Transcript Synced · select any line to jump ▾ 0:00 Chris Titley: You're listening to a Day One FM show. 0:07 Jamie Twiss: Pick My Brain is the podcast where founders pitch me their startup and I try to give them some useful advice so they can connect better with potential co-founders, investors, media, and of course, customers. My name's Alan Jones and I was a founder myself for about 15 years, and after that, an angel investor for another 15 years. So yeah, old. Some of my ventures have been successful and some failed disastrously, But I like to think I've learned a thing or two along the way, and maybe some of that can help you. So if you'd like to learn how to tweak your pitch, subscribe to the Pick My Brain show now, wherever you like to listen to podcasts. 0:40 Chris Titley: Hi, it's Chris Tittley here. As part of the Fintech Fun podcast series, I'm joined by Jamie Twiss, CEO of Beforepay Group. Jamie, thank you so much for being part of the series. 0:50 Speaker C: Thank you for having me. It's a pleasure to be here. 0:52 Chris Titley: Jamie, great to chat again. I've just, for the listeners out there that know nothing about the Beforepay Group, can you give a bit of a 101 about the history of of the business, how it began and, and what it's all about. 1:02 Speaker C: Absolutely. Uh, thanks, Chris. So there are two sides to our business. Um, one part of our business is a mission-driven lending business here in Australia where we directly lend to consumers. Uh, and that operates under the brand name Beforepay, which is how most people know us. And then we have an enterprise software side of our business called Carrington Labs. And what that does is that licenses the capabilities that we've built on the lending side of the business to other lenders, primarily offshore as well. The origins of the business date back about 5 years, so the original vision was to find a more customer-friendly way of helping consumers who just need to borrow a small amount of money for a short period of time. So the most common reason most people need to borrow, it's not to renovate their kitchen, it's not to go on a big holiday, it's because they're fundamentally solvent but they don't have a lot of savings. That's a situation that describes almost half of Australians, depending on the precise definitions you use. And so they have their heads above water, but if something goes wrong and they get a bit back to front, they don't have good options available to them. So a typical use case would be car breaks down and you need $300 to fix it, and you can pay it back when you get paid in a fortnight, but you really need the $300 right now. Or veterinarian bills or things like that. Mm-hmm. So that's an extremely common need. Historically, the financial services industry has not done a good job of providing a product that's fit for purpose and is ethical and customer-friendly in that space. So we set out to provide that product and to disrupt a lot of really, I think, kind of unhealthy and arguably unethical practices that the financial services industry was engaged in, particularly around both the use of revolving debt where it wasn't helpful for customers and also just There's some terrible predatory pricing out there, payday lenders, you know, that sort of stuff. So we offer what's called a pay advance product. It's a short-term unsecured loan. The average size is about $400. The average duration is just under a month. It's all done digitally. It takes about 10 minutes from the moment you download the app to when you can have cash in your account. So it's a very quick and slick digital process. We write about 40,000 of those every week with our team of 50 people. So it's highly automated. And because we price it deliberately in a very affordable way, so we charge a flat 5% fee. So the average all-in cost to the customer is just under $20 for the typical advance. We have to be very, very sharp on our credit and how we make decisions around those loans. So we have a lot of data science and credit capabilities. I'm a data scientist myself. We ingest significant amounts of data with your consent, and we run a lot of rocket science over the top of that to work out whether to lend to you or not, and if so, how much. And so that forms the core of our lending business, but that's also the capability that we resell through Carrington Labs. 4:00 Chris Titley: Fantastic, Jamie. You mentioned the average size around $400. How do people pay that back? And as you mentioned, it's all automated. And how have you gone about sort of managing That's it. 4:12 Speaker C: Yeah, so when you sign up, we get your consent to take a copy of your bank transactions. So we look at the line item transactions that you engage in with your debit card or, you know, whatever else is in your accounts. And we do a number of things with that. We work out how risky you are, we work out what your income is, and a number of other variables. And one of the things we do with that is we determine what the right repayment schedule is for you. And that's based on how often and when you get paid. So if you get paid, let's say, $2,200 a fortnight on Thursdays, and you borrow $300 for us, we might propose a repayment schedule which says, "Well, Chris, why don't we come for the first $150 in 11 days' time when you get paid, and then the other $150 a fortnight later," and line up with that cash flow forecast that we've done for you. We do that through direct debit. So you grant us a debit authority on your account to— Mm-hmm. Recover that directly. You can also push a payment to us if you like. Some people want to pay it off early if they get a bit of money in their account, and of course that's, that's welcome as well. The most important way that we manage bad debts is by lending the right amounts to the right people in the first place. So by design, and again given the mission-driven nature of the business, there is very little sting in the tail of this product. We're not going to come after you with a big stick. It doesn't go on your credit report. And so what we're doing is we're appealing to people's senses of fairness to repay the loan rather than their sense of fear. And that just means we need to really understand you upfront. So in addition to looking