Upstart Holdings, Inc. (UPST) Earnings Call Transcript & Summary

We're just about to get started. Welcome to the Upstart presentation at the Goldman Sachs Communacopia and Technology Conference. I have the privilege of introducing Upstart's Founder and CEO, David Girouard. My name is Mike Ng, and I cover Upstart and fintech here at Goldman. We have about 40 minutes for today's fireside chat inclusive of about 10 minutes of investor Q&A. So if you have a question towards the end, feel free to raise your hand. We'll get a mic run over to you and you can ask your question.

First, Dave, thank you so much for making the time and coming out to speak with us. Upstart has made a tremendous amount of progress in unsecured personal loans and is expanding into auto lending following the acquisition of Prodigy a couple of years ago. Over the next few years, what does success at Upstart look like for you? And can you talk about some of the milestones or hurdles you're looking to achieve that vision.

Yes, sure. Thanks, Mike. Thanks, everybody, for joining. I co-founded the company with the idea of modernizing credit origination, with the idea of improving people's access to credit, but also helping lenders offer credit using a system that's much more modern. I started the company after leaving Google and kind of seeing how that world evolved during my time there. So over time, I think we are looking to transform credit and credit origination by improving who gets approved, at what price, the application of AI to a 5,000-year-old industry. And to do that right, I think we have to do it at a very large scale across most flavors of credit and certainly in the U.S. across most major areas of credit in a way that has a consumer brand. We imagine building a company that will resonate with consumers as a place they can come to know that they will see the very best offer of credit with the very best experience, and at the same time attracting hundreds, if not thousands of lenders and investors who want to participate in the market as well. So that is the essence of what we're doing, applying AI to this industry in a way that's going to dramatically improve credit access for the American consumer, which is enormously important also to American businesses, but also just helping how lenders of the world, particularly banks, credit unions, et cetera, actually originate credit.

Great. That's a fantastic overview. And as you mentioned, much of the Upstart story relates to the technology asset that you've built and the training data that you've amassed. Could you talk a little bit more about how Upstart's underwriting model compares to traditional FICO-based models and other AI lending platforms? What makes Upstart's platform unique and difficult to replicate?

Sure. I mean lending and how both banks or lenders work has been somewhat consistent for a long time. They tend to be rules based. Credit scores are something that came out in 1989. They matter. Your credit score largely can determine whether you're approved for a loan or not. A whole bunch of other simple rules, your level of income, how much debt you have, et cetera. So this what we would just call a rules-based model have been the centerpiece of lending for many, many decades. And even though computerized, meaning computers are now used to make these decisions, that sort of history still dominates lending. And suffice to say, with a lot more data and sophisticated models that can separate that -- interpret that data, we can do a lot better. We can do better at identifying who should we offered a loan, at what price. What is the true risk associated with an applicant. And you don't have to be perfect. There's no such thing as a perfect model. You just have to be significantly better. And that's the heart of what we're building. It's something that looks at not just 10 or 15 or 20, but hundreds or even more than 1,000 variables that is trained every day with 10,000 or 20,000 or 30,000 new repayment data points that can much more accurately reflect the true risk associated with an applicant. So it's a pretty dramatic change to an old industry, and it's not a trivial thing to build or to do. It's an industry that is heavily regulated. It is one that is very conservative by nature. And so it won't change overnight. But we believe the opportunity is vast to do very good things for the American consumer as well as American business, but also for lenders. And we have a great start on it. We've been at this for about 10 years. There's in the range of $30 billion of loans originated through our platform, our models are getting better and better by the day, and we're beginning to expand into new flavors of credit. So we think this is something that can run for a long, long time and will run for a long, long time in terms of the addressable market, the opportunity to improve flavors of credit. And there's kind of like ultimately no losers in this type of market opportunity. It is a straightforward, significant improvement in efficiency and accuracy of lending. And that advantage accrues to both the borrower and the lender. So it's just an enormous change. It's not a trivial change. There's really not a lot of others, we think, in the industry even trying to do what we're trying to do. So we are, for better or worse, in a unique position of building this and very excited about what we've done in the first 10 years of the company's history, but even more excited about what we can do next.

That's great. Maybe you can talk a little bit about some of the shortfalls of traditional credit score-based lending models and how Upstart is trying to help improve the underwriting process. And then along those lines, Upstart just announced their participation in the More Than Fair lending coalition. Could you help talk a little bit about that and what you hope to achieve there?

