Okay, good morning and welcome to the Beforepay Q2 2025 Investor Webinar. My name is Danny Younis, and I handle investor relations for Beforepay. With me this morning, we have the CEO of Beforepay, Jamie Twiss, and we also have the acting CFO, general counsel, and chief risk officer, Elena Chan. Before I hand over to Jamie and Elena, just a note: we will be having a Q&A session at the end of Jamie and Elena's commentary, so if you have any questions, you can type them into the Q&A box on the screen. I would now like to hand over the conference to Jamie and Elena. Please go ahead.
Thank you, Danny, and thank you for everybody who's joining us today. We appreciate you taking the time to attend the Q2 webinar. So we put the release out on the ASX, and you can download it from there or from our website. It would be the numbers that we just released. Look, at a high level, I think the most important takeaway for me is this is just another fantastic quarter for us. I was incredibly pleased with the results that we posted in Q2 of FY25. We had great growth, and we had great profitability. So we got the two key things we were aiming for: great credit outcomes, we were super tight on costs. So really, pretty much every number was as good as we could have hoped for in this quarter. So feeling really, really pleased about how we performed.
I think it's just yet another sign that the business is in a very strong place. This is our sixth quarter in a row of profitability. These are unaudited numbers given it's quarterly. We'll be putting out our audited tax year numbers shortly. So really, really pleased with where we've been. I'll call out a couple of highlights, and I'll hand it over to Elena to go through the details. So probably the single biggest number that jumps out at me from this result is the profit before tax number of AUD 1.1 million for the quarter. That's a record for Q2 in any year, and in fact, it's almost double. It's 95% above where we were last year at this time, so 95% year-on-year increase.
What this means is over the first two quarters of this year, again, on an unaudited basis, we've almost matched the entire full-year profit from the previous year, from FY24. So in terms of a return standpoint, even as we invest in our new growth initiatives, business is absolutely firing on all cylinders financially. So really, really pleased with that profitability outcome. And then also very pleased with a couple of big growth milestones that we've reached. So this is the Q1 where we broke through the $10 million quarterly revenue mark, and that comes off the back of another milestone, getting more than $200 million in pay advances and personal loans issued over the quarter as well. So very pleased with that top line, and that's driven by continued healthy growth in our active user number, which has now surpassed 257,000 active users as of December 31, 2023.
And then finally, our margins were extremely good for Q2 given the seasonality in our business, a 1.75% net default rate. It's a very strong outcome for us throughout that holiday period. So that was the key deliverable that in turn powered those good margins and that great profit outcome. So in short, we're very, very pleased with how the numbers from this quarter went. Elena.
Thanks, Jamie. Again, I'll only reiterate that I've been very pleased with the results that we've been able to present to you today for Q2 FY25. What I'll do is give you a little bit of a meat around the bones of the numbers we've presented. Overall, what I'd say is that, again, that sixth consecutive quarter of positive net profit before tax is something we're very proud of. Our trading has been in line with our expectations as well as the seasonal patterns that we have been expecting. That really is driven by our quarterly advance numbers, which is the AUD 206.5 million total advances in the quarter, which is slightly offset by our seasonality in terms of the higher net default number, which is 1.75%.
That all leads to an increased net transaction margin of $4.8 million for the quarter, which is, when you look at what we were doing in Q2 FY24, which is $4.4 million. We're very pleased with that result, needless to say. Also, really happy to announce a record number of active users. As Jamie said, we have surpassed the 257,000 user mark, as well as thinking about the way our active users have been growing. That's a 3% quarter-on-quarter increase or a 9% year-on-year increase, and we're very pleased with that. In terms of our advances, that's increased by 8% quarter-on-quarter, and that $205.6 million has been predominantly driven by an increase in the total number of advances that we're making, but also the size of the loan on average. And so the average loan size is now at that $398 mark.
Our revenue was also up by 7%, and that's on a quarter-on-quarter basis, as Jamie mentioned. That was a record for us in terms of surpassing AUD 10 million to get to AUD 10.2 million in revenue in Q2 FY25. Then when we look at sort of the revenue against the net default and seasonality, we have a 1.75% net default, which is predominantly driven by seasonality and in line with our expectations. We're still very pleased with our very strong credit performance and our strong recovery management, and that's a great reflection of how we're progressing in the business more generally. That all led to an NTM, as I've mentioned before, which has gone up year-on-year, but obviously quarter-on-quarter has dipped down predominantly due to the seasonality of net default.
