Translink Buzzer Blog

All about managing the transit network: an interview with senior planner Peter Klitz

A detail from the cover of our Managing the Network primer.

This post is first in a series about Managing the Transit Network: all about how TransLink plans transit service in our region. See all the blog posts in the series here.

How do we design our transit network? What principles are behind building a good bus route? And does the layout of a city affect how well transit can serve its citizens?

We’ve aimed to answer all these questions and more as part of a new program called Managing the Transit Network. It’s an evolution of our Service Optimization program, which we’ve talked about before on the blog—and a big part is about helping the public understand the principles and evidence behind our work.

Check out the new Managing the Transit Network section of our website for many more details, especially:

  • the Managing the Transit Network primer: an easy-to-read booklet about our transit planning goals and principles
  • the Bus System Performance Review: a detailed performance report for 2011, plus a series of route-by-route technical summaries identifying just how our system is performing

And you can also join us for a multi-part Managing the Transit Network series on the blog in the coming weeks, exploring the concepts behind transit system planning. Think of it as a mini “Transit Planning 101″ class—we’ll have lots of examples, expert advice from our planning staff, and discussion questions for everyone to think about!

To get us started, I asked senior planner Peter Klitz to walk us through the concept behind Managing the Transit Network, and a bit of context and explanations for the primer and bus system performance reviews.

How did the Managing the Transit Network project come about?

TransLink is responsible for what we call the architecture of the network – deciding where routes go, the places we serve, dealing with issues of chronic overcrowding and looking for areas where the level of service we provide is greater than passenger demand.

We’ve always designed our services based on sound planning principles— we don’t just throw darts at a dartboard and hope for the best. The Transit Service Guidelines are an example of past attempts to capture those principles. In recent years we’ve benefited from access to large amounts of ridership data through our automated passenger counting system. We’re learning more and more what factors drive higher ridership and productivity – those characteristics that contribute to the success of individual transit lines. Putting together the Managing the Transit Network Primer gave us an opportunity to gather those transit service design considerations together in one place and put them in a coherent document for everyone to share and learn from.

Planners are routinely asked questions about why we make the decisions we do…Why can’t a bus route divert onto my street to pick me up outside my house? Why isn’t there a bus route for every subdivision in Metro Vancouver? Why isn’t there bus service to the industrial park where I work? I just bought a house and realized there’s no bus service in my area…when are you planning on putting a bus on my street? The Primer gives us a tool to communicate to people our objectives for transit service, some of the constraints and challenges we face, what impact land use decisions have on transit ridership, and how people can make smart locational decisions. The Managing the Network primer is a summary of who we are, what we do, and what we think about when we design transit services. The Managing the Network primer is a summary of who we are, what we do, and what we think about when we design transit services.

What’s the difference between this and service optimization?

Network management is the evolution of service optimization. Service optimization was a one-time project aimed at achieving the goals of our 2010 Funding Stabilization Plan. We met the objectives of the project and through the process it became obvious that optimizing our service is something we always want to be doing. So we rolled the lessons we learned into an ongoing Network Management program.

What do you hope people understand after reading the primer—the cornerstone booklet outlining our work?

Hopefully you’ll understand three things. First, you’ll have a better understanding of what planners think about when we design services. Second, we hope people – particular people who make decision about land use – will have a better understanding of how land use decisions impact transit service – how certain land use decisions can help the transit system be efficient and productive and how other land use decisions can actually be harmful. Lastly, we hope people will have a better understanding of the importance of location as the key factor in determining your access to transit service.

The last point is the most important. If you take anything away from reading the primer it should be this – your access to transit service is not based on who you are, how old you are, your gender or your income. Your access to transit is based on one thing only – your location.

