So for what it's worth, I spend a lot of my time contemplating the word ecosystem, mostly raging against it because it is one of the less documented wealth transfers. And by the way, the credit card industry in general is a regressive wealth transfer where people who never evolve transactions, like probably both of you and me, when I use credit cards, which is really rare these days, we're paying our bills. At the end of the month, we're not paying a penny of interest. I think Joe on the record saying you've never paid any interest.

And so I think that's a that's a nice place to be if you can afford it, telling people that they should just not borrow and pay it off at the end of the month is an incredibly let them eat cake equivalent for the 21st century, payments. Lots of people in America revolve. Half the country revolves in a $10,000. And so the idea of those people caring about rewards is sort of preposterous. They do not. They care about making their minimum payment.

Hello and welcome to another episode of the Odd Lots podcast. I'm Joe Weisenthal and I'm Tracy Alloway.

Tracy, it feels like, everybody, whether, they want to come after the banks right there is this effort. So many of our episodes, whether we're talking about crypto, whether we're talking about private credit, whether we're talking about payment, it's this goal, this dream of like, let's chip away at some of these bank businesses or these businesses that were associated with some sort of legacy institutions. The banks, they mostly seem to be doing pretty well still. But there's this dream that they can all be sort of like disintermediated away, or that someone could. Each one of these functions that they do can be better done somewhere else.

Yeah. So let's see, I have been in financial journalism for almost 20 years now, which is kind of crazy. It makes me feel very old, but for as long as I can remember, someone has been trying to either reinvent bank lending or reinvent the payments space. And I guess I can see a few reasons. So like, genuinely, some payment architecture is really old fashioned.

Yeah. Especially in the US where, you know, sometimes you still have to write a check for something. Yeah, blows my mind. And the U.S. didn't get chips for a really long time. Chips and credit cards and stuff like that. But also, if you think about the payment and lending and financial market just in general, it's one of the biggest out there, right? Like you're talking about all the economic activity basically. And so if you can get a tiny slice of that total interchange or fees, then you can see how people, you know, people are really interested in that space, know completely.

And we know that there are a lot of like fintech companies for a long time. I mean, some of the most, you know, the earliest internet success stories, period, were, you know, this is definitely nothing new. And of course, one of the first ever was PayPal, which recognized that with the internet was going to come, all kinds of new opportunities for payments and sort of quasi peer to peer literal, peer to peer, transactions, etc.. PayPal is still, of course, extremely important part of the payments, infrastructure, etc.. But yeah, there's so much going on. But to your point, lots of companies have gotten those little slices and made an incredible fortune.

Yeah. We also did an episode on Buy now, pay later, a couple months ago. And there are clearly some interesting questions brought up by the expansion of that space. And you and I talked about how wherever you go on the internet nowadays, you get like, you know, probably at least 2 or 3 little buttons that offer you installment loans on your purchases. And obviously, the concern is whether or not people are taking out credit that they shouldn't necessarily be. There's a big discussion about that. And also the transparency of the credit that they're taking out.

Right. This is we have, you know, and we'll get into this there, there are pros and cons, I guess, relative to credit cards. But one thing with credit cards is that there is decades and decades of data on them, and we know how they're used, and there's really good risk profiling and scores and all this and credit scoring and all this stuff. Bnpl is sort of more novel. And so therefore, to what degree do we know how is as an asset or as a function it performs across different, cycles? Do we know the cohort cohort of people who use Bnpl as well as we know the people, the types of people who use credit cards? It seems a little bit, at a minimum, a bit more ambiguous.

Absolutely. Well, we've sort of hinted at what we're going to be talk about. I mentioned PayPal, we talked about Bnpl, we are going to be speaking with one of the original members of the PayPal Mafia. We're going to be speaking with Max Levchin. He's the founder and CEO of Affirm, one of the biggest Bnpl companies publicly traded. Max. Thank you so much for coming on Odd Lots

Thank you for having me. I'm a big fan of the show. Thank you. It's a bit of a privilege to be here. Thank you for saying that on the recording. So we always love hearing it, but we love when it's on the public record as part of the episode. So, thank you so much.

