The financial services industry has always been data-driven — that’s its nature. But three factors are shifting the game, creating huge opportunities for banks and other financial services firms:
- The volume, the variety and timeliness of that information
- The multiplication of channels (online and mobile)
- The advances in computing power, improvements in software and the proliferation of social media
Retail financial groups have an opportunity to get to know their customers better, and to personalise products and offers to particular needs. They can now direct marketing not to wide segments, but to micro-groups or even individuals. They can be more responsive to events in people’s lives — a holiday, a house move, a wedding, for example. They can improve risk analysis and fraud detection, e.g., by scouring what people are doing on Facebook or Twitter. And, they can gain more insight into how their brand is viewed in the marketplace — both the things people like, and the things they are less keen on.
The following three experts got together to discuss the opportunities and challenges of data analytics in banking:
- Kamran Ashraf, VP, Head of Analytics and Information Services, Visa Europe
- Dr. Kit Skov Hagemann, Head of Customer Experience Measurement & Insight, Santander
- Oliver Werneyer, VP, Data and Distribution Leader, Swiss Re
How is technology and data revolutionizing customer-facing marketing for banks and other financial services firms?
Kamran Ashraf: I sit on a team called Visa Consulting & Analytics whose objective is to help solve the business problems of banks and retailers (such as retention, acquisition, usage). Luckily for me, I’ve been able to see the impact of data and technology on many clients across Europe.
Technology is now more accessible, and simpler to use, in interrogating and analysing large and complex data sets – which is empowering businesses to get much better at delivering personalised service. Ten years ago, most financial service companies really were only doing one-size-fits-all marketing. For example, they would offer the same interest rate on a new mortgage product to everyone who applies. Now you can tailor offers to people’s needs and preferences and this increases response rates. You can target people who have opted-in for marketing with different pricing, value-add offers, incentives and bundled propositions.
The biggest benefit for companies is that they have less wasted budget and improved business performance. Rather than needing to spread a small budget over a large prospect base, you can increase the value of the offer per person by targeting fewer people with the same amount of budget. With the data and tools available now, you can say ‘well, these people are never going to respond to offers that are based on price, because they’re not price-sensitive’. Now, you know only X percent are going to respond well and so you can improve the value proposition to the customer and place your advertising in the right place, at the right time.
Oliver Werneyer: My role encompasses a lot of innovation-driven responsibilities. I am part of a team that looks after the data strategy for Swiss Re Life and Health: anything from how we change technology, to the people we hire for internal purposes, and our clients. Other areas include predictive underwriting projects and some special projects such as mergers and acquisitions.
Technology forces us to think of our customers in a different way. We are a very traditional industry. Probably, not much has changed for last 100 years in the way we do things. Now, with all this data that we can process now, it means we can get new kinds of insights into our clients, and change the conversations we have with our clients and their customers. Our clients are insurance companies. We can now feed data back to them in a way that helps them manage their company, and use data to really provide the best service and products to their customers.
Traditionally, the relationship between an insurer and a re-insurer is very irregular. You re-negotiate terms once a year, you settle claims accounts maybe quarterly, and the feedback that a re-insurer would give a company would involve basic metrics. That’s because of the expense of the technology and the manpower involved. With new technologies, it becomes achievable to run reports nearly on a real-time basis and provide feedback to our clients almost immediately. That means, if something goes wrong in the industry, we pick up on it immediately and warn them that there’s a massive amount of claims on, say, the critical illness side. Then we can investigate what the cause of that problem is.
It means we can be more sensitive to issues as they arise. In fact, the conundrum in the insurance industry is that you cannot always react immediately as some time is needed to let the issue play itself out, because these are often long term risks. We want to avoid making wholesale changes to strategy and product design if there was only a momentary issue or change in customer behaviour. It may be that there is just a short spike at the beginning and not everything’s going to be bad.
Kit Hagemann: We are in the middle of a transition of all our customer measurements, where we are trying to get closer to the customer. That means we can give much swifter feedback, in 24 hours, instead of days as it was before. We are also designing new processes that mean we can go straight back to the customers whenever they have a problem. We are improving the way we measure satisfaction and loyalty in general.
The customer might say there is something unresolved. They state that in an email survey and that is transferred out to the different call centers or branches. We can then contact them and ask them about their issue. In the old days, we didn’t have interactive customer feedback. We just measured a level of satisfaction and left it at that.
The key thing is to be able to go back to the customer as quickly as possible. That’s important because the problem with insights that take too long after the event is that they are very inaccurate. With more timely feedback, the accuracy increases. For example, say you go into a branch and have a bad experience. That experience will stay with you for quite some time, so even if you go in a week later and have a good experience – you will still have a tendency to remember the bad instead of the good. So that last good experience will then be rated as bad. We are trying to tailor what we do to specific incidents, and also making sure that timeliness with which we get the feedback increases accuracy.
These programs are also driving incentives in different channels. So it’s really important that we collect accurate information, so we can drive changes if we need to. If branches hit a target, they get a bonus on a quarterly basis.
We expect to see a drop in negative complaints, because we can capture this feedback and start solving these issues before they become complaints. Many companies collect feedback but they don’t go back to the customer. There’s nothing worse than when customers say they are dissatisfied and then nothing happens. We will track the complaints process later and see how that trends downwards. That means we can pick up big problems before they turn into complaints.