Yes. So More Than Fair, if you want to look at it when you get the chance, morethanfair.com was a coalition announced today. It's essentially a bunch of like-minded different types of consumer advocate groups, financial technology companies, whole different coalition people who have interest in improving how credit works. And as I said, somebody who -- I spent almost a decade at Google before starting the company, the impact of technology in a lot of areas of our lives has been dramatic. If you think about the advent of the smartphone and how it's changed the amount of information, the access you have, if you think about renewable energy, electric vehicles, there's just so many aspects of our life that have just -- compared to, never mind 30 years ago, even 10 years ago, are dramatically different. But if you think about credit access, which is like who has access to loans, how does this work? It largely has not changed in 3 decades. And our view is there's no fundamental reason that has to be true. Credit is so important to the economy and to each of our lives, and how much we pay for credit is just like a fundamental ingredient to our life, yet it hasn't changed. So More Than Fair is a notion that just being fair in credit is not sufficient. We actually need and can make dramatic improvements in who gets access to affordable credit. It's kind of a building block of the American way of life. So it's enormously important. Huge fast progress is possible, though if we have to be honest with ourselves, it's been stagnant and stagnant for several decades. And this like is the heart of why we do what we do.

Great. Could you talk a little bit about the revenue model? How you tend to price your product?

Yes. We've decided -- so first of all, we're not interested in becoming a bank. And there's nothing wrong with banks. Banks are enormously important. They are most fundamental partners in our business, but to be kind of a bank holding company, regulated by the FDIC and/or the OCC or others, it would be a very different business. It definitely is required to be a low-risk, very predictable business. That's the nature of the bank. That's what you have to be. Because you have FDIC insured deposits, you sort of need to operate in that sort of manner. That would fly entirely in the face of what we're trying to do, which is to really, again, change the nature of credit origination, which kind of means we need to be at the forefront of risk, building new models, testing new models, pushing the boundaries, because our goal generally is, we want to bring as many Americans as possible into the sort of domain of the credit system that exists today, which is generally dominated by banks, credit unions, et cetera. We think that's a good thing. We want to bring more people in. But it means -- we don't aim to be a bank ourselves. And over time, as I said, our belief generally is, a lot of the loans in our platform should be funded, held by banks makes perfect sense. Capital markets are another source of funding and really important to us. And we believe there's a role for a place where consumers can reliably depend on the best offer with the best process.

Yes. We'll touch on the funding aspect in a second, but let's just shift gears and talk a little bit about the macroeconomic environment. Could you talk a little bit about how the current environment is impacting your business? What your outlook for the third quarter assumes? And how does the risk of recession or deteriorating consumer health impact demand for Upstart's products?

Yes, I think our view on the macro, what's going on, what has gone on, and what may go on is nuance and a little different from what you might hear in this general purpose business press, if you will. Our view generally is that kind of blue collar America has been through a very significant cycle over the course of the pandemic. And it was a cycle not in the obvious ways that you would see it. It was a cycle brought by enormous government stimulus, propping up consumer balance sheets, reduction in consumer spending in the early stages of the pandemic. That meant, from a financial perspective, super-healthy consumers. That was sort of unceremoniously ripped away in 2021. And that to a lot of America would look like instant unemployment, if you will. Accustomed to receiving a lot of money and not spending a lot, savings rates going up a lot, and then that reverting very quickly. That sort of notion, which is not precedented, I don't think in any of our history, was important and really impactful to the people we serve. So today, we sit at a place where if you listen, a CEO of a large bank might say, "Hey, the consumer looks okay. We don't really know prospects of a recession, don't see signs of it." I think it's because they're living in what I just call the white collar America world and haven't seen or affected that. So our view is there are -- there's been an event that's gone on. We think it's mostly behind us for the -- it is behind us. And then there's a potential recession that everyone is talking about, that in our view, if it happens, is likely what we would term a white collar recession, just a different type of phenomenon. And I think if you look historically at the great financials crisis, you looked at sort of early '90s recession, they all behave differently. The GFC was largely a kind of white collar recession, if you will. So that's our view. There's 2 events that might be a little bit related, but are distinct and separate in time, and unclear whether that future will come. But the bottom line for us is, in the nature of our marketplace, we became funding constrained. There was a lot of concern in either lenders or investors about the state of the economy, what might happen. And I think we're at a place where now the prices are higher in our marketplace. That constrains volume. But our credit is performing exceptionally well. There's still, I think, generally in the market, wariness about the state of the consumer, where things are going. But our business continues to get stronger. The new products continue to get better. We continue to sign up more lenders on the platform, signing up more car dealerships that are putting our software out into car dealerships, which is an important emerging part of our business. And I think we feel very good that we're going to be on the other side of this economic period, and we'll be in a very fast growth mode as we've been over our long history. So for me, in the long history of the company and where we're going, I feel like 2022 is certainly a speed bump. We don't underestimate it. There's certainly parts of our business that we want to improve and make better, and we're very aware of those, mostly on the funding side of our business. But in the long run, we're a business that has generated cash since we've been public, and very, very optimistic about where we're going. And as I said in the long run, we'll view it as a speed bump and nothing more.