Where did you get the idea to start a firm? Why did you what what what prompted it?

But a two sided coin. Both sides bad of personal experience, actually tying neatly to the PayPal story. So I came to the US at 16 from Soviet Union. And so you can imagine I understood very little about borrowing and credit and things like that. And so I got my first credit card on campus at college at 18. Didn't fully understand what I was signing up for, certainly didn't read the fine print where it said 0% asterisk, don't worry about it. Borrowed a bunch of money, financed my first startup from that very same credit card. Promptly got into more that I could afford to pay for. Startup failed, eventually got some nasty calls from collectors, paid it off, and four years later, or five years later, took this little company. PayPal public was basically independently wealthy overnight.

Went to buy a fancy car that you know, as I originally deserved. At the ripe age of 23, I wanted to show off to my my then girlfriend, now wife, and was the client for credit. And not only did I feel completely screwed by the moment when I found out that you're supposed to make the minimum payment. And by the way, interest accrues into principle, and this seemingly low Apr is actually not what it seems to be. Because I missed some date of the specific amount that I was supposed to pay, it sort of bit me again five years later was like, oh, by the way, wreck your credit rating too, so you can't get a loan for anything, even as you sort of proudly sticking a finger at the article and newspaper saying, you know, the company, public, youngest, whatever, whatever.

And so that was the thing that stayed with me for years. My friends would prank me at restaurants. They would ask the waitstaff to tell me that my credit card was declined, which by then was a different credit card, and I would turn beet purple and basically be like, oh my God, like it happened again. Like what? And you know, eventually I knew it was a joke, but I was in my mid 30s when I realized that that's now more likely a joke than that credit score bite from 18.

And so in my late 30s, I sat down with a friend from high school, college and people, you know, I've many of these friends who'd sort of I met when I was just getting my feet wet in America. And ultimately we, you know, we built this amazing company together and asked them the sort of the obvious question, 20 years too late, why didn't we try to fix the yuck that comes with having credit cards? As a young person? And he said, you know, I don't know, but we built so much great. I at PayPal fighting fraud and doing all these really stunning things. Surely we could do a better job scoring credit for young people like you were at 18 and then 23. Then what currently happens? That was sort of the seminal moment. And, you know, a couple of years later we started a firm.

So I have a bunch of questions already. Firstly, what kind of car was it? Did you get it in? And,

I had to pay cash for it. I actually wired money, and it would begin like an 8 a.m. visit to a dealership. Ended up like, it's 5 p.m.. Your wired hasn't cleared up. You can't drive off this lot. And it did. You know, drama, at the time. Don't judge me. It was a black hardtop Mercedes convertible.

Oh, and no judgment at all.

I'm judging a little bit.

I was really young. I was very much trying to impress the girl who's now my wife and the mother.

Well, it clearly works. It works, it works, it works. Okay, Another question. So I just want to. I love those old those Mercedes. I think that there's one of the if I, if I have some rough conception, no judgment. I think those those are great design cars. Anyway.

Should we just talk about cars? Yeah, we could. Okay. Well, you mentioned the underwriting process, and this, in my mind, is supposed to be what makes the buy now, pay later. Bnpl model different, right? Like the underwriting process is more technologically driven, perhaps more nuanced. Talk to us about what you do differently versus, you know, a credit card company or a bank or someone like that.

Sure. And you're you're pulling on a thread that's going to take a long time to unwind. I'll try to be pithy, but, this is where the the rage kicks in. So one of the things that happens with credit cards at the very core, they tell you a bunch of things they don't actually want you to do. So a credit card business model is accrual of interest into principle. So this exponential function of your signing up for Apr x, but you don't actually know what's going to cost you, is all the fact that as you make the minimum payments or whatever the payments you're making, whatever you haven't paid off, the interest that you've accrued folds right into the principal and it compounds and compounds and compounds.

