Upstart Holdings Inc
NASDAQ:UPST
If you don’t study any companies, you have the same success buying stocks as you do in a poker game if you bet without looking at your cards.
Utilize notes to systematically review your investment decisions. By reflecting on past outcomes, you can discern effective strategies and identify those that underperformed. This continuous feedback loop enables you to adapt and refine your approach, optimizing for future success.
Each note serves as a learning point, offering insights into your decision-making processes. Over time, you'll accumulate a personalized database of knowledge, enhancing your ability to make informed decisions quickly and effectively.
With a comprehensive record of your investment history at your fingertips, you can compare current opportunities against past experiences. This not only bolsters your confidence but also ensures that each decision is grounded in a well-documented rationale.
Do you really want to delete this note?
This action cannot be undone.
| 52 Week Range |
21.64
81
|
| Price Target |
|
We'll email you a reminder when the closing price reaches USD.
Choose the stock you wish to monitor with a price alert.
This alert will be permanently deleted.
Upstart Holdings Inc
Earnings Call Analysis
Q2-2024 Analysis
Upstart Holdings Inc
Upstart's second quarter of 2024 marked a significant turning point as the company reported net revenue of $128 million, surpassing guidance by $3 million. Although this figure represented a 6% decrease from the previous year, the company celebrated a 31% increase in loan transactions, totaling approximately 144,000, which highlights growing demand despite a dip in average loan size.
Remarkably, Upstart achieved an automation rate of 91% for processing loans, a stunning improvement from 73% just two years ago. This heightened efficiency contributes not only to cost savings but also enhances customer experience by speeding up approvals without requiring human intervention.
While the company reported a GAAP net loss of $54 million and an adjusted EBITDA of negative $9 million, these were seen as encouraging indicators of progress towards profitability. The firm's adjusted earnings per share came in at negative $0.17, reflecting a diluted weighted average share count of 88 million.
Operating expenses decreased to $183 million, down 6% sequentially, showcasing the effectiveness of a recent workforce restructuring. Although the company experienced pressure from smaller loan sizes impacting loan processing costs, overall, the results indicate effective cost management strategies in a challenging environment.
For the third quarter of 2024, Upstart anticipates total revenues of approximately $150 million, with revenue from fees projected at $155 million and net interest income expected to be negative $5 million. The company is also optimistic about the second half of the year, projecting total revenue from fees of around $320 million and achieving positive adjusted EBITDA by the fourth quarter.
CEO Dave Girouard emphasized the transition from a reliance on balance-sheet funding to long-term funding partnerships, evidencing a shift towards a more resilient capital structure. About 50% of the loans were funded through committed long-term partnerships, indicating a solid pathway forward as the firm aims to reduce overall balance sheet risk.
The introduction of its latest AI model, M18, is a milestone for Upstart, improving accuracy significantly. This model now allows for enhanced borrower risk assessment by incorporating Annual Percentage Rate (APR) as an input, promising to refine loan approvals and subsequently bolster conversion rates.
The management team believes that although macroeconomic conditions have improved, they do not depend on these external factors for future growth. Instead, internal improvements in models and operational efficiencies provide a solid foundation for expansion, regardless of the broader economic climate.
Upstart is not only optimizing its existing model but also keeping a finger on the pulse of the competitive landscape. As some traditional banks pull back in lending, Upstart aims to capture market share by being more competitive in pricing and service offerings in its core personal loan products.
Strategically, Upstart's small-dollar relief loans function as a means to build customer loyalty and trust. These smaller loans often serve as gateways for clients to eventually seek larger loans, effectively allowing the company to enhance its customer base over time.
In summary, Upstart showcased a promising trajectory in Q2 2024, with advancements in technology, effective cost management, and optimistic revenue guidance reflecting a strong foundation for potential growth. As it navigates through the evolving landscape, a dedicated focus on operational efficiency and customer-centric products could provide a competitive edge.
You don't have any saved screeners yet
Good day, everyone, and welcome to the Upstart Second Quarter 2024 Earnings Conference Call. Today's conference is being recorded.
At this time, I would like to turn the conference over to Cynthia Moon, Lead Corporate and Securities Counsel. Please go ahead.
Good afternoon, and thank you for joining us on today's conference call to discuss Upstart's second quarter 2024 financial results. With us on today's call are Dave Girouard, Upstart's Chief Executive Officer; and Sanjay Datta, our Chief Financial Officer.
Before we begin, I want to remind you that shortly after the market closed today, Upstart issued a press release announcing its second quarter 2024 financial results and published an Investor Relations presentation. Both are available on our Investor Relations website, ir.upstart.com.
During the call, we will make forward-looking statements, such as guidance for the third quarter of 2024 and the second half of 2024 related to our business and our plans to expand our platform in the future. These statements are based on our current expectations and information available as of today and are subject to a variety of risks, uncertainties and assumptions. Actual results may differ materially as a result of various risk factors that have been described in our filings with the SEC. As a result, we caution you against placing undue reliance on these forward-looking statements. We assume no obligation to update any forward-looking statements as a result of new information or future events, except as required by law.
In addition, during today's call, unless otherwise stated, references to our results are provided as non-GAAP financial measures and are reconciled to our GAAP results, which can be found in the earnings release and supplemental tables. [Operator Instructions]
Next week, on August 15, Upstart will be participating in the Needham Fintech and Digital Transformation Conference. On September 12, Upstart will participate in B. Riley Securities Consumer and TMT Conference.
Now we'd like to turn it over to Dave Girouard, CEO of Upstart.
Good afternoon, everyone. I'm Dave Girouard, Co-Founder and CEO of Upstart. Thanks for joining us on our earnings call covering our second quarter 2024 results. I've said many times over the last couple of years that I've never lost an ounce of faith or optimism in the future of Upstart, and today you can begin to see why. I'm proud and thankful for the incredible work done by Upstarters in the last 2 years to build a stronger and better company on so many dimensions. The numbers and guidance we released today demonstrate that we're turning a corner. We've made real progress toward returning to sequential growth and EBITDA profitability, and, I believe, toward resuming our role once again as the fintech known for high growth and healthy margins.
