Machine learning in underwriting is all the rage today and has been now for a couple of years. But the first online lender to really hang their hat on this technology is Upstart. It is core to their business and it is why they applied for and received a no action letter from the CFPB back in 2017.
Our latest guest on the Lend Academy Podcast is Dave Girouard, the CEO and co-founder of Upstart. He was last on the show back in 2014, which was their first year of operation, so obviously a lot has changed since then. While Upstart has remained focused on personal loans they have also worked on developing partnerships with banks.
In this podcast you will learn:
How 2018 was for Upstart.
Dave’s view of being a monoline lending business.
An update on their partnership with BankMobile and Customers Bank.
How their bank partnership model works.
The size of their loan book today.
What is different at Upstart that results in a younger demographic on their loans.
How their machine learning model has evolved over time.
Their approach to automation in underwriting.
The impact receiving the CFPB no action letter has had on their business.
The ongoing reporting they have to do with the CFPB.
Why they decided to open a second office in Columbus, Ohio.
Dave’s take on the competitive landscape in personal loans today.
What 2019 holds for Upstart.
This episode of the Lend Academy Podcast is sponsored by Experian’s Clarity Services, the leading sub-prime consumer credit reporting agency.
[expand title=”Click to Read Podcast Transcription (Full Text Version) Below”]
PODCAST TRANSCRIPTION SESSION NO. 182 / DAVE GIROUARD
Welcome to the Lend Academy Podcast, Episode No. 182. This is your host, Peter Renton, Founder of Lend Academy and Co-Founder of LendIt Fintech.
Today’s podcast is sponsored by Experian’s Clarity Services. They are the leading sub-prime consumer credit reporting agency providing innovative risk management solutions to address the full consumer credit life cycle. Clarity leverages the combined power of the largest and most comprehensive alternative credit data source with traditional bureau data to provide a more complete view of the consumers’ credit behavior. Clarity is committed to providing products that address rapidly changing market conditions. You can learn more by visiting clarityservices.com/solutions.
Peter Renton: Today on the show, I’m delighted to welcome Dave Girouard, he is CEO and Co-Founder of Upstart. Now Upstart have been around for several years, we actually had Dave on the show back in 2014, just over four years ago and I wanted to get him back on because Upstart has been making some really impressive progress.
They’re now a profitable company, they have taken on several bank partners which we talk about in some depth. Their bank partnership model is somewhat unique and we talk about the No-Action letter with the CFPB, we talk about their approach to underwriting and automated underwriting and Dave gives us his perspective on the competitive landscape in the personal loan market and we talk about what’s in store for 2019. It was a fascinating episode, hope you enjoy the show.
Welcome back to the podcast, Dave!
Dave Girouard: Great to be here, Peter.
Peter: So people know about Upstart pretty well, you’ve been around for some time and you’ve been on the show, it was four years ago, but you’ve been on the show before, maybe we get this started by just telling us a little bit about your year. I mean, we’re recording this in late December and just give us a little bit about…how has 2018 been for Upstart?
Dave: Yeah, 2018 has been a great year for us, a lot of focus on the fundamentals of the business and you know, for us that kind of always starts with the performance of the credit and with all of the sort of unique approach we take to credit, we’ve definitely moved the ball forward pretty significantly on that front and have been super happy with our credit performance. It means unit economics, you know, so we’re actually generating new profits from loans and that’s also moved forward quite quickly and then, of course, after that growth.
So we’ve doubled in size roughly during 2018, a little bit more than that; we’ve also crossed into being a profitable company. So Q3 of 2018 was our first fully profitable quarter from a gap and a cash perspective and we’re hopeful of the 4th quarter, which is not quite over yet, will be the same. So we’ve kind of reached critical mass, but the growth level’s great and we’re kind of launching a lot of new initiatives, working with bank partners, we’re going to open a second office for R&D so it’s just a lot of things like that going on, but, generally, for us it’s just trying to shore up and strengthen the business. I feel like 2018 was a great year for that.
Peter: Right, yeah, that’s great. So I’m curious, you stuck to your knitting here, your unsecured consumer loans…that’s what you’ve been doing from day one and you’ve seen many others in the industry launch new products in different verticals and you’ve decided to stick with the unsecured consumer loans, at least for now anyway. So how do you view some of your competitors doing different things? Do you feel like the monoline business that you have is still the way to go?
