Ridership vs coverage: The transport planner’s dilemma

Berlin. Image: Getty.

Is your transit agency succeeding? It depends on what it’s trying to do – and most transit agencies haven’t been given clear direction about what they should be trying to do.  Worse, they’re told to do contradictory things.  It’s as if you told your taxi driver to turned left and right at the same time, and then criticised them for turning the wrong direction.

On the one hand, we expect transit agencies to pursue a goal of ridership.  Yet we also demand that provide a little service to everyone, which is called a coverage goal.  The coverage goal requires an agency to run predictably low-ridership services, for non-ridership reasons, so it’s the opposite of a ridership goal

In the fictional town below, the little dots indicate dwellings and commercial buildings and other land uses. The lines indicate roads, and the 18 buses indicate the resources the town has to run transit. Most of the activity in the town is concentrated around a few roads, as in most towns.

A transit agency pursuing only a ridership goal would focus service on the streets where there are large numbers of people – where walking to transit stops is easy, and where the straight routes feel direct and fast to customers. Because service is concentrated into fewer routes, frequency is high and a bus is always coming soon.

This would result in a network like the one below.

All 18 buses are focused on the busiest areas. Waits for service are short but walks to service are longer for people in less populated areas. Frequency and ridership are high, but some places have no service.

Why is this the maximum ridership alternative? It has to do with the non-linear payoff of both high density and high frequency, as explained more fully here.

If the town were pursuing only a coverage goal, on the other hand, the transit agency would spread out services so that every street had a bus route, as in the network at below. Spreading it out sounds great – but it also means spreading it thin.

The 18 buses are spread around so that there is a route on every street. Everyone lives near a stop – but every route is infrequent, even those on main roads, and waits for service are long. Only a few people can bear to wait so long, so ridership is low.

In these two scenarios, the town is using the same number of buses. These two networks cost the same amount to operate, but they deliver very different outcomes.

Ridership-oriented networks serve several popular goals for transit, including:

  • Reducing environmental impact through lower Vehicle Miles Travelled;
  • Achieving low public subsidy per rider, through serving the more riders with the same resources, and through fares collected from more passengers;
  • Supporting continued urban development, at higher densities, without being constrained by traffic congestion;
  • Reducing the cost of for cities to build and maintain road and bridges by replacing automobile trips with transit trips, and by enabling car-free living for some people living near dense, walkable transit corridors.

On the other hand, coverage-oriented networks serve a different set of goals, including:

  • Ensuring that everyone has access to some transit service, no matter where they live;
  • Providing lifeline access to critical services for those who cannot drive;
  • Providing access for people with severe needs;
  • Providing a sense of political equity, by providing service to every municipality or electoral district.

Ridership and coverage goals are both laudable, but they lead us in opposite directions. Within a fixed budget, if a transit agency wants to do more of one, it must do less of the other.

Because of that, cities and transit agencies need to make a clear choice regarding the Ridership-Coverage trade off. In fact, we encourage cities to develop consensus on a Service Allocation Policy, which takes the form of a percentage split of resources between the different goals.

For example, an agency might decide to allocate 60 percent of its service towards the Ridership Goal and 40 percent towards the Coverage Goal.

Major network redesigns often shift this balance, intentionally and consciously.  When we led a redesign of the bus network in Houston, we led a discussion with the elected leaders about their priorities, and they decided to shift the focus of their network from 80 per cent coverage to 5per cent coverage. They knew in advance what the result would be: a more useful network, with the potential to grow more ridership, but also many angry people in areas no longer served.

What about your city? What do you think should be the split between ridership and coverage? The answer will depend on your preferences and values.  For cities, it should be up to elected officials, informed by the public, to decide.

Jarrett Walker is an international consultant in public transit network design and policy, based in Portland, Oregon. Christopher Yuen is an associate at Jarrett Walker+Associates.

Walker is also the author of “Human Transit: How clearer thinking about public transit can enrich our communities and our lives". This article was originally written for his blog, and is reposted here with permission


Why doesn’t London build an RER network, like Paris did?

A commuter walking by a map of the RER B line at the Chatelet-Les Halles station in Paris. Image: Getty.

I’ve heard many people make many different complaints about the Parisian transport system. That it does a bad job of linking a rich, white city with its poorer, more diverse suburbs. That, even as subway systems go, it’s a hostile environment for women. That the whole thing smells distractingly of urine.

I’m familiar with all of these complaints – I’ve often smelt the urine. And I’m aware that, in many ways, London’s is the superior transport network.

And yet I can’t help be jealous of Paris – In large part, because of the RER.

Central Paris. The Metro lines are thinner, and in pastel shades; the RER lines are thicker, and in brighter colours. Image: RATP.