And so the longer you take to pay it back, the more it will cost, which seems obvious, but it's impossible for most mere mortals to predict. And they love it when you're late, because there are late fees which are fixed. And so the less you spend, the higher the percentage the late fee represents. So it's the sort of amazing business model, which is why, you know, one of your recent episodes puzzled over how are these rates so high. The rates themselves are actually not the problem. The problem is the structure. As you fold interest into principle and pay and pay and pay late fees, it can be extraordinarily expensive.

And so from the very beginning of a firm, we asked the question, how can we make a product where you don't have this weird misalignment of interests, where the lender tells you, please pay your bills on time? But what they're saying sort of is, but not to on time and ideally take as long as possible because that's when we make the most money.

So the obvious answer is, of course, obvious. Make a plan that you commit as a lender that you'll never change and don't charge late fees. And that's it. If you agree to those design principles, you are not going to make more money. If someone is late and you're not going to make more money if they take longer to pay you back, obviously you'll make less money because getting money to lend isn't free either. We have to pay for our sources of capital. As you know, as everybody else does, which means that it immediately rotates you sort of 180 degrees where you say, I only want to lend money when I have a lot of conviction. The person is going to pay me back on time, because if they don't, I am just going to lose money.

And so we put those two principles down in the ground very, very early on. It's literally written down to the foundation of the company. We will not misalign ourselves with our borrowers. From that came a lot of things like, hey, we need to understand your cash flow. We don't really care what your credit rating is. Maybe you gamed it, maybe it's not accurate. Maybe your 18 year old immigrant does not matter. We need to understand what your actual capability is. We have to have the right to tell you you are overextending yourself, not just once every few years when you renew your credit card agreement, but for every transaction you have to be able to underwrite and decide yes or no for every single moment that you create these payment plans. Because that allows us to not lose money because we are aligned with you. When you pay us back on time, we'll make some money. If you don't, we're just going to be worse off.

And so that is the foundation of a firm's take on Bnpl. Many people have come along since and sort of change the model a little bit. They introduce late fees, all kinds of other fees we never have. And never will. But even with those modifications, the model is still better than credit cards because of this individual underwriting moment. In our case, no fees of any kind just to be clear.

So let's say I wind up at some website and maybe I want to buy like some Nike's or something like that. And I see the affirm option. How does how do you know how good of a credit I am in that moment?

So it's not a effort free thing for you to use affirm. You'll have to click on affirm, and we will ask you if you've ever seen you before. It will ask you to provide some information for us. That information will allow us to tap into the standard set of records that the credit bureaus aggregated about you. If there's enough data there for us to use our custom credit score, we'll render a decision based on that. If there's not, we'll actually say, hey, we don't fully understand your personal financial situation. We would like to have a peek at your bank account, cash flow, and decide how your cash ebbs and flows relative to your ability to pay us back.

All of this is made possible with really good new technology, so this takes seconds, even though it probably takes about as long to do as it takes to describe it. But we'll look through your personal financial state of right now and make a decision, not just whether it's a yes or no, but also how much of a risk you have of that financial situation changing and that price is the credit.

Do you use any like, creative data points you mentioned? Checking against like the type of item that people are buying. So that's one thing I imagine could be interesting. Like if people are buying groceries on installment, does that mean they're more of a credit risk than someone who is, I don't know, buying a computer or a big ticket item? And then the other thing I'm very curious about, and I think I've told this story before on all lots, but I remember going to see, a startup that eventually failed, and they were in the peer to peer lending space, and they described some of the stuff they were doing with underwriting, where they were like trying to measure people's impulsiveness by how quickly they moved a slider for a loan and stuff like that, which freaked me out a little at the time.

It would freak me out, too. So, no, we do not do, telemetry as a input variable into underwriting. So before I get into what we do use, it bears mentioning that the variables you use in underwriting is a very, very highly regulated domain of lending and underwriting. There's lots and lots of laws going back quite far, things like Fair Credit Act and any other sort of subsequent additions to both federal and state level law that prohibits you from using kind of the obvious things you wouldn't want people to use, like your race or your gender or your age or your creed. Cannot be a factor. It's actually a federal offense to do that.