Dave: Yeah, well I think nobody necessarily wants to be a monoline business forever, but I guess our viewpoint has always been you want to expand from a position of strength where really you have incredibly successful monoline that provides the foundation or the launch pad to do something bigger. And we’ve been very….you know, I just say conservative in that decision making process because first of all, it’s a very small market share and in unsecured personal there’s so much room to grow without moving.
So, obviously, as we just think about what’s the best path forward, it’s obviously easier to add more features and capabilities. We had a trap of borrowers in the product wherein….but, having said that…and I should also add, I mean, there are a lot of companies that have announced products. There’s not a huge track record of companies moving beyond their original product and seeing great success. I mean, certainly, SoFi did super well starting in student loans, launching their personal loans later, that’s probably a notable success, but, you know, for the most part, despite a lot of efforts to expand, it’s not been easy for a lot of the players in the market to do that, so we’re cautious about it.
Because we have a model of partnering with banks, we’re also potentially going…we’ll have products that we work with bank partners on that may or may not end up being in our consumer business; we’re in the testing process with credit cards with a bank partner. So we do see our platform as an origination platform that will certainly work across various consumer products, some may or may not be part of our consumer offering, but we are already, you know, into a second product and would expect to probably be into a third product during 2019 as well. But, I think, it’s fair to say, we’ve stuck to our knitting and really wanted to make sure we have our core product working really well.
Peter: Right, right, fair enough. You just mentioned your bank partners and there’s only been one publicly announced deal and when we chatted a few weeks back, you said you actually have signed five different banks but only one is actually publicly available, as far as announcements go. So why don’t you just tell us a little bit about the BankMobile deal and you announced the deal, I think it was about a year ago, and why don’t you give us an update on that particular partnership.
Dave: Yeah, so that’s been a great partnership for us to help launch the business. They’ve been a great partner, they actually kind of have two distinct banks. BankMobile is a consumer mobile-only brand and then they also have the kind of parent bank, if you will, called Customers Bank and they actually both have programs with Upstart. So it’s been a great, you know, founding bank partner for us. We are beginning to refer people who come to Upstart to a loan to our bank partners so in addition to seeing the Upstart loan, Upstart-branded loan, that people have seen for several years now, they’re beginning to see loans from bank partners like Customers Bank or BankMobile.
That’s a model we’re really excited about and have worked on with this particular partner so they’ve been great, and, of course, getting the first one off the ground, it wasn’t without, you know, a few stumbles here and there. But, generally, it’s been a good partnership and we value them very highly and you just have to also give some kudos to anyone who’s willing to be first, you know, it’s a brave new world how fintechs and banks are working together.
In the early days, you want the banks that like being first because, frankly, most are going to say, you know, who else is doing this so getting that first few going really is the big challenge. As I said, we actually have signed five other banks, in addition to the Customers Bank or BankMobile partnership, so we expect to be rolling them out during 2019, but we’re taking it very cautiously. It is, of course…banks are very conservative institutions, rightly so, and so we absolutely want to make sure these partnerships go right.
Peter: Right, so tell us…I want to delve in a little bit into the….you just talked about the BankMobile branded offering, tell us a little bit more about that and how it works and the offering that you’re providing for your customers. I mean, are you sending people who just come to Upstart through a different funnel or offering or how does it work?
Dave: Well, let’s just say piece number one of our bank partner model is really a white labeled application that includes the entire origination process with all the embedded machine learning models around fraud and credit scoring, automation, etc. in a white label fashion, but where the bank can really control and determine their own credit policy so what type of products they want to offer in terms of fees, rates, durations, what their credit box looks like. So it’s really intended to be the banks’ policy, the banks’ product; we’re really just providing a lot of the plumbing and the pricing engine and the fraud engine, etc. within that. That’s the sort of white label product.
