DAVE - NASDAQ

Good morning, everyone, and thank you for participating in today's Conference Call to Discuss Dave's Financial Results for the Second Quarter Ended June 30, 2025. Joining us today are Dave's CEO, Mr. Jason Wilk; and the company's CFO and COO, Mr. Kyle Beilman. By now, everyone should have access to the second quarter 2025 earnings press release, which was issued this morning. The release is available in the Investor Relations section of Dave's website at investors.dave.com. In addition, this call will be available for webcast replay on the company's website. Following management remarks, we'll open the call to answer your questions. Certain comments made during this conference call and webcast are considered forward-looking statements under the Private Securities Litigation Reform Act of 1995. These forward-looking statements are subject to certain known and unknown risks and uncertainties as well as assumptions that could cause actual results to differ materially from those reflected in these forward-looking statements. These forward-looking statements are also subject to other risks and uncertainties that are described from time-to-time in the company's filings with the SEC. Do not place undue reliance on any forward-looking statements, which are being made only as of the date of this call. Except as required by law, the company undertakes no obligation to revise or update any forward-looking statements. The company's presentation also includes certain non-GAAP financial measures, including adjusted EBITDA, adjusted net income, non-GAAP gross profit, non-GAAP gross margin and compensation expense, excluding stock-based compensation, as supplemental measures of performance of our business. All non-GAAP measures have been reconciled to the most directly comparable GAAP measures in accordance with SEC rules. You'll find reconciliation tables and other important information in the earnings press release and Form 8-K furnished to the SEC. I would like now to turn the call over to Dave's CEO, Mr. Jason Wilk. Please begin.