People have to take more responsibility for their locational decisions, especially as it pertains to transit. If you or someone in your family will use the transit system now or in the future, you have to think about locating in a place that has access to good transit service. Those people that move to the middle of nowhere and then call up TransLink saying “Hey, I just moved to this new subdivision and realized there’s no bus service out here so when are you going to give us a bus route”…well, we’ve all had a great example in the past few years of how hard it is to fund transit expansion and it’s not going to get any easier. Cities and developers need to make good decisions on where to build, and people need to make good decisions on where to live. That’s the key message.

What’s the role of the Bus System Performance Review?

The Bus System Performance Review is one of the most important things we produce. It tells us exactly how our system is performing. It gives us the ability to plan using evidence, and gives everyone else the understanding that we plan our services based on what we observe on our system. We generate summaries on a route-by-route basis, to a high level of detail—we look at a number of key performance indicators, and analyze our service by day type, by direction, by time of day.

At no time in the past have we had as much information about the performance of our system as we do now. Planners at TransLink are taking full advantage of the opportunity. We’re developing new methodologies and practices. We’re also making the information available to everyone in the region. Gone are the days when planners made decisions based on a little bit of data and a lot of professional judgement.

The Bus System Performance Review is planned to be an annual report we produce. Every year, we’ll review the performance of the system on a systemwide basis to give us our baseline. We look at performance on a sub-regional level and on a route by route basis. And it will have a summary of past changes we made and the impact of those changes on ridership. We’re always looking for ways to improve so future performance reviews might have other elements as we go along.

When will it be released annually?

We collect the data for the performance review during the September quarter of every year because that’s generally the heaviest time of year for ridership. The September quarter runs from Labour Day to late December, at which point it takes about three weeks for the data to be transformed and uploaded into the system.

The reports are huge so generating a report from the data can take a few weeks and it usually takes a couple of weeks for us to look through the data reports, examine everything, summarize the data into the format you see in the performance review. I expect that a refreshed version of the performance review will be available in the spring of each new year.

There’s a lot of information on the route performance summaries. If you’re a novice, what should you look at to get a general sense of what’s going on?

We understand people have different levels of comfort looking through technical information so we tried to create different sections that would appeal to different people depending on your level of technical proficiency. The first page of each summary contains the basic line information on key characteristics and route performance. It should be relatively easy to digest.

The graph on the bottom of the first page is what we call the route thermometer. The graph measures productivity on two fairly easy to understand performance measures: boardings per revenue hour and capacity utilization. Both tell us how many people are using our services as compared to how much service we are supplying. The graphs are coded by colour: blue, green, orange and red. The colours indicate how the route is performing compared to the other routes on the system. Blue and green are good – orange and red are not so good.

For example, take the 99 B-Line summary. The 99 B-Line is the best performing line in the system. 99% of its service hours in 2011 are in the top 25% of all service systemwide. This is indicated by the long blue bar showing 99% of its hours were in the top 25% of routes performing in the system. For some of the lines in our system the story is not always as positive. Long red bars tell us that the majority of revenue hours on that line are performing in the bottom 25% of all services systemwide. So the more blue and green bars you see, the better the route is performing. The more orange and red bars, the poorer it is performing.

The second page of each route summary has more detailed information on performance by day type, time of day and direction. So the first page is designed for the interested viewer with less technical knowledge; the real technical detail is contained on the second page.

The primer talks about this briefly, but what does it mean for a bus to be efficient and productive? As a rider, what does it look like?

Think about sitting on your bus. You look around and see that all the seats are full and there’s a few people standing in the aisle. Not so much that it’s too crowded – maybe 4-5 standees. But that’s okay because everytime the bus pulls into the next stop a few people get off and a few people get on. So anyone who was standing can now take a seat.

This happens at almost all the stops on the route. You look across the street and see the bus heading in the reverse direction. That bus has just as many people on it as your bus, and there’s people getting on and off the bus at most of the stops in that direction as well. You remember that the last time you took this bus on a weekend there were about the same number of people on it then as there are now. You’re bus is busier during peak periods and it gets really crowded sometimes but because there’s so many people getting on and off the bus at each stop pass-ups are rare.