We've also rebuilt our funding supply by locking in important long-term funding partnerships and significantly reducing the use of our balance sheet to fund loans. We expect this trend of reduced loan funding from our balance sheet will continue through the remainder of 2024. But this progress is not due to any dramatic improvements in macroeconomic factors or risk. Any such macro wins remain in our future. Rather, our progress is the result of the dedicated efforts of more than 1200 Upstarters. The improvements that are evident in our business today are coming from inside the house. First, significant and even dramatic AI model wins. Second, a revamped and revitalized funding supply. And third, increased operating efficiency. These wins and more are providing the foundation for the Upstart comeback story that I expect we'll share with you in the quarters and years to come. Today, I'll provide some insights to these major initiatives and how they're building on the progress we've made in recent months.
We continue to focus the majority of our efforts on our core personal loan product, where the opportunity for leadership in a fast growing category is clear. Our product today is far superior to what we offered 2 years ago in all the dimensions that matter. Model accuracy, fraud detection, automation, funding resiliency, acquisition costs, and revenue optimization are leaps and bounds better than they were in 2022.
Most importantly, I'm thrilled to share that we very recently launched one of the largest and most impactful improvements to our core credit pricing model in our history. In fact, with this launch, 18% of all accuracy gains in this model since our inception have been delivered by our ML team in the last 12 months. To dive a bit further, Model 18, or M18 as we call it internally, is the first to incorporate APR as a feature or as an input to the model. It's of course common to think of APR as an output of a risk model, at least indirectly. But we know empirically that the APR also affects the repayment risk of a loan. All else being equal, a higher APR will select for a riskier borrower, a notion known as adverse selection. Conversely, a lower APR will select for a less risky borrower. If you have a background in computer science or math, you quickly realize that having APR as both an input and output to the same model presents some challenges. Solving this problem requires running our risk models many times in parallel to arrive at the appropriate answer. In fact, M18 generates approximately 1 million predictions for each applicant in order to converge to the correct APR, which is 6x the number of predictions of the prior model, and we believe the improvement in accuracy is well worth it. Additionally, I'm very happy to report that we expect M18 to substantially improve our funnel conversion rate.
From a competitive standpoint, I believe that significant technical obstacles, such as the one I've described here, are themselves a clear sign of progress. We're pushing the boundaries of computing and AI to build more accurate models, and we've seen few signs that peers in the lending space are far enough along the path of AI-based modeling to even encounter these technical challenges.
We also reached another all-time high on automation of our core unsecured loan product, with 91% of loans in Q2 fully automated. As a reminder, this means no documents, no phone calls, no waiting, and no human involvement whatsoever. Two years ago, this number was 73%, and we weren't sure reaching 90% was even possible. Driving automated approvals up while keeping fraud to minimal levels is an obvious fit for AI, so we would expect Upstart to continue to lead on this front. And automation isn't just a win for cost and efficiency, it also provides the foundation of a fundamentally better product for the consumer.
Ultimately our strategy is to offer the best rates and best process to all, for every credit product that matters. This means continuing to expand our platform to auto loans, small-dollar relief loans, and home equity lines of credit. And we're making great strides in each of these products.
In Q2, our auto team released new underwriting models for both our auto retail and refinance products as well as a new fraud model for auto retail. We've now seen multiple months of calibrated loan performance and are growing confident that our loans are performant and increasingly competitive in the market.
In the interest of continuing to move our auto business to profitability, we increased the monthly fee we charge each dealership for the use of our software. Despite this, we believe we're still quite inexpensive relative to competitive offerings.
We're also investing heavily in servicing and recovery for auto and saw a 33% improvement in roll rates and a 44% increase in recovery rates in the second quarter alone.
Our small dollar relief product continues to grow rapidly, with 57% sequential growth in the number of loans in the second quarter. Our intention with this product is to expand access to bank quality credit rather than to generate enormous profits. Nonetheless, I'm thrilled to say that in Q2, SDL became our second product to reach break-even economics. We also signed our first warehouse for SDL this past quarter. For the current quarter, we've identified opportunities to reduce the variable cost of these loans by more than 40%, which would represent another incredible win and an opportunity to increase approval rates further. Overall, this team continues to execute like pros and is helping Upstart expand its impact on the American consumer rapidly and responsibly.
As of today, our Home Equity Line of Credit is available in 30 states, covering 51% of the U.S. population. We exited Q2 with an instant approval rate for HELOC applicants of 42%, up from 36% in Q1. This means we're able to instantly verify applicants' income and identity without the need for tedious document uploads. Consistent with our experience in personal loans, instantly approved applicants convert almost twice as often as other applicants. With respect to credit performance of our HELOCs, things couldn't be better. With more than 300 HELOCs originated, we have 0 defaults to date.
Finally, we've seen significant interest from Upstart's bank and credit union partners in our HELOC product and hope to launch our first lending partnership before the end of the year.
We continue to invest enormously in servicing and collections. To give you a sense of this, in the last 2 years we've tripled the number of Upstarters on our servicing product and engineering teams. And this investment is paying off. We've made it radically easier for borrowers to make payments in whatever way works for them. We've implemented new channels for reaching borrowers who are delinquent. These efforts and more have helped drive delinquency rates down by 16% year over year and have helped reduce support costs per current loan by 30%. We've also now increased the number of borrowers enrolled in autopay for 36 consecutive weeks. Much of our team's efforts to date have prepared our servicing infrastructure for the deployment of AI models that we believe will enable us to build a significantly differentiated loan-servicing capability.
