Embracing shifting consumer preferences in retail banking
Taking an interest in consumer behaviour
Over the last eight years the Euro Interbank Offered Rate (EURIBOR) has remained close to zero – and even negative – to where consumers have noticeably changed their financial habits. The trend is seeing savers moving away from fixed-term deposit accounts to non-maturing deposits; with some increasing their risk profile in search of higher returns.
The influence of the EURIBOR is coupled with changing consumer tastes, where services have seen greater consumer attention and uptake. Salient examples of this include subscription TV, streaming music, online dating or online shopping and delivery services.
Financial services is not an exception. Compared to a decade ago, it is much easier to acquire and manage financial products without needing to go to a bank branch or to speak with a relationship manager. Anyone can now allocate their savings into investments with just an internet connection.
On top of these changes, COVID-19 is also having an unequal impact on the behaviour of various customer segments. While some individuals have increase “involuntary” savings as a result of remote working and fewer investment opportunities, others have lost their source of regular income.
All these elements challenge the capacity of traditional interest rate risk models to provide reliable results on the impact of interest rate on client’s behaviours and decision making. And is the reason why a large Spanish bank decided to re-evaluate how they assessed deposit behaviour.
Devising an interesting solution
We wanted to understand the customers decision-making process linked with savings and deposits, we relied on the individual customer (in contrast with the traditional “portfolio” approach of using the mass of aggregated deposits) as the unit of analysis. This implied identifying for each customer all the funds maintained within the bank and generating a methodology able to separate these balances in three parts:
- unstable or transactional funds - maintained to cover the customer’s regular transactions (paying the rent, buying groceries, etc);
- stable and insensitive funds – the buffer that each client would like to maintain to cover for unexpected events, security, etc; and
- stable and sensitive funds – the remaining balance or excess funds above the security buffer that would be susceptible to migrate between financial products or entities.
This methodology also needed to identify, out of the stable and sensitive funds, which ones showed a higher / lower propensity to migrate. And it had to be done in an innovative way as we were not able to rely on traditional regressions due to the low and stable interest rates curve over the last few years.
With this in mind, we generated an approach able to tackle this challenge which relied on two core elements: behavioural economics and machine learning.
Our work allowed us to:
- cover the needs linked with interest rate risk analysis at the portfolio level by identifying unstable, stable and insensitive, and stable and sensitive balances;
- increase the banks levels of understanding of customers decisions related to savings, by generating a segmentation of clients based on their capacity to save and their propensity to migrate savings.
Balancing the books
The large Spanish bank applied the methodology we created to the entire customer base and are migrating their interest rate risk models to a customer-based analysis.
The learnings from this analysis have been shared with the commercial team which have shown interest in exploring the potential uses of the behavioural segmentation to achieve certain commercial objectives -i.e. promoting off-balance sheet products- in a cost-efficient manner.