Great. That's a good segue to talk about the funding model and those funding constraints that did hurt origination volumes and revenue this year. So for the uninitiated, could you just talk a little bit about Upstart's funding model today? What's are the primary channels between forward flow and securitizations?

Yes. So our business model looks like a marketplace. We're effectively borrowers on one side, lenders and investors supplying the capital on the other side. And on that capital side, you might just -- we sometimes describe it as the supply chain of money. The lenders, the banks and the credit unions typically will originate and hold maybe 25-ish percent of the loans. The remainder of the loans are originated by banks and then sold forward to some credit institution. And that is sort of by design. Banks and credit unions will always have a low risk tolerance. They will look for things that are primarily less risky and they're willing to accept a lower return because of that. And then the rest of the market has a much more broad and diverse appetite for risk and the required return. So that's the nature of our funding model. It has been largely to date, and to our detriment, to be frank, at will, which basically means every participator in that marketplace can choose what they want to do this month, how much risk they want to take, how much appetite they have, how much volume they want to do, and they could decide something different for the next month. And of course, from a pure marketplace perspective that's all well and good. What I think we've learned and we will improve in our businesses, the market can move in one direction in a very correlated fashion and in ways that sometimes aren't, in our view, entirely just logical, they're just emotional. They're just fear-based or what have you. So the market can react. And that's what we've seen. We've seen, in 2022, concerns of where this is going. So that's the long and the short of it. This is an area that we're going to make improvements in our model and we would like to. And we have said publicly in our last earnings call, we will make moves to having more committed capital on our platform over time. We don't want all of it to be committed. We actually think it's really important to have a diversity of types of funds et cetera. But we would like to have more of it be committed over long periods of time. What we have demonstrated is we've delivered exceptional returns through a cycle to investors and to our bank and credit union partners. We've done nothing but deliver pristine credit to them. So we feel very good about that. But of course, we want to make sure that a lot of that funding is locked in and committed. So that's an area that we are committed to improving on. But over time, like I think, to have the most impact, be the most helpful to both sides of our market, this is a model that we're happy with.

Great. That's very helpful. And could you just talk a little bit more about the progress in reaching out to those long-term committed capital partners and what that more permanent capital structure might look like? And when can we expect to see milestones or markers over the next couple of years?

I mean, it's an evolution. It's not something we're trying to like -- it's not part of us getting into regrowth. It is certainly a goal of ours very quickly to get into growth mode again. And some of that is subject to the economy, of course. But generally speaking, our goal is to be into a long, multiyear growth period, and we aim to do that. It's probably getting into that. This sort of funding of committed capital is not the vehicle to get there. It's more of a longer term. It's more about what happens the next time, 2, 3, 4, 5 years down the road when there's turns in the economy, because we think the way we're back in the growth mode is through funding and relationships that we already have today. We don't need new relationships to go do that. So what I would generally say is over a period of a year, 18 months, we'd like to see probably some double-digit fraction of our capital flowing through our platform, being committed. That is, by definition, going to be somewhat costlier, more costly, meaning the loans will be priced at a bit of a premium because there's some sort of consideration there that makes that capital more permanent. So there's different structures those could come under. We're evaluating a lot of different choices. We're not in a rush to do this. As I said, it's not a sort of near-term urgent thing. It's more of a long-term transitional approach to how we want some of our funding to be. But again, we don't even want the -- I would not say we want the majority of funding to be committed, I don't think that's ideally healthy for the system. We want a much more market-based approach. The other thing is we begin to expand from personal lending into auto lending, which is clearly our next big vector. I think that is, by its own nature, much less volatile. It's a secured loan. It's one that banks' lenders have a lot of history and familiarity with. So that's just another way we'll move towards what we view as less volatility in the platform.