Good morning, everyone, and thank you for joining us. We're pleased to share that we achieved another quarter of record performance as we continue to deliver against our mission of leveling the financial playing field for everyday Americans. Q2 represented a continuation of our momentum with accelerating revenue growth, robust unit economics and strong earnings growth, all tracking ahead of plan. Revenue accelerated 64% year-over-year to $131.7 million, marking our fastest growth rate in over 5 years. Performance was driven by a 16% increase in monthly transacting members and step change ARPU growth of 42%, underscoring our ability to monetize a growing and engaged member base. Adjusted EBITDA demonstrated growing operating leverage, more than tripling year-over-year to $50.9 million. This result represented the largest absolute adjusted EBITDA gain in company history, highlighting strong execution across the business and disciplined expense management. All of this outperformance reflects consistent execution across our team as well as the [ outset ] of our new fee structure, which continues to strengthen monetization and deepen member engagement through margin limits. Given our strong year-to-date performance and clear momentum in the business, we are pleased to once again raise our full year revenue and adjusted EBITDA guidance. Turning now to our three strategic growth pillars: efficient member acquisition, enhanced member engagement through ExtraCash and deepening relationships via the Dave Card. Starting with our first strategic growth pillar of efficient member acquisition. We added 722,000 new members in the quarter, bringing total members to 12.9 million. Our member base grew 14% year-over-year, while CAC modestly increased $1 sequentially to $19. We are increasingly optimizing our marketing investments by device platform and channel prioritizing investments that yield the highest projected gross profit dollar returns rather than the lowest CAC. This recalibration is also closely tied to the higher member lifetime value we are observing following the transition to our new fee model. Importantly, our payback periods on customer acquisition costs have further improved to an estimated 4 months, down from 5 months mid last year. As a result, we expect to scale marketing investment through the back half of the year. This reflects our belief that our enhanced unit economic profile in addition to current market conditions, represents an attractive opportunity to further lean in and drive efficient incremental growth. Our second strategic pillar centers around continuing to strengthen engagement with our members through credit. ExtraCash remains a key entry point for building long-term relationships with our members by addressing what is typically their primary need, short-term liquidity for gas, groceries and bills. In Q2, ExtraCash originations reached $1.8 billion, up 51% year-over-year and 17% sequentially. This represents a new high for the company and reflects both growth in monthly transacting members and an increase in average ExtraCash size. We ended the quarter with 2.6 million monthly transacting members, up 16% year-over-year and 4% sequentially, with both growth rates representing accelerations compared to the prior period. We saw continued gains in new member conversion and dormant member reactivation along with strong retention, all positive signals of strong and consistent demand and the durability of our value proposition. The average ExtraCash origination size in Q2 increased to $206, up 24% year-over-year and 7% sequentially. This growth reflects improved credit segmentation enabled by CashAI, the impact of our new fee model driving higher ExtraCash approval limits and the natural increase in origination sizes as our member base [indiscernible] on the platform. We view this as a win- win, driving higher ARPU for the company while also enhancing our ability to meet our members' liquidity needs. On credit performance, our 28-day delinquency rate increased by approximately 37 basis points year-over-year. A third-party issue, which has since been resolved, resulted in a temporary delay in settlements affecting a limited subset of our ExtraCash receivables. This temporary delay impacted our 28-day delinquency rate in Q2 by an estimated 19 basis points or 9%, implying a delinquency rate of approximately 2.21% had this issue not occurred. Excluding the estimated impact of this issue, the 28-day delinquency rate would have increased roughly 18 basis points year-over- year, which remains within our internal guardrails and aligns with our strategic focus on maximizing gross profit dollars rather than minimizing the loss rate. On a sequential basis, our 28-day delinquency rate also increased as a result of this third-party issue in addition to seasonal normalization following Q1's tax refund season. ExtraCash [ leverages ] CashAI, our proprietary underwriting engine, which enables near real-time identification of credit risk through fully automated analysis of bank account transaction data. Combined with ExtraCash short repayment cycle, this tool creates a rapid feedback loop for optimizing underwriting. This agile framework gives us strong confidence in our ability to manage credit risk across a range of economic scenarios. We're now in the testing phase of our CashAI v5.5, the latest evolution of our underwriting. This next-gen model is designed to fully incorporate the economics of our new fee structure while introducing additional variables to enhance precision. The new model is trained on more than twice the number of features that we used to train our current v5.0 model, which we believe bodes well for future credit performance. We expect to begin deploying the v5.5 model later this year. The third pillar of our strategy is deepening engagement and monetization through Dave Card. In Q2, total card [ sum ] reached $493 million, up 27% year-over-year, reflecting growth in transacting members, increases in card spend per active banking customer and continued synergy between ExtraCash and Dave Card usage. A significant portion of ExtraCash originations continue to be dispersed to the Dave Card. This integration improves member convenience, reduces member cost and strengthens member engagement within our financial ecosystem. Active Dave Card users tend to exhibit stronger retention on ExtraCash and drive higher lifetime value, benefiting both from increased product stickiness and the incremental ARPU associated with the Dave Card usage. As our ecosystem has expanded in value, we have been testing a new monthly subscription price point after nearly 8 years of charging $1 per month. Following several months of testing, we completed the rollout of a $3 monthly subscription fee for all new members. Customer results validated that we could implement the pricing change with minimal impact on conversion or retention and the higher price has proven to be accretive to lifetime value. Our current plan is to grandfather existing MTMs onto the existing $1 price for now. The new monthly fee impact in Q2 is modest as the change was fully implemented in mid-June. We expect a growing contribution in the quarters ahead as an increasing share of our MTM base is acquired under the new monthly pricing structure. Switching gears a bit. Given the importance of cash flow transaction data to CashAI, I want to briefly address recent headlines surrounding the dispute between JPMorgan and open banking data aggregators over potential fees for access to consumers' financial data. First and foremost, we believe it's not a foregone conclusion that prices will increase. We've been encouraged by the strong response from the industry trade groups, policymakers and other key stakeholders who have stepped in to defend consumer rights to free data access. We're also pleased that the CFPB has indicated it will revisit the issue and fast track the resolution. Second, in the event fees do increase, we believe Dave is well-positioned to significantly optimize our use of data while continuing to maintain our existing member experience and business performance. Lastly, given our scale and demonstrated pricing power, we would expect any potential incremental cost to be shared across all stakeholders, further minimizing the potential impact on our expenses. I'd also like to provide an update on our strategic partnership with Coastal Community Bank, which is assuming bank sponsorship for Dave's ExtraCash and banking products from our existing provider. Last month, we began onboarding new members on to Coastal in line with our previously communicated timelines. This marks a key milestone in strengthening our banking infrastructure, adding the risk management rigor and scalability needed to support future product expansion and our broader growth ambitions. Additionally, as Kyle will describe in greater detail, we recently completed an amendment to our program agreement with Coastal, whereby over time, Coastal will serve as the primary funding partner for ExtraCash receivables, which should further unlock the capital efficiency of our business model. More on that in a moment. In closing, Q2 represented another step function change in our profitable growth trajectory. I want to thank our team for their tireless dedication to delivering outstanding value for our members and shareholders. With that, I'll turn it over to Kyle.