Two years ago, we told you that we would upgrade the funding supply on the Upstart platform. We aimed to move a significant portion of our funding from at-will monthly agreements to longer term committed partnerships. Given the importance and complexity of these relationships, we cautioned that this would take some time. I'm pleased to share that we've now accomplished this goal.
We ended Q2 with well over half of the institutional funding on our platform coming from committed capital and other co-investment partnerships. We began with the announcement of our first partnership with Castlelake 15 months ago. This partnership has since been renewed. We've since added significant partnerships with Ares and Centerbridge. Other institutional investors that have been with us for much longer have also returned to the platform. We continue to pursue additional opportunities to broaden and deepen our funding supply as Upstart returns to growth mode.
We also said back then that we'd use our own balance sheet as a transitional bridge to this better state. You can see from the numbers we released today that we've begun to reduce the use of our balance sheet to fund loans. We're hopeful this will continue through the rest of the year, though I'd like to always reserve the option to use our balance sheet to do the right thing for our business.
I'm also pleased to report that banks and credit unions continue to return to the Upstart platform. We've signed 8 new lenders since Q1. Performance and lender demand on the platform are creating a competitive environment which is beginning to reduce prices for Upstart borrowers. In fact, lenders representing about half of the monthly available funding on Upstart from lenders have reduced their target returns recently as their liquidity has improved and their demand for loans has increased. This is the first time in 2 years that we've seen loan prices drop on Upstart.
For many reasons, transforming credit with AI is complex and challenging. Tackling the world's most entrenched problems with AI is difficult and it doesn't happen overnight. But to those who ultimately solve these problems, there comes a tremendous reward.
Today we're tackling problems that we weren't even aware of a couple years ago. My perspective is that, top to bottom, we've gone through a significant reinvention of the company, both from a technology and business model perspective. We're confident we're on the right track and making rapid progress. And this is just beginning to show in our financials.
Despite the fact that many trillions of dollars in credit are originated each year, our competition in AI is scarce. In Generative AI, you have a significant number of well-funded and talented competitors, such as OpenAI, Google, Anthropic, and Meta, at the cutting edge of model building. In AI for lending, you have Upstart.
Thanks. And now I'd like to turn it over to Sanjay, our Chief Financial Officer, to walk through our Q2 2024 financial results and guidance. Sanjay?
Thanks, Dave. Good afternoon to all, and thank you for joining us.
A notable topic for us over the past few quarters has been the macro climate and its impact on both consumer spend and credit loss. The stimuli of 2020 and early 2021 left consumers flush with cash and in retrospect, unleashed a 2-year-plus surge of consumption as consumers plunge to new elevated spending habits well beyond the duration of the stimulus and in our view, also beyond our collective means. These trends were, of course, exacerbated by punishing price inflation. This inflation, which also had its roots in the post-COVID monetary expansion, appears to have mostly run its course as we had anticipated for much of the past year. We now also see signs that the venerable American consumer is reluctantly waving the white flag, acting to moderate outlays and rebalance budgets.
Consumption of goods, both durable and nondurable has actually been falling in real terms over the course of this year. Spending on services has continued to rise, but half this increase over the past year is attributable to skyrocketing healthcare expenditures. Many other subcategories of services consumption growth in our economy have also started to abate. To be unambiguous, we believe this is a welcome development for the American economy, which has been on an unsustainable tear over this broader period of time.
One product of improving fiscal health is that we are seeing credit default trends finally turn a corner, having peaked in aggregate sometime earlier this year and now inflecting back down towards prior lower levels. This dynamic is reflected in our declining Upstart Macro Index, which has now unambiguously fallen for 3 consecutive months and has reached its lowest level since January of 2023. This downward traveling UMI is now a consistent pattern across all borrower segments that we can observe.
While the mobile U.S. borrower continues their rehabilitation, we also note ongoing improvement in the funding markets, both on the institutional side as well as in the banking and credit union sectors. For the second consecutive quarter, we've increased the number of lenders who are active on our platform and have observed reductions in required rates of return. On the institutional side, we have now renewed all of our committed capital deals from last year and are currently in the process of adding new partners to the program in anticipation of future borrower growth. One such recent example is the new agreement we've completed with Centerbridge, a leading global alternative investment firm by which they acquired $400 million of our personal loans.
We are seeing early signs of funding progress in some of our newer products as well. We have secured financing to continue scaling up our auto and small dollar loan offerings and expect to complete our first forward flow sale of HELOC loans in the coming days. These collective funding efforts have allowed us to reduce the overall size of our balance sheet and store up some dry powder in support of any future growth and new product development needs.
With this macro environment as backdrop, here are some financial highlights from the second quarter of 2024. Revenue from fees was $131 million in Q2, down 9% from the prior year as higher pricing for prime loans created downward pressure on origination volumes. Net interest income was negative $3 million, an improvement both year-on-year and sequentially as the larger than typical core loan balance sheet we were carrying until late in the quarter, produced income, which helped to offset excess loss in our R&D portfolio. Taken together, net revenue for Q2 came in at $128 million, $3 million above our guidance, but down 6% year-on-year.
The volume of loan transactions across our platform in Q2 was approximately 144,000 loans, up 31% from the prior year and up 21% sequentially and representing over 89,000 new borrowers. Average loan size of $7,700 was down from $9,500 in the prior quarter, driven lower by continuing robust growth in small dollar loans as well as by pressure from higher pricing on prime loans, which tend to run larger than average.
Our contribution margin, a non-GAAP metric, which we define as revenue from fees, minus variable costs for borrower acquisition, verification and servicing as a percentage of revenue from fees, came in at 58% in Q2, flat sequentially and 2 percentage points above our guidance for the quarter. We continue to benefit from very high levels of loan processing automation with our eighth consecutive quarterly improvement in percentage of loans fully automated, resulting in a new high of 91%.
