Should We Redesign Our Bus Network? How?

When planners develop proposals for redesigning a bus network, how do they do it?   And when is it necessary?

In 20 years of doing bus network planning around the world, I’ve encountered few systems that don’t have some major obsolete features in their design.  Most public transit authorities continually fix small problems in the network but have trouble fixing the big ones.

Making superficial changes to a network is like adding little bits to a house.  One by one these bits make sense, but over time they can destroy the design of the house,You may also be doing these little remodels because you can’t face the fact that the foundation is rotting.

Cities change, and every 20 years or so, the bus network should be comprehensively reviewed. Such a project should really include an exercise where you study the city’s demand patterns and design a network on a “blank slate” i.e. deliberately not considering what your network does now.  Such a thought process will retain the strongest features of your existing network, but will let you discover new patterns of flow that are a better fit for your system as it is today.

Network design is a process of creative thought, not just analysis.  When we rethink a network, we’re doing what a scientist does as he tries to form a new theory:

1.  Data Presentation.  Look at all the data and try to see NEW patterns in it that the current theory/network misses.  (Geographical representations of the data are crucial at this stage.  The quality of your data presentation will limit the range of ideas you’ll have.)

2.  Creative Thought.  Form new ideas that work with those patterns.  (This step is creative rather than analytic, and proceeds in unpredictable bursts of inspiration.)

3.  Analysis.  Test those ideas against the data.  Revise or discard those that don’t fit the data well.  This is the analytic step, and must not be confused or conflated with the creative step that precedes it.

4.  Go to step 2 until you have something you like.

This process is important because great network design ideas solve many problems at once, just as a good scientific theory explains lots of data at once.  The single most common mistake in network planning is to think about only one problem at a time, and look only for solutions to that problem.  That kind of narrow thinking may be necessary to get from one day to the next, but every 20 years or so (or more often if your city is changing fast) you need to do the larger process I’ve described.

You must also control altitude.  The first stage of the work must look at the entire city or service area, to be sure that you solve problems that can be solved only at that scale.  Only then can you look at details.  More on this here.

My preference is to do this process in an intensive professional workshop setting, similar to the design charrette process in urban design.  Typically, about 15 professionals set aside 2-5 days of their time.  The people need to be a mix of roving consultants like me and staff of the agency being studied, but they all commit to be open-minded, and are encouraged to think about opportunities and not just constraints.

Sometimes, when running these processes, I’ll ask everyone to name one network design idea that they’d really like to do but that they assume is politically impossible.

The point is to break out of the constraint-dominated thought process that often prevails in the daily life of transit agencies.  This isn’t a comment on anyone’s creativity, but rather an observation about the daily experience of bus network planners.  Bus agency staffs get very little appreciation, and lots of criticism, no matter what they do, so they tend to become risk-averse cultures.  When I run a network planning workshop, my first job is to break through that, and encourage the client’s staff to welcome their own insights even if they may seem politically impossible at first glance.

Core Design Workshop in Yekaterinburg, Russia, 2015.  

The workshop room has a large table where we can sit around a drawing in progress.  On the table is a base map of our study area, covered with either tracing paper or clear acetate, so that we can draw and revise over and over.  The walls are covered with maps of relevant data about our project.  There’s also a whiteboard where anyone can sketch out ideas that aren’t ready to go on the map.

(In 2020, of course, we figured out how to do all of this online!  It works much better than you’d expect.)

The workshop proceeds through the iterative steps that listed above. We start with a half day or day of just reviewing the data.  Then we start brainstorming possible “big moves,” large structural ideas which, if pursued, would lead the rest of the network to take a new form.  If we’re planning around a new rapid transit project, then that’s already the “big move.”  If we’re just redesigning a bus network, we may think of big moves of our own, such as installing a grid system if that’s appropriate.

Once we have an agreement about big moves, we proceed top-down to address the more local design issues that follow from them.  This sequence is important, because the big moves will have the biggest impact.  The design of the Frequent Network, for example, is always a big move, and I insist that we hammer out this design before we turn to the less frequent local routes that connect to it.

As we develop ideas, we may do some quick analysis to help us verify them.  We also sometimes break for field exploration, as different members of our group go out in cars to check various routings that we’re thinking of.

By the end of the workshop, we may not have total consensus, but we always have a lot more consensus than the client expected going in; we’ve always come up with ideas that were new to everyone in the room.  The core workshop isn’t the end of the process, but after this point we usually have a map that bears at least a 70% resemblance to what we’ll finally recommend.  More importantly, everyone in the workshop will have ownership of that process, and will see their own influence on it.  This, finally, is crucial.  One of the hardest jobs of a consultant is to transfer the results of his work to the client agency, so that they see the ideas as theirs rather than as “what that consultant recommended.”  Ultimately, the interactive design process is the best way I’ve found of doing that.

7 Responses to Should We Redesign Our Bus Network? How?

  1. anonymouse May 23, 2010 at 1:02 am #

    It’s definitely a good idea to do this sort of thing every once in a while to flush out the stale bits from the network design and to realign the network with the way the city has grown. But just because you’ve redesigned the network from scratch doesn’t mean that it has to be presented that way to the public. Chances are, many of the routes will look the same, and others will look like modifications or extensions of existing routes, and presenting it that way will keep people who use the system and are used to it from getting too confused by the changes. Also, I think this sort of exercise is still applicable for systems that rely on fixed infrastructure, like trolleybuses, streetcars, and even subways. Obviously, the options are constrained by the infrastructure, but it can still be useful to make sure you’re making the best use of what you have, or to find opportunities to make small investments with large benefits in terms of providing better service.

