Learning Objectives:
- Learn where fintech lenders have excelled to gain market share and how banks can leverage these learnings
- Understand the incredible advantages that remain with traditional lenders, and how to leverage them
- Gain insights on how data science holds the key to exploiting these advantages
- Discover the lessons learned from this collaborative pilot program between TNB and Lendio, and what comes next
Phillip (00:07):
All right. Thank you everyone for joining us today in this exciting discussion about small business lending. My name's Philip Toliver and I'm with Lendio. This is my friend and client, Ray Garcia from Texas National Bank. He'll be introducing the bank in just a moment. Maybe just to share a little bit about Lendio and background, who we are and what we do and how we came to work with Texas National Bank. So Lendio is the country's largest marketplace for small business lending. We've been doing this for 12 years. We work with 75 lenders on a nationwide basis, and we're funding on the order of a hundred to 120 million in loans per month. We're also very active in the tax credit space, and we've built out a lender SaaS offering, which is part of what we're discussing today. So one of the things that we've observed traditionally with our marketplace business is the difficulty with which traditional lenders, whether they're banks or credit unions, have in competing on even ground with fintechs. And one of the big advantages that fintechs NBFIs, if you will have had in this space is by using technology and data science to assess borrower credit risk, and in particular using transaction data. So that's a big part of what we're going to be talking about today. Our partnership with Texas National Bank began through PPP as a lot of FinTech and bank partnerships began over the last few years. And so while that original partnership was just around PPP and borrower acquisition, it's evolved substantially now into more about cross-selling and upselling the bank's existing depositors using some principles around data science, which we're going to go into in just a moment. So Ray, I'll hand it over to you and maybe you can talk to us a little bit about Texas National Bank, who you are and what you're seeking to achieve.
Ray Garcia (01:53):
Well, thank you Philip, and good afternoon everyone, and thank you to the American Banker and all the sponsors putting this great event together. So good afternoon everyone. My name is Ray Garcia and I'm the strategic growth Officer for Texas National Bank. Texas National Bank is a community bank. We're located in South Texas. We are a minority owned depository institution and a community development financial institution. The bank was established back in 1920 with a mission to provide excellent customer service, high quality banking products, and with an emphasis on relationship banking. Today, Texas National Bank has over a hundred years of serving the community of the Rio Grande Valley. So again, just really to put it in a map, we are right next to the border. We're 10 minutes away from the border with the US and Mexico. So our area, of course, is largely represented by minority communities and first and second generation immigrants that of course, chasing the American dream. Yet historically, our area has been significantly underserved and underrepresented when it comes to banking and access to the right financial services.
(03:02)
So I think at Texas National Bank, we've always taken pride in being a part of the solution. We, we've been guided by a steadfast commitment to really give back to our community and help our community. And in response, the community response has been great. The bank has experienced rapid growth over the years. For over the last 10 years alone, the bank has gone from about a hundred million to 700 million in assets. And although we offer a variety of banking products and services, small business lending has always been at the core of what we do. And for us, I think it's especially important because of what small businesses bring to our community. They help create jobs, they spare economic development, and they really serve as the economic engines for our local community. And in 2020, of course, during the pandemic and as a result of the execution of the paycheck protection program, right? PPP, the bank really started its digital journey. And it was interesting, I think for us in an effort to help out our small business owners and really help preserve jobs for our community, we understood that we needed to embrace technology and just do something a little differently than when we've done in the past. Of course, our partnership and really just leveraging this first beginning to leveraging digital capabilities really allowed us to facilitate the funding for these PPP loans. So for a community bank in South Texas, we were able to originate 15,000 loans, about 400 million worth of funding for small