Operating expenses were $183 million in Q2, down 6% sequentially from Q1 as the workforce restructuring we underwent yielded lower payroll costs across all of our functions. These savings were somewhat offset by the impact that higher loan volumes and smaller loan sizes are having on our loan processing costs.
Altogether, Q2 GAAP net loss was $54 million and adjusted EBITDA was negative $9 million, both comfortably ahead of guidance and encouraging proof points on our path back to profitability. Adjusted earnings per share was negative $0.17 based on a diluted weighted average share count of 88 million.
We ended the second quarter with loans on our balance sheet of $686 million before the consolidation of securitized loans, down from $924 million in the prior quarter. Of that balance, loans made for the purposes of R&D, principally auto loans stood at $396 million. In addition to loans held directly, we have consolidated $135 million of loans from an ABS transaction completed in 2023 from which we retained a total net equity exposure of $21 million. We ended the quarter with $375 million of unrestricted cash on the balance sheet and approximately $449 million in net loan equity at fair value.
We have long maintained that once the macro environment ceases to be a headwind, we'll have the opportunity to generate conversion growth through improvements to our models and acquisition campaigns. With loss rates that have now collectively appeared to plateau, this is precisely what we are expecting for the duration of this year. Last quarter, this nascent trend gave us the foundation to provide guidance for the back half of the year, which was based on an assumption that our model gains would deliver their historical pace of growth. Our model launches since that time have in fact produced enough uplift to put us ahead of schedule. Note that despite our relative optimism on the macro climate as it relates to credit performance, our guidance for the rest of the year in no way relies on either further improvements to the macro environment nor on falling interest rates. Either of those eventualities, should they occur, would likely show up as tailwinds to our forecast.
With that in mind, for Q3 of 2024, we are currently expecting total revenues of approximately $150 million, consisting of revenue from fees of $155 million and net interest income of approximately negative $5 million, contribution margin of approximately 57%, net income of approximately negative $49 million, adjusted net income of approximately negative $14 million, adjusted EBITDA of approximately negative $5 million and a diluted weighted average share count of approximately 90 million shares.
For the second half of 2024, we expect revenue from fees of approximately $320 million and positive adjusted EBITDA in Q4.
Overall, we would like to say that we feel good about how we've managed financially through this challenging period. We emerged with expanded margins and a reduced cost base, underpinning the tangible progress we've made on the road back to profitability and successfully reimagining our funding model has created a more resilient capital base and a shrinking balance sheet. More importantly, we are optimistic about the strength and direction of the business as we look ahead. While we are wary of prematurely sounding the all clear, the macro no longer appears to be a direct impediment to our business. An improving macro climate is not contemplated in our forward numbers and it's not something we need in order to thrive. But if and when that does materialize, it should be wind in our sales.
I would like to conclude by acknowledging the entire Upstart team for persevering together through this long metaphorical winter and also to all of our departed teammates who have been a part of the cause, even if they are no longer able to. I'm looking forward to a time in the near future when we all will have to refasten our seatbelts.
With that, Dave and I are happy to open the call up to any questions. Operator?
[Operator Instructions] We will take our first question from Mihir Bhatia with Bank of America.
I wanted to start by just asking if you could comment a little bit more about just the intra-quarter trend than what you saw in July. It sounds like you're quite positive on the back half of the year. And maybe if you could just comment a little bit on what you saw both in terms of loan demand and also just credit performance as you went through the months in the quarter and to the extent you're willing to about July?
Great to hear from you. So you're asking about credit trends and loan trends in July? And through the quarter, month by month in the quarter, like did loan demand accelerate? Did you see more demand in June than April? I see. I mean, at a high level, I guess, to the extent you can hear optimism both in our guide and in our comments, it's probably reflective of a quarter that obviously is leading into Q3 on a good footing and a positive trajectory. And to the extent that we are guiding Q3 on an upward trajectory, I would say that July is representative of that as well.
Okay. And then maybe just switching a little bit to the expense structure a little bit more. What I'm really trying to understand is the fixed versus variable cost of the model. So as top line expands, what kind of impact will that have on profitability? And how much should we expect to drop to the bottom line versus maybe you reinvest in growth or product expansion or additional growth. How should we be thinking about that equation?
In rough terms, as the business expands, I would expect our contribution margins, which really capture our variable cost base to shrink somewhat, mainly due to reductions in take rates. As the business becomes more profitable, we will probably invest more in volume and in lifetime value. I think the cost components of our contribution margin should be fairly consistent because we essentially attempt to originate up to the point of marginal cost breakeven, and I don't think those points will dramatically change as we scale.
As for the fixed cost base, well, the intention is that it will grow slower than the top line of this business, meaning we should achieve operating leverage as the business scales. And so between those 2, I think that scale should drop pretty efficiently to the bottom line as we rescale.
We will take our next question from Ramsey El-Assal with Barclays.
The conversion rate increased quarter-over-quarter and obviously a lot more year-over-year. I know you mentioned some pretty exciting model improvements. I guess, what should we expect on conversion rate for the next couple of quarters? Are your model improvements driving maybe further conversion rate improvements? Or should it plateau at a certain point? What should we be looking for?
Ramsey, great to hear from you. I would say that to the extent our guidance is indicating upward trajectory, almost all of that is coming from conversion gain, and the underlying model accuracy driving funnel improvement over time. And I would say for the foreseeable future, that will be the growth model. There is potentially a rate at which those conversion rates plateau, but I don't think we're close to those rates at this time. So there's still a lot of runway to improve those conversion rates and drive the growth of the top line.
Okay. A follow-up for me. On the smaller dollar relief loans, can you talk about these loans in the context of being like an acquisition channel for larger, longer-duration borrowers or loans. In other words, are you seeing any of these small dollar customers come back and apply for larger loans that you can now kind of underwrite sort of like a training wheels type of a scenario in terms of being a channel into your core business?