So that’s what a productive bus route looks like to us. Good loads that fill the seats with a few standees, a high degree of passenger turnover along the way, strong ridership in the reverse peak direction and during off-peak hours, busier during peak periods but relatively few pass-ups and the bus manages to keep a competitive speed versus the other traffic on the road.

So what’s an appropriate length of time for a bus route to mature?

It really depends on the service area. Different parts of our region absorb capacity at different rates.

For a bus route in a heavily urbanized area we can expect any new seats we put out there will be filled in a few months – sometimes immediately after a service improvement has been made.

Other parts of our region may require more time for additional capacity to be absorbed. A new line in one of our suburban areas may take two to three years to generate productive ridership. In those cases building a transit network is a longer-term investment.

The idea that we have different goals for our service across the region speaks to our three network design objectives: maximizing short-term ridership, encouraging long-term ridership, and providing access to transit service around the region. The question remains, how much of our resources do we spend on each objective? That’s a question for another discussion.

What are the next steps?

Based on the system performance review, we know how the system is performing. We also know the characteristics that make for a successful transit service. The next step is to go through our system, look for services that are underperforming and propose changes to those services to bring them into better alignment with our objectives and service design considerations.

Can be make a circuitous route more direct and legible? Can we serve an area of higher demand? Can we re-design a service to be more efficient and more effective? Can we make our services more useful to a greater number of people? What are the issues? What are the opportunities? What are the tradeoffs? Who might we impact? Are the potential benefits greater than the costs?

These are the questions we’ll be asking. We hope to have some ideas for people to comment on in the near future.

Thanks Peter!

Questions, comments and more

First, please do take a look through the Managing the Transit Network section of our website, and get familiar with the documents there. The primer is particularly an easy, informative read!

And please leave any feedback and comments below — I’ll pass them along to Peter and the rest of the planning team for a response.

I’m looking forward to digging into these topics more over the next few weeks!


23 Comments

  • By Eugene Wong, June 28, 2012 @ 1:34 pm

    Thank you, Peter, for the hard hitting facts. I appreciate how straight forward you were.

    Jhenifer, do you have any information on whether or not the marketing department can speed up how quickly a route matures? Aren’t there good marketing techniques for this? I see no attempts to promote new routes a year or months in advance, and I assume that this could lower the net costs of Translink. With Apple, people line up to buy the first few. With the buses, it doesn’t seem very hyped up.

    Community members, do you think that there might be a way for Translink to start a new route in the suburbs and rural areas that might lower the cost of long term investment?

    I was under the impression that Translink was going to send out cars, many years ago. Maybe car service for a year would build up support.

  • By JKKT - Kyle, June 28, 2012 @ 3:01 pm

    I have a question about the Break-even point for the Cost/Boarded Passenger. From my studies, this would mean the cost each time a passenger gets on a bus, and excludes transfers. The 99 obviously is at full cost recovery at $.61 per passenger. The 312, ranked #8 in the system, must also be at full cost recovery.

    Lets assume that a passenger uses 2 to 3 or an average of 2.5 buses in their trip, with $2 the average fare. That would mean that to break even, a bus must have a $0.8 cost per boarded passenger cost. I know it’s not this linear and simple, but am I anywhere close? According to another doc, Annual Service Cost = [Annual Revenue Hours] x [Average Cost per Service Hour]. Does this include: 1)Driving to and from the depot, 2) maintenance of the bus? Are the Cost/passenger MUCH higher in Richmond because of the less fuel efficient buses, or because of lower ridership, I think it is the former.

    For some routes, like #134, 136 which fail in both cost/passenger and capacity utilization, I have heard a statement by one of translink’s planners that the only reason why comm. shuttles are not used is due to the higher costs in getting the shuttles TO the route than the lower operating costs. I thought that from now until 2015, there must be a huge amount of space at BTC.

