00:25:50 · Guillaume Moubeche, Jason Wilk
[00:25:50] S1: And on your side, like, how do you decide, like, to launch a new product when you have something that's already working, already growing? Is it because you want to get, like, more growth? Is it because you want to get better retention? Is it a bit of both? [00:26:04] S0: Well, one, it just comes back to what our our customers want. And we have the benefit of having millions of customers. We can talk to them. And going back to my days at Y Combinator, the the number one lesson from Paul Graham was always to talk to users, and we, you know, do that all the time through user research. Really important to understand what they want. We also have the benefit of having a connection to our customer's primary checking account. And so we can actually see all the products that they're using. And so that's amazing market intelligence for us to go decide what else we need to be doing. But in the end, yeah, it's, you know, building customer happiness. And if we do well there, that's gonna lead to better lifetime value through, you know, either more ARPU or better better retention or both. And that's sort of the name of the game. [00:26:47] S1: And is it like like, how exactly do you do you decide, like, about a new product? I know, like, you're gonna talk to customers, etcetera, but why did you choose a new credit card and not something else?
00:27:00 · Jason Wilk
[00:27:00] S0: We wanted to lean in on where we thought our differentiation was, versus our scale, neo bank competitors. And to us, that is our AI based underwriting of cash flow information. And it was working incredibly well for our extra cash product, which is giving people up to $500 between paychecks. The beauty of that product is, is the payback is very short. The downside is that the customer's willingness to use that for multiple different purchases is quite low. You tend to use our core product for gas, grocery, and rent, all these sort of non discretionary expenses. But we aspire, and we could see through the cash flow data that our customers are using things like BNPL and Sun Private Credit Card for discretionary purchases, of which we had 0% market share. And so, when I think about areas where we think we have the right to win, our customers love us for credit. We think we can expand the credit, and give them a different use case of how to use the company, made a lot of sense. And so, it felt very clear to us as to what to build. We just had to go back to sort of the core principles of the business of what we wanna disrupt. And it is these expensive fee streams from banks and credit card companies. And just like overdraft fees are really hurting Americans, revolving credit card APRs and credit card late fees are over 100,000,000,000 a year. And so, I like the idea of eating other people's margin through better products. And we're gonna do the exact same thing with this product.
00:28:27 · Guillaume Moubeche
[00:28:27] S1: Yeah. And you have already like the the user base and, you know, like the... So so it's whenever you deploy, like, you know, you're gonna have like a a large addressable market from the start, so it's it's quite smart. And I was wondering because earlier you mentioned, you know, that the interest rates started to to kind of change and it affected, like, your your company. So how exactly do you plan? Because you're essentially offering like short term loans. So obviously you're a bit dependent on the macroeconomics. So how exactly do you plan on things that are a bit outside of your control?
00:29:06 · Jason Wilk, Guillaume Moubeche
[00:29:06] S0: That's hard because the the market just tends to swing naturally from things that are out of your control every day. You know, things like the war impact the stock, even though that should actually be a tailwind for our business because things getting more expensive, more people turn to our products. But it's hard, you know, in world where the stock market is largely traded automatically at this point, it's nearly impossible to avoid being, you know, pulled by the by the macro. With respect to interest rates though, you know, because our capital turns over so quickly, even with our new credit card product, which is a pay in forward duration, we're not really reliant at all on private credit markets or the cost at which it takes to borrow, because we're not taking long duration risk. And that's one of the thesis we think that investors would really pay attention to in our story, is that we don't have that risk whatsoever because our average duration of credit is about eight days. And Okay. The return profile we get on our credit transactions far exceeds, you know, even the highest of of interest rates to borrow over the last twenty five years. [00:30:11] S1: So what's what's kind of like the the default rates you try to to focus on? Like, what's like the the highest default rates you can go to in order to have like a business that works?
00:30:11 · Guillaume Moubeche, Jason Wilk
[00:30:11] S1: So what's what's kind of like the the default rates you try to to focus on? Like, what's like the the highest default rates you can go to in order to have like a business that works? [00:30:23] S0: Well, I think we just talked about our last earnings call, our our gross spreads on every credit transaction are over 5%, and we return that capital every eight to ten days. So, you know, that's arguably our ceiling of what we could go to on credit. And if you look at our hundred and twenty day loss rate at this point, it's nearly 1%. So the spreads are there. We've... We originate millions of these micro credit transactions per per quarter. So it's very... You know, we've got a lot of visibility into performance, and don't expect that to erode in any kind of any kind of cycle. [00:30:57] S1: And with the new, like, AI models, because you mentioned that you are leveraging AI, like, to... And machine learning, I assume, to, like, kind of predict, you know, like, the default of one customer. So therefore, you can know, like, who you're gonna pay, where you can you... We can loan, like, money. So how how does the new model, like, affect, you know, like, your company? Is it, like, in a positive manner? Like, do you see your algorithm getting better, like, every single month, or do you see, like, no real impact because you already had something that was working?
00:31:30 · Jason Wilk
[00:31:30] S0: It's all incremental. I mean, there's two things going on with our business. One, because the duration of these... Our credit transactions are so short, the models merge really quickly as to what actually works. In comparison to like a, you know, an installment lending business that's lending capital for a couple of years, so hard to use AI or cash flow data to predict performance, because so many things could happen. The economy can shift. There can be changes in the customer's employment. But with such short duration credit, we're able to learn so quickly on what actually works to full maturity of the entire book. And that is so valuable. And so every time we launch a new version of our models, we're inputting another 100 or 200 features that we can test to see if it's incremental to credit and know within a matter of weeks or, you know, a few months, the performance outcomes. And so it's just a really powerful place to sit. You combine that with the fact that almost 98% of our originations per quarter are going to repeat users. We can always start to, you know, increase credit per user just by the fact we start to know our members more. So with both those dynamics happening at the same time, leads to a really stellar financial performance.