Ramsey, this is Dave. I think that's a pretty good description of how that product works and why we have it. It's really to push deeper with small amounts of dollars at risk to be able to approve somebody on a shorter-term loan is just an opportunity for the models to learn faster and go faster and to acquire customers that are then eligible for other Upstart products later. So it is doing a super nice job of pushing the boundaries of our models, both in terms of the automation as well as in the selection and pricing. So that's gone extremely well. We have seen quite a bit of return for other loans. So that's also improving well. And as we said on the call, it's become economically strong for us. It's not a drain on us in any way. So it's been just frankly, all around a great win for us, and we would expect it to continue to be.
We will take our next question from Kyle Peterson with Needham.
I wanted to start off on the size of the balance sheet here. It was going to see some nice runoffs just in the core personal side. I guess, how should we think about the pace of runoff over the next few quarters, especially now that seem to continue to bolster your funding?
Kyle, great to hear from you. The answer to that question is a bit about the outcome of how fast the borrower side of the platform is scaling up due to model improvements and how quickly we're signing up new capital agreements. Obviously, intention continues to be delivering that those borrowers and that yields to our lending partners and to the institutional markets. But there's always going to be a bit of mismatch in timing. We may get a model win and not have the capital available or we may sign the capital up and the model win may come afterwards. And so I think in the give and take between those 2 sides of our platform, that's where we've historically used our balance sheet to step in. And so all that to say, I do believe that the medium-term direction here will continue to be a reduction in our balance sheet, at least as far as core loans are concerned. But there may be some timing mismatch along the way such that there may be some sort of swings up and down as we do that. So it's a bit hard to really calculate a very accurate pacing, if you will, given the volatility of those 2 sides of the business.
Okay. That's helpful. And then I guess just a follow-up on expenses, particularly with the fixed cost base, I think you guys have said kind of in the past that in the fixed cost base you guys have today can support a lot more volume than you guys have been doing, call it, over the past 4 to 6 quarters here. How much if we do get a better environment for originations, I guess how much more volume can you guys support with the fixed cost structure that you guys have today? I know the contribution margin you guys gave was helpful earlier. Just trying to think about the fixed cost leverage.
Well, I guess I'll say that through the end of this year and with the growth plans we have, we feel pretty good at where our cost base is. If the business were to start to really take off beyond that, I think there are some areas on the margin that we would like to reinforce. But nothing on the level of what we anticipate the growth of the business itself could be. So I guess the main takeaway is there will be improving operating leverage as the top line grows.
We will take our next question from Peter Christiansen with Citigroup.
I want to dig into your comment about some of the at-will supplies of funding coming back. Just wondering if you could give us a barometer where we are perhaps compared to maybe, I don't know, '21, part of 2022 in terms of some of those levels or at least indication of funding level that we saw back then. And then, I guess, well, back then, we also had 40% of your funding volume was through the ABS market. Would you expect to be returning to the ABS market for issuance in the near future?
I would say that the recovery of what we think of as the at-will funding markets, real large, that's the world of credit funds and hedge funds that predominantly depend on ABS as a liquidity channel. It's early days for the recovery. I don't think we're near the scale that we were at a couple of years ago. And that's, of course, reflective of the fact that the ABS markets are certainly not at the level of volume and liquidity that they were back then. But I do think that those markets are rapidly improving, and we have plans to be back in the ABS market certainly before the end of the year. So I think those things continue to be on a good trajectory.
That's good to hear. I recognize that period is not a fair comparison, unique era. But secondly, in terms of the co-investment, how should we think about that level progressing over the next, I don't know, 1 or 2 quarters. Is that still, do you think, going to be a portion or tied to your front-end principle?
Pete, this is Dave. I think the co-investment partnerships are definitely key to our future. I mean that was what we've been working on for some time to go from almost entirely at-will funding a couple of years ago to having longer-term committed partnership. So that is very important to us. The at-will funding can be useful in a lot of ways, but overdependence on ABS, particularly when those markets can ebb and flow quite a bit, isn't healthy for us. So as we said, we have well over half of our funding at the end of Q2 in these longer-term partnerships. And we think we would like to maintain that percentage. So I think where we want to be with more long-term committed capital, less reliance on ABS and that sort of structure as we grow back, we would like to sort of keep things as they are now.
We will take our next question from James Faucette with Morgan Stanley.
I wanted to follow up there on the committed capital. How should we be thinking about what that looks like in terms of unit economics or accounting treatment in those partnerships versus kind of at-will generally?
James, in terms of unit economics, the loans that are being funded through that channel look very similar to the broader institutional loans. They differ from the lending partner channel in that the risk aperture is a little broader, and the returns are a little commensurately higher. But in terms of our unit economics, there's really very little difference between that channel and maybe what you might think of as more of the at-will institutional channel. In terms of the accounting, look, these deals, I would say, are still becoming more and more standardized or templatized as we do more of them. I think historically, they've shown up in a couple of different places on our balance sheet. But increasingly, we're going to look to sort of standardize the structure of the deals that we do. And we do pull the holistic view of it together on our investor earnings deck, which gives you a glimpse of the total exposure.
Got it. And then quickly, last quarter, you alluded to the fact that you had indexed more to prime than you've historically had. And given some of the prior actions you took. Just wondering if you can give us an update in terms of what you're seeing in prime versus subprime this quarter and what you anticipate getting back to more normalized mix?
James, this is Dave. Our mix has swung towards prime. And I think generally that we would see, as we regrow we would like to be very balanced across the credit spectrum. And we think that's best for our brand. It's best for stability of the business, et cetera. So one thing we would anticipate in the coming quarters is a stronger position at the primary end of the credit spectrum than we've had traditionally where we really have not had funding appropriate to compete in that part of the market, but we think that's changing. So I think you'll see us be more balanced in the future than we've been in the past with regard to the credit spectrum.