Part Two really is we can help those banks acquire customers. We have a model that tends to be leaning heavily toward millennial age borrowers, it’s a great demographic. We’ve done a lot of acquisition through various channels to upstart.com and now we can begin to actually direct some of that traffic to our bank partners who, I think, value that demographic and it just sort of completes the equation for us. So the white label product really is….you know, essentially, you think of it as their channels, their distribution, but additionally, through our kind of referral platform we can bring customers to them.
Peter: So you’re bringing customers to them. So you’ve got BankMobile, do they have a consumer loan offering that people going their site, or are you just sort of referring people that come to Upstart?
Dave: It’s really both. Yes, you can go to BankMobile and get a loan from BankMobile. It’s an Upstart-powered loan and you see our logo at the bottom of the page, it’s a very familiar experience if you know us, but it’s their channel, it’s their customer, etc., but we can also refer people from our website to bank partners and that’s increasing. That’s probably….you know, as of last month, somewhere between 5 and 10% of our loans that originated via upstart.com actually went to bank partners, rather than to the original Upstart-branded loan that we’ve had for four years now.
Peter: Right, right, got it. So let’s talk about your loan book, I know you’ve already mentioned that you’re pretty happy with the way losses have been trending, maybe you could give us a little more color on that and what’s the size of your loan book these days?
Dave: We’ve been originating since May 2014 so I guess that puts about 4.5 years, it’s all unsecured personal loans, all US-based business and it’s about $2.8 billion in loans thus far. You know, they’re, unsurprisingly, probably about $12,000 on average or something in that range so, you know, it’s maybe $230,000 in total in the loans. We’ve been nationally 3-year and 5-year, we have a small test in 7-year.
The majority of our book is actually 5-year, so that’s kind of, roughly speaking, what it looks like. Our borrowers tend to be younger than you see in a lot of other platforms, the average borrower being probably 28 or 29 years old, in that range, and I think our credit model and our platform itself sort of have an advantage in that demographic which tends to just…..it doesn’t necessarily mean we target them per se, but we have a competitive advantage due the nature of the underwriting model that attracts that demographic.
Peter: Okay, we’ll get into that in a little bit. So just on that, do you think that’s the reason that you’re attracting those people? Is that really because they get a better deal, or others find it difficult to underwrite, I’m just trying to get a…what is different at Upstart that seems to skew to a younger demographic than say at Lending Club or Prosper or SoFi?
Dave: Sure, well we have a lot of information we use in our model and what we’re known for from a very early days was using education-related data, highest degree of education obtained at schools somebody went to, the area of study, things about their work experience, the industry they’re in, the company they work for and that’s kind of a small part. But, on a relative basis, the credit data which is absolutely part of our model…in fact, we use full files from two different credit bureaus in our core model, but on a relative basis, it weighs less, it’s less strong in the model than it would be for almost any other lender.
And that kind of means we’re un-correlated to how others underwrite oftentimes and for a certain demographic, you know, well educated, younger, with a thinner credit file, very often they’ll find Upstart understands them better, can underwrite them better and therefore, will offer them a better rate. And that’s the heart of our model, it’s just knowing more about the borrower, not necessarily just this demographic, not just about education, but having a more complete 360-degree view of the borrower allows us, in our view, to generate some advantages in underwriting
Peter: Okay, so then…I know you guys were pioneers when it comes to AI and machine learning. When you’re creating your model….the model that you have in existence today, I presume that’s all built around that kind of technology, I guess, so tell us how it’s changed because one thing about machine learning is it’s best when it works with huge amounts of data and when you’re getting started it’s hard to have huge amounts of data, so tell us how that’s evolved over time.
Dave: Yeah, that’s absolutely true. You know, when you begin to use these technologies in the earlier stages, I mean, at the beginning you have no data, you’re really using third party proxy data to build Version 1 of your model and, of course, any lender has to do that to get started. Beyond that, you begin to generate performance data and, of course, you need a significant amount of it and seasoned loan, etc. I would say, generally, that we feel like our model really found its footing in terms of just really reaching critical mass of data and really learning super quickly in the last year.
You know, we have thousands of payment transactions happening everyday now and the model is learning super quickly. That’s sort of the core of the machine learning model is every time a payment transaction happens, whether that’s an on time payment a delinquency, a pre-payment, etc., the model just gets a tiny bit smarter. Now as that data accrues, the model is getting smarter, you can begin to introduce more sophisticated techniques that really weren’t feasible, you know, when you had half the number or tenth the number of transactions to model or to train on.