businesses. It was a big endeavor. Our team kind of rallied and came up and really came together to make that happen. But the reality is that without embracing digital banking and digital capabilities, it was an effort that couldn't have been met. And just to tell you a little bit about in terms of what we do traditionally as a small business owner. So annually, we fund, I would say on average about 75 million worth of small business loans to over 200 small business owners. And a lot of those, of course, that are minority business owners and in entrepreneurs or just all sorts of small businesses. So when we started on this digital banking journey, we set out some goals. And I don't know, there's a lot of community bankers or smaller institutions here in the room, but I think you'll relate to some of the goals that we put together. They centered around innovation. We wanted to make sure that, innovation was important to us. We wanted to make sure that we provided access, we wanted to make sure that we created value, and we wanted to make sure that we improved performance. So just to start with the innovation side, we really asked ourselves how can we foster responsible innovation? We saw what the FinTech space was doing, and we knew we needed to stay competitive and relevant strategically. It was going to be very, very important for us. The next part is about access. So as a community development financial institution, we're always looking to provide access to our community and especially our small business owners. And so our goal was how can we continue to expand access but do it in a safe and sound manner? See, my prior experience is actually as a bank regulator, and please don't hold that against me again in a room full of bankers. But that is my background. So obviously when we were looking at providing access, we wanted to make sure that we paired granting access with the right risk management capabilities. We knew that in order to have to continue to provide access and do it in a safe and sound manner, we needed to have a good corporate and risk governance structure to really support that endeavor. The other part that we looked at is expertise. We wanted to make sure that we had the right expertise to take on this action. So although we pride ourselves in our bank, we have a great team, lots of experience and expertise. The reality is that we had to kind of humble ourselves a little bit and ask, do we have the right expertise and the right resources to do something differently? And so for us, it was about identifying a partner that can bring in the right technology, that can really help us identify non-traditional lending alternatives. And so obviously finding the right expertise and the right resources was important. I think another big part of this, is when it comes to when we have these ambitious goals around innovation and growth, they often come with challenges as well. And so as we looked at our goals and we looked at some of the challenges that we were facing, it was important for us to also think through some of those challenges. And thank you for jumping onto the next slide. And I think in the challenges we had a couple of things that I think will also resonate with the room. The first one was technology. We all have legacy systems, technology stacks that can be very rigid at times and really lack the flexibility to do things a little differently. We also knew the cost was going to be a challenge. I think with cost, first of all, there's a high cost to developing and implementing new technologies, but then there's also a high administrative cost with just keeping up with all the applications and the administrative side of all these things. So cost was another important factor to us. At the same time, we also knew that we would be dealing in the small business side with very unique small business profiles. As you all know, a lot of our small businesses, they often deal with informal bookkeeping systems, and they might go to the most cost effective accountant there, which then it gets overcrowded. And as you all know, that might lead to untimely or inaccurate financial reporting. In addition, a lot of our small business owners operate in cash or they just rarely borrow. So credit history and evaluating credit worthiness was another big challenge. So I think for us, one of the things that we set out to do is, okay, how if we're going to take on this challenge, then we need to make sure that we find a trusted partner that can help us with this. And this trusted partner should bring the right technology, should bring the right expertise, should do it in a cost effective way, should help us analyze our application process, streamline the way we do things, especially when dealing with our unique borrower profiles. And this is why we decided to find a solution and a trusted partner that made sense for us. So Philip, do you want to tell us a little bit about the solution?
