Royal Bank of Scotland goes back to 1970s values with big data

Blurred vision

In the 1980s, RBS, like most other banks, lost focus on its customers and became more interested in creating products to meet sales targets.

In RBS’s case, that meant selling credit cards to customers, whether they really needed them or not, said Nelissen, who is known in RBS as “the data guy”.

“We had to get hundreds of thousands of credit card mailers out of the door. That was all that mattered,” he said.  

The tactic may have boosted profits, but it left most customers with a feeling that RBS did not understand or care about their needs, said Nelissen.

IT systems in a poor state

The bank set about restoring its reputation for customer service two years ago, setting a target to become the number one UK bank for customer service and trust by 2020.

But it faced significant hurdles, including out-of-date IT infrastructure, poor data quality and limited funds to invest in new technology, Nelissen revealed.

“When we started, everything was a mess, things were terrible. There was a conclusion: your data quality is terrible, you can’t get access, you can’t do the analytics, the business does not engage, it’s a difficult environment and unless we can invest a lot of money we are not going to get anywhere,” he said.

Getting personal with account holders

The bank has created a “personology” team to analyse customer data so that it can meet their real needs, even if that means smaller profits in the short term.

The team realised it could focus on projects that would make a significant difference to customers without major investment in technology.

“We got close to customers and found out what the problems were. We realised the solutions did not need to be perfect, they just needed to be better than before,” said Nelissen.

Data analysts also spent time talking with the most successful bankers to find out how they prepared for meetings with clients. The team discovered, for example, that the best bankers would check whether customers with packaged bank accounts were paying twice for mobile phone or car breakdown cover, before meeting them in person.

RBS set about restoring its reputation for customer service two years ago, setting a target to become the number one UK bank for customer service and trust by 2020

The bank used analytics technology to identify and contact every customer paying for insurance products twice to alert them they were already covered. The majority decided to stop paying for extra third-party cover.

One of the most effective initiatives has been to wish customers a happy birthday if they turn up to a branch or contact the callcentre on their anniversary. “It sounds trivial, but wishing customers a happy birthday gives us fantastic service,” said Nelissen.

Another simple service but effective service is sending customers a text message if they forget to take their money when using a cash machine. Now, they receive a text within 24 hours, letting them know that their cash is safe. The bank plans to introduce more advanced analytics technology that will deliver messages instantly in the future.

The bank is also using data analytics to identify customers who could save money by consolidating their loans. “We look for customers who have unsecured borrowings with us across a number of different asset classes, where, if they came into us, we could lend them that money at a cheaper rate and save them money,” said Nelissen.

“It genuinely costs us money to do that, but we do it because it’s the right thing to do for our customers,” he said.

RBS has made its biggest improvements to customer service by changing its approach to mortgages. In the past, the bank used to profit from customers who forgot to renew their mortgage deal, and ended up defaulting to the more expensive standard variable rate.

Now the bank is using analytics to remind customers when they log in, or access banking services through a mobile app, how much they will save if they sign up for a deal. The project helped the bank reduce mortgage attrition by 5% in 2014, said Nelissen, adding: “I don’t think we have even got started yet.”

Inside the data warehouse

To make all of this possible, the bank has brought data from multiple IT systems into a data warehouse.

The personology team are using technology from Chordiant, now owned by Pegasystems, to develop customer care tools that can make recommendations to staff in branches and callcentres that will offer benefits for clients, even it if means losing out on short-term profits.

The bank has built dashboards using SAS visual analytics tools to show how these simple changes are affecting customer satisfaction. “In the past couple of years, the difference we are making to customer satisfaction, and net promoter scores, has been significant,” said RBS’s director of analytics and decision-making, Andrew McMullan.

Changing the banking culture

One of the biggest challenges in the project has been shifting the bank’s head-office culture away from developing new products and meeting sales targets towards improving customer service.

“That is a big cultural shift to make – for people to believe this approach will be right in the long term and to set aside some of their personal objectives,” said Nelissen.

We want to really help customers get the most out of their banking relationships

Christian Nelissen, RBS

Banking staff have also taken some convincing that it is possible for RBS to use its data to benefit customers in this way. There were a number of projects that were really good, but which did not take off because staff did not believe in them, Nelissen revealed.

Some senior banking staff, for example, were initially hostile to the idea of alerting customers to the fact they were paying for some services twice. “I remember having a conversation with a manager who told me, ‘You are mad, we are going to lose our P&L’,” said McMullan. “We have not. Every one of those customers has cancelled the other [non-RBS] product.”

One way RBS tackled this problem was by ensuring data analysts sit together with their business colleagues to work jointly on projects. “I have a Christmas party test – I know they are doing a great job if the guy they support does not feel he can have a Christmas party without inviting them, because they are working so closely together,” said Nelissen.

Existing technology at its limit

RBS has now reached a point where it has done as much as it can with its existing IT systems, said Nelissen. “We were driving the existing environment as hard as we could, and to make the front end look good we had a lot of people in the background doing some fairly tightly controlled, but vital processes,” he said.