Thanks, Jason. Q2 was another record quarter, highlighted by accelerating revenue growth, continued margin expansion, disciplined marketing spend and increased operating leverage. These factors collectively drove outsized growth in adjusted EBITDA, further underscoring the strength of our business model. Let me walk through the financials in more detail. Starting with revenue. Total revenue was $131.7 million, up 64% year-over-year and 22% sequentially. Growth was driven by a 16% increase in MTMs and an acceleration in ARPU growth of 42% to [ 200 ]. These metrics reflect a full quarter of monetization from our new fee structure, larger ExtraCash sizes and deeper member engagement across ExtraCash and Dave Card. Before turning to expenses, I want to note that we've expanded the view of certain operating expense line items on our P&L to help provide greater transparency into our cost structure, specifically by distinguishing between variable and fixed components. Under this revised classification, variable costs include provision for credit losses, processing and servicing costs and financial network and transaction costs. In this new view, readers can directly reconcile total revenue to non-GAAP gross profit using specific line items on our statement of operations. Compensation and benefits, technology and infrastructure and other operating expenses represent our fixed operating costs. Finally, advertising and activation costs reflect our previously reported advertising and marketing line item and now include member activation costs, which were previously a part of processing and servicing costs and other operating expenses. It's also important to note that only the advertising and marketing components of this line are used to calculate our customer acquisition costs, given that activation-related expenses may apply to both new and existing members. With that framework in place, let's walk through each of the operating expense categories. During the second quarter, our provision for credit losses was $25.2 million, up approximately $10.8 million year-over-year, primarily due to increased origination volumes, which grew 51% over the same period. Additionally, as Jason mentioned earlier, a third-party issue, which has since been resolved, caused a temporary delay in settlements affecting a limited subset of our ExtraCash receivables. The estimated impact of this issue was approximately $3 million in Q2, which is reflected in the provision for credit losses. Excluding this impact, provision for credit losses would have represented 1.2% of originations, roughly in line with the year-ago period and consistent with our plan to manage credit performance to maximize gross profit dollars. It's worth noting that provision for credit losses was also up on a sequential basis as expected, given the favorable repayment trends we experienced in the first quarter as a result of tax refund season. We anticipate provision for credit losses as a percentage of originations will reach its high point in Q3 since the quarter ends on a Tuesday, which is typically the intra-week peak for outstanding receivables. The higher gross receivables balance will itself cause the provision to increase regardless of any potential changes in credit performance. Using Q2 as an example, had the quarter ended on Tuesday, July 1, our provision for credit losses would have been approximately $1.7 million higher. Had it ended on the Friday prior to quarter end, it would have been approximately $4.5 million lower. This illustrates that a 4-day difference in the day of the week on which the second quarter ended could have driven a variance of over $6 million in our provision for credit losses. Processing and servicing costs decreased 4% year-over-year to $7.2 million, driven primarily by efficiencies gained from two significant vendor contracts renegotiated last year as well as the scale economies inherent in most of our processing vendor contracts. As a percentage of ExtraCash origination volume, these costs improved to 0.4% from 0.6% in Q2 of last year. Financial network and transaction costs previously included as a component of other operating expenses increased 11% year-over-year to $7.2 million, which was largely attributable to increased Dave Card spending volume. As a percentage of revenue, financial network and transaction costs decreased to 5% from 8% in the year-ago period. This brings us to non-GAAP gross profit, which we previously referred to as non-GAAP variable profit, which grew 78% year-over- year to $92 million. We've changed the name of this metric to better align with industry norms, though the definition and calculation remain the same as in prior disclosures. Non-GAAP gross margin, which we previously referred to as non-GAAP variable margin, came in at 70% for Q2, in line with our expected gross margin range of high 60s to low 70s that we outlined last quarter to reflect credit performance normalization following tax refund season. Relative to last year, our gross margin expanded approximately 500 basis points as a result of processing cost optimizations and key vendor renegotiations. Advertising and activation costs