Phillip (09:38):
Yeah, absolutely. Thanks for that background, Ray. And I think for any of us in any industry growing seven X over a 10 year period would be viewed as phenomenal growth. So kudos to you and the rest of the team at Texas National Bank. And I also might add, Ray has been an incredible advocate for what we're doing, which is interesting coming from your background as a former regulator. I'm sure you're struggling to break out of that. Yeah, that DNA, that the OCC built up in you over those years. So let's talk a little bit about the solution. So what specifically is this? What does it do? How does it work? So firstly, I'll start with this premise and the premise is the following transaction data is a better indicator of a borrower's ability to repay than credit scores. So set that in the back of your mind as you listen to the remainder of the presentation. And why is that? Well, mostly because small businesses independently do not have a great track record of credit. The credit agencies tend to not know a lot about the historical ability of them to repay, and as a result, we use individual or owner credit as a surrogate to assess the credit worthiness of the business. The problem with that is there are a lot of very highly successful people who have independently in their own personal lives, great financial management who start businesses. Businesses of course with limited liability. And those businesses struggle to get off the ground. And it's very common as we know for small businesses to fail. But transaction data really holds the key to knowing how good these businesses are from a credit worthiness standpoint. So what do we actually do? How do the mechanics work? So firstly, we source bank transaction data hopefully, and that data can come from one of three places. First, it can come from an aggregator like Plaid or Yodlee or Finesity. Secondly, we can take uploaded bank statements and we can OCR optical character recognition and bring in that raw data. And the third is what we did with Texas National, is we bring that data in directly from a core, and it doesn't matter what core it is. This is, we're able to pull that data in and then run it through a transaction classifier. And so every transaction that a business is undertaken over a long period of time, it can be a year, it can be longer, is classified. This is a deposit that was revenue related. This is a deposit, this is an NSF, not just a bank charged NSF, but an intraday drop below a certain figure. It could be zero or it could be higher. And so we go through and we classify all those and then we tabulate them. Here's the statistics over a 30 or a 60 or a 90 day period. And those are then run against the bank's credit policy. And I use the term credit policy a little bit in air quotes. We don't typically think of credit policies quite in these terms, but we can say, here's the minimum number of revenue generating deposits, or here's our tolerance, if any, for suspicious transactions. And so that information is then used to create initially a set of DNQs, and then secondarily a set of risk tolerances associated with the credit worthiness of this business. What's most typical at this point is you assess annual gross revenue without any tax information necessarily at this point. You don't have to have it, but you can. And we are using the bank transaction data to assess annual gross revenue. And then so you can risk adjust that. We're going to prove 15% for a business with a FICO above a certain level, we'll approve 12% of AGR above a certain level, or you could do the same math using number of monthly deposits and so on. And in any case, what we've done now is using just transaction data, we've weeded out a huge swath of DGS. We've established the set of borrowers who we think are going to be eligible for this product, and now we've assessed what we're willing to offer them. And so at this standpoint, you can imagine Boca LLC has now been approved for $75,000 of a line of credit or a term loan product. So what do you do with this? So in the case where this is an existing depositor, you actually go and market directly to that potential borrower with a real offer. Now, you and I in our personal lives, I'm sure get these all the time, you get an offer for a HELOC. The one I'm most familiar with is getting periodically getting an offer from American Express. If you use your American Express a lot, inevitably you've gotten an offer over the last two years that says you've been pre-approved for a personal line of credit of 6% APR and $40,000 or whatever. This is precisely the same thing. It frankly doesn't exist in any real mature degree in small business lending, but it's been around the consumer for the last few years. And so imagine the uptake you get as a small business owner when rather than seeing a flier that says, Hey, come into the branch and apply for a $50,000 loan, or this product is eligible for up to 50, it's personalized to you, it's personalized to your business, and it says you have been pre-approved up to this amount, the uptake on that is much, much higher.To give you some statistics on it, direct mail is sub 1% response rate, right? Emails has been trending lower and lower over time. What we've seen with this product is that 20, and in some cases as high as 30% with some other lenders of small business depositors are pre-qualifying for the product using the framework that I just laid out. And then you go to them with this direct targeted outreach, and they're coming back at a rate of eight to 10% and eight to 10%. Those are all not just warm leads, but really hot leads that are coming back in through the application experience. There's another important part of this, which is that when you're targeting your existing depositors, they expect you to know already everything you already know about them through the deposit relationship. Michelle was just talking a moment ago about the importance of cutting