To customers, the bank might look seamless, but the technology behind the scenes is messy, he said. “Our customers don’t understand how complicated we are. We have lots of systems, the data is rubbish.” he said. “They think they have a relationship with one entity, but from the back end it does not look like that at all.”

The bank is now investing in open-source big data technology, including Cloudera’s implementation of Hadoop. It is working with small startup companies offering innovative systems for dealing with unstructured data.

“We want to really help customers get the most out of their banking relationships and step in to help when we see them struggling with something,” said Nelissen.

RBS plans to update its recommendation engine with technology from US-based customer relationship management (CRM) and BPM specialist Pegasystems. It chose the company’s technology in late 2014, following a competitive tender, and plans to go live in August 2015.  

“There is a huge amount of alignment between the way Pega thinks and what we are trying to do,” said Nelissen.

RBS plans to use Pega to create decision tools that will allow callcentre staff to make tailored recommendations for customers. The technology will combine decision trees created by business specialists, with predictive and adaptive analytics technology – which learns with experience – to select the best offers for each customer.

The technology, which will be integrated into mobile apps and online banking, will be able to spot patterns and make different recommendations according to the customer’s location and what they click on.

“We are using the data we have and our decision-making ability to replace our branch managers, to have conversations with them,” said Nelissen.

Getting into the minds of customers

The project is part of a bigger programme, which includes investment in big data technology to analyse unstructured data such as customer feedback.

“On the unstructured side we want to explore the best of breed. The [startups] have got the new technology, so we are trying to work with the best in every part of the stack, so we have an open-source stack,” said Nelissen.

The bank’s ambition, he said, is to better understand what customers are thinking about and what they need. For example, when young people start saving, looking at the bank’s online mortgage calculator, or reach the end of a fixed-term mortgage deal, the system will alert staff to talk to those customers about mortgage options.

On the other hand, if a customer has made a complaint to the bank, it’s more important for RBS staff to deal with that. “The most important thing is what is in the customer’s head, what is worrying them,” said Nellison.

The bank is also considering extending data analytics to corporate banking. That would give business advisors the ability to check the company’s cash flow, or foreign currency payments, and receive automatic recommendations to help the client.

“You start to get the bankers to the place where they are as good as the best people we have, and you make it easy for them,” said Nelissen.

Winning loyalty

Rather than running marketing campaigns, the bank plans to use technology to offer advice that will benefit clients, even if it means losing out on short-term sales of financial products.

But the strategy will pay dividends by building up customer loyalty, said McMullan. “We reckon that over 90% of customers will stay with us because we did the right thing when we had to,” he said.

The bottom line

Nelissen is reluctant to discuss the return on investment the project is likely to create, but he said the bank would not be funding it if it did not make a difference. “When I went for dinner with the CEO he told me that for every £1 he spends with us, he gets a tenner back,” he said.

The science of being personal

RBS has renamed its data analytics department the “personology” department to help the bank’s staff understand why customer data is important.

Personology – the science of being personal – is all about using data to help the bank understand each of its customers, said Andrew McMullan, RBS director of analytics and decision-making.

“It’s given my team a real identity,” he said. “They come into work each day and they are really engaged in making things work for our customers.”

RBS’s strategy is to use data to offer genuine benefits to its customers, rather than to make a quick sale. In the long term, this will give customers a sense of loyalty to the bank and, over time, make it more profitable, RBS claims.

McMullan told a story about his mother-in-law, who took create delight in telling him that she would never bank with anyone other than RBS’s Scottish rival, the Bank of Scotland. The reason – her bank manager spotted a problem with her daughter’s account, helped her solve it, and agreed to waive all banking charges.

“She said, ‘I will never bank with anyone else’. Why would you if someone knows you so well, and is prepared to help you?” he said.

McMullan hit on the idea of sending customers a text if they accidentally leave their money at a cash machine after he made that very mistake himself. He was told by his colleagues that it would take two weeks to find out whether he had lost the money or whether the ATM had reclaimed it.

“I phoned the operational guys and they said, ‘No that’s wrong, you know straight away’. I said, ‘Wouldn’t it be good if you got a text, saying your money is okay’.”

He approached the board, who initially had difficulty believing that anyone would be remiss enough to leave their money in a cash machine. But when he presented statistics showing that the problem affects half a million customers a year, they were convinced.

Now customers receive a text within 24 hours, reassuring them about their money. When Pega goes live later in 2015, the text will be sent instantly.

The ability for branch and callcentre staff to wish customers happy birthday is helping the bank to shed its image as the least trusted bank in the UK. In one case, an 85-year-old man was in tears after staff wished him happy birthday.

“He was crying because no one else knew him, and some of the staff were crying too. It was very emotional,” said McMullan.

RBS is also developing ways to make it easy for customers to apply for loans. The bank has a policy of lending only to customers who have a cheque account. Now it is able to use that information to ensure that when eligible customers do a Google search for loans, NatWest and RBS appear at the top of the search ranking.

Similarly, customers who have a high risk profile, or do not have cheque accounts, will be nudged away from RBS when they search, McMullan revealed.