Danish investment bank Saxo Bank has managed to drastically reduce the time it takes to onboard new customers and get them trading on its platform thanks to some recent investments in data science and advanced machine learning technologies, and now it is planning a broader cultural change aimed towards becoming a cloud-first bank.
Saxo Bank specialises in brokering online trading for clients across 170 countries through its core technology platform, making it a "fintech before that term was even invented," according to its founder and CEO Kim Fournais.
This platform is currently a multi-tenant private cloud system but is in the process of moving to the public cloud with technology partner Microsoft Azure. Speaking at the TBM Executive Strategy Forum last week, Global COO of group IT at Saxo Bank, Richard Douglas, spoke about what the bank hopes to achieve with this shift and the dividends it is already getting from some recent technology investments.
Douglas talked about how these technology investments are all aimed towards three broad strategic goals: digitising the entire value chain, creating a world-class sales and service organisation, and industrialising the bank's white-label, wholesale offering.
"The more that we can digitise and automate our value chain means we can take away manual processing or ad hoc processing and give our sales and service organisation more timely, relevant information, so that when we're interacting with customers, we're interacting with relevant information, and insightful information to make it a world-class experience," he explained.
To support these broad aims the bank has increased its annual technology investment by 40 percent. Douglas gave three tangible examples of how the bank is leveraging modern technology to deliver on these three strategic goals, starting with digitising its onboarding process.
Of course onboarding in financial services requires some stringent know your customer (KYC) and anti-money laundering (AML) checks, which can introduce friction to the process.
"We have to have an enhanced client risk framework, which is really around understanding our direct retail clients to more complex institutional clients," Douglas explained, with the process taking on an average of five days across an average of 1,500 new clients a month. "That doesn't sit with the ambition levels of the organisation."
What Saxo aimed to do was leverage robotic process automation (RPA) technology to speed up the process. "We looked at integrating with a lot of different third parties around being able to ID customers on a live basis," said Douglas, while "also enhancing the data from external sources."
The result was an average onboarding journey for a "non-complex customer" being reduced from five days down to less than one hour on average, leading to a record month in April of 18,000 customers onboarded, up from a fairly static average of 1,500 a month.
"So we achieved 12x scalability in terms of being able to onboard clients without adding one additional headcount into the onboarding team," he added.
Lead quality scoring
The next step after onboarding clients is to actually get them funding their accounts and trading on the platform, a task which still relies on human sales staff calling clients to talk them through the process and nudge them in the right direction. Technology can help here too, by enriching the onboarding process with data and some machine learning algorithms.
"If we 12x scaled the number of clients coming into the ecosystem we have a historical issue of only a small percentage of those clients traditionally going on to actually fund their accounts and start trading," Douglas explained. "So how do our sales staff start understanding which clients are worth contacting to ensure they go on to fund?"
So instead of sales staff using their gut and experience to guide them towards which clients to contact, Douglas wanted to put in place some analytics around lead quality scoring to try to ramp up the success of these client interactions.
By pulling together its historic CRM data, information captured during onboarding, and some external data sources, the bank is quickly able to understand "that if someone puts in an address that is part of the wealthiest part of the city then there is a good chance they're going to be a client that we want to do business with," for example.
These algorithms were all built in-house by a data science team of 13 and a data engineering team of 18 people, both of which have grown significantly in the past two years.
"During this period we have brought in new leaders with fresh perspectives and the gravitas to take the organisation on the journey to appreciate and focus on the opportunities which data science and data engineering can open up," Douglas added.
Naturally, however, sales staff pushed back on being told how to do their jobs. "The sales guys weren't too happy," Douglas said.
Easing those concerns took hard work, from "identifying the impacted stakeholders to lots of communication around data science and machine learning algorithms so that they will understand the real-life applications you can make with this stuff and the difference it can make to their day jobs," he explained.
Then, again, the results spoke for themselves (see below), from a conversion rate hovering around 10 percent rising to well over 60 percent of the clients with a high lead quality score whom Saxo contacted going on to convert part of their accounts and start trading.
In a broader sense Saxo Bank currently runs a single core technology platform. "It's global, it's multi-tenancy and multi-asset," Douglas said, "and we are in the process now of wanting to take that entire stack to the cloud."
Today that stack resides in a private cloud instance in Denmark, leading to some latency issues with Asian clients, so the bank wants to further leverage the scalability, flexibility and advanced machine learning capabilities afforded by hosting in the public cloud, specifically with the vendor Microsoft Azure.
"We're striving for continuous integration, continuous delivery, we want to increase the levels of automation and not have any manual configuration through that process," he explained. "The commercial outcome of that is if we can automate more and get faster cycle times for software delivery, then we can reduce the time to market of features and new value-add items with customers, but also it improves your governance and control in the platform as well."
Douglas wanted to stress that this move wasn't about cost savings as much as it was about remaining competitive in the future.
"At no point am I talking about costs or cost reduction, except for more cost-effective scalability," he said. "I'm not talking about this being a business case.
"We're very much on this journey right now. It's going to take us a year or two. But I guess some of the learnings that we've found so far is that this is not a shoestring initiative; you need to be willing to invest if you haven't grown up in the cloud, there's a lot of fundamentals that you need to go back and work through. Whether it's those fundamentals around your software delivery life cycles, whether it's your architecture, there is a lot that you need to invest in."
Of course a seismic culture shift of this kind from a 27-year-old financial services organisation will require some change management.
"This is really about making sure that you as a leader and the leaders below you are helping to foster the right culture," he said. "The way that we define it at Saxo is culture equals the sum of the behaviours of the organisation."
These key behaviours are: personality, knowledge, skills, attitude and motives. These all need to be as aligned as soon as possible if your organisation is going to make a successful digital transformation, according to Douglas.
"Just because you had people in the past who have been doing a role, if you want to move that organisation forward, then potentially you need to look at bringing in new knowledge and skills," he said.