    The C23/C21, rated #3 in Cost/passenger, but yet standard buses are not used. the argument from that planner was that there is simply not enough buses. It’s clear that the distance from Oakridge => downtown is > or = to the distance from BTC to downtown.

    I think that the cost/ boarded passenger should be linearly related to service levels and capacity. Transit will never be profitable, but since the commissioner sees it as a business, you should try to show him how many protests will result when the least profitable routes are removed in favour for ones with lower cost/passenger.

  • By Jhenifer Pabillano - Buzzer Editor, June 28, 2012 @ 4:01 pm

    Passing these all on to the planning team! Answers soon.

  • By mike0123, June 28, 2012 @ 8:57 pm

    A couple questions:

    Would it be possible to post all of the BSPR route summaries in a format that is easily machine-readable, for example in tab-delimited text, .ods, or .xls? This would make it simple for people to view and analyze the data in a spreadsheet program.

    For each route, would it be possible to add the route length, number of stops, and number of turning movements to the summary and the average trip time and the average time spent at timing points to the current performance tables (i.e. by time of day). These measures could be used to calculate the average speed of the route, which I think should be a key performance measure, taking into account some of the things that slow buses down. It could also be used to monitor this aspect of performance from year to year, for example from removing unnecessary stops, adding transit priority measures, and making routes straighter, and could help in advocating for these types of transit improvements with municipalities.

    I imagine the route summaries will become very useful for determining how best to organize bus routes and to monitor the results of those changes. Given the state of the bus network in some municipalities like Burnaby and New Westminster, I assume this information will only be used for re-organizing bus routes during the sub-area reviews which come up every 5 or 10 years. Is this correct?

  • By Jhenifer Pabillano - Buzzer Editor, June 29, 2012 @ 9:06 am

    Hi Kyle: here’s the response from planning to your comment.

    The costs shown are allocated costs, not the real cost for that specific service (ie. they’re based on an average cost per service hour for a given vehicle type across the entire system). The average cost per service hour is calculated by TransLink finance and is done simply by dividing the total operating cost of all services of that type by the total service hours of that type. This average cost includes variable costs, semi-variable/fixed costs and allocated costs, but not TransLink administrative costs. The Annual Service Costs shown are also the cost of in-service time only, so they don’t include costs associated with deadhead (when buses are driving to/from/between their routes). Given this, variations across routes in terms of Cost per Boarded Passenger are strictly the result of differences in ridership, service hours and/or vehicle type (conventional vs community shuttle) and don’t account for differences in vehicle efficiency or in deadhead requirements. From a Network Management perspective, we’re interested in the fundamental performance of the route itself and don’t want to penalize certain routes because of where they happen to operate out of or how old the buses are that happen to be running on them. Though these are certainly issues that our operating company takes into account.

  • By Jhenifer Pabillano - Buzzer Editor, June 29, 2012 @ 10:08 am

    Eugene: you make a great point about the marketing of services. Duly noted for the future!

  • By Eugene Wong, June 29, 2012 @ 12:02 pm

    @mike0123

    I assume, that you’ll download the data in other formats, and then digest it.

    I’d like to see a list/table of the routes according to rank, a list/table of which routes are profitable, and a list/table of which routes have an increase in ridership on the weekend.

    If you are going to digest the information in that manner, then would you upload copies, please?

  • By SS, June 29, 2012 @ 4:32 pm

    The data for route length, stops served, average scheduled speed, timing points, and schedule hours can actually be parsed using the GTFS data provided by TransLink. However, I don’t trust the data for route length and speed just yet as there are lots of errors in the shape coordinates in GTFS the last time I checked.

  • By JKKT - Kyle, June 29, 2012 @ 6:14 pm

    As I did a calculation according to that theory a few months ago on the cost of a community shuttle here: http://cityinaction.wordpress.com/2011/11/05/the-cost-of-a-community-shuttle/ I got $66 instead of the $56. Do langley community shuttles cost more, or did I make a suicidal flaw in my theory?

