"How much?" A question about imageability and seams in transport fare systems

(Something of a follow-up post to the recent transport informatics survey.) A recent conversation with Jarrett Walker, a consultant based here in Sydney, popped up the following thought. Jarrett, experienced in metropolitan transit systems, was thinking through ideas around fare pricing given the new possibilities enabled by fully automated systems.

Design of fare structures have been fundamental to transit planning for years, attempting to define charges for journeys in equitable yet efficient fashion. Balancing those last two factors mean that the basic problem is often shot through with tensions - e,g, richer suburbs paying less than poorer, due to zoning often based on radial principles emanating from a central core, and so on. Jarrett knows more about that than I. Whatever, existing systems based around zones etc. do at least usually have a stated, consistent pricing for journeys that can be easily communicated, even if not necessarily agreed with.

However, Jarrett was wondering about some emerging thinking he'd heard around the possibilities of new smart card-based, integrated ticketing systems, and the sense that varying prices could be generated in real-time, based on variables like distance, time of day, number of passengers on board, overall running costs of the system at that point, demand etc. That you wouldn't know the actual price you'd been charged for that particular journey and that actually, you needn't. You just swipe the card and conduct your journey, in the knowledge you'll have pre-paid to a certain amount, or pay a monthly bill subsequently. You'll trust the system will charge you fairly, of course, and you could see the breakdown of costs at the end of the month, or when your pre-pay card needs topping up, and so on.

(As an analogy, you'd contend that few people really know/care the exact cost of each one of their phone calls, for instance. The payment is represented by sometimes complex monthly plans, based around a number of free minutes/texts that are bundled, a certain number free within a network and so on. Obviously, some do know what they pay each time, but hasn't the general tendency has been towards bundling into monthly packages, abstracting away from pricing the actual individual calls at time of connection? With a pay-as-you-go model for those without financial security.)

So the drift towards an ongoing service model of variable pricing bundled into pre-paid or direct-debited packages seems an option. With an increasing deployment of GPS devices in all vehicles and RFID-based tracking of passenger entry/egress, it seems likely that some transit systems will try this out, in effect neatly hiding the complexity of pricing from the citizen.

This is partly also due to the sheer complexity of pricing systems e.g. Sydney's train system alone has over 120 'fare products', apparently. Multiply that by ferry, bus and light rail. This fare complexity is largely a result of attempting to be equitable, and at the moment the complexity is shared by both system operator and customers.

However, Jarrett wondered whether citizens might actually want to understand, or engage with, their public transport system a little more deeply. That pricing is one way of perceiving the structure of the transport system, and that's something that customers might innately want to do. He thought that it might be important to perceive how the system works, at least as expressed in fare structures. It's a map of the city, in a sense. In a city like London, the topography is overlaid with a mental model of the zones, which take on a kind of meaning over and above fare products (I proudly lived in zone 1, would more or less travel to zone 2, and so on.)

When he asked me the question about whether perceiving the sysetm was important, I immediately thought of the importance of seams and imageability.

Seamfulness, some long time readers will know, is a particular interest here (and of others, like Anne Galloway and Adam Greenfield.) It holds that a desire to hide complexity via an apparently perfect, hermetically-sealed product can actually mitigate against a successful informational system.

A classic example here is the iPod, which given its undoubted success also indicates how complex the argument is. That success is down to its carefully linked system architecture with iTunes as well as its rigorously reduced interface and seductive aesthetic. And yet its alleged undoing is also to do with its 'perfect' design, in that batteries are difficult to replace (meaning most people don't) and that it's a music experience that can't be tweaked or modified much. You could argue that if the iPod showed its seams a little more, it would be more malleable as a device, and even more engaging as a product experience. Doing that without damaging its seductive sheen and usability would be tricky but potentially rewarding. That old "beautiful seams" ambition.

The other reference is of Kevin Lynch's concept of imageablity, from his pivotal book The Image of the City, which I've always thought should apply to system design - the ability to perceive the system around you (visually, spatially, intellectually) and be left with a strong 'image' of its structure. Also known as legibility. A few years ago Peter Lindberg developed the idea specifically around software architecture, and I've subsequently thought it an essential feature of good system design (whether the system is a building, a music-playing device, a transport system or indeed Grand Theft Auto.)

So it seems to me that the ability to show/hide structural detail is fundamentally important element of a system. It enables the legibility of the system. And that showing a bit more detail, if carefully and sensitively articulated, can only engage the user further. Not necessarily exposing minute technical detail - though a handful will always want that - but enabling perception of the basic ambit, structure, joints, seams, influences, and so on. It certainly enables that form of engagement known as adaptation or even hacking - not in the pejorative sense of the word - but in the sense of building upon systems and extending them - as we've seen with transit systems that do begin to expose their behaviour.

And of course, as these and similar pervasive systems migrate into many spheres of life, deciding how visible to make parameters, motives and controls becomes even more important. Will hiding such intricacies reduce civic engagement in urban information systems? Or conversely, will its seamless design lead to increased take-up of services like public transport and thereby greater civic engagement?