increased 20% year-over-year and 30% sequentially to $15.5 million. We typically moderate marketing spend in Q1, which tends to be less efficient given that tax refunds reduce our members' liquidity needs. In Q2, we ramped investment to capitalize on continued strong demand for ExtraCash and to take advantage of the stronger LTV to CAC returns we've unlocked through the new ExtraCash fee structure and the higher subscription fee. Looking ahead, we plan to continue increasing marketing investment throughout the remainder of the year as our outlook for new member growth and lifetime value expansion remains strong. More specifically, we expect year-over-year growth in marketing spend in Q3 and Q4 to track at or above the pace we observed in Q2. Compensation-related expenses rose 9% year-over-year to $26.4 million. As a percentage of revenue, compensation expense declined to 20% in Q2 from 25% last quarter and 30% in the year-ago period. Additionally, our annualized run rate revenue per employee expanded 66% to $1.9 million, up from $1.1 million in Q2 of last year. These improvements highlight the scalability of our business model and the productivity gains resulting from our investments in AI and our broader technology platform. Technology and infrastructure expenses and other operating expenses, which primarily consist of platform compute infrastructure costs and third-party software expenses increased 3% and 1% year-over-year, respectively. Over the same period, revenue grew 64%, further underscoring the scalability of our platform. During the quarter, we recorded noncash expenses from mark-to-market changes in the value of the earn-out and warrant securities that are outstanding. The $7.9 million earn-out expense this quarter reflects the higher value of those potential shares, while the $20.5 million warrant expense is tied to the increased value of outstanding warrants, both driven by the strong performance of our stock and warrant prices. To be clear, these are noncash expenses and not a reflection of the underlying business fundamentals. That said, we anticipate some volatility in these figures as our stock price changes in the future. GAAP net income increased 42% to $9.1 million from $6.4 million in Q2 of last year. Our year-to-date effective tax rate was approximately 17%, and we estimate our 2025 annual effective tax rate to range between 19% and 21%. Adjusted net income, which excludes nonrecurring items, stock- based compensation and noncash fair value adjustments to the warrant and earn-out securities increased 233% year-over-year to $45.7 million. Similarly, adjusted EBITDA reached $50.9 million, more than tripling compared to Q2 of last year, with flow-through from gross profit to EBITDA of approximately 90%. Turning to the balance sheet; we ended the quarter with $104.7 million in cash and cash equivalents, marketable securities, investments and restricted cash, up from $89.7 million at the end of Q1. This $15 million increase was attributable to free cash flow generation, offset by an increase in the ExtraCash receivables balance, which on a gross basis increased by $43.4 million over the last quarter. As Jason mentioned earlier, following the recent amendment to our program agreement with Coastal, we expect to move a significant portion of our ExtraCash receivables off balance sheet. We believe this shift will meaningfully reduce our direct funding obligations, lower our cost of capital and unlock substantial liquidity to pursue capital allocation opportunities going forward, all while allowing us to eliminate the warehouse line debt from our balance sheet by mid-2026. In addition, the new arrangement provides a total funding capacity of $225 million, representing $75 million more capacity than our current credit facility. We anticipate beginning to transition ExtraCash receivables under the new program by early next year. From a capital allocation perspective, we remain focused on flexibility. Our priorities continue to be reinvesting in organic growth opportunities to drive future growth, increasing our dry powder to facilitate potential M&A and opportunistically returning capital to shareholders via share repurchases. Given our strong performance through the first half of the year, we are once again raising our full year outlook. We now expect revenue of $505 million to $515 million, up from our prior range of $460 million to $475 million and adjusted EBITDA of $180 million to $190 million, up from our prior range of $155 million to $165 million. The midpoint of our revised outlook implies annual revenue growth of 47% and adjusted EBITDA growth of 114%, and we continue to expect gross margins to be in the upper 60s to low 70s for the remainder of the year. We're proud of the financial and strategic progress we've made in the first half of 2025. We're delivering durable growth, expanding margins and innovating for the benefit of our members. With continued focus on execution, we're confident in our ability to create long-term shareholder value while advancing our mission to build a better banking experience for everyday Americans. And with that, we'll open the line for questions.