Of course, we know that satisfied customers stay longer with the business and that we can do more up-selling – they buy more products – and they start recommending us to other people.
How do you think the business should be organized to get the most out of customer data?
Kit Hagemann: Recently, we had a big organizational change. Before, we sat in an independent organization within finance. That made us objective. But we lacked access to customer data in general. Now, we are now part of CRM.
That means we can make the best use of data across functions. In the past, you would see people measuring customer satisfaction sitting in marketing and they would work with communications. But they would be too far away from all the other types of data – for example, CRM data.
Organizations should be set up in a way there is an easy flow of data between elements, whether it’s satisfaction, CRM, or market information. It should all be centralized in a way that eases the flow to generate greater insights.
We’ve also been working with HR on employee engagement to see the impact on customer satisfaction. That requires a lot of stakeholder management and trust, because employee data is confidential. If the employees think I’m going to compromise them, I’m never going to get the data, even though we are all internal stakeholders. The biggest constraint generally is that data is stored in silos. Getting this information together is one place is very difficult.
Oliver Werneyer: The older and bigger the company, the more difficult it is to change it. You’ve got the money but you’ve also got the legacy systems. The biggest problem we have is how to integrate the systems we have. We might have five or six different data warehouses that hold data on everything from underwriting to claims.
Changes in technology mean that everybody’s evolved to a more inclusive approach. That means technologies can handle more than one source of data at a single time. They can actually go out and fetch data. And they can use real time data and stationary data at the same time. This new technology has actually made it much easier for big companies like ours to absorb that and actually make use of our data, rather than take all of our systems, get rid of them and implement one new system. That accelerates the delivery of big data as well.
Kamran Ashraf: You need a cross-divisional team to work on it and drive a return on data investment. A lot of organisations spend money to build a data warehouse and employ a few analysts to extract data from it. But they usually end up with pockets of people using it, and the whole company doesn’t benefit at an enterprise level.
Particularly, you need a cross-divisional committee overseeing two things. One is the governance of that data to make sure people adhere to data compliance policies and regulations so data is not being used in ways that would be against the customer’s wishes. From a physical security point of view, the data should be secured with restricted access and protected so it cannot be breached.
The other half of it is to apply the data and analytics to relevant business problems. For example, you may be an organisation that’s losing customers. You can look at the data and find out not just who or when customers are leaving, but also, digging deeper into the data, why they are leaving, and what can be done to solve the problem. Such root cause analysis may show, say, attrition is highest in month 11 before they pay an annual fee. That, in turn, may change how sales contracts are structured.
How can social media be a valuable source of customer data and insights for banks and insurance companies?
Kamran Ashraf: In the big data world, people talk about the three vs: volume, variety and velocity. Variety is concerned with all the new types of data that is now decipherable – unstructured and semi structured data that is fuelling techniques like text and sentiment analysis. People are making use of social media such as Twitter and Facebook. Customers are openly posting information about your organisation and your competitors telling us about bad experiences in branches, with ATMs, or alerting us to possible fraud.
Historically, banks have used sample based research or very small customer panels to represent an entire portfolio or market. There is no way that can be representative of the broad set of business issues that impact the brand. Now, you have a better picture of your customers, what they want, and what they require from you as an organisation. They are talking about it online with their peers on a near daily basis. Potentially, that holds a goldmine of information to draw insights from.
The key thing organisations are trying to achieve especially in financial services is to develop an omni-channel service and analytics. What are customers experiencing across branches, letters, mobile, or when they ring the call centre? Customers expect banks to be multi-channel with integrated systems, available 24/7. Looking closely at the data shines a light on whether the customer experience matches the brand promise.
Oliver Werneyer: I would ask: What are we trying to achieve with big data? It’s really to understand the insured life. We don’t immediately see the value of social media in terms of health. But in the longer term, it may be possible to see from social media if people are, say, checking in at places like, the gym or on a cycling route. If that person is a diabetic, we can compare that information to what he said on his questionnaire. If he travels a lot and goes to the gym, it may be that he’s managing his disease well and we don’t need to ask him about it. But if he stops going to the gym, we may know he’s not looking after his health because he’s less active. That might be interesting for us.
The most immediate value of social media would be mainly from a fraud and non-disclosure perspective. If people say they don’t smoke and drink, but they like five smoking brands on Facebook, that could mean he probably lied about his smoker status, and we can ask him to confirm that before it gets to claims stage. Then at the claims stage, if someone is claiming for disability, and he’s travelling a lot to Greece and Spain, it probably means they are probably not bed-ridden. Then, you can say ‘well, maybe we should audit that claim’. The aim is to understand the client, so we only have to bother them if something is really up. You don’t want to go to somebody’s house who is actually sick and say ‘please prove that you’re still sick’. That is just annoying to the person who is already struggling with illness or incapacity.
Kit Hagemann: It’s a valuable source of data and insights. What’s great about social media is the timeliness. That can be really beneficial. But I also think you have to be very careful what you use it for. You can end up reacting to the loudest voice. Someone can say something on social media and we can blow it out of proportion. I certainly wouldn’t take social media and build a company-wide strategy around it. It’s better to combine it as an add-on to other data that we have.