With the world in turmoil, capital market firms must innovate fast. We spoke to industry veteran John van Uden, VP Infrastructure Hosting & Delivery at TD Bank, to find out why digital innovation is challenging and how ambitious firms can do it better.
After heading up major innovation projects at Citi for many years, John van Uden is now responsible for delivering an ambitious cloud migration at TD Bank. We sat down with him to get a leader’s view of the state of digital innovation in the capital markets, and what organisations can do to ensure their transformation initiatives succeed.
ThoughtWorks: Which technology trends are having the biggest impact on the capital markets sector?
John van Uden: What we’re seeing is a shift towards a commodity-based approach to technology, where organisations are buying the standard pieces off the shelf and then focusing their IP on where they can add value. So, for example, you might have commodity software to process trades, and then put your efforts into improving risk calculation models or finding ways to more intelligently price, based on the huge amount of data that’s available to you.
Another big change we’ve seen on our public cloud journey at TD is that, where previously we might buy into a technology for five to seven years, we’re now taking the view that we may only use technologies for 24 months. We buy into products we like – the ones that can deliver value for us – but on our cloud journey, we want to be able to pivot whenever we need to.
ThoughtWorks: The global COVID-19 crisis has had a huge impact on all of us, but how prepared were capital markets organisations for that level of uncertainty and disruption?
John van Uden: Well, if you look at trading floors as an example, across the industry many of the big players had a lot of problems. That was quite interesting, because we’ve all been through quite a few seismic events over the last 20 years, so you’d think banks would be better prepared. We all went through 9/11, and 7/7 in London, and had periods where access to trade floors was quite complex and we relied on contingency trade floors.
The COVID situation has thrown a curveball at trade floors, but it’s one we coped with pretty well at TD. We were ready to move to trading from home and made the shift over a couple of weekends. We had a kind of Frank Whittle moment, where we took two unconnected technologies and glued them together to create a great solution for our traders. We also sent the traders care packages with all the kit they needed, so they could just open it up, plug everything in, and get up and trading within half an hour.
ThoughtWorks: That’s a great example of some fast-paced innovation, but what do you see as the biggest barrier to digital innovation in this industry?
John van Uden: The biggest barrier is COVID-related – it’s having the conviction to continue on the journey in the face of adversity. Digital transformations require a concerted effort around investment that must be maintained; you can’t start investing and then change your mind and try to slim it down.
To help maintain that focus on the investment, it’s important to identify the areas that you think will add the most value. And it’s not just cost reductions that are important – there are many other business outcomes that are just as important, like increased revenue generation, entering new markets, and so on.
So, it’s vital to make sure it’s clear what outcomes you need to get from your investment in innovation, and then you have to have measures that show incremental progress. There’s no point in that journey where some magic happens and value miraculously turns up – it’s hard effort, day in, day out.
ThoughtWorks: What impact do legacy systems have on organisations’ ability to innovate at speed?
John van Uden: Legacy systems still have a role to play, and they don’t have to hold back innovation. Especially in the current climate, nobody’s going to sign off a massive legacy replacement, nobody’s going to ditch their mainframes. You have to accept that you’ll have a bunch of what we call ‘heritage systems’ that do a job perfectly well – there’s no value in making that shift for the sake of it.
The first thing you need is a good view of which applications you want to keep, which you want to get rid of, and which you just want to maintain. Then you can plan how you’re going to migrate those to the newer platforms in a straightforward path that adds value along the way.
On the infrastructure side, it’s really difficult to try to blend heritage infrastructure support with newer infrastructure as code platforms. So, in my experience, those teams have been maintained separately, with one team to the heritage systems and another team with skills on the new platforms to help them move forward at pace.
The thing you’ve really got to keep an eye on is how you look after the people who are in the heritage space as it starts to dwindle. How will you give them opportunities to move into the new technologies? How will you provide an educational path to new platforms for someone who’s been looking after a 20-year-old system that’s three years from expiry?
ThoughtWorks: You mentioned earlier the huge amount of data that the large players have at their disposal – but how can organisations ensure they’re getting maximum business value from their data?
John van Uden: You need to invest in the groundwork to make it happen, and make sure you have a good data platform defined. But a lot of people spend too much time building the framework for collecting the data and not enough time on extracting value.
Obviously, you’ve got to have that framework for ingesting and normalising the data; you’ve got to have good data governance, privacy teams, and data stewards. But what’s really important is how you enable the federation of that data so people across the business can extract value from it.
Sometimes, executives focus more on centralising everything into a massive data lake that they lose sight of, and they don’t know how all these data scientists – across multiple organisations – are going to extract the value that’s in there.
It’s also useful to remember that, as a colleague of mine puts it, there are plenty of bars of gold on the floor without getting too sophisticated; you just need to harness the data you’ve got. You don’t need to be on the far end of AI and ML to extract large value from data, as long as you’ve got a good, rich source of data on your customers and on the market.