going to analyze how service levels relate to ridership, you have to think about speed. Speed is not just a disadvantage for the customer; slower speeds are also a service cut.

You, the customer, want to go a distance, but the transit agency will pay for your service by the hour. So the quantity of service you experience will be governed by how easily hours turn into distance — in other words, by speed.

So if a transit agency budget grows by 1% a year but speed is falling 1% a year, the customer should expect slower speed and no other growth in service.

When measuring the service quantity that affects ridership, then, look at service miles or km, not service hours.

16 Responses to Basics: Slower Speed is a Service Cut

  1. John D March 26, 2018 at 9:15 am #

    we needs more priority lanes for buses that offer a consistent speed all day long, every day. Why this is so hard to achive?

    • Mike March 26, 2018 at 12:03 pm #

      Because it often requires taking parking lanes or general-purpose lanes to fit the transit lanes into a narrow right of way, and drivers have a lot of political clout. In Paris and other cities the government would just do it, but American cities are generally afraid to. Shops also complain that they’d lose most of their customers if street parking is reduced, ignoring the walk-up customers who would come if transit were better. In cities with wide rights of way where you can add transit lanes without reducing general-purpose lanes or eliminating parking you’re starting to see more transit lanes.

      • el_slapper March 28, 2018 at 4:49 am #

        It was easy in Paris because most car drivers in Paris come from the Suburb. The successive leftist mayors had no problem sacrificiing the poor suburban right voters to improve the comfort of the wealthy parisian left voters(I’m simplifying a lot, but that’s the overall picture).

        Adding bus lanes(and later replacing them with big trams that are always overcrowded) was the thing to do, but it required a specific, favorable political configuration to be successfully done. Had most of the car users been Parisian voters, it would never have succeeded.

  2. Poisson Volant March 27, 2018 at 7:42 am #

    Another circularly negative effect of slower transit speeds is that a fraction of riders will decide to use automobiles in the form of personal cars or ride-hailing – which of course adds to the congestion, further slowing down travel times.

  3. mike dunn March 27, 2018 at 11:08 am #

    As far back as 2000 I could travel home from the MTA building to West Hollywood in half the time by auto including a short shopping trip to Ralphs as opposed to taking the RED Line and transferring to the 704 at Vermont or continuing on to Highland and utilizing the 780 and 704 lines to get home.

    Early in the morning I still see total gridlock on Santa Monica Bl. westbound in West Hollywood and again eastbound in the afternoon. This does not just exist in West Hollywood but from Downtown L.A. clear to the Santa Monica border. And Sunset Bl. is no better if not worse.

    Has the MTA addressed the problem or planned to address the problem with Light Rail? NO!! They would rather extend the Gold Line to Montclair Plaza outside of Los Angeles County which also includes duplicating the current METRO LINK service. The east-west travel corridor thru west Hollywood was recently address in this MTA site and how the proposed extension of the Route 2 freeway was killed by Governor Brown. Why not take the bold move and restore the freeway funding including inflation and build the much needed Line Rail Line?

  4. asdf2 March 28, 2018 at 6:17 pm #

    It’s actually worse than this. In many areas, traffic is highly variable – one day, a road is free flowing – the next day, at the same time, it’s bumper to bumper. In order for buses to reliably follow a schedule, the schedule has to be set for the worst-case traffic conditions, so even if traffic is only bad 20% of the time, the agency has to pay the service hours for a bad traffic day 100% of the time. On a day when traffic is light, you don’t get more service – the bus just sits longer in layover.

    Even worse, variable traffic often means a lot of time points in the schedule where the bus is going to be early on days when traffic is light. Which means, on days when traffic light, the bus is going to be doing a lot of sitting idle in the middle of the route, and passenger travel times are now based on bad traffic days every single day – the only way to move faster on a good traffic day is to abandon the bus and drive.

    I saw this first-hand once, riding the bus on Christmas Day. The roads were empty, and every mile or so, the bus would sit and wait for a few minutes before continuing along its route. It was quite depressing.

    • Jack March 30, 2018 at 4:37 am #

      I just finished retiming our fixed route system, and this is exactly the frustration I had. Currently, we are more or less late much of the time, so the bus is always moving. Our system was last retimed about 15 years ago. In that time, ridership has increased 60%, resulting in buses stopping more, and of course development along the routes has slowed service down.

      Travel times, even at the same time of day, vary significantly along many routes. Therefore, a given schedule is either always on time, but often has down time along routes, or it moves more and is sometimes late. Deciding where to place the schedule to maximize speed, on time performance, as well as maintaining maximum service without a budget increase was a difficult decision. I did some research to see if there was a magic formula I was missing, but it was clear everyone, even the big guys, have to make this same decision.

      To maintain the budget, we had two routes that lost one loop per day to prevent a significant increase in cost, despite not reducing service hours.

      All our fixed route service is by bus in mixed traffic. Jarrett’s post, and your comment asdf2, were very appropriately timed for us.

      • Steve Lax April 1, 2018 at 2:29 pm #

        There is another variability, the bus operator. If you add running time, the good operator will adjust so that he does not run hot. The operator who doesn’t care (or who is a day off or vacation relief operator who doesn’t have a feel for the route) will sometimes run hot and sometimes run cold. But overall, there will be an additional increase in running time after one adds running time in many instances.

        While GPS tracking might identify a few slowpoke operators, it is generally hard to modify their behaviors.