And so we don't do any of those things. We go as far as to say if it correlates to some of those things, we also couldn't use it. And that too is a subsequent law as well. So, so so first of all, we're very, very thoughtful about making sure we don't step into a prohibited territory. That's called prohibited basis as the, the fancy term in the industry.

What we can use and what we do use, as you mentioned, merchants share with us what's being purchased despite the burrito gate that sometimes, comes back in the press, people don't actually use affirm to finance Mexican food or other cuisines. They do sometimes use affirm for groceries, but it mostly is, or almost entirely really is for things like I'm throwing a giant party and I need several hundred dollars worth of foodstuffs, and that's a lot to pay down with my debit card. So I'm going to use a firm instead. So the average sized transactions for us is roughly $300. So it gives you a sense for like what people use a firm for.

So that's just, you know, apropos apropos where it's being used, what we glean from the information about what's being purchased, one good mental model, which this is like an approximation, but but a decent one if the useful life of the item is meaningfully shorter than the time it takes you to pay it back, you may find yourself questioning the quality of the item long afterwards disappeared. And so a natural question we should be asking is what do we understand about this item? But also maybe these items history of quality. Before we say, oh yeah, sure, buy this thing that's going to be out in six months and pay for it over two years.

So maybe the most useful thing I can say about actually what variables, what new variables we've introduced into our underwriting a vast, vast, vast majority of them, in fact, all of them actually in influence, a decision we make by 1 or 2% at most. Anytime someone tells you I found this magic variable, and if I just look at that, it's like 30% better underwriting or 5% better defaults. They're lying probably to themselves more than to you, but they're definitely lying. Anything like that just immediately becomes brittle as the macroeconomic reality changes, as people suddenly don't care about the item or whatever it is that they found. As this magic bullet of like, suddenly I know who's a good risk, it just disappears.

And so you actually want lots and lots of subtle factors that are also compliant with all the various regulations helping you shape a score that will tell you how likely this person is to pay back. And that's what we use.

Let's go back to the just the core of the business model. So let's imagine there is a shirt that costs $100 and I have an opportunity to okay, I'm going to buy it, with a firm. So to be for payments of $25 and there's no interest. And so for me, there is almost a free lunch because the time value of money, etc.. All right. I'm spreading this out over four payments. You're going to get that 100 in an index. And then I'm going to get that extra, that extra $75 after the first payment I'm going to know. And so you make money because the retailer is paying you a cut because that you've brought someone into the door. So talk about how is that gut determined? What how much are they paying you for that? And how competitive is that against other potential Bnpl companies, or perhaps new entrants that haven't come into the market yet? Talk to us about how that prices for them.

This is my question too, because when you go to check out, you see buttons. And the buttons, I presume don't mean that much to people. You just click one of them. So how do you differentiate yourself?

We've been around for nearly 15 years, and the way we've differentiated ourselves by the way we basically don't advertise. When you see us a checkout, that's when you learn about a firm at some point and then you come back to use again. And we have incredible retention rates. People who have used a firm a couple of times, basically with 90% probability, will come back to use a firm again, not necessarily the next week. In fact, we average something like five plus transactions per year. But when it matters to them, they will use a firm. And the way we differentiate ourselves is we were there for them when they needed us and if they ever stumbled. Our best customers are the ones that email us or call us and say, hey, I was late to pay my bill, something happened. What's the late fee? And we tell them there isn't one. That's the moment when they grasp how we're different from the rest of the industry, and that that's probably why they come back to us as often, as much as they do.