So what we’ve really seen over time is we went from a fairly simple model to one that’s more sophisticated and over time we just see opportunities to create even a tighter fit in the model by upgrading various aspects of it. It’s not a singular model, it’s really an ensemble of different models that come together to create one unified model. There just continues to be more opportunities.
Just to give you a single example, I mean, we’re working on right now…every lender has a timing curve they associate with their models for, your know, for their 3-year product, for their 5-year product, where there’s an opportunity to actually build a completely custom timing curve for every single loan and create a lot more accuracy in doing that. That’s a fairly far reaching concept, the idea of having a custom timing curve per loan, but we have fairly strong evidence that it will create a much more accurate model and these are the kind of things that get our data science team excited.
Peter: Interesting, really interesting. So then how are now approaching automation because you’ve built this great model and you were formerly with Google and many others from Upstart have come from Google with obviously…very strong on the technology side. Automation is something that I know is….no one’s really been able to crack it in the consumer space and I’m curious about what your approach is to automating your underwriting and what percentage of your loans have a human involved or what percentage is it completely automated?
Dave: Sure, so on the sort of rate decisioning, the offer of credit or the decline, that has been always, since the beginning, a 100% automated so there is no human intervention involved in deciding whether we can offer somebody a loan and what size loan, etc., or whether there is an adverse action notice and we need to decline them so that process is always been 100% automated. The interesting part comes….you know, when someone’s approved for a loan and they decide to move forward, what is the sort of steps that have to happen between then and when they can receive the funds and this is where…..we’re not, of course, alone in this, every lender wants to make that as smooth as possible.
In our first couple of years, our rule is very simple, we verify 100% of everything and that means we have to make sure of the identity of the person is true and accurate, we have to make sure their income statement is accurate; in our case, the educational background they stated is accurate. And so we quickly began to bring in automated ways to do that behind the scenes, but they never had full coverage, so, invariably, almost everybody would have to upload two or three documents. We had a phone call with every single borrower which was just a verification call as a sort of belts and suspenders check on everything.
We did this for the first…oh, I don’t know how many, but two to three years of origination and then at some point, the degree of automation now became such that we’re very close to people that had almost everything entirely verified before they even took a step. It was done behind the scenes in a few seconds so we decided to build models that would say, okay, if we know these nine things about this borrower, what’s the likelihood that that tenth thing is true and if it is or isn’t true, how much does it matter? So those two questions which are sort of intuitively interesting questions, you can build really sophisticated models around.
So at the end of 2016, about exactly two years ago, we began to have a model that would say, can we instantly approve this person right now? So think of it as you request your rate, you like this rate so you say you want to move forward, you give us your social security number, you connect your bank account and this will say, great, you’re approved, accept right now. So that sort of in-session instant approval for us didn’t exist until December 2016, so exactly two years ago and to a very small test for very small loans, we began to AB test that and the obvious thing is we knew they would convert better. What wasn’t as clear was how would the credit perform so we really started it at a small scale and then extremely slowly rolled it out as the models have improved.
If you fast forward to today, about 2/3 of our loans, not quite, between 60 and 65% of our loans this month are approved in-session immediately without the borrower doing any sort of upload or phone call or anything and that’s up from 0% two years ago. The good thing about that is those loans convert about twice as much because basically when you can make it super easy, you don’t give them a lot of reasons to shop elsewhere. It’s a very positive effect. Lo and behold, the credit actually performs equally good as well.
So we’ve been able to do that while at the same time keeping fraud rates at super low rates. Our fraud rates are in the range of 20 basis points, if you include it all sort of what we call first payment defaults. So very low fraud rates, very high conversion rates and, of course, it does save money on the cost of originating a loan. So all of those are obviously great things, it’s just part of being a data science and technology-centric platform.
Peter: So the early AB tests that you did, has the credit always performed roughly equivalently or did you have to tweak some things initially?