        • Jack April 2, 2018 at 4:42 am #

          True, operators add a human variable that can be good or bad. This is a supervision issue, and there are good supervision practices that can effectively address it. Unfortunately, transit systems with a budget crunch often see the supervision level of the organization as an attractive place to trim, since it doesn’t have an immediate impact on service level on the street or ability to comply with regulations. However, this often results in long term service problems and poor labor relations, both of which are much harder to fix than they are to break.

          I’d still rather deal with the human element than concede that “if we give them enough time in the schedule, they’ll just sit around lazy and I want them to have to keep working.” I’ve actually heard that said by a transit manager responsible for a system’s timing. To me, that seems like terrible motivation to give a poor product to the customers.

      • asdf2 April 2, 2018 at 6:41 pm #

        I personally prefer the option of putting all the padding into the layover. This means buses will usually fall behind schedule, but they’ll catch up between runs and at least begin each run on time. Then, add a good real-time feed so that people can check their phones to see when the bus is actually coming.

        For frequent service, it doesn’t matter if every bus is 5-15 minutes late, as long as there’s a steady stream. In fact, when headways get to 15-minutes or better, it’s usually best to just ignore the schedule completely and focus only on the real-time.

        The only drawback of this approach is that it can cause trip-planning software to give unreasonable travel time estimates, or suggest connections that cannot be relied upon. Ideally, this would be solved on the software layer, by having the trip planner software itself analyze the real-time feeds day after day along each route, and figure out how long a trip will actually take.

        • Jack April 9, 2018 at 4:30 am #

          In a high frequency environment, I think you are right, asdf2. However, in a low frequency environment (every hour or more, as much of my service is), this means making promises we know we aren’t going to keep, and people being late to work. When taking the earlier bus means arriving to work 1.25 hours early, people will risk a bus that shows up to work 5 or 8 minutes early on the schedule. (Same with connections). Therefore, for trip planning in a low frequency environment, the arrival time is just as if not more important than the departure time.

          We also don’t have the luxury of apps, due to funding.

  5. visitor April 2, 2018 at 2:40 am #

    Notice the operating speeds in the chart – 10-15 mph. Most people can achieve or exceed that on a bicycle, at least on flat ground. Bicycling is good exercise and a great alternative to buses.

    • Jack April 2, 2018 at 4:43 am #

      It is! Biking also has limitations due to poor bike infrastructure, weather, an aversion to showing up to work sweaty, and many people’s physical limitations. However, when possible, I think biking is a great addition to the transportation options list.

  6. Transit Planner April 2, 2018 at 2:28 pm #

    Probably the travel speed was calculated from the timetable of the bus lines. For correct interpretation the punctuality of the lines and how it developed over the years would be needed – maybe in 2009 the speed on paper was faster, but punctuality was down so more time was added into the timetable – so that the timetable reflects now reality better than before. Or maybe extra stops were added in the meantime and ridership went up. So the given data can also mislead. Yet the main point is true: travel speed, real travel speed, matters.

  7. Dexter Wong April 3, 2018 at 3:12 pm #

    Yes, biking can be faster than driving. When Metro opened their Santa Monica line, a local radio station held a contest between a car, Metro rail, Metro bus and a bicycle over who would get there first. Suriprise, the biker got to Santa Monica first, then the car, Metro Rail and then Metro Bus.

  8. Georgist Economist April 7, 2018 at 12:39 pm #

    It is not explicit in the article, that a decreasing speed is also an increasing headway, if the same number of buses serve the route. This also means a decreasing capacity, thus if ridership were to stay constant and high, it would also exacerbate crowding.

    In fact, this is largely how the cars killed the streetcar. Traffic lights introduced for the cars (previously, no such device was needed) and general congestion slowed down the streetcars. This hurt their usefulness in at least three ways: speed, wait time, crowding. As passengers increasingly left the service that was less and less useful, the companies went bankrupt, eventually replaced by buses.

    The bustitution also had an “optical illusion” behind it. Because the streetcars of old stopped in the middle of the road and had passengers walk on the road surface to/from the curb, a stopped streetcar blocked the road, and cars waited behind it. When the passengers were done alighting and boarding, and the streetcar pulled away, there was some stretch of empty road in front of it. Buses, on the other hand, pulled to the curb and cars could pass them while passengers were alighting and boarding. As a result, car drivers saw an illusion: it looked as though the streetcars were an obstacle to car traffic in a way that buses weren’t. Consequently, they supported the bustitution.

    Of course, it didn’t help that modernist architects (who had a huge influence on land use design) hated streetcars and adored cars, either. Their ideas for what the city should look like (such as Le Corbusier’s Radiant City) consisted of enormous towers with freeways for cars between them. The modernists’ designs were not only mobility-geometrically unsound, they were also revoltingly ugly blocks of concrete and glass, as well as uncomfortable to live in, and dependent on mechanical ventilation&lighting. With as little as 8 feet of ceiling height, windows were barely 4 feet tall, and there’s perhaps 2′-2’6″ headroom. Even +25% height, 10′, gives +50% light and natural ventilation, as well as twice the headroom. That makes the room feel very different, and much more pleasant. +50% ceiling height admits fully twice as much sunlight and air, and close to three times as much headroom. How on Earth did the modernists manage to do EVERYTHING wrong, even though they had working examples of transit and land use synergy (streetcar suburbs), beautiful neoclassicist architecture, and sufficient ceiling height?