The next bus will come when it comes

Data is not a service. Service is not data. Knowledge is power, but it’s not always very much power.

My default mode of transport in London is the bus. Having information about when the next bus is coming has been exciting and empowering, even if the shine has worn off a little bit.

But I’m interested in the data because I want to catch a bus, not because I have a deep inner yearning for it in its own right. So here are two small stories about how that can be harder than it should be.

I am at a bus stop. It’s not looking good: there seem to be two buses in a clump eleven minutes away. As it turns out a bus actually comes in seven minutes – but almost certainly not the one which was eleven minutes away. And over those seven minutes, the data tells me a remarkably incoherent story.

Beyond observing the fault in the space-time continuum at 1710, it’s quite hard to make sense of what is actually going on. But in a way it doesn’t matter. What matters is that I can’t trust the data: it appears to be telling me something, but I can’t be sure that it is telling me anything.

Another time, the other side of London. I am on a bus. A hundred yards ahead, I can see another bus on the same route. The driver announces that the bus I am on is to stop short of the terminus and we must all get off at the next stop. As we get off, the bus in front has just pulled away – and it turns out that the next bus is 19 minutes behind. There is, as far as I know, nothing wrong with that data, but the wrong decision has been made based on it. Turning round the first bus rather than the second, or even holding the first bus for the thirty seconds needed to allow the second bus to catch up would have allowed a smooth transfer with nobody waiting. As it was, the wait was long, tedious, and completely unnecessary.

This isn’t really about buses. As so often on this blog, it’s about better connecting the front end with the back end and focusing on the service not the process. There is a lot of excitement about open data, and so there should be. The openness of TfL’s data allowed me to watch those two stories unfold on two excellent third party apps, and it’s really useful to be able to do so. But easing the transmission of dodgy data, as in the first story, isn’t really what’s wanted. And in the second story, the problem was not the data but that TfL didn’t seem to be paying any attention to it.

In the end, service design is unavoidably about designing the service.

Original post