Data can help create credit for the underbanked

Millions of Americans are living their lives largely outside the banking system.

Lisa Fischer, chief growth and lending officer at Mission Lane, joins us to discuss how data can be used to provide more financial access to the underbanked.

A few takeaways from the conversation:

  • Alternative measures, like spending habits and bill payment history, may offer insight into their true credit risk.
  • Artificial intelligence and machine learning can process entire data sets to yield a better understanding of individual borrowers.
  • The ability to analyze large sets of individual-level data lines up with the industry’s emphasis on personalizing the customer experience.

Subscribe to the BAI Banking Strategies podcast:

Apple Podcasts     Spotify    Google Podcasts    Amazon Music

Sign up for the free BAI Banking Strategies newsletter and get industry insights delivered to your inbox.

By clicking the Subscribe button, you acknowledge that you have read our Privacy Policy and Terms of Use and agree to be bound by them.

Below is a full transcript of my interview with Lisa Fischer.

Increasing attention is being directed toward Americans who have little to no contact with banks or other financial services – as a group, they are commonly called the “unbanked” or the “underbanked.” Our guest on the podcast this week to discuss how data can help banking institutions reach the underbanked is Lisa Fischer, chief growth and lending officer at Mission Lane. Lisa, welcome to the BAI Banking Strategies podcast.

Thank you for having me. I’m really excited for the conversation.

Lisa, for our listeners who may be less familiar with Mission Lane, can you tell us more about the company and where it fits into the financial services landscape? Maybe include a little bit about where the name comes from, assuming it’s not your street address?

Well, our mission is to help underbanked customers find financial freedom. And we work in the FinTech industry, and that’s our objective. And thus, you get the name, our “Mission,” but it takes a journey, which you get “Lane,” because you have to go over time and go on a journey.

When it comes to the underbanked in the U.S., I know we’re talking about millions of people across all demographics. Could you put some numbers to this group to give us a better idea of the size and the diversity of this opportunity that you’re working with?

Yeah. There’s over 100 million customers, people in this population. There are people that are just starting out in credit, they’ve had some sort of personal event, they’ve struggled with credit or lending for some time. It’s a variety of different types of people, but it’s a very large part of our population.

We’re hearing a fair bit these days about alternative credit scoring and how people whose financial lives can’t be neatly reduced to a three-digit number from a traditional credit bureau. How do you at Mission Lane think about the traditional credit score, and how do you use it in your work, if you use it at all?

We do use it. It’s the foundation for the most depth of information for banking and lending. But it’s just one piece. And what we’re learning and what we’ve grown to understand is it’s not enough to understand the nuances of a person. It’s not enough, that one line or that three-digit number and the information associated with it, is not enough to know everything about a customer so that you can understand their needs and deliver great products for them.

A common industry term applied to the unbanked or the underbanked seeking credit is that they are “thin file,” or they are “credit invisible.” I want to ask you, is this too limiting a shorthand to be using to rank or to categorize these applicants? They have bills, they have other obligations that they’re paying. Maybe they don’t have much of a borrowing record, but they do have a financial presence.

These terms are much too limiting, and they’re just the tip of the iceberg. You need to look at the customer holistically. You mentioned some of the things. They’re checking transactions, their bills, rent, demographics about these customers, so you can understand the full financial landscape of each of the customers.

Mission Lane is a purpose-driven organization, and I would guess that one of those purposes is trying to get to “Yes” for a credit applicant. Is that how you look at it as well? And if so, you’re still a business, so how do you think about your risks and think about managing those risks?

Yeah, the goal is absolutely to get to “Yes” for all customers. But that’s to get to “Yes” for a product that meets the needs of the customer, and the customer can afford, because it is in both of our interests for those things to come together. And that’s how you best manage your risks outside of fraudulent activity. But is to get a product with customers that is really going to meet their needs and that they can afford. And therefore, they want to use the product, and therefore, it’s going to serve a great purpose in getting them to financial freedom. But that’s the goal. And therefore, by staying true to that goal, you manage your risks in that way.

Let’s bring the conversation around to the role that data can play in making it easier for the underbanked to access credit, while at the same time protecting those who are offering the credit. What kinds of data are we talking about that would-be lenders don’t have that they should have.