[Operator Instructions] Our first question comes from Devin Ryan of Citizens Bank.

Really nice results. I guess first question here, just on the new -- well, I guess, the transition to the fee model that you just made. I'm just curious because obviously, you executed that through the second quarter. Is there any remaining benefit related as we look forward to that kind of transition? And then as we look out over the next couple of years, can you maybe just talk a little bit about how you see revenue per advance trending? It just kind of continues to move higher. But what levers do you have to drive that higher? How much room is there to drive average advance sizes higher? Is there room maybe over the intermediate term to still tweak the fee model because it sounds like you still have pretty good pricing power. So I just want to think about kind of the trajectory there, but then also even short-term here, is everything already baked into the 2Q?

Yeah, Devin, so I'd say if you recall from our last earnings call, we mentioned that March was the first full month of the new fee model. And so we did get the full benefit of the new fee structure in Q2 here. That said, if you think about future monetization, we're rolling out our new v5.5 model so do expect our ability to keep growing originations per user and earnings given our pricing power and spreads there, feeling good about the ability to keep growing with the member base.

Got it. Okay. And then just a follow-up. I just want to come back to the point on moving the receivables to Coastal. I think that's pretty interesting, obviously, kind of freeing up capital as well. So can you just tell us what the direct financial impact of that is? Like what's the cost of that relative to the current arrangement? And then as you think about freeing up capital and you have more excess capital and the company is obviously generating a lot of excess capital as we move forward here. What are your priorities there? What type of opportunities are you looking at with excess capital?

Yeah. Thanks, Devin. So yeah, I mean, there's a couple of impacts to the financials. First and foremost is that the overwhelming majority of our receivables will go off balance sheet and will just be held at the bank. We do -- we are, as part of that arrangement, paying Coastal for the balance sheet usage, but it is a 200 basis point reduction relative to our current cost of funds with our existing warehouse line. So awesome sort of win-win there for us and Coastal as part of that partnership. But yeah, you're right. It is going to free up a substantial amount of cash. Our expectations of well north of $100 million, including paying down the existing warehouse line. And as I talked about, our approach to capital allocation right now is just maintaining flexibility, and we do want to position ourselves to take advantage of opportunistic M&A. So we want to have sufficient dry powder, and we'll look at share repurchases and capital return alternatives as well. But like I said, our approach is to kind of build up some cash right now to give ourselves dry powder.

Yeah. Okay, great. And then if I can just sneak one more in here. The $3 monthly subscription that you're moving to with new members, how much data do you have on that in terms of how that's affected customer acquisition and then even how customers behave once they're on the platform in terms of repeat use? And then as you kind of bump up the subscription level, do you plan on adding more services or anything else into that where maybe there's features that you could drive incentives to get more usage of other products like the Dave Card or something else? Just trying to think about kind of the evolution here because you haven't touched the monthly subscription at some time.

Yeah, Devin, as mentioned, we've had the same $1 fee since the company launched in 2017. So it definitely it was time for a refresh given all the increased value we delivered to the customer. As mentioned, the dollar fee will be grandfathered in for existing users. it is $3 for new members starting in June. So the full Q3 will have the benefit of the new fee structure for new customers. We're feeling good about the value we're delivering. If we think about new features that we can add in here, it really would be things we think are incremental to drive more retention of the subscribers, not necessarily things we feel like need to have increased monetization for the $3.

But, just to add to that, we -- we're not at all in a rush to get this subscription pricing change out. We've been testing extensively over the last couple of quarters to assess impact at the top of the funnel as well as subsequent period retention and the like. And we just saw nothing negative really on that side of the equation, which gives us -- gave us a lot of confidence that we could roll this out and sort of maintain the conversion and retention characteristics. So the -- that we had previously. And so the pricing change is effectively fully accretive to lifetime value. Beyond that, I think the higher sub fee does give us a promotional lever that we didn't have before, at least not as powerful one. So we could think about waiving the fee for certain activities within the app and just create -- it's just more of a promotion lever that we didn't have in our toolkit before. So excited to put that to work in different use cases.

Our next question comes from Joseph Vafi of Canaccord.

Yes, great results here once again here in Q2. Just maybe we kind of double-click a little bit on that third-party issue that drove up the delinquencies a little bit, if there's any other color to provide there and perhaps what occurred there and if there's any measures you've taken to make sure that, that doesn't happen again. Just if we could get a little more color there, that would be great. And I have a quick follow-up after that.

