Started. So welcome everybody to the Affirm Fireside Chat at the Goldman Sachs Communicopia and Technology Conference. I have the privilege of introducing Max Levchin, founder and CEO of Affirm, the leading US Buy Now, Pay Later provider. My name is Michael Ng, and I cover Affirm as part of our fintech coverage here at Goldman. We have about 35 minutes for today's presentation, so if you have any questions at any point during the session, just raise your hand and we'll get a mic runner over to you. So first I want to thank you, Max, for making the time out to be here with us today. Really appreciate all your time and insights. To start things off, maybe you could just talk about, you know, what are the key things that you're working on and focusing on these days. What's most top of mind for you at Affirm? So we announced the card at the last earnings in a big way. We announced it quite some time ago, but, we sure took our time getting it exactly right, and, it is now. So I have been very focused on making sure that the team that works on it has all the resources they need. You know, as much as I can, participating in product design sessions, and has a very long, very exciting roadmap, and so just trying to understand how far can we take this thing and just how big it would make it. So, very, very focused on the card. Job number one is always credit, and so at any given time, I'm, these days, our credit team is extraordinary, to put it mildly, and so I mostly just cheer them on and ask interesting questions, but I spend a lot of time still looking at that. And then as much as I can, try to figure out where the next S-curve is going to come from. Like, we showed the bottom part of the S-curve that the card is currently demonstrating, and so my primary job, in addition to hiring the best team possible, is to come up with the next S-curve, hopefully a couple years from now. Great. We're going to spend a ton of time talking about the Affirm Card during this session, and a lot of the KPIs that you guys shared during earnings were really surprisingly positive. Before we get into that, you know, I want to talk about the balance between growth and profitability. You know, Affirm has demonstrated really good growth this past year, right over $20 billion of GMV, 30% growth year-over-year, revenue of nearly $1.6 billion, and all while maintaining RLTC margins of 3.3%, which you guys have said you expect to continue next year. Could you just talk a little bit about how you're balancing growth and profitability at Affirm? You know, what toggles are most top of mind for you as CEO? So probably the most important thing that we've always cared about, or for the most recent 10 years, is making sure that the unit economics are consistently strong and just keep on printing that 3%-4% number, and that is crucial to long-term planning of the business, long-term resource allocation, et cetera. And so I think that that's always been the case. That's something I cared about a lot, and we are very serious when we say that's what we're going to do. And the, the business works really, really well in this really high profit for a payments company, margin structure, and that's what we intend to maintain. In terms of operating expenses, I think we demonstrated over the last six months quite a lot of discipline, and like any company, sort of with a little bit of the pandemic heyday, we had a few lessons to learn, but we quite like the education, so we will continue being very careful. You saw in our outlook, we are not projecting any significant increases. We think we still have leverage to gain in a bunch of places. Sales and marketing is kind of a, y ou know, we built our business with the help and support of our merchant partners, so we don't spend much money acquiring consumers, or none whatsoever, really. So we'll continue being very disciplined about OpEx. And then the trade-off between growth of the business and profitability in my mind is a false trade-off. Like, you should be building. I stole this line, it's not mine, but I forgot where I stole it. "The only businesses worth building are the ones with network effects." And kind of the superficial answer is, well, those are the easiest to defend because you just end up sort of having these self-reinforcing systems. The other part of it that's a little bit more kind of investor-friendly version of the same adage, is that network businesses inevitably exhibit increasing leverage. And so really good ones have kind of exponentially rising leverage, but even the ones that are not that good, if they're true networks, whereas people join, it becomes easier to justify joining and participating and staying within the network. You just see really, really strong leverage that just accretes to the operator of the network. And so because we are a payment network, we have the natural growth is good adage, versus growth has to be moderated until we grow into the expense, footprint, et cetera. Assuming really strong unit economics, which is why I started there. But we are very much pro growth. I spent a fair amount of my time asking our teams, now that we have this really cool card that seems to be resonating, what can we do to bring it into more people's hands faster? And not by spending money marketing it or figuring out any sort of a inorganic ways, but we do have a massive audience. We have a lot of fans that use our product