Paris, you see, has not one but two underground railway systems. The more famous one is the original Paris Metro, opened in 1900: that’s the one with those fancy green portals with the word “metropolitain” written above them in a vaguely kooky font.

The Metro, though, mostly serves Paris Intra-muros: the official city, inside the Boulevard Périphérique ring road, site of the city’s last set of walls. As a result, it’s of very little use in most of the city’s suburbs. Its stations are very close together, which places a limit on how fast its trains can cross town. It was also, by the mid 20th century, becoming annoyingly overcrowded.

So starting in the 1960s, the city transport authorities began planning a second underground railway network. The Réseau Express Régional – Regional Express Network – would link suburban lines on either side of Paris, through new heavy rail tunnels beneath the city. Its stations would be much further apart than those of the metro – roughly one every 3km, rather than every 600m – so its trains can run faster.

And fifty years and five lines later, it means that 224 stations in the suburbs of Paris are served by trains which, rather than terminating on the edge of the city, now continue directly through tunnels to its centre.

The RER network today. Image: RATP.

London is, belatedly, doing something similar. The Elizabeth Line, due to open in stages from later this year, will offer express-tube style services linking the suburban lines which run west from Paddington to those which run east from Liverpool Street. And Thameslink has offered cross-town services for 30 years now (albeit not at tube-level frequencies). That, too, is going to add more routes to its network over the next few years, meaning direct trains from the southern suburbs to north London and vice versa.

Yet the vast majority of suburban National Rail services in London still terminate at big mainline stations, most of which are on the edge of the centre. For many journeys, especially from the south of the city, you still need to change to the London Underground.

So, could London ape Paris – and make Thameslink and Crossrail the first element of its own RER network?

In a limited way, of course, it’s doing just that. The next big project after Crossrail is likely to be (original name, this) Crossrail 2. If that gets funding, it’ll be a new south-west to north-east route, connecting some of the suburban lines into Waterloo to those in the Lea Valley.

The proposed route of Crossrail 2. Click to expand.

But it’s not immediately obvious where you could go next – what Crossails 3, 4 or 5 should cover.

That’s because there’s an imbalance in the distribution of the remaining mainline rail services in London. Anyone who’s even remotely familiar with the geography of the city will know that there are far more tube lines to its north. But the corollary of that is that there are far more mainlines to the south.

To usefully absorb some of those, Crossrail 3 would probably need to run south to south in some way. There is actually an obvious way of doing this: build a new tunnel from roughly Battersea to roughly Bermondsey, and take over the Richmond lines in the west and North Kent lines in the east, as a sort of London equivalent of RER C:

Our suggestion for Crossrail 3. Image: Google Maps/CityMetric.

But that still leaves a whole load of lines in south and south east London with nowhere to send them beyond their current terminal stations.

In fact, there are reasons for thinking that the whole RER concept doesn’t really fit the British capital. It was designed, remember, for a city in which the Metro only served the centre (roughly equivalent of London’s zones 1 & 2).

But London Underground wasn’t like that. From very early in its history, it served outer London too: it was not just a way of getting people around the centre, but for getting them there from their suburban homes too.

This is turn is at least in part a function of the economic geography of the two cities. Rich Parisians have generally wanted to live in the centre, pushing poorer people out to the banlieues. In London, though, the suburbs were where the good life was to be found.

To that end, the original operators of some lines weren’t just railway companies, but housing developers, too. The Metropolitan Railway effectively built large chunks of north west London (“Metroland”), partly to guarantee the market for its trains, but partly too because, well, housing is profitable.

In other parts of town, existing main line railways were simply added to the new underground lines. The Central line swallowed routes originally built by the Great Western Railway and London & North Eastern Railway. The District line absorbed part of the London, Tilbury & Southend Railway.

At any rate: the Tube was playing the same role as the RER as early as the 1930s. London could still benefit from some RER-type services, so hopefully the Elizbaeth Line won’t be the last. But it doesn’t need an entire second metro network in the way 1960s Paris did.

There is another idea we could more profitably steal from Paris. Those suburban railways which aren’t connected to the RER are still run by the national rail operator, SNCF. But it uses the Transilien brand name, to mark them out as a part of the Parisian transport network, and – as with the RER – each route has its own letter and its own colour.

The Transilien & RER networks in Paris. Image: Maximilian Dörrbecker/Wikimedia Commons.

This would not have the transformative effect on London that building another half a dozen Crossrails would. But it would make the network much easier to navigate, and would be almost infinitely cheaper. Perhaps we should be starting there.

Jonn Elledge is the editor of CityMetric. He is on Twitter as @jonnelledge and on Facebook as JonnElledgeWrites

Want more of this stuff? Follow CityMetric on Twitter or Facebook