at a lot of financial variables to understand your P&L, if you will, we also look at a lot of behavioral factors to understand, are you the sort of person who's going to honor this obligation and repay us? if for reasons often beyond your control, you're unable to repay it, we will keep trying to collect that, but we will generally do so in a kind of friendly, on-brand kind of way, more about reminding you that we're still out there. And we find the same behavioral modeling that helps us find customers who want to repay is also the behavioral modeling that helps us understand who's likely to repay a defaulted loan if it does indeed come to that. 6:27 Chris Titley: Jamie, you mentioned some numbers of transactions that go through your books. Could you want to give some high-level numbers around the business and the adoption and I suppose some of the feedback that customers have given you? 6:37 Speaker C: Yeah, so, you know, last year in FY24, we did about just over $700 million of advances. So that's, you know, just under 2 million individual loans that we made to customers. And again, this is with a team of 50 people. So very, very automated indeed. Given our fee structure, that works out to about $35 million in revenue. We are profitable, so, One of the benefits of running very, very lean on overheads, uh, with that high level of automation, and also really having a good handle on our credit, uh, is that we have good unit economics and very flat costs. So last year we had a profit before tax number, uh, of $3.9 million as well. Um, we have well over a million registered users, uh, 250,000, uh, active users as of the end of the first quarter. So it is a pretty scaled business. We do play a pretty significant role. Mm-hmm. In the lives of many Australians. 7:28 Chris Titley: And Jamie, you talked about the unexpected expenses around potentially a car breaking down or medical expenses, etc., or pets or vet expenses, etc. But as the numbers of customers are growing, do you see more people using them for one-off things, or are you finding, broadly speaking, that times are a little bit tougher out there and people just need a little top-up? 7:49 Speaker C: Yeah, so we do have a very good source of data as to what's happening in the economy, uh, because we do see hundreds of thousands of transactions, millions of transactions a day. We have over 1.5 billion lines of transaction data, uh, in our database at this point. Um, and so we've got a good sense of where I think the average Australian is at. And, and you're absolutely right, times are a little bit tough. We are seeing, uh, those inflation— inflationary pressures have been pushing through. We've seen a displacement of spend from discretionary to non-discretionary. We've seen especially volatility in some hard-to-control prices, costs like utility costs and petrol costs that has pushed people back. Now, recently wages have, for most people, been keeping up, but that's very sector-specific. So you have people who are getting very healthy income increases and you have people who are not. And as we're starting to see inflation moderate, we are also seeing those— that upward wage movement start to moderate. So it continues to be a difficult time for people. I think in terms of the use of our product, so we certainly do continue to see a lot of one-off expenses. And then there's probably a spectrum, well, there is a spectrum ranging from genuine unpredictable one-offs like the flat tire, like the sick pet, then through things that are perhaps foreseeable, but still you understand where the strain comes from, such as, you know, heading into the holidays, it's often a time that people get a bit stretched, or we have a lot of young families where, back-to-school shopping is— that's a meaningful expense for, for many Australians. And then you see that all the way through to more general kind of household budget, household expense type, uh, type costs as well. 9:28 Chris Titley: And Jamie, with, uh, with Carrington Labs, uh, I believe there's, um, expansion into the US. Can you talk about that and maybe some of the, the future plans for the business? 9:37 Speaker C: Yeah, so Carrington Labs in particular, that's our, again, our enterprise software subsidiary, and our Our target market there, although we will work globally there, but I'd say our priority market really is the United States. And in particular, mid-sized lenders in the United States seem to be a bit of a sweet spot for us. And that's obviously a very large market and it's one where I think we genuinely have a pretty distinctive capability that's serving us in good stead there. So that's gone quite well, I think. I was pleasantly, I wouldn't say surprised, but I thought there was every chance that when I started traveling internationally and speaking at conferences and selling this capability to lenders, I would find half a dozen others doing the same thing. I was on a call, sales call this morning for Carrington Labs, and I think the comment was, "We've been looking for someone who can do this, but there's a lot more hype than actual capability." And I think we at Carrington Labs think that we genuinely have a pretty distinctive capability. So I think that's a very promising avenue for us. It is high margin because it's essentially a fixed cost model that we then retrain on client data. So I think we feel pretty optimistic about the future of that business, noting of course that banks can have slow sales cycles. So I'm always reluctant to promise any particular milestone on any particular timeframe. So that's the Carrington Lab side. On the Beforepay side, I think you maybe divide it further into a couple of different areas. The pay advance business, uh, continues to perform well. That's a very— Mm-hmm. Steady performer. I think if you look at customers' reactions to the product, once they've tried and used the product, they tend to really like it and they tend to be quite loyal to it. And so if they don't have the need, we don't want them to borrow. We are a mission-driven business. Uh, but if they do have the need, we want them to remember that we're there and, and if it's right for them to come to us to, to do so. And so I think that, that pay advance business continues to go well. We've also recently launched a personal loan product. So that's a regulated credit product. Um, so it operates under a somewhat different framework. That's a larger loan with a longer duration, and we're gradually going to extend the limits and duration on that. And that's obviously a very significant market already in Australia where we think we have a pretty good right to play and win based on our capabilities in digital lending and credit assessment. 