What do you think? I'm aware that I'm posing the question to a particular audience, but do you think that, in this case, a transport system that has a choice to hide the potential complexity of a fare system should do that? Or should it reveal its complexity either through having set fares or by displaying the calculated fare on the spot? Does convenience trump legibility?

Comments

[lauren]

great post dan.

given that the pricing structure is a pivotal power base, i think complete transparency can be a slightly dangerous thing. i also think that assuming people will just accept what's fed to them is equally as dangerous. oyster has a nice mix - you get a sense of how some of their more complex calculations are working, when you see the second half of the trip come off the balance. but at the same time, all you really do is see how much you've got left on the card, which satisfies most people. you have a basic understanding that it's cheaper to use the card (than to buy a straight ticket).

i think the decision about how much to show needs to be based on what people feel gives them a choice. if you know how much it's going to cost, you can choose to walk. or ride. you know how much to beg for if you don't have enough. and you know how much to give if someone needs a hand. plus if the complexity of it is evident to a certain point, businesses will be able to track the cost of transport more accurately, making it a viable budget line (accountability being quite attractive to those crunching numbers).

a good example outside the transport system is the food industry and putting all the ingredients on the packaging. most people don't want the full complexity of the manufacturing process, but like to know that they can make choices based on how much of the good stuff and the bad stuff is in it. even if they ultimately choose on convenience and/or taste.

i'd like to see a transport pricing structure like innocent's ingredients: 1 apple, 3 oranges, 25 grapes and a squeeze of lemon = to and from work in the city everyday, with the bonus for not travelling home before 6pm.

[Desmond Bliek]

Excellent post - system legibility does matter, especially as we move towards more dynamic pricing systems that reward users for travelling off-peak or charge them according to distance travelled. The phone analogy is instructive - while we might not see the price with every call, consumers do need to know that evening calls are charged at a lower (or flat) rate, as opposed to peak period, day-time calls. The more we use variable pricing to manage scarce resources (transit space, road space, bandwidth) the more knowledge and especially real-time knowledge of pricing structures becomes important.

[Nama Henderson]

First i think the wireless phone analogy is quite right.
Why even pay for individual trip just for the convenience of the public transport.
Yet, i do think legibility is important particularly for legitimacy. We want to know what we are being charged. Although we may not want to be bombarded with info, we prefer the option.
Finally, having the information might quite quickly lead to changes in usage habits. It would certainly provide better data both to the end user and system "engineer".

[Jarrett]

The thought came up because I was thinking about the questions of "equity" that come up in fare structures -- the political issue on which most proposals founder -- and particularly trying to imagine a system that would universally be understood as "everyone paying their fair share." The problem is that equity is inversely related to complexity. The simplest fare is to charge everyone the same regardless of how far they ride, but that's also the least equitable.

So I was trying to imagine the opposite. Well, what you're paying for is the cost of driving your transit vehicle right now, plus your share of the capital costs, plus your share of the company overhead. Your fare would be that amount -- for the distance you travelled -- divided by the number of other people on the vehicle with you at the time. To be precise, your fare would be the sum of the cost of each segment you rode -- where a segment goes from one stop/station to the next -- each divided by the number of other passengers on the same vehicle and segment.

Technologically, it's clearly possible. We can imagine a very smart card that would not only know how far you're travelling, biut also know how many other smart cards were on the same vehicle at the same time.

(As I write this, I'm visualising it working on a bus where smart cards require rider to "tag on" and "tag off." It would be different on gate-controlled systems, but not fatally.)

Would this fare be a good idea? Well, I like it personally, but the real point of presenting it is to convince you of several other points:

1. If your idea of simplicity or legibility involves being able to calculate or fare before you make the trip -- or even for a computer to be able to do so -- then this is the ultimate complexity. You can't know the cost of the trip until you take it.

2. I'm not sure if this complexity is "illegible" exactly. In a way, it's very clear. You have to trust the computer to figure your fare, but the principle behind the fare is easily explained. It has the particular virtue of educating citizens about the actual cost of every increment of service, which could cause people to respect their infrastructure and services rather than take them for granted.

3. "Equity" in fares increases with greater complexity.

4. Virtually no real-world fare systems, however "equitable" they may pretend to be, include the colossal real-world factor that we all understand if we've ever shared a taxi. Our fair share of the cost of service depends entirely on how many people we're sharing the ride with. I find this interesting, and telling.

Again, I'm not advocating it, just offering it as a thought experiment. I leave it to readers to speculate on how this would affect patronage. We'd certainly all be more patient with overcrowding!

[Ben]

My first reaction would be to say that transparent is better, but your mobile phone argument is a strong one.

At the start of last year (in London) I switched from a monthly Travelcard (cost approx £90) to Oyster PrePay. For 3 months I put £90 a month on the card.

I'd always preferred a Travelcard because of the implied security. Yet I became convinced that PrePay would save me money - I never use the Tube at weekends, I often walk to work etc.

At the end of 3 months I had a balance of £120 left on the card.