Something like 95% of our transactions come from repeat customers. So just to give you a sense for the loyalty that we've engendered in our user base without ever making promises, you know, in on broadcast media, that's sort of why they come back to the business model point. So first of all, probably worth expanding the aperture a little bit. So we offer both interest free and not interest free loans. The paying for the Joe described is exactly as it sounds. So in our case there are no fees of any kind. So if you see $100 shirt and you say it's going to be for payments of $25, that is it. You cannot you can take a decade and you wouldn't be able to use affirm pretty soon after you went well past the window of repayment. So the punishment for for being late for too long is, hey, you need to pay us back actually, before you can do this again, which I think stands to reason much more than actually let us hit you with a $30 late fee. But that's beside the point.

So that's one form of transaction. We also have what we call longer term zeros, which is for $1,000 fancy suit. You could potentially get a loan from a firm for 24 months or even 36 months, which will also be zero. In both of those cases, it's the retailer that's effectively paying your interest on value of money becomes even more expensive in six months. So you've sort of you've nailed the business model there. In some situations. The retailer just doesn't have the margin or isn't interested in paying your interest. And so they'll say, look, you're going to use credit card, you'll pay some interest. If you're going to use a firm, it's okay for for us to have this transaction happen. If you are willing to pay interest to a firm.

And in that scenario, you will see both the principal and interest will calculate for you both the rate and the dollars. And we'll show you the schedule. And before you commit to any of that, you will see and agree to the schedule that we agree on. In the latter case, the consumer is paying us something. The merchant is probably paying us something significantly less than what they're paying us. If they're absorbing the cost of interest, the cost of time, value of money.

The business model is really, really simple. In our case, to understand the value is fixed, so whoever is paying us almost doesn't matter. That has to cover the fixed cost of underwriting, servicing all the usual bits leave a little bit for us to actually be a profitable company, which we are. And also absorb the probability of default, which, by the way, sort of all the way back to this alignment of incentives and easy check. Like so did it work like this whole idea of like, oh, you know, you're just not going to profit when people stumble. Our delinquency rates are about half the industry of credit cards. That should give you a sense for we don't make nearly as many mistakes. Or perhaps we are not willing to let people go late because we don't benefit from it. Back to the business model. That's how it works.

The competitiveness of this industry is just like any other competitiveness of any payment. Industry payments are notoriously a competitive world. There's actually never been a monopoly in payments. Full stop is a fun puzzle. I used to do cocktail parties like name a monopoly in payments. It turns out there has never been one.

I go to talking to all about cocktail parties. Yeah,

I, you know, there's a reason I'm a fan of the show, right? I listen to, to, to you opine and all sorts of esoteric things with, with great gusto. And so I think that's we'll tell you everything you want to know about the competitiveness of the, of the space. But we have decided from the very beginning that we want transparency and honesty with everyone involved, with our borrowers. That's why we say everything upfront. That's why we tell them, here's your disclosure on exactly the rates and the schedules with our merchants. We tell them, here is the price you will have to pay us if that's where you're paying. If you can't, don't want to, that's fine. We will probably not be the right one for you. We won't make it up elsewhere. We're not going to take it out of the consumers hide when they least expect it. That's just not the brand.

There are definitely cheaper providers out there. There are plenty of brands that will say we're just like a firm, but way less expensive for merchants. What that really means is that's because we charge all sorts of hidden fees. We have all sorts of other ways of taking it out of consumers skin. What this means practically, with 15 years of experience is consumers as they cast their eye over the list of buttons, they say, well, the one I want to use is the one that's not going to get me. If I stumble. And that makes for 24 million active users in the last 12 months. And we feel like we've we've kept our promise and the consumers keep coming back.

Just reiterating Joe's question, though, on the merchant side again, I imagine they're getting offers from a bunch of different, installment lenders, Bnpl companies, and I don't think they do. They care much about the underwriting. It's not really their problem. It's your problem. Right. So I assume the thing they care most about is the fee,

which seems like they actually care about three things. Okay. So they absolutely care about the fee like that. Without it, which which can range from as low as less than credit card. So a typical credit card rate today is on the order of between two and a half and 3%, depending on the size of the merchant. But that's kind of the average in the industry. We will definitely meet you around those numbers at the absolute lowest, and goes up all the way to, let's say, low single digit percentage points when you are subsidizing consumer interest. So we are not a cheaper than credit cards provider like that. That's not you know, they want anybody to have that illusion.