We will take our next question from Dan Dolev with Mizuho.
Great quarter, great results. Very happy to see that. I want to know what's going to happen assuming interest rate cuts, how much torque do you think there is in the business that you can actually expand growth, expand loans as the environment gets more easier for you to do that? That's pretty much my only question is like how much upside can we dream to dream at this point?
Great to hear from you as always. Look, reducing rates, benchmark rates and market rates are unambiguously good for the business. They haven't obviously been the main headwind to our business. Default rates have been much more punitive in how they've evolved over the last 2 years or so. But definitely having the benchmark rates go up from 0 to 5-ish percent has been a headwind as well. And as that reverses, it would presumably be a tailwind. It's a bit hard to quantify the exact nature of the tailwind as rates reduce. And it obviously depends on how far back down they go. But each quarter point will result in lower financing costs for the institutional investors. And as that creates lower hurdle rates, those will result in lower rates to our borrowers. And I guess I'll just say that I think each cut would be a noticeable benefit in terms of its impact on our conversion rate.
Got it. Well, it definitely looks like you're upstarting a new cycle. So congrats again. That's good.
We will take our next question from Giuliano Bologna with Compass Point.
All right. Congrats on the results and some of the new funding announcements. One thing I'd be curious about digging into a little bit is your marketing expenses, so you had some improvement in your marketing efficiencies this quarter. And in the past, what you've kind of said is that there were some challenges with some loans being priced about 36% that you couldn't necessarily convert. And I'm curious, when you think about the improvement in your marketing efficiency this quarter, how much of it was driven by being able to approve or underwrite more loans under 36%? And I'm curious kind of how that could evolve over the next few quarters and how that's kind of factored into your outlook at this point?
Sure. So the marketing efficiency is a function of our funnel conversion, most generally so when the funnel converts better, our marketing tends to get more efficient, et cetera. So that's a dynamic that's always in play. The 36% kind of rate cap on Upstart means that as base rates go up and as risk goes up, fewer and fewer people are approved, and we've seen that in spades in the last couple of years. We went through a 2-year period where rates almost constantly were going in an upward trajectory. And every time that happened, a bunch more people would not be approved because effectively, the system requires of them goes over 36%. So that's a little bit unwinding going the other way now, which is a good thing. Partially or most of it actually is due to model accuracy and the newest versions of the models. We're able to sort of identify more people who fit under that envelope of 36%. And the result of that is that you see marketing efficiency improving. So that's a dynamic we would expect to continue in the coming months and quarters.
Got it. And maybe picking away at that point. I'm curious in a sense of where things are in the sense of when we think about funnel conversions and kind of the improvement, where do you think we are kind of improved 10%, 20% of normalization? And is there a lot more to go with 100 basis point or 200 basis point decrease in interest rates.
Giuliano, I would think of this as an ongoing journey. I think the model accuracy has systematically improved since the beginning of our company and each improvement has a commensurate improvement on our conversion rates. Those can obviously be temporary setback on the macro, but as the macro normalizes, so all our conversion rates. And the question to how much better they can get is sort of the same answer to the question of how much more accurate can your models get at approving good borrowers and avoiding bad ones. And we've talked about the fact that we think we've really just kind of scratched the surface in terms of our model, the ability to improve explainability in credit default. And so we believe that the longer-term road map of this company continues to be improving models and improving conversion rates over the years. So we don't think of it as sort of normalizing right now. We think we're back on the journey of improving models and improving conversion rates now that the macro is no longer a direct headwind.
Maybe one very quick question. You're obviously going around 50% or rolled off 50% forward committed capital. That's kind of a percentage of your funding. I think in the past you referred to that as where you'd wanted it to be, close to the higher end of the range. I'm curious, would you look to overshoot that and then grow kind of the spot or uncommitted business to catch up with that? Is there any structural limitation in the near term to what percentage of volume or funding you'd want to have come from forward committed capital sources at this point?
Yes. I think that given that we are feeling increasingly optimistic about the road map of model improvements and the lack of macro headwinds, I think it's in our interest to put some more capital deals in place now and in your words to try and overshoot a little bit in anticipation of that growth materializing over the coming quarters, just given that these deals are relatively heavily negotiated and they take some time to put in place. So I think we want to err on the side of having those partnerships in place in anticipation of where we see the borrower side of the platform growing.
We will take our next question from Rob Wildhack with Autonomous Research.
Question on the outlook. Updated guidance suggests better trend on originations. You guys sound pretty positive overall. Could you maybe break down how much of the improved outlook is coming from maybe mechanically from lower interest rates versus a better model versus maybe better funding? How would you quantify each of those or any additional drivers into the better outlook?
Rob, this is Dave. I think that there is no assumption of improving interest rates or reduction in kind of macro risk built into that. So the guidance is based on really what we're seeing based on improvements we've made internally. And maybe the way to think of that is better model means better conversion rate. The other important input is we have to, of course, have sufficient funding supply to keep up with that growth. But the gating item in terms of like what's really gating where our guidance sits today, it really is just about economic funnel conversion. And it's improved a lot really through model improvements primarily. And at this point, we feel comfortable that we, on the funding side can make things match well. So that's a long-winded way of saying it's really through things we've done ourselves. It is not based on any assumptions about improvement in rates or risk in the environment.
Okay. And then a question on the small dollar loans. I mean could you give us some color on how much the growth in small dollar loans may or may not have impacted the conversion rate quarter-over-quarter? And the same question going forward, as you grow in small dollar loans, does that drive the conversion rate a lot higher?
Yes. Rob, the SDL product is having an impact on overall conversion rates. I think it's on the order of maybe a 2% or 3% impact at the scale that it's at. So it's not insignificant, but it's also relatively minor.
We will take our next question from Simon Clinch with Redburn Atlantic.