That's great. So let's shift gears and talk about some of the loan products. The U.S. unsecured personal lending market was about $130 billion in 2021. I think Upstart's share of that was about 10%, really impressive. Could you just describe your views on unsecured personal lending? What are your expectations for Upstart as it relates to market share gains and competition?

Yes. So unsecured personal lending, an installment loan you would get from your bank or from anywhere, is a market that is not much more than 10 years old. If you went back further than a decade, maybe a bit more, you might be able to get that loan from your bank as a favor, if you will. It was never an important category. And there are several reasons. One is, historically, for a bank, it's not very profitable to offer this product. And it's hard to underwrite an unsecured loan in to a consumer in many ways, like the idea of making a $10,000 loan to somebody and all you can do is report it to the credit bureau if they don't pay it back. It's sort of like flies in the face of someone who might have been in banking for the last 50 years. So it's a relatively new product, largely invented, if you will, not invented entirely, but really brought to the forefront by fintechs, not us, I mean, some of the earlier sort of innovators in that area. But having said that, consumers love this product because it's efficient, it's easier to get, it's simple to understand. And as many companies, including Upstart, have pushed on that, it's a product that you can be approved for virtually instantly. You could have the money the next day. Compare that to a HELOC, which might take many weeks or months to actually secure through a process that can be really difficult. Now if you're borrowing $15,000 for a redo of your kitchen, a HELOC could make sense, but you might be willing to pay a little bit more if you could have that loan instantly with no process. And so I guess my point is, unsecured personal lending has a little bit of a dislocation that consumers absolutely love it. It's like the duct tape of, I can pay for a wedding, I can pay for my new kitchen, I can pay off my credit card, I could pay for that coding course I want to take, anything you want. So it's kind of super nice. Banks have always still historically been like, "Wow, can we make any money on these loans. They're unsecured, is it risky, et cetera. And so I think it's a category that the consumers will drag and platforms like ours will drag forward and will continue to grow. It has been right up until COVID, the fastest-growing flavor of credit in the U.S., because it's very popular. And now banks are figuring out, working with people like fintechs, how to do it right, how to do it efficiently, how to do it profitably. So I think it's a category that's going to keep taking market share. BNPL is itself a flavor of unsecured personal lending. It's just a very short-duration product and it's taking market share from credit cards at that end. So I think in many ways it's just a simple product, but now that it's technically possible with modern technology and the risk can be managed appropriately, I would expect it will continue to keep growing. And we also are confident we'll continue to gather market share. So it's itself a very, very large opportunity.

Yes. I mean, I've certainly been impressed by the market share gains in U.S. personal lending that Upstart has been able to achieve. But can you talk about how you balance that with differentiation and underwriting, right? Somebody might say, at some point, you become the market and it's harder to become or stay differentiated. So maybe you can talk a little bit about that, or whether you're actually expanding the borrower box in the market with borrowers that may not be served.

Yes. I think the grain of market is most people probably don't even know these products are available. So despite the fact that it's gotten quite large, as you mentioned, it's still very small compared to other flavors of lending. It's 1/8th or 1/7th the size of auto lending. It's not even a rounding there compared to mortgages. So it's still modest, but it is growing fast and I think just consumer awareness of, hey, this might be the best way to pay for my wedding as opposed to taking 4 credit cards or whatever machinations they might do. So anyway, there is still, even after all this time, low consumer awareness of the product. But I think that we believe the advantages of having sophisticated AI-based models just continue to accrue over time. One of the things we measure and we set high really aggressive goals for is model accuracy. We want to make a certain amount of improvements to model accuracy every quarter that we measure very carefully. And we use these things called OKRs, objectives and key results, the thing sort of Google invented many, many years ago, but very specific to saying we want to put upgrades to our model such that they will be more accurate. When they're more accurate, they tend to approve more people at lower rates, all else being equal. So that sort of like technology push forward means we can offer better rates to more parts of the U.S. for more reasons. And this is just a one product category, but it can go elsewhere. And so I mean, I don't worry about market saturation. We're going to move into new categories. At some point, we may move beyond the U.S. So addressable market to us is never, frankly, a concern. And I think the long run for us is we would like our technology to be used to support origination of as much credit as possible across the U.S., which is trillions of dollars every year.

