You’re the company’s first Chief Digital Officer, which is a relatively new term many corporations have yet to embrace. How do you explain what it is you do to people with little point of reference for the role?
Well, to distinguish what I do from the role of a Chief Information Officer, which supports all the basics of IT for a business, there are perhaps four things which have given rise to a new era in technology that together explain the role I currently occupy. One is the reduction in cost of compute, and the movement of that compute from the computer to the cloud. Second is the ubiquity of bandwidth. Third is the growth in low-cost sensors and devices that can collect data from pretty much anywhere. Fourth is a range of technology-led use cases that feed off the first three–AI, automation, drones, augmented reality, robotics. Every industry is seeing a type of disruption they haven’t experienced in the last 50 to 60 years since the early days of the information revolution. This means you need a very different business response to compete. Most companies in competitive markets now face challengers in the form of start-ups, and so find they must respond in like manner. A Chief Digital Officer thinks about the various elements I’ve mentioned in terms of both processes and customer outcomes, and the reality is that at some point in time, the role of CIO and CDO will fuse in any truly integrated business. Our CEO Steve Robertson has a strong tech background–having set up BT Openreach–and our focus is very much on the digitalisation of our business and getting information on our customers to generate insight and action.
Start-ups often struggle to scale up to the point where they are self-sufficient or can be bought by larger firms. How can start-ups get in front of potential buyers and investors more easily and effectively?
I don’t see Thames Water buying start-ups, but I can see us partnering with small companies that are offering a product we could apply to our business. Personally, I like working with smaller technology companies because they’re more agile and you can shape what they develop and have them work well with you. The community of investors with angel capital that these start-ups need to tap into is very well funded and extraordinarily brutal. Getting funding is tremendously difficult, track-records are extremely important, and who actually makes these decisions can often be opaque. If you haven’t got the right network, it’s very difficult to succeed. The right board of directors need to be connected to the right group of angel investors, who need to be connected to right investors, all of which gives start-ups the right exposure they need. It’s really all about tapping into that community to make sure you’re funded and can showcase products to companies and further investors.
Is Thames Water planning to partner with any start-ups soon?
In our supply chain, we have a couple of smaller software companies that are developing for us. We’re also working with a small data analytics start-up called Explore AI. Those businesses are quite easy to find however, because they scale with demand. When it comes to hardware or tech related start-ups, it’s much harder, and you need to have the funding to develop products. We’re not working with anyone that is super small and in the field of leading-edge technology, but rather those that have the balance sheet to invest in sensors and technologies that will help us get better insight about our networks. For example, Microsoft and IBM have innovation funds that they are willing to use to invest in problem-solving issues with us. One thing we’re working on is a low-cost sensor to measure headroom in a sewer. Depending on the geometry of a sewer, there has been a minimum headroom space to avoid a flood. If we can measure this, we can know where there might be a blockage, which might lead to a flood or pollution. Some networks will be 20 feet deep, others two feet deep, but whichever you’re working with, if you can detect a problem digitally, you can deal with that problem much quicker. Some of those sensors cost around £1000 each, because no-one has yet scaled them. If you can get that down to £10, then the economics for deploying them nationwide change dramatically. Large firms are instrumental in scaling such products, so we work with them on these projects for long-term supply. I don’t think that its right for us to fund start-ups with our customer’s money, but I do think there needs to be wider, industry-led response.
Thames Water is investing £11.7 billion in new infrastructure in its next five-year business plan. How does your digital strategy come into the various tranches of that investment?
Well, we have 15 million customers, 150,000 kilometres of water, waste and collection piping, and 500-orso factories treating either water or waste. We have a whole heap of customer-type solutions, and a maximum network which needs monitoring, and factories producing stuff meanwhile. There aren’t many businesses in the UK with that sort of spectrum of assets, if I can use that in the broader sense of the word. The regulator in the UK has decided to split wholesale retail for most infrastructure businesses, except for water, so we don’t have that same firm boundary. If you think of some of the more painful problems our customers experience, they not only impact on the customer but on our networks and production as well. Getting real-time data on what happens between customer and supply is therefore critical. We had a pipe burst on Goldhawk Road in London in February this year. Water everywhere. It took us between 10 and 12 hours to identify and shut the main. However, this was a water network that like many has been built over hundreds of years and is interconnected. Out of hundreds of thousands of valves, of which you’re not sure those that will turn off the right main, you have to act on at least two at the same time. All this in midst of a flood and traffic moving through the country’s capital. With data that can pinpoint exactly which you’re looking for and that allows you to isolate the right valves, you’ll get a better customer outcome very quickly. So, we need to stitch together the data about our customers, network and production. We also want to connect our data to our website, so that our customers can see where there are traffic works to re-plan their route.