Our operating expenses for the quarter was at 3.6 million, which is down from the 5.3 million that we saw in Q1 FY25. That's predominantly driven by the one-off adjustment that we had in Q1 FY25 relating to a share-based payment adjustment. What we're seeing for our operational expenses, of course, is number one, that we continue to invest in our growth, and we're seeing gentle increases in those, and we continue to expect gentle increases as we invest in our strategic priorities. We're also seeing that in terms of our core product, our cap maintains at that $40 mark.
In last quarter, it was $39, so roughly the same, but we're seeing good growth coming from that cap, so we're very pleased with that. Our reported cash at bank number was AUD 18.3 million, and then that excludes the AUD 4.8 million that we hold at a third-party settlement to fund our future growth in loans.
And so if you add those together, that's a really nice AUD 23.1 million cash position that we have. In terms of funding, we have undrawn capacity of AUD 16.6 million, which allows us to really fund and have lots of capacity to fund future growth as well. I think overall, what I'd say, that culminated in a AUD 1.1 million net profit for Q2 FY25, and we're really, really pleased with that result. And I think we're very positive about the way that the business is going. The business is responding to all of the different customer demands, and I might actually hand over to Jamie for a quick segue to talk about some of our progress on strategic initiatives.
Excellent, so you've heard about the core business, which again is essentially the entire economics of the business at this point. The top line is driven by that pay advance product, but as many of you would know, we have two significant strategic initiatives that are underway. The first one being Carrington Labs, the second one being our personal loan product. With Carrington Labs, we're delighted that we have a client in the U.S. fully up and running and operational in Q2, so that's fantastic news, and we continue to feel very comfortable with the pipeline that we've got. It's a pretty compelling product that I think is showing a lot of interest, so I think we feel really good about where that one is at, and when we have additional announcements to make, of course, we will make them.
On the personal loan side, as many of you would be aware, we've launched that product in Q2, so those first loans have been issued. As we said at the time, we're currently going through a thin testing phase in order to ensure that we're comfortable with the credit and that it performs in the way that we expect, not just in terms of the overall default rate, but also looking at the detailed analytics of those loans and are they performing in line with the way that we expect, which is actually, I think, even more important than the overall default rate. So far, everything looks pretty good, but we continue to be in that testing phase, so it's not that material at this point. But of course, when we're in a position to start to grow that business more significantly, we will communicate clearly around that as well.
And so that's where we stand. So I think very happy with both our strategic initiatives and extremely happy with the performance of the core business this quarter. We are delighted with our ability to put out solid, clean profit numbers every quarter alongside solid growth, and I think we feel very confident about the future.
Okay, thank you, Jamie and Elena. We will now move to the Q&A session. Once again, if you have any questions, please type them into the screen right at the bottom there that you can see in the Q&A box. We do have questions coming through. We've got four or five coming through right now, so I'll start on those. So the first question, Jamie, there seems to be an acceleration in active user numbers. Have you made any changes to your marketing strategies or upped the spend?
So let me start by talking about how we approach this overall. The short answer is we've made lots of minor tweaks, but nothing terribly different in this quarter. The way that we think about this is we think, what's the marginal value of one additional customer? And that will depend on who the customer is and the channel they've come through and the risk profile and so on. So we think about what's the expected lifetime value of that customer. And then on a marginal basis, not on an average basis, and that's a very important distinction that platforms aren't very keen for people to make. But on a marginal basis, we want to pay meaningfully below that lifetime value. We'd like a reasonable payback on the target, say, a one-year payback or faster on that.
So what we're always watching is for that last dollar spent on marketing, do we think we've got an adequate return on that dollar? And we will keep spending as long as that is the case, and then when we feel like we're not getting that adequate return, we stop spending. So in practice, that means that the marketing spend in any given week, month, or quarter, or even day really is an outcome of our ability to ensure that each one of those dollars is working hard for us and being productive. We've not made a conscious decision to try to spend more on consumer marketing, but in periods of higher growth, that number will creep up. That's because what we're very focused on is that marginal cost of acquisition.
In terms of specific changes that we've made, I think what we find is we've gotten, I think every month we are getting sharper on, particularly around Google, how to think about how those campaigns work, how to think about budget management, how to disaggregate your average CPA cost per acquisition into a marginal one, which is really important. So is that marginal cost very flat and then suddenly it goes up sharply when Google runs out of likely customers, or is it sort of a steady, smooth hill climb? And then we manage that budget accordingly to manage towards that marginal CPA outcome that we're happy with. I think we've continued to get sharper on how we did that, and we made some tweaks and changes here and there. I think Facebook is an interesting one as well.