I was wondering if you could talk about what it takes or what levers you can pull and what macro tailwinds you might need to see the sort of gross inquiries that come into the Upstart network before conversion. How do you drive that higher over time because that is down quite materially from where it's been in the past. And I'm just wondering if that was just overstated previously and whether there's actually quite a lot of upside in this coming cycle for that?
Just about the top of the funnel inquiries that sort of the...
Before conversion yes.
We've not published sort of traffic to the site. So that's not something that we've discussed publicly or trying to track in that means. Is there something different you mean by that?
So I'll just take your volumes and then kind of back out from the conversion rate sort of what it was before you've converted. I just use that as a mean sort of track approximately what volumes would be...
Yes, it's not exactly the same thing, but it's directionally correct. Generally speaking, a lot of times, we are controlling that by how much we're spending in various marketing channels and also just generally how competitive our rates are. So that's part of whether we're doing direct mail or some sort of digital acquisition or whether we're kind of remarketing to our own customer base or through partner channels that are responsive and can vary how much traffic they send us based on the quality of our rates, et cetera. So those are things that are a function of the market in some sense or how strong our product is, how much we're actively marketing. So I hope that fills in some of the blanks for you.
Okay. And maybe you could talk a bit more about the Model 18, M18. And just I guess, can you give us a sense for those of us who aren't educated in machine learning and stuff like that, but just really how unique something like that is and ultimately, how quickly a model like that really starts to have an impact on your business?
Well, we're in a sort of never-ending quest to accurately price each and every loan offer that's made on our system. And one of the things we've known and I think most lenders of some sort know is that the quality of the offer you make to the market, meaning the level of the APR has an impact on who accepts it and therefore, how that loan performs. So the APR, which is most people would think about the output of the model actually affects the performance of the loan. So this is something, again, most people would tell you they have an intuitive sense of. But mechanically answering it and having models that are sophisticated enough to handle that is very important, particularly in the modern world where consumers have lots of choices. They compare rates all over the place. This is something that even 10, 15 years ago, hardly existed. But today, consumers have a lot of ways they can compare and find the best rates. So having a lot of savvy around that notion of adverse selection in positive selection is really important. In solving it from a technical perspective really comes down to trying to converge to the appropriate APR. And what that amounts to technically for us is running our risk models many, many times in parallel in order to converge to the right number. And it's a significant challenge that we've gotten over. And I think we're just beginning to reap the benefits of it. The guidance that you're seeing for the second half of the year, a significant fraction of what you're seeing in terms of our optimism for the second half of the year comes directly through the improvements in that model. And also, we see a lot of continued opportunity in that domain in that area to improve the models. And again, that's what we're in business to do. It's generally where all the advantages of Upstart are is when we can build better risk models and we're having some really good success in that area right now.
We will take our next question from Vincent Caintic with BTIG.
First, I just wanted to follow up on the funding partnership discussion. It's good to see that the credit investor demand is increasing. Just if you could maybe talk about some of the discussions you're having, what are those credit investors focus on? What's changed where you're now getting more sign-ups? If you can give a sense of how pricing has changed or improved? And maybe how much of your annual origination volume is now covered by all these new sign-ups?
Vincent, welcome back. On the funding partnerships that we are engaging in, I mean, there's sort of 2 general vectors. One is increasing comfort or confidence with credit trends in general and maybe sort of macro risk. And then second, we're sort of being innovative in some of the financial structures that we're coming up with and discussing with some of these partners and prospective partners. And it's sort of, I would say, a learning curve for all of us in terms of how to get these partnerships implemented and put in place and managed. And so a lot of the journey with the prospective partner is just about understanding the model and the structure and how it all works. And then the recognition that there's definitely ways of creating win-win partnerships here for us as the issuer and for these counterparties who are interested in the yield. And so I wouldn't say beyond that, there's been dramatic changes in preferences over rates, and sort of supply and demand dynamics. It's mostly been just an ongoing education for all of us around how these structures work. And I think it's going in a very good direction.
In terms of capacity, as Dave said, we're sort of a bit north of 50% of all the institutional money that's going to to fund the loans on our platform in the past quarter came from these types of arrangements, and we'll aim to maintain that kind of coverage or that kind of capacity over the long term, in the medium term, we'll maybe overbuild a little bit in anticipation of some growth that may happen in the coming quarters.
Okay. That's great color. And my second question, just if you could talk about the competitive environment for consumer financing. It seems like others in this environment might be pulling back when you hear about some of the traditional banks on their earnings calls were talking about seeing stress in the low end in the middle consumer. So it seems like a lot of competition is pulling back. But I just wanted to get that sense from you what you're seeing with that competitive environment.
Well, I think our position on the consumer, I like to think we've been ahead of the crowd a bit in the sense that it was clear there was deterioration of credit at the sort of less affluent part last year and then later last year into the more affluent part. But as Sanjay said in his remarks earlier, we're seeing sort of uniform improvement now across the board. So we sort of feel like we've been kind of signaling this for some time that we're nearing the end of the cycle. And I think we just have clear indications that credit is actually in a normalization period, not in a deterioration period. Now what others are seeing or saying and where their data is coming from, I obviously can't speak to. But I think we feel pretty good about that.
With regard to banks and lenders can either be a partner of ours or they can be a competitor of ours. But I know the ones that are partners of ours are tending to see increasing liquidity and that sort of swung to the place where they're needing more assets, they're needing more loans. And we talked a bit about that. So they are coming in a little bit more competitively, lowering their return targets and really wanting to sort of swing the dial a little bit. So I don't think there's any sort of caution to the wind like environment, but I do think the sort of lack of liquidity that was really serious a year ago and it's probably carried on through the end of 2023 has really improved a lot. And for us, that means the lending partners and banks and the credit unions have definitely returned, and that's been very helpful for us.
