Alternative Data for Improving Financial Services

Dr. José Murillo, Chief Analytics Officer at Banorte, sat down recently with IAssesments on the importance and potential of alternative data and psychometrics for improving financial services.

Alternative data solutions are being adopted widely in financial services across Latin America and around the world. Such solutions are already considered by many to be the future of credit scoring, especially as a means to service unbanked groups, and those without traditional credit scores. In Mexico alone, for example, there are roughly 60 million people who are unbanked, according to World Bank statistics, and globally there are as many as 1.7 billion. Clearly, there is a tremendous need to service this portion of the population, and the use of alternative data solutions is one potential way to achieve that.

That is why we are particularly excited to have had the opportunity to sit down with Dr. José Murillo, a leading expert in the use of alternative data in financial services. Dr. Murillo is the Chief Analytics Officer at Grupo Financiero Banorte, in Mexico, one of the largest and most profitable commercial banks in Mexico. Banorte is also well-known for its progressive approach towards using and testing new data solutions, and José has been a big part of that. In fact, according to José, Banorte has been undergoing “transformational efforts over the past 6 years to enhance its use of information, as a way to better serve its customers,” and has successfully surpassed many of its competitors in this respect.

The Big Revolution

José starts our conversation by explaining that there is a “big revolution on how to better understand a customer’s risk profile, which has to do with the data you collect, and the way you process that data.” He breaks this down into 2 main types of data: First, banks already have a lot of available information, but don’t always leverage that information. For example, a bank’s transactional data, the time and place of transactions, etc. The other type is external information that can be accessed, such as telco and social media data, as well as information that can be learned through direct interventions, such as through psychometric surveys.

When asked if all data is fair game to be used a credit data, José considers this carefully, and explains as follows: “The broader the information sets you have, and the more you use them, the better you are able to understand the connections. There might be information that you thought was not relevant, but perhaps you used it on a small sample set. That’s the power of artificial intelligence. Some of that information will need to stand the test of time, proving that they are not just spurious relationships, but that they are really explaining causality. And then, the challenge is how you integrate all of that information.

The Right Mindset

Indeed, integrating and adopting new data sources is a definite challenge. What Dr. Murillo describes requires a great deal of careful research, and he and Banorte really seem to have embraced that approach. “If you really want to exponentiate the power of your data,” he explains, “you need to do lots of testing. And that leads to the first pre-requisite – you need to have a mindset to be open to experimentation and testing. You need to be investing and using part of your time to discover what new datasets are relevant.”

José briefly refers to his experience with psychometric solutions in this respect. “Among those sources that have potential are psychometric tests. The preliminary results are encouraging,” he adds. Dr. Murillo then draws an interesting corollary for this concept: “The potential is rooted in pre-modern economies, when local banks got to know the people they were serving after life-long personal interactions, and would know if a customer was responsible. Today, small neighborhood shops are similar, knowing their customers personally, and deciding whether to lend them before their paydays. In modern times, and for large corporations, it’s difficult to have that type of personal interaction, but you can substitute that knowledge with alternative data sets. Among them are psychometric tests, which can help understand the psychological makeup of the customer, and their intrinsic motivation to honor their debts.”

“You need to have a mindset to be open to experimentation and testing. You need to be investing and using part of your time to discover what new datasets are relevant.”

It’s Going Mainstream

It can be argued that fintechs can more easily risk failing with new models, while banks have a different type of responsibility to their customers, making them naturally more cautious and slower to adopt. José agrees, but clarifies that “legacy systems can slow you down in some cases, and in some cases it helps. Banks can take on small scale pilots to see what works and what doesn’t, but eventually they are looking to scale up the value for the customers they are serving. They [banks] do need to have a swifter response, and a farther understanding of the usage of alternative datasets for credit, and sometimes behave more like fintechs. But, in a nutshell, that is already happening.

It is well known that big banks pilot new technologies often, but are these just endeavors to test solutions for relatively marginal populations, such as the unbanked, or will these solutions eventually be adopted more widely, and even replace core legacy systems? “It’s going mainstream,” Dr. Murillo answers assuredly. “And the pandemic has accelerated the need to incorporate alternative datasets for all customers. COVID has significantly increased the pace of digitalization of the customer base, and this affects 3 main aspects: 1) increased scale – for serving more customers, at almost zero marginal costs; 2) increased scope –  for providing better products and services; and 3) increased learning – the more customers and products, the faster you can learn to serve them better, and understand the connections between the products and the way they are being used.”

“If I can understand you better, and I use the power of that understanding to help you be financially stronger, then that is something that you are going to welcome and embrace.”

The Key is to Benefit the Customer

As these services are evolving and being adopted, there might be concerns regarding customers’ own viewpoints around the use of alternative data. José agrees that this is an important consideration, and feels that customers’ expectations have changed in recent years towards having more efficient and effective banking products available. “That’s the key,” José concludes. “You need to use alternative data to benefit the customer. There are moral and ethical limits as to how you can use that data, but basically, at the end of the day alternative data, AI, and the data analytics revolution only makes sense if you are able to better serve the customer, and approach the customer with better value propositions. If I can understand you better, and I use the power of that understanding to help you be financially stronger, then that is something that you are going to welcome and embrace.”

We couldn’t agree with you more, Dr. Murillo!  Thanks so much for speaking with us today, and for sharing your insights with us.