    You say that you are interested in the “fundamental performance” of a route. This would likely mean the Cost/boarded passenger, or the average capacity utilization?

    The problem with the data is that it doesn’t give an accurate representation of the niche of each route. The “Contribution to the FTN” be the most reasonable piece of info. Some buses may be profitable, and have a high performance, (#44), yet those routes turn into a non grid shape, serving only one passenger group (downtown to UBC), and those routes can be easily substituted by others like the 84.

    Can you give me a more clear response for the break even point in the cost/boarded passenger?

  • By Graeme, June 29, 2012 @ 9:35 pm

    I was looking over the Bus Performance Review and I’m confused about the numbers for the 480. I took the 480 every day to UBC during the 2010-2011 school year. The bus was definitely running at far over 100% every morning and afternoon. The aisles were always packed with standing passengers. Your numbers, however, are showing much less than that.

    There may have been a drop off after the service was truncated to Bridgeport in the Fall of 2011, but the numbers for 2010 are also lower than what I experienced.

  • By Graeme, June 29, 2012 @ 9:52 pm

    I was thinking about the numbers regarding the 480 and I think I figured it out. The 480 is a classic one-way commuting line. It is full in the morning on the way to UBC, but empty on the way back to Richmond; correspondingly, it is full in the afternoon on the way to Richmond, but empty on the way back to UBC. This effect would reduce the overall capacity utilization percentage and not reflect the ridership during peak periods in certain directions.

    I’m glad I was able to answer my own question!

  • By Marvin B, July 2, 2012 @ 6:33 pm

    For the 480, wouldn’t using the same bus for 2 routes make sense? From Richmond as a 480 then do something like a 49, 44 or some other route? Then stay out as that other route until the afternoon peak when it does the Richmond run again? While the 480 is doing something else, the community shuttles could do the run from UBC to Richmond. It may take a bit of tweaking to find the exact times, but I think this makes sense.

  • By Chris, Public Transport, July 5, 2012 @ 12:43 pm

    Hi, the route summaries are very informative. I didn’t realize how crowded some of the buses on the weekends were. Since the 43 and 44 do not operate on weekends, has thought been given to operate the articulated buses that they would use weekdays on weekends to reduce overcrowding on 22, 25, 41, and 49? Obviously you may have to extend some of the stop zones on 22 and 25, but I’m sure Translink could do that. It seems like that would be an inexpensive way to improve conditions for the passengers.

  • By Jhenifer Pabillano - Buzzer Editor, July 5, 2012 @ 2:19 pm

    Hi Kyle, here’s the response to your comment:

    Theoretically the break even point for a service would be the point at which the cost per boarded passenger equals the average fare per boarded passenger. We have the ability to measure the cost per boarded passenger because we know the number of annual revenue hours, the cost per revenue hour and the number of annual boardings. What we don’t know with as much certainty is the average fare per boarded passenger. We know what the systemwide average fare per boarded passenger is because we know how much annual revenue our services generate and we know how much annual ridership is generated. What we don’t know with a high degree of accuracy is the average fare per boarded passenger on a route-by-route basis. People make transfers between different services, people are using different fare media, travelling different distances, etc…If a person starts their journey on a 502, transfers to the SkyTrain and ends up on the 99 to UBC what portion of the revenue generated from that individual do you assign to each of those services? There’s also the fact that the transit network provides a regional benefit to the people of Metro Vancouver and we can’t or shouldn’t separate out different services from the whole. Perhaps the introduction of the Compass Card system will help us better understand people’s places of origin, destination and points of transfer in between. This may give us better insight or ability to assign revenue on a route-by-route basis.

  • By Jhenifer Pabillano - Buzzer Editor, July 5, 2012 @ 2:21 pm

    Graeme: here’s a follow-up from planning to your comment.

