In some recent conversations I have been putting forward the idea that we might need to re-consider what we class as infrastructure. I am not sure that I have always carried the room with me, which had led me to take the time to try to clearly articulate what I think, and why I think it. To start with a clarification – if we look at the National Infrastructure Commission’s home page then we see they already have responsibility for data infrastructure (i.e. the infrastructure for data)…that’s not what I’m talking about. Equally National Highways have a lot of data about their infrastructure (for example lots of data about the condition of such-and-such a road at point x)…that’s not what I’m talking about either. My point is that we need to start thinking of data as infrastructure which is to say the data that describes how a piece of physical infrastructure is being used is becoming just as important as the physical infrastructure itself and we should be thinking about it in the same way.
The traditional model of infrastructure is of fixed, physical assets that support key underpinning parts of national activity. The remit of the National Infrastructure Commission, for example, covers transport, energy, water, and telecoms (i.e. the infrastructure that enables the collections and communication of data) and all of these are of critical national importance, and therefore political in the most fundamental sense, in that they affect the economic and social outcomes that we enjoy. Of course each of these sectors is organised differently: most parts of most of them are not in public ownership, although all are regulated in some degree. All of these sectors are dependent on all of the others.
Transport, I would argue, is particularly complicated because of its multi-modal nature and also because, for road in particular, it is uniquely personal. There are a few reasons for this: the actual physical infrastructure (mostly publicly owned and managed) is used by a plethora of individuals and organisations who have a wide amount of freedom to use it when, and how they wish. This differs from power, or water, for instance, where we are all customers of large organisations who own and use the infrastructure to deliver us services…but only on their terms. Secondly the road system is the only major public benefits provided free at the point of delivery out of general taxation…other than the NHS. Thirdly roads are not just corridors that we use to get from one place to another…they are also the immediate environment outside our windows, and surround and border all the offices, factories, schools, pubs, hospitals and other places we spend our lives in. Power cuts are deeply serious and have massive impacts; the water quality crisis is a big issue for many people…but lockdowns during the pandemic (which were, in effect, just bans on making journeys: or the removal of transport) affected everyone and everything in ways we are still talking about years later.
Now, what has this got to do with data? Well transport (and transport infrastructure) has been relatively unaffected by the digital revolution that has transformed sectors of the economy like retail, entertainment, finance and so on. I’m of a certain age so for much of my life the music I listened to was (in the form of records and tapes) an important part of my stock of possessions and selecting and buying it was very important to me. Nowadays, like most people, I stream music and I have a wider choice at less cost and fuss: ownership turned out not to be the important thing I thought it was. It turned out that what mattered to me what the data (i.e. the music stream) and not the physical artifacts (the records and CDs). What would it be like if we, as a society, made a similar shift in transport. What if it’s not the roads themselves that matter as much as the way we are able to use them?
The reasons for why transport has been slow to change are worth at least one other article, but it is not the case that there is no potential for transport to be changed and improved by digital technology. The transport technologies we use today have been developed over millennia within the constraint that we don’t know where anything is, or where it is going. The bus stop and the timetable, for example, are technologies to enable you to coincide in time and place with your means of transport. But the development of cheap, mobile computing and geolocation technology means that has all changed: almost everyone and everything is now instrumented and able to report its location and plans in real time. For the most part we have not taken advantage of the implications of this yet. The growth of Uber is an interesting case study here. Uber is not, fundamentally a new business model: their model of recruiting drivers who bring their own capital asset (a car) and work piece-rates, is shared by thousands of minicab forms. Instead of using a telephone-based human despatcher they built a high tech job allocation system but the economic gains from this are minimal - their real innovation was to recognise that the two traditional markets of taxis and minicabs were based on out of date technological assumptions. Taxis were allowed to charge higher prices because they offered greater convenience: you could hail them on the street. But the advent of mobile phones with geolocation technology meant that the traditional disadvantage of the minicab (that you had to book in advance and agree to meet in a pre-arranged spot) had, in practice gone away. It is interesting that the vested interests and inertia of the status quo mean that little or no change has been made to these out of date market structures even now. Notwithstanding their growth and scale, Uber’s Silicon Valley invest-to-monopolise-and-then-harvest-rents model may turn out not be such a great investment after all: they have got cumulative losses of $30 billion or so and the very low costs to enter their market may mean they have poor prospects of ever making that money back…but the underlying point about the market structure stands and the transport industry is full of similar potential areas of improvement. For the most part the reasons why change has been so slow are nothing to do with the technology but relate instead to market structures (i.e. regulation), standards, and other non-technical barriers to change. But it would be unwise to bet against change never happening…the economic, social and environmental needs to change are too strong.