all the time. Let's get them all onto the card platform. And so I think there's a lot more to do in terms of gaining operating leverage over the next few years, and that's actually quite exciting. Like, we've definitely, as a company, matured into this world of operating leverage is cool, and it's something to brag about. Cool. I was wondering if you could talk a little bit about how Affirm's value proposition to consumers, to merchants, you know, has evolved along with your product portfolio, right? You started out with a lot of longer-dated loans, and now you have, you know, paying for short-term zeros, installment-bearing debit payment tools. Mm-hmm. So, you know, the value prop has changed along with that. So I was just wondering if you could describe that. I think the value proposition of Affirm to consumer has always been more time to pay. You can reframe it in all kinds of fancy language, and buy now, pay later is a catchy moniker, but fundamentally, if you look at the Affirm logo, by now, you know that if you click on that, you will have more time to pay for the thing you're buying. And that value prop is universal. It's the easiest to understand and justify when it's a particularly unique set of circumstances. It's a big cash outlay, and you're not paying any interest, and it's really nice in your cash flow spread over 36 months. And so where we started was sort of the canonical use case of Peloton and Casper, and sort of all the nice things you'd like to have, don't really want to revolve on the merchant is willing to subsidize your interest, Affirm. And so that was the origin story of where we came from. It turns out that the same value proposition can be offered for someone who's buying a $100 dress or $200 bicycle, and we march down from the higher AOV to significantly lower AOV products. The more time to pay value prop becomes has to become more nuanced as you get down into those territories, because in some cases, the merchant doesn't have the margin to make it zero. So that's how we ended up with interest-bearing products. In some cases, it really doesn't make sense to spread something over 18 months. You're probably looking at four biweekly payments. And so we grew our product portfolio as we saw merchants come in and say, "We think our consumers could benefit from more time to pay, help us out, but we need to offer the value to consumer in a way that they understand versus, you know, whatever someone else is doing." On the merchant side, the value prop has always been better conversion. Like, when you come to Affirm as a merchant, the question you're asking is, "If I added your button to my checkout, how many more sales will I have? Like, what's the incremental gain on my sales volume?" And again, that evolved too. When we came in, I remember one of the coolest conversations to date with a merchant CEO was this woman called me. She ran a pre-owned apparel, kind of a high-end apparel marketplace, and she said, "I'm calling this the Affirm pop. You guys added 35% more volume to my GMV." And it was because people were buying these really cool ball gowns, but they were really expensive, and you don't need a ball gown with very, very few exceptions. And so they sort of. The unattainable suddenly became very attainable, and, you know, the rest is history. And so the GMV gain looks quite different, merchant to merchant. In the case of Peloton, you're not going to buy two bikes instead of one, but you're going to decide to go for it versus postpone or save up, or whatever the alternative may be. In the case of apparel, you might go for the shoes and the dress instead of just the shoes. And so we widened the collection of offerings by looking at both AOV and circumstance of the merchandise. And so if you imagine kind of a two-dimensional matrix, at least that's what I imagine when I'm bored, dimensional matrices. And you can see that there's still a lot of white space actually in the kind of even lower AOVs and transaction types where we haven't played, and we have all the intention to. So kind of back to the what am I working on, a lot of my time is actually spent visualizing this two-dimensional matrix of all the things we could do for all the transaction types and all the AOVs, and asking the question, "What is the more time to pay product looks like for that?" And so there's a lot more white blanks in that matrix than there are check boxes right now. I was wondering if you could give any examples that, you know, won't spoil any surprises, but might just put some texture around what you're thinking about? So a good one that's available today is the card. So the card is this really neat, unique, if I do say so myself, device that is a debit transaction, so settles basically immediately, but also allows you to say, "That transaction, I want it to be a loan." And some of these transactions are 0% because the merchant is willing to subsidize the interest or the time value of money, and some things are proper interest-bearing loans. And we saw at launch, basically, that something like 42% of the transaction count, not the GMV, because typically, Pay Now transactions are much smaller. Average, 42% of the GMV, or 42% of the transaction count is Pay Now. So if you sort of imagine my two-dimensional matrix, in the very, very bottom is the tiny GMV of my Starbucks coffee from yesterday. I used my Affirm Card. It's not exactly a, like, today, it