Dave: I mean, sometimes there’s very subtle small differences, nothing that caused us to go backwards in any sense, but if we would see things that are in the order of like five basis points or something, we would learn something or the model would see a trend that would allow it to….you know, generally speaking, we would observe, upgrade the model, observe some more and then sort of basically loosen the reins a bit over time. So I think if we had just turned it on full blast the first month, frankly, would have been irresponsible and God knows what would have happened…
Peter: (laughs) Right.
Dave: …so we very purposely were putting the scientific method to practice to figure out how to make sure these models were sound and that could…of course, when we started these were loans that we were balance sheeting ourselves so we, of course, weren’t putting our loan buyers at risk in the early days. That’s generally the way we roll is we want to test things quickly, we like to be able to do some of that on our own dime and then bring it along as an advance for our loan buyers and then eventually, to our bank partners, which hopefully benefit from all this downstream.
Peter: So then when do you think…or do you have a goal to be 100% automated at some point?
Dave: No, we don’t anticipate being 100% automated, there’ll always be…first of all, there are just always bad guys out there and so some fraction of your traffic is just always going to be, generally, people trying to fool you one way or another. So where it ends is not totally clear, right, we feel pretty confident even by the end of 2019 we should be at about 75%. I think we intuitively believe, just based on the experience of seeing what sort of fraud rates really are out there that, you know, it probably won’t get higher than 90%, but we think we have a very clear shot to 75% this year and then how much it takes to close the rest of that gap is not entirely clear to us. There is still for us a long way to go.
Peter: Right, right, Okay, I want to just talk briefly about the No-Action letter that the CFPB issued. It’s over a year now, back in 2017, and this was the first time they’ve done a No-Action letter and I know that you guys have always been pushing the envelope on using alternative data and different kinds of things we’ve already talked about, so from a commercial perspective, what impact has that had on your business, the fact that you have received this No-Action letter?
Dave: Well, I think, obviously, for somebody like us whose sort of main reason for being, if you will, is to further the notion of machine learning, credit use of alternative data, it’s really what’s special and unique about Upstart, making sure you’re doing it within the law, is very central to our existence. Our view, generally, is the types of technologies we’re building will be used in a decade across all flavors of lending, not just unsecured personal but cards and mortgages and business lending and real estate lending, etc. So it’s just fairly obvious to us that ultimately computers are going to do this better than the human can and there’s enormous winds across almost any type of credit business in the use of these technologies.
Having said that, you know, it’s really hard for regulators to keep up with that and we wanted to really be out there first and say, at least in the U.S., the pre-eminent regulator with respect to consumer protection, the CFPB was really there. We kind of opened the kimono and said, what we’re doing is number one, very good for consumers. We’re making credit, more affordable credit available to more people, we’re actually lowering rates and having higher approval rates than a traditional credit model would for any demographic you can name.
And so we felt very good about what we’re doing and, fortunately, found in the CFPB, who has a new name, by the way, now (laughs), but, in any case we found them to be receptive to the idea. You know, they also wanted to learn and see where the new world of lending was going and that was 2-1/2, maybe almost 3 years ago now that we’ve began with them.
So for us it really says, look, this stuff is not scary or wrong or illegal; it actually improves access to consumer credit and the regulators have looked carefully at it and we’ve come to an agreement with the regulators about how to build these models properly and how to monitor them and report on them to make sure they don’t end up in violation of fair lending laws.
So it was really important for that reason, it gave comfort to our loan buyers, to people that would buy bonds in our securitizations and, ultimately, to our bank partners who were all going to be looking towards these types of models. So we’d like to think we did the industry some good because we certainly won’t be the only type of lending platform using these types of models and breaking some ground, paving a bit of road with regulators, I think, is hopefully helpful for a lot of others.
Peter: So then do you have ongoing contact with the CFPB? Is there like a regular reporting, monthly/quarterly reporting type thing that you’re doing?
Dave: Yeah, it’s a little more machine to machine, if you will, in the sense of our system has to generate a report that is dropped to CFPB quarterly. So we do that for reporting requirement, we have to run certain tests, a waterfall of test to ensure that our model’s compliant. So it’s sort of a standardized process at this point, if you will, so we’re friendly with them.