I think historically, when you have worked with your partner banks, they've provided some parameters around underwriting standards or FICO scores that they would be willing to originate. Some of those partner banks have removed those bumpers and are completely reliant on the Upstart credit underwriting platform. Could you talk a little bit more about that? What are the benefits to you? What are the benefits to the partner bank when something like that happens?

Yes. What that really represents is the transition from a dependence on hard rules that a human can just write down and understand, to having confidence that the model will do the right thing without hard rules, but using a lot of information. And of course, that sounds a little scary. People like to say, okay, here's who we approve and here's who we don't. I could describe it in 6 words or 10 words or what have you. And my credit committee -- and this is a lot of history why a rules-based decision is just -- it's certainly in the world of lending how most of the world is operated. A model-based approach really says, look, we don't have to worry about exactly what their credit score is. Of course, there's no question, credit scores are useful. People with super high credit scores are definitely more likely to be approved than people with super low credit scores. But we don't need to put a hard rule in there that says, it's at 650 or 680 or 620. We actually can let the model figure out and we can make a business-based input, which is like maybe we don't want loss rates above 4% a year, right? Or maybe we don't want APRs above 20%, or we can set a bunch of parameters that are really business inputs. And the model will do the right thing. And that's hard at first blush for someone who's been in the industry forever to just accept. So of course, we don't force this on them. We say, look, if you want to have a minimum credit score of 680, so be it, you get to decide that. If you want to have a maximum debt to income of x, you decide that. So as you're trying to win over the industry, we don't want to like force down this idea that you have to just trust the model, it's going to do all the right things. But we let them put as many guardrails as they want. And over time, they can begin to take those off and begin to build trust in the model. So that's the idea. There's just no doubt in our mind, given what we know and where we are, that you can do much better by letting the software do its thing and figure out the nuances of the price and who to approve, not to prove, et cetera.

These last couple of quarters, as you mentioned earlier, Upstart has faced some funding constraints. And in an effort to maximize revenues, margins, profits, Upstart has taken price, right, and the take rates have gotten better. Could you just talk a little bit about this decision, some of the feedback you've gotten from your lenders and also end consumer borrowers?

Yes. So look, I mean, this is a constrained environment. For any lender out there, prices are higher than they used to be. Approval rates are lower than they used to be. So there's a bit of just the basic macro input to the realities of our business or any business that's involved in lending. Go get a car loan at a dealership, look at mortgage rates, whatever, take your pick, rates are up. And rates being up in our business come from the return demand that's either from a lender or from an investor, the perceived risk the model sees out there. And then also our own take rate, which is how much we're taking. So in reality, when rates are moving up, volumes are moving down, and we're naturally going to be in more of a defensive mode. So our take rates are -- we fortunately have control over our take rates, and we can raise our take rates and effectively continue to be a fluid platform with no risk, but able to sort of navigate through the turbulence that you see in industries like this, where rates are moving really quickly. So I view that as a positive. I view it as a feature of the business is that our take rates can move up when rates are moving up. And that means the offers are not as good for consumers, that means as many consumers aren't being approved. And certainly, as we move beyond this, we would expect to see rates come down, assumptions of risk and the environment come down, and our take rates come down. So if you looked at our economics, our contribution to margins have been super high. And that's just that. I view that as a feature. Again, we can increase our take rates when necessary. But certainly, we don't aim to keep them that high. We want to drive them lower. We long-term want to approve as many people at the best rates we can with our partners. But it's a bit of a flexing that this is a way to navigate the environment of 2022.

So let's shift gears and talk a little bit about the growth opportunity. You guys have certainly been investing in auto lending. I think you've facilitated $400 million of originations year-to-date. What are some of the most important things that you've learned in pursuit of this category to date? And what are some of the key priorities that you're focusing on as you invest across both re-fi and purchase?