is not especially exciting to say, Well, I used my Affirm debit card instead of my Wells Fargo debit card or wherever my primary bank account is, but that's why we're product developers. There will be a feature, I won't spoil the surprise, as you put it, that will make it a more compelling product than the one that my Wells account gives me. That's a product that hasn't yet been fully created, and an AOV that is the most common AOV there is. Like, we spend a lot more time buying coffee and water bottles and sandwiches, and so that's where we intend to play as well. It has to be better and more compelling than the alternative, and that's what we're working on. You know, one unique thing about Affirm, and there are many, is that, you know, this is a company that really has underwriting in its, in its DNA. And, and some of that I think is because of the, you know, history of underwriting longer term loans, you know, Peloton, Casper, as you kind of mentioned. But could you just talk a little bit about Affirm's underwriting practices and how that might be a point of differentiation, and why, you know, relative to other buy now, pay later customers or companies? Several earnings ago, maybe a year ago now, I attempted to write a poor man's introduction to Affirm's underwriting advantage, and I think the market punished me severely on that one. We've been punished severely for several other letters that I've written, but that one I thought it was most unsatisfying because I put so much work into really trying to explain the extreme coolness that goes into underwriting. We see underwriting as sort of the foundation on which the house of Affirm sits. It's always job number one. It's the thing that every employee is expected to understand. It is not a sort of a domain of crazy quants that are rarely seen. People who work in underwriting at Affirm are celebrated, that their names are known, they're heroes. We know exactly who the brightest bulbs are and just how important they are to the business, and the cultural foundation is actually really important. If you typically go to any number of buy now, pay later, sort of writ large Silicon Valley or global Silicon Valley payment startups, you'll find pretty frequently that underwriting is like a part of IT, because IT is part of risk. And that, you know, sort of makes sense, but it's insane. Or you have an analyst group, like in analytics, you know, you check your web logs and make sure that people are signing up really quickly, and then sometimes build some underwriting models. That is not what we do. Underwriting, risk management or quant risk management team for us is a major part of engineering. It's the best engineers that we have want to work in our ML group, and we have both machine learning and a generative AI group that's quite separate and works on something very different. But both of those groups employ some of our brightest people. They are extraordinarily valuable, and they know it. The approach we take is probably not super different from most really disciplined underwriters. We're just unique in a sense that we elevated it to kind of extreme discipline. It's very much of a every data matters, every type of data source we can get our hands on, everything we can possibly quantify, we do. We'll. You know, one of the things that I do is I get spammed by every data vendor under the sun about, you know, "We have this cool new signal," and I delete 99.9% of the spam that I get, but that particular one gets forwarded to Nitesh, who leads our ML group, and says, "Here's a data feed. Do you want to give it a try?" So we will ingest anything. We will run it in an AB test mode, or we'll run it basically in a dry mode, where a model will sit alongside the deployed model and see if it predicts outcomes better. One of the things that makes us very different as well in our approach to underwriting, we take a lot of care to be explainable. So at our scale, regulators are, you know. The eye of the regulator is upon us, and from the very beginning, our point of view was you either hide, but then you can't grow, or you embrace it, and then you try to build bridges and communicate exactly who you are and how you do what you do. That means you have to be willing to get examined for things like fair lending and disparate impact, and anything and everything, including, these days, explainability of your models. If you tell someone, "Sorry, you're declined," you can't just say, "And we're not going to explain why." It has to be fairly understandable and human readable. So we spend a lot of time these days looking into explainability and traceability of decisions, and in the ideal world, which we're a couple of steps short of, but that's another thing I'm personally very excited about. I'd love to be able to tell someone, "You've been declined. Here's the very specific two, three things you can do to get approved." You have to be able to stand by that promise, and that itself has a sort of quasi-firm offer of credit built into it, so you have to be very thoughtful about how you approach it from the regulatory perspective, but that's ultimately what people want. No one is too upset about being declined if they're told, "Here's how you achieve your, your next credit goal." So anyway, I can also. If you want to get into how do we actually do it, there's a really good shareholder letter that I wrote. It's, like, seven