I wouldn’t say we’re sort of actively trying to dig to the next level of anything, but we have an agreement with them that obligates Upstart to report to them regularly. We’re also beginning to do that on behalf of bank partners so if we’re helping them originate credit using our system, they want to have the same sort of coverage so that’s another aspect to what we’re doing.
Peter: Right, right, got it. So I read recently you opened up a new office in Columbus, Ohio. Obviously, there’s many Bay-area companies doing this these days, going out to less expensive locales. I’m curious, why did you choose Columbus and what’s the office actually going to be doing?
Dave: Sure, let me first just say, as a University of Michigan graduate, it’s about the last place I would have started (Peter laughs) and I think when we announced it my Michigan friends thought it must have been some sort of cruel joke or something or that I was taken hostage or something of that nature, so, yeah, I had resisted this a long time.
When I was leading quite a large team at Google, I had teams in many parts of the country, as well as, frankly, the world and I kind of knew the tax you pay in having engineering teams, software development teams, etc. in various parts of the world trying to coordinate. It’s not trivial, so I really avoided it for a long time here and finally realized, you know, growing to seize the opportunity in front of us here in the Bay Area is just really going to be difficult.
It’s just so much competition for so few software engineers and data scientists, so we finally sort of came to the conclusion several months ago that it was time. We really needed to look beyond the Bay Area, there’s so much talent out there that for many good reasons don’t want to live in the Bay Area. So we began to really think about where there was sort of a mismatch of available talent, but without the Facebooks and the Googles and the Ubers hoovering up every really strong data scientists or engineers because that was really what we wanted to avoid, just heading out to another market where the competition may not be the scale of Silicon Valley, but felt the same and Columbus really fit the bill for us.
I mean, enormous amount of talent there, it’s actually the second largest city in the Midwest after Chicago, a huge and talented university there, as much as I hate to admit that (Peter laughs) and just a place where we can have data science and engineering as well as operations, there is a huge wealth of financial services talent in that area for lots of reasons, also banks there, a lot of our bank partners we’re talking to and working with are in Ohio.
Probably the clincher for us was I got a guy named Grant Schneider who was one of our first and most talented data scientist, actually has spent the prior ten years of his life in Columbus and did want to move back to the Midwest with his wife and his young family.
To me, that was actually a formula for success at Google, not that you just pick the city and try to whip it into an office, but that you actually had a leader who was ready to go transfer the culture, make it a success and we have that in Grant. So when we put it all together, for us, Columbus made a lot of sense so we’re going to be up and running there in just a few weeks and have our first handful of employees there even before end of January. So it’s a good start and our hope is it’s going to help us build this team faster and pursue a lot of the opportunities we see in front of us.
Peter: Right, you probably have a bunch of Ohio State fans, I imagine, that’ll be coming to join your firm, that won’t bother you? (laughs)
Dave: You know, sometimes you just have to swallow the pill, but, in any case, I’ve made my peace. I did promise that if I ever go to a game in the horseshoe, I will be wearing Maize and Blue so that I will certainly hold true to.
Peter: (laughs) Okay, okay, we’re running out of time, just a couple of more things before we go. I’d like to get your take on the competitive landscape today when you’re talking about personal loans and….obviously, we’ve seen banks enter the space, even Marcus in the space; we’ve also seen some of the online lenders fall by the wayside, so when you’re looking at…what is your perspective now on the competition that we have between…obviously, you’re working with bank partners, but you’re obviously aware that banks are competing as well, so how do you view the space, who has an edge and what’s your perspective?
Dave: Sure, so I guess my overarching statement would be….you know, it’s a conclusion we came to a couple of years ago…was that banks will, for as long as any of us will be around, will continue to dominate consumer lending in the United States. There will always be non-bank parts of the market, but generally speaking, there are advantages with respect to cost of funding, permanence of funding, etc. are so significant that for us the choice was either number one, become a bank, or number two, partner with banks and we chose the latter. So that’s my first statement, we are big believers that banks can and will be around.
There may be a lot of changes in market share, some banks may not change with the times and not do as well as others, there may be completely new banks formed out of nothing and certainly some are trying to do that and that can all be true. But, generally speaking, it will be banks who dominate consumer lending.