We've been spending more time looking at how those campaigns work, similar questions around targeting and what do you put in front of people, thinking about the different types of creative and how you match that with audiences. So lots of little tweaks, but no big strategic shift in how we approach consumer marketing.
Okay. The next question is around default rates. So congratulations on another solid result. I noticed the seasonal default rate was in line with last year's results. What are your expectations for recovery rates, and will this be reflected in the next quarterly results?
Thank you for the kind words and for the question. This is a net default rate, so it includes the recoveries received in this quarter. The way we do it is we provision for loans that are outstanding. We fully provision any pay advance that has reached 62 days and moves into default. We have estimates based on in-quarter loans extended, and then that's netted off against actual cash received in the quarter, which could be from a loan of any duration. I think we would expect the sort of recovery performance we've seen to broadly continue. We haven't made any big changes in those recovery rates.
If you look at the spread between gross and net defaults, you can see that we've gotten better over time as we have in virtually everything, but they tend to fluctuate a little bit but not all that much. So I think we should expect that looking forward into Q3, we will see those holiday loans coming through, and so that's part of the natural seasonality of the business, and so there will be a reasonable level of defaults off the back of that. And then we'd expect a broadly consistent recovery performance. There is some lag in that, but to be honest, the effect of that lag is a little bit lost in just the rest of it.
Okay, thank you. Now there's a couple of questions that have come through on Carrington Labs, and they're nested with multiple questions within, so I'm going to ask these probably one after the other, Jamie and Elena. So on Carrington Labs, the first one is, is the client that came on board a paying customer, or are they under trial?
So it is paying customer.
Yeah. Beforepay, do we have any indication of the timeframe and the sort of revenue this will represent annually, or what the growth strategy forecast is for Carrington Labs?
So I think on the first one, we haven't put out any specifics, and we probably won't put out any specifics around any individual client arrangements. I think at this point we've said that the overall Carrington Labs revenue is not that material to the group as a whole, and at some point when it becomes material, we'll talk about it and disclose it accordingly. But in terms of the growth strategy, I think we, as I said a moment ago, we feel very confident in the quality of the product. And not only that, we feel very confident in really three things. One, the quality of the product. Two, the distinctiveness of the product. And as I've said in the past, this was almost a bit of a surprise to me that there weren't more offerings out there that felt competitive with what we do.
The actual building of AI and machine learning-powered models for lenders is a field much talked about but very rarely actually participated in. And so the nature of competitors in the space is pretty thin on the ground, and I think we can take them. And then I think the other thing, which we expected to be the case, is it's a good time to be out there offering AI/ML solutions to credit risk. I think just we are in the right place at the right time on that one. So we feel extremely confident about the product. As we've said in the past, these tend to be slow sales cycles. There's a complex enterprise software sale often to a bank or other highly regulated entity.
The thing that we're very focused on is do we feel good about the pipeline and how we are working through that pipeline? I think the answer to that is unambiguously yes. I think upon reflection, we're probably at the point of confidence where we can push a bit harder and put a bit more resource into getting those deals through the pipeline. As we've noted in the 4C, we will be looking to selectively add some headcounts to help us accelerate that.
Okay, continuing on with Carrington Labs. The report mentions Carrington Labs completed a fully implemented solution for a U.S. client. Could you clarify?
It just means that it's up and running now with live animation, so client is seeing it in a few moments.
A follow-on from that, what is the average end-to-end timeline from initial client engagement to full implementation?
I can say the fastest one is probably six to nine months-ish, depending on how exactly you define. The longest time is probably infinite with some of these larger banks. But I think there are some lenders who are kind of decisive and relatively nimble, especially those that are a bit more startup year, a bit more fintechy, and then they're large banks. This could be a multi-year process, certainly.
Yeah. You sort of answered the follow-on there just then, but what factors most influence the duration of implementation? For example, customization, regulatory compliance?
Yeah, so regulatory compliance, so the way the product is designed is one where we don't, for a given implementation, be disappointed if we had to do a lot of compliance work on the product. The product is an analytic software product, and obviously if the client needs to think about it from a certain way in terms of compliance, we'll, within a reasonable time, be accommodated with that. But it's not the case that there's a lengthy licensing process or anything like that with these. I'd say the biggest two factors, one is just the natural speed of client decision-making, and these are at the very core of what a lender does, and so I think understandably a lot of them really want to kind of be thoughtful about that, and many of these are quite large organizations with many decision steps to get through.