down that friction in the borrower experience and reusing information the bank already knows nothing is more frustrating to you or I as consumers. And certainly the same would apply to us in our business lives of having to re-key information that a bank knows about us. Hey, what's your business name? What's your address? No, you already have that, right? And so a big part of this is that pre-qualified offer includes a unique URL with some authentication, and then once the borrower goes into that, the entire thing is pre-populated. The only things remaining are a few verification steps and it's some adjustment to the approved loan amount. So a big part of this is highly targeted, low cost marketing, low friction borrower experience, and much lower cost of underwriting. So if you look at the statistics in community lenders right now, you'll typically find there is virtually no difference in the underwriting expense for a $50,000 loan as a million dollar loan. And the main reason why community banks have largely pulled back from this market is because of the difficulty of cost. Effectively underwriting loans sub $250,000. With these tools in place, those costs drop precipitously and make for a much better borrower experience. I want to share with you all a few of the key learnings we've had as we roll this out. And by the way, we're going to save a few minutes for questions at the end. So it'd definitely encourage you to think of what, if you have questions, what those questions might be. Firstly is that SMB lending really is changing dramatically. That demand for a highly digitized low friction environment is not going away. It's accelerating and finding ways of delivering that to your deposit relationships as well as your lending relationships is exceedingly important. McKinsey did a study towards the latter end of 2020 talking about digital trends in the pandemic. They found that the propensity of an individual consumer in their personal lives to use a digitally native experience for more than 80% of their interactions, not going to a hundred but 80% that increased threefold in one year. There was 10 years of digital adoption in 2020. Now we have known this for a long time in our personal lives as consumers. What I want to leave you with today is far as this goes is that in small business, the requirements are a lot more like consumer than they are like corporate or commercial. The second is transaction data. This really is providing new opportunities. And as I referenced a moment ago, it really does change the way we look at risk from the standpoint of SMB borrowing. There's been extensive research over the last few years. There was a report done with Bank for International Settlements and Partnership with Lending Club and Funding Circle that established this as the importance of transaction data. And more recently researched that Lendio participated in using PPP data with NYU a stern professor there who talked about the importance not just for indicating propensity to repay and likelihood of default, but also about the importance from the standpoint of financial inclusion. So there is plenty of research about the importance of doing this and we're excited to have more case studies. Ray, maybe you could talk a little bit about your experience here and in particular the product cross-sell upsell, which is such an important topic right now given the banking mini crisis that we've had over the last few months.
Ray Garcia (19:37):
Yeah, absolutely. So I think for us, one of the great things about transactional data is that it helps us understand how these small business owners operate. You gather data on how they operate, and then at the same time, this allows you to really understand their business model a little bit better. And then after that you sort of shift the mentality from being another lender to more serving as a consultant for this particular small business. And so I think that adds a lot of value. I think the data helps to again understand when perhaps the business is going to be experiencing working capital issues, temporary cashflow imbalances, seasonality factors, or even when they're just ready to graduate into another type of loan that is more suitable for them. So I think on the one side, it gives you an opportunity to leverage additional lending opportunities. But I think the other beauty of this, and I think it's especially important in today's environment when liquidity is such a hot topic, transactional data it helps you improve the stickiness of those deposits. See, our customers are rewarded by using their deposit accounts because we can evaluate their transactions. So the use of transactions essentially allows us to, allows a customer to continue to benefit from products like this that are very convenient, ease of access, but through good deposit behavior. So I think from that standpoint, it really increases customer loyalty and just the stickiness of the deposits. And I think it's just another great tool in terms of sustainable low-cost funds moving forward. And I think just another tool in the arsenal, and I think I want to share a quick kind of personal story because I think one of the first loans that we funded was to a local snack, snack land. It's called a snack land drive through. Okay, again, keep in mind we're in the border. So the things that we as Hispanics enjoy are things like fruit. We love chips and salsa. I mean you name it, all these tasty, savory snacks. We're all about that. So we have a local business that we know to be very, very popular because again, I frequent this business pretty often and there's always a ton of traffic, so they're doing something well. However, and we maintained their deposit relationship for years, however, never for some reason we never had done any lending opportunities with them. And so when we evaluated the transactional data, one thing that I guess should have come into no surprise was that they qualified for