What set you off thinking about this challenge?

It all started from the fact that our younger clients were struggling to get consumer credit right when they need help to move into their first apartment, buy a car, etc. Banks grant consumer credits on the basis of a credit score. To obtain a loan, clients are rated by statistical algorithms that prioritize risks. The higher the score corresponding to a client’s statistical risk profile, the greater the chance he will actually repay his loan.
However, the analysis criteria currently used rely mainly on client banking history data. For instance, you have a better credit score if you’ve been working for the same company for a long time or if you have been with the same bank for several years.
These algorithms reach their limits for some clients, notably young people, whose banking history is still limited.
We also have to factor in new modes of behavior. Young active people are hyper connected, have high-speed access to information, and are constantly networking online. All these elements can help us know more about them and adapt our credit offers accordingly.

How can young people's access to credit be improved in practice?

Clearly, conventional data on professional experience or the length of the banking relationship cannot reliably identify young clients who will actually repay their bank loan. We therefore wanted to determine new, relevant data to give young clients a chance of fulfilling their projects.
So we tried to contextualize the information available in the best possible way: for example, earning €1,500 a month in Paris or in a village is not the same thing. The young person’s education also provides valuable information.

How did you proceed?

Ideation is born from listening to the market and our clients to understand their “pain points”. In a nutshell, our young clients’ problem is access to credit and ours is to find young people who will repay their loans.
Once the issue was put simply, we harnessed the skills of different teams from Natixis and BPCE. We also worked with young clients and with the marketing teams of Banque Populaire du Nord.
The multidisciplinary team focused on the following question: how can we enhance our credit score with non-banking data? The good news is that today, through hyper connectivity, we leave data tracks online rather like those we leave in the snow on a ski slope. This is also the case for government agencies. This is what is known as open data.
We’re all familiar with the tactic consisting in targeting our areas of interest to send us appropriate offers. This time, the goal is different. The idea is to improve the client experience by allowing a young person to access credit thanks to an enhanced score. Information available on the web is therefore used for the client’s benefit.

Why did you launch a Big Data challenge, and how did your collaboration with a start-up help?

We needed help to learn how to process large amounts of disparate data (big data). We therefore had to combine our knowledge of the banking industry with the expertise of data processing specialists.
So we staged a competitive bid between several start-ups specializing in this field to see if our ambition could be implemented operationally. We decided to work with Cap Digital, a competitiveness cluster we have partnered since 2015, to benefit from its ecosystem of start-ups specializing in big data. A request for proposals was launched in the form of a challenge in which around one hundred start-ups took part.

What conclusions did you reach?

The first findings are very interesting and promising. We found that publicly accessible open data is effective to increase credit acceptance by close to 25% and reduce risks by the same amount.

All this sounds easy, right?

It’s not quite as simple as it appears. The start-ups provided expertise that we do not yet have in-house, i.e. the ability to process disparate data and new ways of analyzing it to turn it into results for the bank.
The challenge wasn’t easy for the start-up either. We operate in a regulated framework where personal data protection is vital. Our partner therefore had to adapt to a highly secure technical environment.

What happens next?

The combination of a start-up’s expertise with our business-specific know-how is an excellent formula. And that's one of the main lessons we learned from this challenge.
We’re going to test our conclusions on a real scope, in pilot mode, to validate our customer promise of both increasing lending and reducing risk by 25%.
We’ll then continue the process to onboard the findings and continue to enhance our score with new web data.
In future, we could very well use open data more broadly so as to anticipate our clients’ behavior way upstream and fine-tune our marketing proposals.