11:58 Chris Titley: You just touched on capabilities, and before we get to some personal stuff about you, I'm interested to know around the use of AI and the use of technology within the financial services, broadly speaking, because to to me from the outside, it seems like there's a lot of talk around AI, a lot of tools that are being used in various different industries for efficiency gains and productivity tools. But when it comes to sort of consumer-based financial services, probably not so much yet. But how's AI going with Beforepay? 12:23 Speaker C: So it's a great observation. And the question I often ask bankers in particular, they talk about AI this and AI that, and I always say, well, what are you actually doing in production at the core What is the core of your business? And the answer is always nothing. So the vast majority of banks are talking quite a bit about AI, but when you look under the hood, they are using AI coding assistants and they're using large language models to draft low-stakes text like press releases. And they've got an internal tool, you know, a RAG tool to help their staff look up, you know, policy and product information. It's really this stuff sort of nibbling around the edges, which is nice, but not really very distinctive. We have deployed AI really at the very core of our business. And again, it depends how broad your definition of AI is. I think, you know, the, the heaviest lifting in our business is done through machine learning. So we ingest that, you know, 1.5 billion lines of transaction data. We have a pool of about 50,000 candidate attributes that we can measure off the back of that data. So everything from your income to— how much you spend on groceries to the velocity of money once it hits your account and how you move money around, you know, tens of thousands of those. And we select the best 400 to 500. So that's primarily a machine learning process. We use some neural net-based generative AI capabilities to do transaction categorization. So what we call BERT embeddings in the text descriptions alongside your statement to help us understand what exactly you are spending your money on here or there. 13:59 Jamie Twiss: Mm-hmm. 13:59 Speaker C: And then we're experimenting with generative AI to create those candidate features. So we've built some pretty interesting technology where we can, we have a code that will essentially self-generate new coded features. It'll come up with ideas through a variety of different prompts and then code them up, debug itself. That's a whole interesting capability in itself. Run testing and filter them for predicted value and send some through to the— Yeah. To the main pool. We are using quite a bit of AI actually in the core of that lending process. The key thing for us is that because generative AI and more broadly any neural net-based technology is inherently unpredictable and difficult to control, we would never use it to actually make an individual decision. We would only ever use it to come up with ideas and build out the framework of the model. And the framework of the model is, is is driven by controlled statistical techniques. So that's a very big emphasis and focus for us, which is that using explainable AI where you can stand in front of a decision and say, "Here's why we did it." Jamie, when you're not talking about personal loans and wage advances and AI over credit decisions, do you get to switch off at all? 15:14 Chris Titley: And if you do, what's maybe a hobby or something that the listeners may not know about you? 15:19 Speaker C: Ah, yeah, man, I wish I were more interesting. You're putting me on the spot here. So I am a big reader. I'm a big reader, so I actually studied Russian literature at university before I retrained as a data scientist. And there is nothing I love more than just a fantastic piece of fiction. I think the ability to really get to the core of the human experience is still fundamentally the domain of great writers rather than philosophers or technicians or anything like that. And so, you know, a lazy Sunday afternoon with, you know, a heavy 19th century novel is, Absolutely floats my boat. I've got, well, I was going to say teenagers, but actually they're now 18 and 20, so only one teenager left. So I apply to spend time with them and my application is considered and periodically granted, and that's always nice and you get a bit of time with them. I would call myself a serious runner, but I don't know if you can be a serious runner and very slow at the same time. So, you know, I run the Sydney Marathon every year and— 16:16 Chris Titley: Oh, wow. Okay. 16:17 Speaker C: Maybe plod or stagger is a better verb to use. Well, when I say every year, I mean I've done it for 2 years in a row. So And I think the saying is past performance is a guarantee of future results, right? So yeah, exactly. Every year for the last 2 years. 16:30 Chris Titley: Definitely in financial services for sure. 16:32 Speaker C: Yeah, so you can extrapolate that. I'm going to do it forever. I think that's the obvious conclusion. So a mix of things. 16:37 Chris Titley: Yeah, so a Russian literature fanatical data scientist that runs a publicly listed company that also does marathons every year. There you go. So first, you've nailed it in one as the first. I don't think I've ever had that before. But Jamie, thank you so much for giving the listeners out there an understanding of the Beforepay business, the history of the business, how the business is progressing thus far, and also the new tools and new product suites that you're rolling out, including the personal loan product, and a little bit about you and the CEO that runs the business. So thank you so much, and looking forward to catching up. 17:11 Speaker C: No, thanks, Chris. Really appreciate your time.