If I gave you every detail, you can’t go into it going every detailed data element, because it’s how they work together and how you can understand a population substantially more detailed. You need to look beyond the traditional data sources and look at they’re checking data information on us, data information so that you can extend the relationship in their transactions there. You need to continue to look holistically at the customer, and how do you create a picture fully of the customer? But if you try to say, “Oh, it’s these three elements,” then you’re looking much too narrow. It’s how the information works together to define who the customer is.

Banking institutions have vast repositories of data and other information about their customers, but we often hear stories about how challenging it is to leverage that data for business purposes. For the underbanked, we’re talking about adding in more data from other sources. Tell us more about the process of harnessing data and getting past those obstacles that banks are dealing with now. How do you do that?

Not to date myself, but I’ve been in this industry a really long time, and about 10 or 15 years ago, we actually didn’t have the ability to do that. And because the processing time to get through the information was substantially too long, but with the advances in technology, advances with AI and machine learning, you are able to process so much more information in such a shorter period of time, you have the ability to add much more information about a customer holistically so that you can understand them better. Back when I started my career in this area, you must sample data. There was no possibility of using whole data sets, and you would have to pick and choose what you think should go into it. Now you have the ability with machine learning and AI to look across all of the data and look at it holistically because of the power of those tools.

I’ve read reports from researchers that spend a lot of time thinking about these sorts of things, and they say that while AI and machine learning can help remove barriers to credit access, as you’ve been talking about, the technology also has the potential to reinforce those barriers. How do you ensure that bringing in and analyzing more and more varied data, how do you know that that will lead to fairer decision making and outcomes? Because models are only as good as their inputs, right?

Absolutely. And it is something we think about regularly, and the goal is to have the inputs be as broad and as unbiased as possible. Therefore, we do a thing we refer to as “universe testing,” testing across the entire universe to remove any potential biases that goes into the model development sample. And then on the production side, you must have rigorous monitoring controls in your production environment. You have to be looking at what is going on with the different variables and the transformations as you are using it in production and know that you understand the choices the model is making and how it is affecting various different populations.

Earlier we discussed the breadth of Americans who fall into the underbanked category, including low-income families across all races and ethnicities, young adults, older people, recent immigrants. It’s a long list. How do you sort out the categories of the underbanked in terms of the data needed to be able to reach each one of these discrete segments? And is there such a thing as personalization in the data as they’re analyzed and applied?

I don’t know if I would refer to it as personalization, but I would refer to it as individuals. We don’t look at people in categories because the way you can actually really get to a better place with providing customers with products that will meet their needs and they can afford is to understand the individual. Even in some of the categories you refer to, there’s a wide range of economic statuses and how people interact with their products. And if you stay at the segmentation level, you do not understand those nuances, therefore you won’t be as refined at meeting their needs. I would call it getting to the individual level. That might be what you’re referring to as personalization, but it’s really understanding the individuals, not the category, because they flux throughout each of the different categories based on their financial needs.

If I had to ask you for a single piece or a single category of impactful data that’s not part of the traditional credit bureau analysis, this is data that so powerfully predicts the ability and willingness of an underbanked borrower to make their loan payments, and in doing so, still best reflects the risk to lenders. What would you say that data might be?

That implies that it’s simple and that there is one thing, and it is an extremely complex. We’ve talked over this time about the magnitude of data, but it’s also how those data elements interact with each other to help you understand who the individual is. But if I were to say one tool, it’s testing. And it’s broad testing so that you can gain the information on how people are going to perform with you on the various types of products you are offering so that you can have a complete view of who the customer is.

Lisa, I’d like to wrap up our conversation by putting some faces to the data. I’ve spent some time on the Mission Lane blog on your website. It’s an interesting, really readable mix of basic financial education and real-life testimonials by people who have overcome financial challenges and there’s always a moral or there’s always a lesson from their personal story that others can learn from. I’m curious about how the blog fits into your mission and what kind of value you’re getting from it.

The blog just puts color around our mission and what we’re trying to accomplish for customers. Mission Lane is here to help customers and meet them where they’re at and deliver them to financial freedom. They represent everybody in the world that has gone through struggles or is trying to get through this, and Mission Lane wants to meet them where they’re at and help deliver them to a place where they feel like they have financial freedom.

We all find ourselves in different places financially, so it’s a meaningful mission to try to provide those underserved by the industry with more credit options. Lisa Fischer, chief growth and lending officer at Mission Lane, we appreciate you joining us on the BAI Banking Strategies podcast.

Thank you so much for having me. This was a great time.

Terry Badger is the managing editor at BAI.