Early on this year, the United Kingdom recorded multiple incidents of burst pipes after temperatures rose following a prolonged freeze. The national water regulator Ofwat published a report naming several firms that weren’t adequately prepared to respond. How did these events motivate you as Chief Digital Officer, given your relatively new start at the company?
Well, we certainly weren’t walking blind before those incidents. We had a lot of data, but we weren’t always tapping into it. If you think about servers and production facilities, there are a lot of operational technologies measuring reservoir levels and so on. There is a system of control on those levels which triggers an alarm when those levels are exceeded. That is a system that continually uses data to predict an issue, but to do the same for piping networks puts a much bigger strain on the system’s computing power. The question we now have is how we should exploit our existing data to give the operational teams more insight and allow them to respond better. Since the big freeze thaw, we’ve been trying to build visualisations of real-time data flows in a pragmatic way for the operational folks. I think doing that will have, as it has already throughout the summer, a dramatic impact. Personally, I think this has crystallised for me the importance of the work we do with data.
The UK’s Environment Agency has warned that climate change, population growth and land use are exerting stress on England’s water supply and could result in shortages by 2050. Meanwhile, Thames Water plans to build a storage reservoir connecting Oxfordshire to the Thames that could take some 15 years to build. What is your understanding of a water scarcity threat to the UK and how should data be used to mitigate it over the coming decade?
The reservoir issue isn’t one I’m heavily involved in, but if customers use less water, or if we fix more leaks, then we have more supply. There is a range of technological responses to that issue also. By the end of this asset management period, we will have around 300,000 smart meters installed, 700,000 in the next, and 1.5 million in the one after. The intent is to be as fully metered as we possibly can be by 2035. All our customers will know the extent of their consumption, and we can help them reduce it further through awareness and smart tariffs. Typically, when a customer is metered, the reduction in the use of water is somewhere between 12 to 20 percent. One customer’s smart meter was inexplicably ticking all the time. Turned out it is was leak, something like £10,000 worth of water lost. We were able, as part of our customer-side leakage service, to identify that leakage and fix it for free. Around a quarter of the total leakage is on customer pipe work, not Thames Water pipe work, and being able to identify which is ours and which is customer side is important. In any case, we need to get much smarter about where these leaks are, so we can fix them much quicker.
Do you agree that anything that can be digitalised can be hacked, and that Thames Water’s drive to digitalise its operations as the country’s biggest water supplier is a rather daunting prospect?
There is no question that as our business becomes increasingly digitalised security of our systems and data is of paramount importance, especially given the criticality of our services in people’s lives.
In my new structure we appointed a Chief Information Security Officer who is looking at the security of all our IT and OT systems.
As a general point we will make sure we employ best practices using a pragmatic risk-based approach. On the IT side we are investing a lot of money to ensure our network, data centre, applications and end user devices are up to date and secure. This is the primary method of defence. Behind this we have a series of processes and protocols to reduce risk and to respond appropriately to a major global incident. On the OT (connected asset) side the risk is different as we have a number of distributed yet integrated technology systems and sensors across our network.
In this domain we need to ensure that we design the system in a compartmentalised manner so that any incident is isolated and can be contained. If we do this then the impact of any hacking will be less severe. For example, if you take a sewer monitor, the worst that can happen is that it spits our errant data. The response is to send an engineer to investigate who will see that the sensor is misbehaving, and we know not to rely on it. Ultimately in a worst-case scenario we would have to manually control all our key infrastructure, which is part of our incident response process anyway. I don’t doubt we will continue to remain a target for hacking and we will continue to invest in a set of capabilities to appropriately respond.
Interview by Jack Aldane