this for our product. And I think it just validated obviously what we were doing with transactional data, but it also sort of helped identify these small business credit invisibles, if you want to call them that were already depositors of the bank. So of course we started a loan marketing campaign when we had our pre-selection offers and we sent out emails to our customers. And so you can imagine as a community bank that's never done a loan marketing campaign, we start getting calls, right? Hey, is this you first of all, right? And kudos to them because with fraud being what it is today, I think it was smart on their part, but we said, yes, it's us. It's a new product, a new solution that we're trying out. Please give it a try, you qualify. So that of course, she starts her customer experience and really enjoys it. Very frictionless, just kind of gets to the point. We fund her within 24 hours and then we called her for some feedback and said, Hey, what did you think about this new product? She's like, oh my God, it's great. Really enjoyed the experience. It was easy, convenient, accessible. I was said, great. I said, so how come in the past you never did any lending with us? And she said, well, we just enjoyed. I just like the convenience of not having to provide all that financial documentation. My accountant's always like a year behind and you all are requiring most current financial information. So it was just such a convenient way of doing it. And she's like, I've done loan like this before, but with a merchant provider. And so this is an opportunity for us to really tap into those opportunities with that particular customer. And again, I mentioned one of our goals was create value and really improve the customer experience and provide that convenience. And I think our ultimate goal has always been as a community bank, is to deepen relationships and really embrace that. And kind of funny story about this, and I did not know this at the time, but fast forward like two, three weeks later after we funded this loan, I'm having a conversation with my mother about her day and she's telling me she went and looked out for one of our Ts and she was very ill, but she's doing better now. She just had a lot going on at work and she starts telling me about how she owns a small local drive-through a snack land that does all these Mexican fruits and cubs and all this other good stuff. And I asked her, of course, what's the name of the business? And it happened to be exactly the business that we had just funded, so leave it to the Hispanics who have these big families. But I think it was just a great experience about how we can just deepen our relationships. It all started with just embracing digital technology. We evaluated the transactional activities. We started this loan marketing campaign against something we have never done before as a bank and got a great response and really just deepen our relationship with customers. And now we have customers asking, Hey, you know, am I going to continue to qualify for this product? I said, well, if you maintain your deposit relationship with us, of course your business has done well, and we'll be able to capture that in the data. So just another example of how really we've used this product so far.
Phillip (25:22):
That's really tremendous. And I think it's easy to get wrapped up in the stats, the 20% offer rate and the eight to 10% acceptance rates on the offer. But being able to personalize it really means something to that individual borrower. And then when you step back and look at it, what behavior that drives for the financial institution, particularly given the challenges we've seen in the market recently, that cross-sell and upsell is exceedingly important. Absolutely. So that's terrific to hear. We hope we've left you with a little bit of a taste of what the solution is, but more importantly about the opportunity that it drives for community lenders. SMB depositors are an extremely important part of the community of our engine for economic growth in this country. And hopefully this is creating a new opportunity for traditional lenders to compete on even grounds with FinTech and alt lenders who have really been driving most of the growth in the market since the financial crisis. So with that, we'd like to open it up to questions. We've got about four minutes and we'd love to hear some questions from you all. This gentleman here, and we've got some microphones, so just hold your hand up when you would like, and we'll get the mic over it so everyone can hear the question. Please go ahead, sir.
Audience Member 1 (26:36):
How do you make sure that the transactional data that you have about a small business is a complete set with Texas National Bank? What if you have partial data and there is no full disclosure about the rest of the transactional data could be maintain at another bank, for example?
Phillip (26:55):
So it's a fantastic question and the approach that we've taken to date basically says this one is we can detect intra bank transfers and determine whether or not the bank account that we're looking at is the primary account. So that's one premise. The second is if we have enough deposit related, a revenue related deposit, so we can detect whether it's revenue or some type of interbank transfer, we believe that this is an operating account and then we can treat it as such. So a typical rule of thumb with some of the lending policies we've dealt with is like 10 revenue related deposits in a month. They need to be varying in value, you need to see them from different sources. But the data science can more or less detect that. I think I saw another hand over this way, this gentleman.
Audience Member 2 (27:58):
Because of close proximity that you build with your clients and depending those relationships, I'm looking at all the location of your branches. How scalable do you think is the business model beyond the border, beyond those few cities?