Yes. I think we have a sense that auto could be our largest business within a few years and maybe by a significant amount because it is a very large industry. Of course, most people that own cars also have car loans or leases. It is a very inefficient and dysfunctional industry in many ways. Who's paying what on a loan is a function of many odd things about this industry. So there's just a lot of fixing to be done. I think almost any of you would say, the process of buying a car certainly, and financing it is not a pleasant one. So there's a lot of fixing to be done. And we have what I would describe as extremely good early product market fit between the software -- the company we bought, Prodigy, 1.5 years ago, that's really getting pushed into dealerships. We're pretty quickly improving how the process of selling and buying cars works. Getting our loans in there is at the next phase of it. And it really comes down to having better loan offers to more people with instant approval, so they don't need to go home and get a pay stub or come back next day, et cetera. And that sort of improvement, though it's at the very early stage, is pretty dramatic. And in fact, in many ways, we think in our personal loan business, we've had a proprietary product, but we're kind of out there in a nonproprietary world to win over consumers, meaning we don't necessarily control all where the consumers are and how they find us. We're out in the wild. As this business, auto business moves forward, we have our software in the dealerships. So that's an exceptionally valuable place to be, valuable real estate to have to be where the loans are happening. So the auto lending market is in one way to put it, it is the largest BNPL market in the world where you go out and you buy something and you get financing at the place you're making the purchase. And we have what we think is just an incredible opportunity there. So that's why we're so bullish on that market. This 2022 environment has slowed it down. Rates are going up. Dealerships don't have a lot of cars to sell. So it's not exactly a favorable environment for us to be emerging with the new product in. But having said that, if you just look at the chart of adding dealerships to our platform, it's incredibly rapid growth. So we're very bullish on it.

That's great. Could you talk a little bit more about how your go-to-market for auto differs relative to personal loans? You mentioned that with personal loans, you have kind of this nonproprietary go-to-market with auto. You obviously have the software and the dealership. Maybe you can expand on that a little bit more? What's the addressable opportunity here? Is it mostly independent dealerships, OEMs, new, used auto purchase or re-fi, kind of...

Yes. I think over the long haul, we'll see 2 principal points of aggregation, and maybe some additional, but the most principal ones being Upstart itself, meaning consumer hears about us somewhere. That's how our personal loan business has largely been built. Also, the auto re-fi business today is largely being built that way. And then another point of aggregation really is in these dealerships. And I think, we just see continuation to invest in those can take us a long, long way with the products that we have. It's good at upstart.com to better like cross-sell and bring people in who might have had a personal loan. They might want to refinance a car loan, or someday they may get a mortgage from us, which is another product we're working on. So there's a lot of obvious value to that presence. But I think being in the dealerships, where such an important financial transaction takes place, we just see as enormously important over time. So I would say, we do aim to be a brand, a brand that is not a bank per se, but is more of an Internet adviser-type brand. You might think of Google like, Amazon like, but where you know you can get the best rate and the best process. And I think it will spot into the dealership, where you can see that an Upstart powered dealership is going to be a better experience for you. And I think if that's the brand we can build and sort of accrue value to over time, we'll do it through the experience we can deliver, which again, in lending, the good news is, it's really simple, like best rate, best process. There's not that much else to it. And we think it's going to be a shopped market forever. When you're looking for a mortgage or a car loan or any type of loan, credit card, you're going to look around, you're going to check rates, you're going to do comparisons. And I think that's a market that we can do really well in over time.

Great. I do want to leave some time for investor questions. So if you have a question, just raise your hand and we'll get a mic over to you. We have one over here. But while the mic is coming over, maybe I'll just sneak one in. Could you just talk a little bit about how you're balancing growth versus profitability? Are there certain margin levels that you're optimizing for where are the key areas of investment?

Yes. I mean, the simple way to think of like the structure of priorities is, number one is always accuracy of the credit model. This is something that can disrupt our efforts to try to always make that more accurate. Second would really be unit economics, meaning like we're not into like doing a lot of activity, making a lot of loans where we aren't making some sort of gross profit on that loan. And then third is growth. So whatever is left over goes to growth. Or maybe just put suddenly differently, we've been profitable in generating cash since and before we went public, and we like that. I think it's a strong statement about the model. But we aren't certainly looking to maximize profits. The growth opportunity is so vast that I would say if we could just dial the things perfectly, and you can't always dial them perfectly, we would be marginally profitable and pouring everything else into investment in products and into growth. And that would be a model we'd be really happy with to run with for a very long time.

Great. Thanks for the question.

Thanks, Dave. I'm just wondering, in terms of what kind of KPIs or data points we're looking at to see that the model, the AI is effective even kind of in a tougher credit environment? Because it's always been very good during the kind of boom period for credit, but as we move to a tougher period now, I guess there's some question marks about that and anything you can give to give us conviction about the resilience of the AI into this sort of environment, please?