The other challenge, which ties back to what I said a moment ago, is just our capacity to actually move folks through the pipeline. So I think what we've found, and we've said this in the past, is that it's relatively easy, almost too easy, to fill the very top of the pipelines. People are very interested in talking about this. You can start a conversation, but at a certain point in the process, then you have to do some real work about actually ingesting their data. The model training itself is fairly automated, but you still have to kind of run the wires from them to us and so on. Our team that does that is quite small, and so the decision we're thinking about is, should we invest more to try to move folks through that pipeline a bit faster?
Yeah. The questions are still coming through on Carrington Labs. I will get to the other questions as well. So the next question on Carrington Labs, as you activate more clients for Carrington Labs, do you think the revenue, as well as the margins associated with it, would end up surpassing the core product? If so, how do you think about capability and resource allocation?
That's a great question. So I think if you just define the core product as pay advance versus defining the core product as Australian domestic lending, including personal loans, because that's a huge market, you might get different answers. I think what I'd say is I don't think we have a firm point of view on whether one or the other will be larger at any point in time. The Australian domestic consumer lending market is a huge market. The global banking market is even larger. So depending on the level of penetration Carrington Labs gets to, it could be a very, very significant business indeed. Now, what the timeframe is and how that compares to the balance sheet product, I think that's hard to say. But we've launched Carrington Labs because we do think the product market fit is very, very strong.
The quality of the product is very strong, and it is a very, very large market, so we do see quite significant potential for that business over time.
Okay. There's a couple more on Carrington Labs. The next one is, have clients observed measurable improvements in default rates or other credit performance metrics since adopting Carrington Labs' solutions?
I'd say it's probably too early to say that tentatively, given it's still early phase. What I can say is that, as a general rule, if you talk to a bank that is using a traditional credit score plus scorecard model, which is basically, are you up to date on your credit card payments, that sort of thing, and you replace it with a several hundred-factor AI/ML-powered model, you will generally get a significantly better result off the back of that. Or put another way, of all the things I worry about with Carrington Labs, whether the product works and is better than the old way of doing things is not my topic.
Okay. The final question on Carrington Labs. Does it make sense to use channel partners such as core banking software or loan origination software partners as a go-to market for Carrington Labs?
Absolutely, it does. We are thinking about different paths. There's a path where we go straight out to a lender directly and work out a way that we can fit something into their stack. There's a different path where you work through a channel partner that already is providing part of the technology stack. So it could be someone that provides the data environment or the banking software platform or the loan management system, and we integrate into that. I think they're both viable options. I think our challenge is just getting our head around all the different possibilities and actually investing the time to progress them. So far, we've probably had more traction on the credit sales just because that's a simpler process.
You're talking one-to-one at the end organization versus where you have three different parties that need to be involved, but we see a lot of potential in both paths.
Okay. The next question is about personal loan products. So has that product moved beyond the initial phase, and when will you offer it to all customers?
It has not moved beyond the initial phase. In terms of the rollout plan, so we will take the time that we need to get it right from a process technology but most important credit and analytics point of view. When we feel comfortable with that, we will then open it up more widely. We're still only offering it to a portion of the existing customer base. We will open it up more widely within our existing customer base, and then after that, we presume we'll then open it up more broadly to people who are not currently customers of Beforepay. That timing will be guided entirely by when the product is ready for that.
That means both, obviously, systems, processes, risk models, and analytics, as well as when we have the appropriate funding in place, given that that will likely be a pretty significant increase in size of the business.
Yeah. The next question is also around the personal loan product, but I think, Jamie, you've just answered it, but I'll ask it anyway if you want to add anything further. When will you roll out the personal loan product at full capacity?
Yeah. I think if I prefer vacuum.
The next question is around your cash balance. With plus or minus AUD 80 million cash at hand, do you feel you have adequate funding available to grow the personal loan facility further?
So we absolutely have the capacity to grow it further. I think when we look at the long-term potential of that product, the personal lending market in Australia is a multi-billion dollar market, so we're not going to self-fund that entirely out of equity. So at some point, we will be, you would imagine we would be looking to add additional funding capacity in the form of a debt facility or a warehouse or a similar at a spurt-like growth, as we have on the pay advance side.
Okay. Just a reminder, if you've got any questions, please type them into the Q&A box at the bottom. The next question is around profitability. So now that you have shown consistent profitability, is there an opportunity to source a better funding rate from your existing lender or an alternative lender?