  2. Peter Parker May 23, 2010 at 4:22 am #

    Jarrett, Having just gone through this process in Melbourne, I’m curious as to the point public consultation is brought in.
    It is early on in the piece (in which you might only have an evening of their time – rather than several days as you would with professionals) or much later on when you’re testing your draft network for acceptance?
    Although some of the public who attend these things might be very parochial, anti-bus or have other agendas (eg local school issues), our experience is that as a whole they do generally provide sensible and worthwhile input.
    Most of our reviews (which covered anywhere between 15 and about 70 routes) actually included 2 rounds of public meeting, but the last 2 or 3 cut it down to 1.
    Another point relates to the frequent network, and whether there is scope for another step between when you’ve planned your frequent network and when you starting on the individual local routes.
    This immediate step (especially in low density suburbia) could be for short corridors that don’t justify their own frequent route but would need the frequency provided by two or three carefully timed overlapping local routes.
    Here in Melbourne we’ve just had a bus review (City of Wyndham) where the reviwers understood this but it got missed when the new network was put into practice.
    Hence we got a very lumpy timetable, with six bus departures from the railway station in a few minutes followed by none for 37 minutes (towards the area’s biggest shopping centre that’s remote from rail).

  3. EngineerScotty May 23, 2010 at 9:53 am #

    A common complaint about how public involvement is done (from the public, and often from those cynical about the transit authority’s competence or motivations) is that it’s all a show–a presentation of “here’s what we are going to do, like it or not” disguised as the seeking of feedback. In many cases, if an agency has gone through several rounds of design already, there’s little room for a late public review to result in any significant changes.
    Here’s a question: Were a TA to a new network design (here lets assume that the scope of the re-design is limited to route planning, not major new infrastructure–so no environmental impact statement or local equivalent is needed)–and receive overwhelmingly negative feedback from the public–under what conditions where the design be scrapped?
    In my profession (SW development), there are many different levels of customer/user involvement during the development process. In so-called “rapid development” methodologies, users are involved early and often, and given prototypes (something which doesn’t have an equivalent in transit system design, I realize). Even in more traditional development models, where customers aren’t directly involved until later, the feedback is important. I’ve been on projects before that underwent significant re-design late in the cycle after customers objected strenuously to what we were preparing to ship.
    Cross-discipline advice is often useful–though its use is frequently limited, so take it with a grain of salt.

  4. Joseph E May 23, 2010 at 12:58 pm #

    @EngineerScotty; you said: ” In so-called “rapid development” methodologies, users are involved early and often, and given prototypes (something which doesn’t have an equivalent in transit system design, I realize)”
    The technology underlying Google Maps and many local transit trip planners could give us a virtual “prototype”. The transit planner could make a prototype schedule, and then put it into a version of the trip planner. The public would get a chance to see how their trips would change, at least if the buses are on schedule.
    The next step would be to build a piece of software which would allow the public to shift around bus service, and see the results. I know I would pay a good money for a program like that! It might even sell as a Wii or PC game. On a more limited scale, the transit planners could provide several different alternative networks, based on different funding levels or different priorities, or showing the addition of a new rapid transit system.

  5. Wad May 23, 2010 at 3:25 pm #

    @Joseph E,
    There’s a freeware game that does something like what you describe. It’s a German game called “Mobility,” and it was developed in part by the Frankfurt transport commission.
    It’s like a rudimentary Sim City, but the key point is you get to draw transit routes and stops and your goal is to move the city’s passengers efficiently.
    See this site:

  6. Nathanael May 23, 2010 at 8:41 pm #

    Data is critically important.
    One thing to do is to get the data together — trying to figure out where people are coming from and going to, and where (and when) they want to —
    Then start community meetings for brainstorming, *providing* the data, and simultaneously collecting the anecdotal information which may fill in the gaps in the data. (They have some crazy name for these sort of meetings which I can’t remember.)
    This all comes right at the beginning of the process. At this point, a really effective transport agency should have a sense of public priorities and desires. At that point a major redesign becomes *politically* possible, though several rounds of feedback remain necessary to *keep* it politically possible. It’ll still need to be sold as “changes” most likely, of course.
    At least this is what seems to have happened in the cases where there was the least public objection and the most approval of changes.

  7. Jarrett at May 25, 2010 at 3:40 pm #

    I think I agree with all comments to this point. I write a little about “build your own system” planning games, which I use as a consultation tool, here:
    Peter, the process I’m describing in the post is the core part of a service review where you actually hammer out the proposal. You really do need 3-5 experts in a room, uninterrupted, to do that. Of course those experts include operations representatives who know a lot about what customers think, and data on public opinion may be part of the data that we study at the beginning.
    I’m a big advocate of doing consultation shortly after this stage, where the ideas are still soft, before we’ve all “fallen in love” with our own work.
    Where there’s a lot of cynicism — people saying “this is all for show; you’ll do what you want to do anyway” — I recommend drawing two service plans that differ in some fundamental way, such as the balance between service designed for ridership and service designed for coverage. When people see that there are still two plans, it defeats some of those skeptics.