Yes. The question, if you didn't quite get it, is like what are the right KPIs to measure quality of credit, particularly in an environment like we see? This is a really hard question because credit is incredibly complicated, particularly trying to compare credit across products or companies or what have you. There are just many, many ways to measure it and none of them are perfect. So I will say, for us, what we believe is the most important measure is performance of credit relative to expectations of the lender at time of origination. So if this loan was supposed to return 5% ROA, is it returning 5% ROA or is it returning above or below that. That is the essence of what matters most is performance according to expectations at time of origination. Now of course, the macro is going to affect that. Our models do not and cannot predict what happens in the economy in the next 5 years. So we have to make some hopefully conservative assumptions about that. But we will certainly -- if the economy goes much worse than we would have expected, you're going to underperform. But generally speaking, that's the heart of the model is performance according to expectations. Hopefully, when times are reasonably good, you have enough buffer in there to absorb some deterioration in the economy. And I think we've seen that. We have not had lenders and/or investors lose money in our loans at all in our history. So the returns have ebbed and flowed a bit, particularly in the last year, as we've sort of talked about. But generally speaking, our returns have been extraordinary over the cycle, and we feel really good about that. But that is, in our view, the thing we care about the most and the one that we are very transparent with our investors and with our lenders.

Great. So in closing, could you talk -- oh, we have a question over there behind you.

Just on the transition to a more committed funding base. I mean, what feedback are you getting from your investor base? What do you need to show them, or what are some of the hurdles for them to move to become a more committed partner? And the second part of that, as you transition there, how do you envision your balance sheet becoming a larger part of your funding base. Historically, I do remember that's more for testing and sort of like a pilot phase, but I think there has been some change over the past 2 quarters. I want to see how much balance sheet retention risk are we expecting until you complete the transition into seeing more committed funding base?

Sure. So on the first part of your question, committed capital partners, we're basically looking for partners that want to invest through the cycle, generally looking to hold to maturity. And the trade-off really is they're getting some guarantee of access to yield. And in return, they have some committed capital. So that's sort of a structure that we haven't pursued to date. And that's the short version of it. We're not talking about the investors that are commonly on our platform today. Many of them are types that they might be buying bonds from our platform or something, but they're not the central investors on our platform. You can think of sovereign wealth funds, insurance companies, something with long cycles and need for yield over long periods of time. So that's the basic trade-off. It's commitment of yield and commitment of capital. Now can there be some incentive? We would expect that committed capital will come at a somewhat higher cost, which means the prices of any particular loan is a little bit higher. Or maybe put another way, the price is a little higher during the good times for it to be there at a lower price and then not as good economic times, but that's the trade-off that makes that work. And that's something we think we need more of. And it's just sort of a -- we're not moving wholesale to this. We're just saying this is a new sort of addition and a type of relationship we should have built historically and we intend to. With regard to use of our balance sheet, we generally say, look, what we're trying to do in terms of the sort of wholesale change of an industry and how credit works is very hard. And our intent is to use our balance sheet in whatever way will forward the business. So we don't want to have any hard and fast rules. We've said in the past we want to use it to incubate new products where you can't expect either lenders or investors to sort of participate on day 1 on an unproven model. So that's maybe an obvious one. But maybe there's also dislocations in the market. But our whole view is, look, our balance sheet is not so large that we will ever be, in any sense, a balance sheet lender. It's not our intention to be. It's not the nature of our model. We've also been incredibly, I would say, prudent stewards of capital as a company overall. As a private company, we raised maybe $165 million -- $160 million. And when we went public, we probably still had $90 million of that. We've generated a lot of profits from then. We have retained earnings. We've been adding to our cash balance since then. So I say that only in the sense of I think there's reason to trust us that we are prudent sort of stewards of capital, that we don't intend to become wholesale of a balance sheet lender by any stretch. We don't have the capital even if we wanted to do. That's not the nature of who we are and what we're doing. But we're going to use our balance sheet to move the market in ways that we think will move it expeditiously. And that's, I think, in the interest of our shareholders and the interest of the company to get to the place we want to get to.

Well, that's a great way to cap off the session. Would you please join me in thanking David Girouard for his time, thoughts and insights.

Thanks. It's my pleasure. Thank you so much.