pages long, and I'm confident at least the first page was read by some people. But it's available, I think it's from May of last year. Okay, great. You know, that was a really interesting point and thought, and I was wondering if you could talk a little bit about Affirm's competitive position, you know, against the backdrop of, you know, the macro cyclicality that we've seen over the last couple of years. But also, as you mentioned, arguably, the greater regulatory oversight as well. Mm-hmm. And you know, why that positions Affirm well from a competitive standpoint, because of everything that you guys have done from an underwriting perspective, a disclosure perspective, and so on. Yeah. We took a lot of care from the very beginning to be, as they say, at the table, not on the menu. I spent three years on CFPB's advisory board because I wanted to understand what really matters to the bureau. Obviously, their agenda changes, so I don't pretend to know what, you know, they're working on right this second, but I do know the foundations of the consumer protection side of the regulatory regime. We spent a lot of time talking to our banks. So we obviously are not a bank ourselves, but as we lend, we're, at least some percentage of time, are doing it on a bank partner's paper. They're very tightly regulated these days, certainly by FDIC, who've just gone through a couple of close shaves with a couple of regional banks. So all of those regulators are pass-through regulatory relationships. Just because we're not a bank, I very much care about what a DSE thinks of the quality of credit that we produce, and our fair lending results are known to them as they examine the banks directly underneath them. If those banks work with us, they see all of our information, and we take great pride in being very transparent to that. So in terms of preparedness for more regulatory attention, we feel very good about it in the sense that we've always lived in the assumption that we are going to be, we are regulated. There's not a world in which we just sort of will be so small that we won't matter. I think a fair amount of our competitors believe that they'll either talk their way out of it or hide their way out of it, and certainly not under this administration, but frankly, it's also just a good thing. Like, regulators, generally speaking, are doing the regulating for the right reasons. They're not trying to just create more work. They're actually trying to protect consumers or protect deposit holders at banks, et cetera. So feel just fine about our ability to navigate that well. It is a cost, however, like, you have to be willing to go hire compliance people, in which we have a fair number. Anyway, so and that's how I feel, feels very good. Sort of broadening the competitive question. I think in the US and Canada, where we're active, we've pretty well proven that we are the preferred partner for most people that care about things like consistency, quality of approvals, consumer relationship. We won't be the reason why your shoppers hate your credit provider, and that's pretty important. I think with plenty of competitors that say, "Well, what do you care if we overcharge them junk fees?" As the regulators will call them these days. We do. We don't charge late fees. We don't charge deferred interest. We've taken the position that the consumer that we serve at a merchant, whoever that merchant is, from Amazon to a brand you've never heard of, has to look at us and say, "That was really amazing. That's a great experience. I will come back here. I'll use Affirm, but I'll also use this merchant because they use Affirm." And so long as we're accretive to the merchant, both through helping them sell more and brand halo, I think that we were doing our job really well. I take a lot of pride that if you look on Amazon, you can find ratings on anything, including payment methods. We have a really good rating, which is very important to me. I check it a lot. And so that's a pretty key component of our competitive differentiation. I think it's not an easy business if you're underwriting day in and day out, there's lots of opportunity. And you treat your underwriting people like they're working in IT. It's ample opportunity to stumble. We've always taken the view that underwriting is our core competitive advantage, and so over the years, we've proven that to be a pretty important thing, and that's another reason why we've won as much as we did. I, on the flip side, don't think there will ever be a monopoly in payments, and so I fully expect to continue earning a right to be at every checkout that we're at. Right. And, Adaptive Checkout has also been a point of differentiation, both for merchants and consumers as well, right? Yeah. You'll actually, you'll see some really cool Adaptive Checkout innovation from us in the sense that we've gotten so smart about what makes sense for consumers. We can start very soon predicting what is the best financial program for you. Like, if, if I want to get dangerously close to AI, we're never going to use AI for things like underwriting. We'll use ML, but AI in today's parlance means something entirely different, and nobody wants to have their underwriting model hallucinate sometimes. And so we're, we're very, very far away from that sort of thing. But having AI or at least a really good model that knows how to advise you on the best