Data analysis to prevent customers from getting into financial trouble

“Rabobank is the market leader in mortgages in the Netherlands, so mortgage customers are extremely important for our bank. If too many of these customers weren’t able to pay their mortgage payments, then the bank could start to have financial problems as well. So we’re required to maintain a reserve; a ‘capital cushion’. As data analysts, we use the Residential Mortgage Model to calculate the amount of that buffer. Rabobank developed the model based on EU legislation.”

Early warning

“We also use the Residential Mortgage Model to discover if and when mortgage customers will develop financial problems. That’s at least as important as the other purpose of the model. After an early warning signal, the bank can contact a customer to talk about the situation. The earlier we receive the signal, the more likely the bank can help the customer to avoid the problem, for example by temporarily suspending payments or calling in a budget coach. It’s amazing how we can use data to help customers.”

Tricks of the trade

“We’re currently redeveloping the mortgage model. On the one hand, that involves a more efficient way of coding and creating analyses so that we don’t burden the server as much and can improve the layout of our data warehouse. On the other, we think about improvements that could help predict potential problems for mortgage customers better and earlier. We solve those analytical problems using code. I’m often impressed by how my colleagues can use tricks of the trade to find answers to extremely technical problems while under pressure.”

Choosing for yourself

“As a data analyst, you can choose which parts of the project you want to work on. Some choose to focus entirely on data analysis, for example. I personally enjoy working on everything, including discussing the changes with the Dutch and European central banks. I love the variety of the work. If I spend too long on the same analysis, then I start to lose focus. But I also love listening to music while I dive deep into a coding or analysis job.”

Billions of data

“If you love sinking your teeth into analyzing data, like I do, then working at Rabobank is a party. At mortgages alone we work with billions of data. Customer data like location, purchase price, original mortgage amount, amortization, residual amount, payment habits; you name it. Plus a lot of external data about the economy, for example. We bring all those data together into source tables, then we filter and compress them into data sets. And we have millions of data sets.”

Initiatives for improvement

“A lot of data requires a lot of experts: there are around 80 people from the Netherlands and abroad working in this team. We work together in projects, but there are also a lot of improvement projects. We data analysts initiate those ourselves. Things like building smarter dashboards, setting up a better data quality framework, or exchanging knowledge between teams more efficiently. Thanks to one of those improvement projects, we now have a new tool to use. Before, we data analysts couldn’t look over each other’s shoulders when using coding languages like SQL and Python. But now we can see what other people are doing live, and we can work in each other’s scripts online. That makes it easier to work together.”

Completely done

“You can decide whether or not to participate in improvement projects or work groups. I never feel pushed to think along; I just enjoy doing it. I participate in the work groups for storing standard documentation and for recruiting loss specialists; experts in the field of dealing with losses. Those activities make my work more dynamic. Sometimes I’ll be completely done at the end of the day, but I’ll still feel like working on a piece of code I’m writing for an improvement initiative within the department.”