So how can we realise the greatest benefits from these changes, when they come? One approach is to let the private sector get on with it: the market will ensure that the optimum answer is reached, surely? This may have some truth in it if we are confident that we have designed our markets correctly: because markets, themselves, are bits of technology we use to get the outcomes we desire. Let’s take the logistics industry, for example: we do not allow companies to compete with each by having ever faster HGVs, or by allowing them to save money by putting cheaper, but dangerous lorries on the road: we have designed the market to reward competition on some things, but not others. But digitised transport products and services will find new ways to do things and there may be a whole new set of undesirable ways to compete that can be created. What is more there may be whole classes of benefit that we cannot capture unless we have some layer of collaboration that we enforce. The classic example is the early days of the mobile phone industry: in the US companies were allowed to compete by fencing people into “walled gardens” – if you bought a phone on one network it could not be used on another. In Europe the market was built around the principle of consumer choice and the GSM standard meant that phones could inter-operate on different networks. The European mobile phone market grew much faster than the US market…what was good for the consumer was also good for the provider but it took government to take the lead and avoid the unproductive competition that happens when no-one is prepared be the first to share.
The point about data is that it could enable new ways of managing transport systems that could increase capacity by better planning and organisation. For example there could be much cheaper ways of reducing congestion than building new roads. In our road system congestion is, as the software people say, a feature not a bug: roads are free at the point of use, paid for out of general taxation which means we have, as a society, chosen to ration them by queueing (which is, not surprisingly, the same way we ration healthcare in the NHS…the other great public service offered free at point of use). There is no vehicle or device than can remove congestion: it is designed into the system. However we could use the capacity we have in more intelligent ways (25% of the road system is practically unused, for example: the bit between 11:30pm and 05:30 in the morning).
Interestingly enough we already think roughly this way in the case of railways. Network Rail doesn’t just build and manage metal rails: they manage a signalling system as well that enables the capacity on those rails to be safely maximised. A signalling system is data: it manages the status of each segment of line (i.e. is it occupied or not) and communicates it to the vehicles in the system to control the safe use of the system). The National Highways control rooms, Variable Messaging Signs and other operational systems do something of the same…but imagine how much better these systems could be if they set out to optimise the system as a whole, and made use of fine-gained geo-location data from individuals and vehicles, and similar new data sources. By the way I am not arguing for “control” as such: but there is a whole range of information and incentive, cost and benefit, that can be marshalled if we have the data, that can preserve choice and personal autonomy but generate much greater benefits from the existing infrastructure.
If we could, as a society, use the data that shows where everyone and everything is, and where it is going, to help people to see quicker, easier or cheaper ways to get where they are going then we might make big improvements. Funnily enough a lot of this data already exists: your mobile phone provider knows about all of the journeys you take already, but the market incentives do not encourage the use of that data to optimise transport. This does not require the government to take control of all data (or any, actually) but it does require us, as a society, to think about how we might go about getting access to data and use it to improve transport outcomes, whilst preserving people’s security and privacy. This will not be easy, and it certainly requires very different skills to the ones that we have traditionally needed to run our transport system…but I don’t believe it is impossible. And if we were to do that what would be the value of that data use? Would it seem as important as the roads themselves? Transport is a social good and my choices affect others to it is right and proper that a digitised transport system should involve some legal and social limits and, surely, the way the data enables the system will be as vital part of our infrastructure as the tarmac under our wheels?
We have already taken the first steps in this direction (National Highways, for instance, is thinking of how to be a service provider, rather than just the maintainer of physical infrastructure) but, to generalise about the hundreds of highway authorities, and supporting supply chain, and the whole ecosystem of regulators and providers that enable our transport networks.. we have barely started. Let’s accelerate the change and grasp the opportunity by thinking differently.