    I’m glad to see Graeme is understanding the nature of route productivity. If we only look at the average capacity utilization on a route we not be accurately capturing people’s experience in riding that route. A service that generates 100% capacity utilization in the peak direction and 0% capacity utilization in the reverse peak direction will only show a 50% capacity utilization overall. The customer traveling in the peak direction will wonder where we’re getting our numbers from. The second page of the route summaries addresses this issue. On the second page we look at route performance in a much more detailed level – by day type, time period and direction. For example, the average peak load in the westbound direction during the AM peak period is 71 versus 12 in the eastbound direction during the same time period. The 71 probably better reflects your personal experience riding the 480 to UBC every morning. The tricky part is that we all pay for those buses heading in the opposite direction as well. That’s why land use is so important. We need to give people a reason to ride in the reverse peak direction to fill up those buses as well.

  • By Jhenifer Pabillano - Buzzer Editor, July 5, 2012 @ 2:22 pm

    Marvin B: here’s a response from planning to your comment.

    This is essentially exactly what we do. We call this interlining – bringing buses into and exchange as one route which then leave as another. This is particularly prevalent for our articulated fleet at UBC for example. We don’t use conventional buses in one direction and minibuses in the other – the logistics are too difficult. We do operate conventional buses during peak periods and minibuses off-peak. The organization can realize significant operational savings in that manner. Thanks for your question!!

  • By Jhenifer Pabillano - Buzzer Editor, July 5, 2012 @ 2:23 pm

    Chris: here’s the response from planning to your comment.

    We have given some thought to alternative service delivery models on some of our higher ridership corridors. This could include off-peak service on the 43 for example. Everything depends on having resources available for service expansion of course. We’re planning on taking some of our ideas out to the public for discussion this fall so stay tuned and perhaps you can tell us your ideas in person (or via e-mail).

  • By Graeme, July 5, 2012 @ 7:05 pm

    Jhenifer,

    For a route with such one-way demand like the 480, would it not be possible to just not run routes in the opposite direction when there is no demand? The 480 actually duplicates other services, so nobody would be completely out of a ride as a result.

    I’ve also always wondered why the 480 isn’t a direct express service from Richmond to UBC without any stops in between. If this were the case, the 480 would be able to travel along Marine Drive, avoiding Kerrisdale (so slow!), and make the journey a lot faster. It does pick up some passengers on Granville, but as soon as it is travelling on 41st the 480 is just duplicating existing services. I know there are express services within the system, but are there any that just serve A to B and nothing in between? Is it possible to offer this service or is it politically difficult?

  • By Andrew, July 7, 2012 @ 4:05 am

    i’m surprised that the 16 is #1 for avg capacity utilization, it seemed overcrowded ever…
    the 25 being 2nd doesn’t surprise me though, especially since the peak capacity times for it line up with school hours (a lot of schools along that route)

  • By Marvin B, July 9, 2012 @ 12:50 am

    @ Graeme:

    I like what you’re saying about Richmond to UBC via Marine Drive instead of 41st Ave. Problem here is this:

    Do you think the folks living on Marine Drive want buses running down their street? I think that’d be a fight for both Translink and Vancouver city hall.

  • By Jhenifer Pabillano - Buzzer Editor, July 30, 2012 @ 10:21 am

    Hi guys — just wanted to add that I have been passing your answers onto the planning team, and they’ll take your specific suggestions into account through their ongoing Network Management Program. Always good to hear from you!

Other Links to this Post

  1. The Buzzer blog » Big goals, big challenges: what we think about when planning the transit network — July 6, 2012 @ 8:01 am

  2. The Buzzer blog » Guiding themes for planning a transit network — July 13, 2012 @ 12:17 pm

RSS feed for comments on this post. TrackBack URI

Leave a comment

Please read our Participation Guidelines before you comment.