Number two, you know, with respect to more competition, there are certainly no shortage of it and we know a whole bunch of banks are going to launch consumer lending programs in 2019. I generally think that if you have the same credit model, the same distribution, roughly the same funnel as everybody else, it’s going to be pretty damn hard for anybody, including any bank, to do anything meaningful in this market. It’s become obvious to us that you need to have an edge, you need to have something about your business that is completely distinct that you can build on and accrue value to over time.
You know, for us it’s automation in the credit model, in the machine learning, and, you know, others have different approaches and some are really experts in capital markets for example, but generally speaking, if you don’t have an edge, you can put a program out there, but it just may be not very impactful to you. So our view, generally, is …I don’t worry about, you know, the number of competitors out there; what I worry about is anybody getting ahead of us in the areas that we consider vitally important?
For us, again, the automation, the modeling, and we don’t see that so we feel good about that position, but I think everybody in this industry, bank or non-bank, have to took at themselves and say, you know, what’s the podium we’re going to stand on, what’s going to make us very different from the 50 others in this industry? If you don’t have a good answer to that question, you know, I think it’s only a matter of time before you got to give up, but everybody has to be honest with themselves and say, what can we do that’s unique and different? We feel good about that.
Peter: Right, good points, good points. So last question then, what are you looking forward to in 2019? What does 2019 hold for Upstart?
Dave: Well, certainly, the area that we haven’t done nearly as well as I would have liked is building our team. I mean, our headcount plans for 2018, we frankly came woefully short of …which is it partly the inspiration, if not all the inspiration for the Columbus, Ohio office.
So, first of all, we really have to be better at building and growing the team faster and we’re really focused on that. There’s just so many things, we have such a backlog of opportunities, of things we want to do if we only had another 20 or 30 data scientists or another 30 or 40 software engineers, etc. so that’s really the number one focus for us and I feel really good about that one.
So I think making our first bank partners successful is enormously important to us and we’re pretty much going to put everything we have to behind making sure those that take the risk with us, that are early players in this round of partnerships, that they don’t burn their fingers, that they aren’t made to look bad. So that’s going to be hugely important for us in 2019 and getting other products out there. We have been, as you rightly noted, very monoline-focused, but there’s certainly more we can and should do and I think 2019 will be a big year for that.
Peter: Okay, we’ll have to leave to there on that note. I appreciate your coming on the show today, Dave.
Dave: Peter, it was great to speak with you.
Peter: Okay, see you.
Peter: You know, the personal loan market, I would say, is more competitive today than it’s ever been. Dave and I were just chatting after I hit the record button to stop recording that, he said that if banks are going to come into the space and they create an online funnel for their loans and if it’s like 10 or 15% more friction than one of the top platforms, they’re going to get very little business because there are so many good options out there now for people.
I think what the online lenders have taught people is that we want efficient funnels, we want things to happen quickly and we want it to be easy and if it’s not that way …these days you’ve got very little chance of becoming a successful operation. I think, along with that, you need to do all the other things right, as far as customer acquisition and many other things, it is not a simple business and it is very competitive and you need to have something unique, as Dave said, about what differentiates you from everybody else.
Anyway on that note, I will sign off. I very much appreciate your listening and I’ll catch you next time. Bye.
Today’s episode was sponsored by Experian’s Clarity Services. Clarity’s suite of FCRA-regulated reports and predictive scores yield significant insight into a consumer’s financial behavior throughout the alternative financial services industry. Clarity delivers data-driven risk management solutions that address prospecting, credit evaluation, fraud detection, portfolio management and collections. You can learn more by visiting clarityservices.com/solutions.[/expand]
You can subscribe to the Lend Academy Podcast via iTunes or Stitcher. To listen to this podcast episode there is an audio player directly below or you can download the MP3 file here.
Peter Renton is the chairman and co-founder of Fintech Nexus, the world’s first and largest digital media and events company focused on fintech. Peter has been writing about fintech since 2010 and he is the author and creator of the Fintech One-on-One Podcast, the first and longest-running fintech interview series. Peter has been interviewed by the Wall Street Journal, Bloomberg, The New York Times, CNBC, CNN, Fortune, NPR, Fox Business News, the Financial Times, and dozens of other publications.