financial decision is a really cool idea, and so I'm very excited about some ideas in that domain. Great. And if we could go back to, Affirm Card. You know, maybe just talk about, you know, the strategic rationale in a little bit more detail. You know, how has the rollout looked like so far, you know, relative to your expectations? You know, where do you think the volumes and revenues can get over time? Probably best not predicting numbers over time, given various restrictions and that sort of thing. But in terms of launch, you know, I am probably the harshest critic of all things Affirm, and I do have to say it was pretty awesome to see the hockey stick that, you know, as soon as we took the leash off, it just took off. And so that was, that was pretty great. And, you know, we still have yet to push it, as they say in marketing. Today, it's a thing that you can find in your app if you're eligible. We have maintained so far a fairly stringent underwriting standard, so hopefully all of you can see it in your app. But if you don't, don't blame me, blame our underwriting team. You have to be in good standing. You have to have used Affirm for several transactions, and so it's, it's a product that's available to you if you are a at least somewhat committed Affirm customer. We will obviously open it up to everybody at some point, but for the moment, with this kind of growth, we're actually quite, quite comfortable with the just letting people who are clearly excited about Affirm for all the right reasons. It's growing well. I'm, I'm quite happy with the, the way it's growing, quite happy. Was a little bit positively surprised, even at the launch, so I'll, I'll go that far. The reason I was positively surprised is in kind of the. Product roadmap for the card, we're like 10% in, is, is sort of maybe, maybe less, but there's just a lot of work to do on. For example, today there's not a really great reason to throw away your debit card and just use Affirm, even though it will settle against your existing bank account. We will soon give you some reasons to do that. All the really good cards out there, like the, the really successful new credit or debit cards, they all follow this idea that using this product makes it better over time. The easiest way to do it is to say, "Well, you know, your, your rewards go up more," or, you know, whatever some of the basic reasons. But if you do a really good job thinking through product ideas, you can really build a better product that just improves as you go, as a user. And so we're very, very tuned into that notion, and we're working very hard on how to make the card better over time. And there's a bunch of really cool things that you can unlock the more you use it, and sort of there's. If you look for that today, you may be one of the chosen 5% that are being A/B tested into that particular program. But there's a lot of very interesting things happening in how to make sure that we get frequency of use. The reason we launched the card, and what I'm focused on, what the company's focused on, is frequency. We have access to 65% of all e-commerce in the US. We have designs to go beyond US. We have extraordinary partners that have been generally very good for us in terms of our ability to grow and while maintaining our unit economics. What we need the most is more transactions with the users we have, and I think we have a lot more ideas than just the card, but the card is the most obvious one. If I can persuade you that the only card you really need in your wallet, or at least on the top of your wallet, is the Affirm Card, that's going to make it more than four transactions per year, which is what we reported last quarter was our frequency. And so frequency, frequency, frequency. Our CFO likes to say that if I repeat the same word over and over again, eventually the company just makes it true, so I'm just going to keep on saying frequency. Well, I've been impressed by the share of transactions that are pay now, which are very high, but also the share of GMV that's interest-bearing, which is actually bigger than the consolidated company, at I think it was 80% of GMV on Affirm Card being interest-bearing, which is great for RLTC margins. Mm-hmm. On the Affirm Card. Yeah. You should expect that to go down some. Sure. because over time we want more donuts and more breakfast, and we don't think you should be paying interest on those. Maybe shifting gears a little bit and talking about some of your merchant partnerships. You know, focusing on the relationship with Amazon. You know, how has Affirm's relationship with Amazon evolved since exclusivity ended earlier this year? How penetrated do you feel like you are in that specific opportunity? And, you know, what are some of the things on the roadmap to drive growth going forward, to the extent that you can talk about it? So a good factoid that you can see in the numbers, and I think we said it out loud, but it's, it's definitely in the numbers that we see. Shopify, which is a three-year, three years since availability, is accelerating. So while I can't exactly tell you what's on the roadmap for Amazon. Sure. I can tell you that when you work with really, really huge, multifaceted partners that chose us because of scalability, approvals, technical acumen, ability to develop new products with them, and have about a, you know, metric ton of other things going on in their companies, be