Ray Garcia (28:17):
Yeah, so I think one of the, I guess if you look at where this can go, the idea is that you can then set up connections to evaluate even other transactions, other bank accounts that are not necessarily at our institution. And I think that's where the scalable part really starts. And I think again, another tool to maybe drive deposit growth to your bank. So the idea is that if you can connect to another institution and if that's where they hold their primary operating account, then that allows you to underwrite that particular account and then with the idea of hopefully gaining retention through that as well.
Phillip (28:51):
Yeah, so just to hone in on that point, at the onset I was talking about sort of three sources for these transactions. One is aggregated, the second is OCR on top of bank statements. And then the third is from the lender's core. In those first two scenarios, that applies to borrowers coming in the door for the first time. And then the second part I'd referenced there, whether it's through Lendio marketplace or some other marketplace or banking as a service type of model, there's lots of opportunities at acquiring other customers, but you have to be able to respond quickly and assess quickly. And we're finding speed to funding and frictionless borrower application is becoming so important in decisioning. So whether depending you've got sort of an independent decision about where you acquire the borrower and how you do the underwriting, but if you don't have a way of doing the underwriting quickly and cost effectively, then it's a challenge to meet the growth objectives you're talking about.
Audience Member 3 (29:54):
Hi, thanks. There's a fantastic case study. You talked a lot about some of the revenue and cost side of it. What about the risk side? So you mentioned a BIS study. How does using transactional data affect the way that you price credit or have to hold back capital? What's the default rate impact, things like that? Has that changed since you've rolled this out?
Ray Garcia (30:16):
Yeah, I guess I can can still ahead so far. Obviously with any other product, we take a very measured approach on how we do things. Again, I think responsible innovation sort of dictates that you have to do it that way and also do it in a sort of an agile approach, right? Change as you see it appropriate and adjust and iterate. I'll say that since we start our product, we have had no defaults. The loans have performed very well. I think one important part of this is obviously the revenue estimate, but also the liquidity reserves. And so if I can really see who has liquidity reserves and keep up with at least this loan for a prolonged period of time based on that history of liquidity reserves with our bank. So far we've had really, really good response. We do have risk limits of course, and tight risk limits to start, but the idea is that eventually we'll over time build more history, adjust pricing as we start seeing, getting more default data at that point. But so far so good. Let's hope it stays that way.
Phillip (31:16):
And just a time check. We'll have one more question I think is that, okay,
Audience Member 4 (31:22):
So in the previous presentation we heard a little bit about how FICO score is not the only determining factor about credit, and you touched on credit policies a little bit. So how has that evolved over time in your case where you see just having a FICO score is not a true representation of the consumer's entire credit, and then there's so many more aspects of it. I'd love to hear your response to that.
Phillip (31:47):
Yeah, I mean I think there's a couple aspects to that question, and I'd love your take on this as well, Ray. One is that particularly in small business, is extremely tricky to figure out the credit of the small business versus the credit of the principal owners. And so most lenders have taken the approach of, well, you borrow in the name of the principal owners of the business, and that's the credit that we're relying on. That tells you almost nothing about the business itself. And so we view transaction data as being a better lead indicator of credit, but there's also sort of the track record, if you will, of the owners and what their mentality is about credit. And so we don't think transaction data replaces. Everything you can get from FICO, but we find that it's a really strong indicator of the business as a standalone proposition. Yeah.
Ray Garcia (32:47):
I guess I'll just add that I think the model, in addition to the transactional data, it looks for other things like ownership information. It looks for any BSA AML KYC items as well. Times in business, certificate of good standing, all these things that are important that give you some insight into the character of the principal owners. And I'll say that, take the example of our snack land a borrower, the initial underwriting criteria that we put out almost kind of punished her a little bit. So we realized that we had to change that a little bit. So we had a trade line requirement. Well, she's never borrowed just because she hasn't had an to borrow. She does very well. The data shows that she does very well. So I think if we would apply traditional sort of standards for underwriting that particular business, she wouldn't have qualified. So we've been tweaking along the way. We still have a lot to learn, but I think it's been a great experience just learning and evaluating how the impact of transactional data. Yeah.
Phillip (33:53):
Thank you all for the great questions and for the time today. We really appreciate it. Thank you so much.