Yeah. So thanks, Joe. So basically, what happened was sort of a reporting issue on a small set of the receivables that caused us to delay settlements or repayments on those receivables. We sort of caught it as part of our audit process, and we put in additional steps to ensure that something like this doesn't happen again in the future, and we feel confident that we've sort of closed that gap. But yeah, it basically caused a delay in the collections on the receivables, which ultimately has a sort of impact -- negative impact or adverse impact on the ultimate collectability of those receivables, which is about a $3 million adverse impact in the provision. So that's sort of a high-level summary. But again, we feel really good about our -- the steps that we've taken to ensure that something like that doesn't happen again.

Sure. And then if we kind of look at the rollout of the new AI engine, maybe we get a little more detail there in terms of, I guess, we're going to be, I guess, focused on size of ExtraCash advances as a driver of more ARPU using the AI engine and then maybe how you're looking at that double set of -- the doubling of the set of data points that you're evaluating, what that may mean also in your delinquency? If you can share any kind of initial thoughts on how that might help both on size and on delinquency.

Yeah. So effectively, with every new model release, we're looking at areas for just better risk splitting. And so moving good risk up higher in the limit curve, increasing the value prop for those users and finding pockets of bad risk that we downgrade or kind of eliminate from the portfolio altogether. And that's certainly what our simulations are suggesting where the ultimate kind of impact of that is higher overall average origination sizes per customer, but also lower delinquency rates, so kind of a win-win for the business there. We just started testing the model about 1.5 weeks ago, but the simulations, like I said, indicate both the upside performance on delinquencies as well as average origination size, which is -- supports overall levels of increased net monetization for us.

Our next question comes from Jeff Cantwell of Seaport Research.

Congrats on the results. I just want to focus in on the updated revenue guidance you provided for the full year, which was taken up to $505 million to $515 million. Can you just explain where the incremental enthusiasm is coming from for you guys as far as your outlook for revenue? In other words, is the raise due to the updated subscription fee or is it greater ExtraCash demand, et cetera? Can you maybe break that out for everyone or how should we be thinking about it?

Jeff, so I think we're still very bullish on our new member adds. As you saw, we added 722,000 new members in Q2. We expect to keep ramping up marketing given the efficient trends we're seeing in our paybacks. As noted, our payback periods have improved just to 4 months as we've improved LTV with the new fee model and improvements in retention. And so if you factor in both new member acquisition, improvements in ExtraCash with respect to the spreads as well as the new fee model, we're feeling very good about the update to the guide.

Got it. Okay. I wanted to circle back to what you spoke about in your prepared remarks with regards to data aggregator fees. Can you just explain how they work for you guys currently in terms of who bears what cost between yourselves and the aggregator for any data? And obviously, it sounds like there's a lot of variables potential outcome. So my question is, do you have any range of estimates at this point as to how this might impact your P&L or is this simply not going to be material? Just want to see if you can help us out on that front.

Yeah, Jeff, I'm just going to point back to my comment I made in the script really that we think that there's a few things going on here. One, this is being thought on the policy level. So we're happy to see the CFPB step in here because we do not think it's a foregone conclusion that prices will be going up for us. We also have, we think, significant pricing power, both with respect to how big Dave is, so our pricing power with our aggregator partners as well as our pricing power with consumers. And so we don't expect even in a world where prices do go up that we would bear the 100% cost of that increase. Overall, we're feeling good about our position. And in a world, again, where fees were to go up, we feel like we have a major lever to significantly optimize our data aggregation in general and that we don't need to pull as much data to feed cash [ out ] that we are right now and still continue to make major strides in our performance.

Okay. Understood. And then lastly, just to squeeze one more. How are you thinking about M&A? You mentioned increasing your dry powder and perhaps considering being opportunistic. Can you maybe give us some thoughts, high-level thoughts on what would help you build the Dave ecosystem out further?

We really think about M&A in two ways. One is, can it increase distribution to expand and diversify customer acquisition for the company and/or can it be something that can be accretive to ARPU for the existing member base, both as a bolt-on for new subscription products and/or for expanded credit products. The company continues to assess opportunities in the market, not near on anything right now, but we're continuing to keep our ear close to the ground and having conversations.