We will take our next question from Reggie Smith with JPMorgan.
I've got 2 quick ones. I guess you guys called out model improvement and a better UMI, which is great to hear and see. My question is, how should we think about those 2 things in the context of the deterrence in the co-investment portfolio? And I guess, specifically, I'm trying to figure out, I mean, should that manifest in better performance there? If not, like where do these gains and model efficiency accrue. Obviously, consumers are getting approved more loans. But how do you think about how that flows through to your business? And I have a follow-up.
Reggie, it's a great question. In general, model gains or model accuracy improvements such as the one that we highlighted, generally improve our ability to accurately separate risk and that generally shows up mainly in our improved conversion funnel. So it would create business expansion. It wouldn't necessarily improve the calibration of the model in how it assesses an average pool of loans. So it wouldn't necessarily be expected to have a huge impact on the performance of the co-investment positions we have. The UMI to the extent it continues to fall would have a direct impact on the performance of loan pools, such as the ones that our co-investment partnerships have invested in because it essentially means that credit trends are improving in real time. And as they do the performance of those loans, any loans that are outstanding would be expected to improve and potentially overperform, and that would result in higher returns to our investment positions. So I think that would have a pretty direct impact.
Got it. Understood. And then I guess a follow-up on the unit economics. I'm not sure how much you guys can share here. But curious with somebody's committed structures, I assume you're selling these loans, maybe at a slight discount to par or maybe like where are you in that? Where are you in terms of that? And is the thinking that over time you could get to a place where you do some of at a premium to par or is kind of par the aspirational goal there? Or am I completely off and maybe you selling them at a gain right now, I'm not sure.
Yes. The committed partnerships we are in as with all of the at-will capital, the traffic in the institutional markets, all of those loans are trafficked at par. And I think that's our goal. We're not necessarily looking to create a business model from gain from sale. I think our goal is to traffic loans at par that are correctly priced to the borrowers. And to the extent we're co-invested we'll participate in the yield.
Would be one last one. Okay. I just have one other question, and I wanted to give you guys [ flowers ] for return to EBITDA positivity in the fourth quarter. Just curious how you're thinking about stock compensation expense longer term. I noticed that it's been up well above where it was when you guys were much more profitable. So just curious like what's the pain there.
Well, we're happy to take our flowers for the return to profitability. I appreciate that. How we're thinking about stock compensation, I don't think it's dramatically different than how we thought of it in the past as a tech company that's set forth in the valley, it's important to us for our employees to have a stake in the mission and the outcomes of the business. And I think we're at a pretty comfortable balance between cash compensation and equity compensation depending on roll and level. So I don't necessarily see a dramatic departure from how we've managed it to date.
We will take our next question from Arvind Ramnani with Piper Sandler.
I wanted to ask on this call and in the prior calls, you all have talked about some of the big investments you have made in kind of improving your model and sort of like capabilities. And as we get into a better like kind of operating environment or lending environment, now that you have like kind of a better model. I mean, how do you expect the business to kind of perform in a more conducive environment just given the backdrop of kind of better offering.
Well, I think as I kind of said in my remarks, I think we've gone through a pretty significant transformation of the business over the last couple of years, both from a technology perspective and from a business model perspective. On the technology side, we feel much, much better at the quality of the models, how quickly they can react to changes in the environment, the amount of separation we're getting. So it's just the normal trajectory of an AI model, where it's getting more and more data, more and more variables. We're putting more sophisticated software and as we talked about Model 18. So higher degrees of automation as we talked, we have a record high on that front. So the technology side has just really improved a lot and just made us more efficient. And I think on the business model side, one of the things we clearly identified is we needed to have a funding structure that had permanence to it so that when we grow and even if there's bumps in the road along the way, which there inevitably will be, we can grow through them. And that's kind of what we've done on the business model side has really changed the nature of funding from completely at-will to dedicated partnerships, and we have some skin in the game in these partnerships as co-investors, which we think, given our role and our aims in the market is a structure that makes sense. So of course, in the good times when rates are dropping and the consumer is getting financially healthier, that's all very easy, and we're hopeful that's what we're headed into. But of course, the test is when the market is not so easy, but that's what we're designing for. We're designing for a future with less volatility and more ability to thrive through whatever economic climate we find ourselves in.
Yes. That's really helpful. And I know like I mean, willing to give you the benefit of doubt that that you're like our models are better. But I wanted to ask, like, have they been validated by some client feedback, some banking partner feedback? Or like, I mean, what is your sort of comfort level in saying that, hey, we have a better model like? I mean, are you looking at internal data and coming to a conclusion? Or are you getting that from external validation, kind of what really gives you sort of comfort that you have proof that you have a better model.
Yes. I mean there's very, very well-understood statistical techniques to actually describe and quantify accuracy of a model, and there are several different ones, and we use generally all of them. So it's not hard for us to assess ourselves whether our model is getting more accurate or not relative to prior versions of our model. So it's not hypothetical in any sense. It's something very straightforward in terms of building more accuracy into a model. Certainly, every lending partner and credit investor on our platform sees all the data that is coming out in terms of all month-by-month performance data, et cetera. They have their own means of evaluating whether they think the credit is performing well or not what have you. But they're not looking at the software, if you will, trying to assess our model, but they care about the results, of course. But I don't think there's any reason to question that we can accurately identify the level of improvement in accuracy that we see in each subsequent version of our model, it's kind of the nature of the system to do so.
There are no further questions at this time. And Mr. Girouard, I will turn the conference back to you for any additional or closing remarks, sir.
All right. Thanks, everybody, for joining us today. As we discuss the actions we've taken over the last few years are beginning to pay off, and we believe we're well set up for the remainder of 2024 and into next year. So I hope you all enjoy the rest of your summer. We look forward to speaking with you all in the fall.
This concludes today's call. Thank you for your participation. You may now disconnect.