That's a great question. I'd say we have nothing to announce at this time. That's a good question.
Okay. The next question is around something that's quite topical at the moment, AI and deepfake. Can you talk about the AI utilization in your business and any remnants from the deepfake commentary in the press recently?
Yeah. I've seen questions about DeepSeek, the Chinese AI model. To quickly cover off the first one, we've obviously spoken about this a number of times in the past, so our models are, at their core, powered by machine learning and artificial intelligence, noting that AI has many different definitions, and so we use different things in different ways. That's very much at the core of what we do in our core business in terms of how we make decisions around loans, as well as what the Carrington Labs product is. It's very much at the core of what we do. I think just a couple of commentaries on DeepSeek. As some of you would know, but it's better said than done, recently a Chinese company released a large language model, something like ChatGPT and Gemini and Llama and the others.
And they reported that they had trained it very cheaply using a relatively limited number of the chips that are used to train these models. And so whereas some of the big models people are using today, like the latest versions of ChatGPT, have cost hundreds of millions of dollars still, they said this one is a tiny fraction of that. And so there's a lot of commentary in the press about, is this a threat to those big companies? Does it open up AI more broadly and so on? The first thing I'd say is that this is, if anything, good news for us. So we don't build foundational models. That's a Google-scale problem. So we use some of those models as inputs into things that we are building, and there's a lot of complexity around how we make sure we do that in an explainable way.
We have some good IP around that, but we are not trying to compete with OpenAI or Google or DeepSeek for that matter. So if DeepSeek has in fact built an amazing new model at a fraction of the cost, that suggests that, because we are buyers of that technology, that that might open up more options for us. I don't know if we would use a Chinese model until the security situation is clear, but in general, if it turns out to be cheaper and easier to train these models, that will probably give us more choice in relative costs as consumers. I think just talking more broadly on DeepSeek, it was not specific to our business. Actually, there's a lot of good data about whether it's real.
So did they actually train it this cheaply, and did they steal some of the weights from an open-source model like Llama? And I don't know. I don't think it's not impossible that they do what they claim to do. I think there's. I wrote a piece on this maybe a year ago, and Sam Altman was saying we need $5 trillion for the next generation. I think there's pretty compelling evidence. There's a good paper that came out in, I think, 2022. It's called the Chinchilla Paper, like the animal about a Chinchilla family of models, where they trained the models on different amounts of computing power and different amounts of data. And the short conclusion of that was, we as society and the economy are massively over computing these models, given that we actually only have a relatively small amount of high-quality training data.
I think it's very possible. Meta even said this in their release. I think it was Llama 3.1. They said, "We know we overtrained this, but we think that's probably a good thing to do anyway." I think it's very, very possible that actually the Emperor has no clothes, and huge CapEx price tags actually have not been money well spent. Because if you're overtraining on a certain amount of data, you're just kind of churning it over and over again. It's like stirring something that's already fully stirred. I think it's not impossible. It would be very interesting if it came to pass.
Excellent summary, Jamie. It looks like we're on the final question. Here's your last chance for participants on the call. If you have a question, here's your last chance to type it into the Q&A box. The last question is around your customer acquisition cost of AUD 40 with the new personal loan product. Will that change materially in the future?
In the short to medium term, probably not, because we have over a million registered users, and that's a very meaningful portion of the addressable market already. That's where we will start. I don't see that number changing as we're going through that. When we open it up to new users, to people who are not existing Beforepay customers, you could imagine we will adopt the same philosophy, which is on a marginal spending on a marginal basis up to the marginal value of that customer. And if, as we expect, the personal loans have significantly more margin, better unit economics than the pay advance product, you could imagine us being willing to spend more in order to acquire them. Whether that materially moves group-wide CAC, I don't know.
But I think that the core thing to take away is that we will always limit our marketing spend to only those dollars that are clearly earning their keep when we send it out into the world.
Excellent. Liz, let me double-check. It looks like there are no further questions. So that concludes the Q&A session then. I will now hand back to Jamie and Elena for any closing remarks.
Great. Well, thank you very much again for joining us. Again, just really pleased with how this quarter has gone. I think we're just steadily delivering good quarter after good quarter, and I think we've got our heads down working towards the next one. We will see you for the half-year results towards the end of February. I'll apologize in advance. I'm going to actually be dialing into that call from the United States. I'm doing some Carrington Labs work over there, so I'll be remote for that one. But thank you for joining us today, and we look forward to seeing you at the next one.
Yes. Thank you.