it Shopify or Walmart or Amazon, Target, or all the, all the extraordinary brands that trust us, it takes a while to scale it, but everything they do is very high leverage. When you're that big, you can't afford to tinker. You have to commit and build something that's going to scale and scale for years. And so whatever percentage of any measurable thing we have at Amazon, Amazon is sort of endlessly large. So long as we earn a right to be there, and we try very hard every day to be a great partner, to earn a right to approve people that can pay us back to create more volume, there will be long-term investments that both companies will make and continue scaling. I would caution against thinking, "Oh, so that's what that means next quarter," because it took three years to get to this next wave of acceleration with Shopify. But by no means am I trying to suggest that. Finally, it's accelerating. I think there will be more accelerations. It's not an accident. It's another thing that we've built together. It always takes a while, because if you're playing to a huge audience of be it merchants or consumers or both, you want to test it, you want to make sure it's accretive, you want to make sure it doesn't cannibalize anything you don't want cannibalized. But that said, we don't do things lightly. But when we do, we tend to get them going and going for a long time. That's also why the card is so exciting. Like, now that it's out there, I know the DNA of the company is to build something for a long term and make it huge. Great. I wanted to go back to something you mentioned earlier, which was about profitable growth and how, you know, growth and profitability are not necessarily trade-offs. But, you know, you talked to managing the business to this 3%-4% RLTC margin. There's obviously a lot in there, you know- Mm-hmm. MDRs, interests, you know, could you just talk a little bit about some of the things that we should consider as we just think about RLTC margins over the next few years? So today, we'll be very disclosive about all these things. As the business changes, we will always be very, very clear about what is and isn't changing. Obviously, the economics of pay now transactions or effectively pay now transactions are not the same thing as an interest-bearing loan over a long period of time. And so at some point, when it becomes material, we'll start breaking out pay now and perhaps other kinds of transactions and sort of making a point about where we make more and less money. For the moment, I think what we have said stays. We expect to stay within this three to four range. There will be extraordinary moments in either direction. For a time, we were over four, and that felt great, and one or two times we were under three, and that felt less great. Generally speaking, we feel that we can run the business at scale to those numbers. The business is not static. We'll build big things. If those big things become things like pay now, it's probably going to look a little bit different. I'm pretty sure it'll take a little while before we have to change our outlook on that. That said, we made it very clear in our, in my letter anyway, the current fully baked card economics are on par with the rest of the business. Right. We probably have time for one question, if there is any in the audience. Just one up here, please. I think if you say it loudly enough, I'll just repeat it. Yeah, I would like to understand a little bit. You mentioned that part on you yourself looking into new data sources for better underwriting. Would love to understand a little bit more what type of data are you looking for? What's the budget at the moment at Affirm for buying said data? And everything you could elaborate on that would be interesting. We're pretty cheap, so whenever we buy data, we pay very little for it. And we do buy a lot of data, but we'll look at anything. The test is very simple, Can you extract incremental, differentiable, and detectable gain to approvals? So keeping approval constant, can you have a slightly lower loss or maybe significantly lower loss? And if the answer is yes, there's a bunch of hurdles to go through. Most important ones of which is, are you discriminating on a variable that you can't discriminate on? If so, can't use it, and so on. But we are omnivorous and as proud as we are of our results, we're not proud to the extent that we were always looking for more improvements in the system, and that includes the data sources that we buy. At this point, we have an extraordinary advantage in data in two ways. We have a decade of history of proprietary data that's entirely just transaction history that we've generated. The reason it's so important is because it has the outcomes, so every one of our transactions is annotated with, "And that's what the repayment schedule looked like." And that's unique and can't buy that because we are who we are. And the other one is, for a vast majority of our transactions, we have incremental data that card networks don't pass through, and so that's where we get some really, really interesting bits of alpha, and we'll continue looking for it. On balance, we probably buy less data than we did in the beginning because we found ways of generating our own or found that it wasn't as useful to us as we thought it would. Well, it's a great way to cap off the session. Thank you, Max, for your time. This was amazing.