Our next question comes from Mark Palmer of The Benchmark Company.

During the second quarter, what portion of the ExtraCash advances that were extended were extended to those who had already been on the platform? I know last quarter, that figure was in the high 90s percent. Was that consistent with what you saw in the second quarter as well?

Mark, thanks for joining. Yeah, pretty consistent trends there. I mean, as we ramp up user acquisition in any given period, you tend to see that number come down a little bit, but we're in the, call it, 95% to 96% of units originated to existing customers. And as you might imagine, the average limits for existing customers tend to be higher than brand-new customers. So if you look at it on a dollar basis of originations, it's, call it, 97% to 98% of originations to repeat customers. So no real change in the dynamic there. The overwhelming majority of our base is repeat usage, and that's something that we continue to see as we've made improvements to retention and customers are sticking around longer, that number has ticked up over time.

And with regard to the average ExtraCash advance size, obviously, now with the mandatory fee structure, you have more incentive to increase the average size of the advances. Any observations from the first full quarter of the mandatory fee structure in that regard? And how do you see the average loan size trending over time? How should we be thinking about just how much potential increase we could see over the next number of quarters?

I mean as we talked about last quarter, we had done a significant amount of testing before we implemented the new fee structure just to understand customer impact and receptivity to the change. And I'd say in Q2, things just played out accordingly as we expected them to. So nothing really to point to as far as surprises, just kind of coming in line with our expectations. As I mentioned and we've talked about, we are rolling out our new underwriting model, which has better -- indicated better risk splitting capabilities, and we would expect, based on that to be able to continue to drive average origination sizes up over time and we feel good about our ability to do that. We haven't provided any specific guidance around what we would -- what we expect that to look like, though we do expect that to be a lever for ARPU expansion moving forward.

[Operator Instructions] Our next question comes from Jacob Stephan from Lake Street Capital Markets.

Congrats again on a nice quarter here. Maybe just touching on the kind of $3 per month sub fee. I know you talked a little bit about this, mentioned an increase in LTV, but maybe could you help us understand retention metrics a little bit? I know you've only had these customers for two months. And is the LTV uplift, is that all from the increase in sub fee or are customers more inclined to use it given that they're paying more? Any help there?

Well, the LTV expansion has been a multiple of things, right? It's the ExtraCash origination volume going up per user. This is, I think, mostly helpful in how fast the paybacks are being generated. So we are paying back customer acquisition in now 4 months, which is, I believe, a record low for the business. So really excited to see that. As far as the retention and conversion dynamics, we've been monitoring the situation since last year. And so there's been a significant amount of testing we've seen to really make sure this is the right price point. And we were testing everything from $0 all the way up to $5 to really test the efficacy of the pricing and $3 was really the sweet spot to really see no impact to conversion or retention. And so that does really have a positive impact on LTV and also the shorter payback periods.

Maybe reiterate that. Reiterate that point is we have done extensive testing on this new subs price point to ensure that we've not rocked the boat either from a conversion perspective or retention and feel really good about the results that we've seen from a testing and supported us putting -- implementing the fee as we've talked about.

Okay. And then second one, I just wanted to touch more on CashAI, 5.5 here. It sounds like you've trained it for the new fee structure. Maybe you could help us understand what's different about the model now and also provision for credit losses has improved so much already. Where can it actually -- where can it go to?

I mean it's just as I talked about, double the model features of the prior model and the goal is to just have better risk splitting capabilities, which we feel like we've accomplished with this latest version of the model. So feeling good about driving originations up from here while also finding some opportunities to cut bad risk to drive delinquency rates down. So we expect that to play out with the new model release and like I said, increase -- have the potential to drive ARPU expansion as a result of the new model getting into production sometime later this quarter. In terms of the provision, as we've talked about in the past as well, our goal isn't to drive the lowest levels of loss rate is to maximize gross profit and LTV for our cohorts. And we think this latest quarter is kind of demonstration of having a slightly higher level of provision, but also driving gross profit dollar expansion at a very attractive level. And we would expect that dynamic to continue to play out where loss rates are sort of in a healthy place at this level, and we can now optimize the model around the new fee structure and drive just gross profit dollar expansion off of this new base.