They care about the rate more than the rates. They care about incremental sales. So if your next marginal buyer is either a no thanks or it will cost you 5%, if you have the margin of better than 5%, you don't want the item staying on your shelves. You would like to pay this 5% to get the marginal buyer to say yes and so because of that incentive, which I think is easy to follow, you do have lots and lots of merchants that say, all right, so what I really care about then is the approval rates, which is the second variable we care about.

So as people come through and say, hey, I'm not going to use a credit card, not interested in buying unless I have an installment loan, let's say from a firm. They need to know that that rate is going to be high enough. Otherwise why have a button that just sits there and looks pretty? And so our rates are typically higher than the rest of the industry because we are so obsessed with underwriting, because we have no crutches e.g. late fees. And so this underwriting thing, I don't we don't really try to explain it to merchants at all. What do we do? Say, look at our approval rates. If you see their approvals reach a meaningfully higher than the industry, you know you're going to get value from this. That's why it's worth paying more than credit cards. And so that is true. And that's that's what they do.

The third piece they do care about every merchant knows that. The second transaction is they want they want the first transaction cost them probably more than the margin in it. They're paying Google, they're paying Facebook. They'll soon be paying an Elm provider or Chatterbot provider to drive that first transaction in. They want to sell you another thing. So they very much care about if they had sold that first transaction using a buy now, pay later provider, it better not harm their brand because if they buy now, pay later provider is harassing the consumer for paying them back, or has late fees or has unexpected fees. It's going to accrete negatively to the merchants brand. And so they actually do care about lead fees in a way that has less to do with financial consequences, be more with will people come back to us and use our firm again? Or will they be like, oh, I hated that merchant. I got into some bad transaction. So they care about all these things through their own lens.

That makes a lot of sense. So let's say I start accepting your premise. Okay? This is per se better than, credit card. One reason that people can't just sort of quit credit cards generally is not every place takes buy now, pay later. People want something in their wallet, right? They want something at the restaurant. And wherever it is that they know it's going to be accepted. Now, if you are launching a card or you have a card, right,

we have a card

okay around for a couple years. So you have a card. But I'm curious, you know, you you mentioned you referenced that recent episode that we did with our Drexler. And one of the things that he said about cards is they're very costly to advertise. The marketing expenditure for the card companies is very high. You mentioned that you don't do a lot of advertising. I'm curious, though, on the card component specifically, is that going to be, you know, if you want to be, quote, top of wallet to consumers? The first thing I think I've heard that phrase, I sorry.

That's right. I'm laughing because I'm really I'm trying to claim myself with the industry lingo, your true insider, that if you do that, that's, you. So if you want to be top of wallet for the consumer and in the card, can you do that with an advertising lite model, or do you really need to spend to get significant wallet?

Sure you can. The card is available to our users. So this is not a thing that you'll see on times Square displays. This is a thing you find out from affirm once you are a borrower. Once you're actually in good standing, you qualify for the affirm card as the hey, you really liked us, didn't you? This was great. You're transacting five times a year. If you want to take us with you to your retail, shopping experiences, restaurant or otherwise, we have a way. It's called the affirm card. It's actually pretty magical. The card is a dual mode card. It switches from debit to credit explicitly as you tell it to. And so when you're buying a burrito, we expect you to use the debit mode, which just literally takes the money from your account and that's it. There's no interest, obviously, it's just a paynow transaction.

If you're buying a TV, you can buy it now with affirm in the app, you say, hey, my next transaction. I want that to be a 12 months to 0% loan because the retailer wants you to have it interest free. Maybe you're paying some interest, but you're setting up that transaction in the app and then your card is ready when you tap it. Next time it becomes a loan automatically. It's a pretty magical experience. It's super popular. It's a it's a thing that we haven't needed to advertise because we went from, we have this card idea to about 12% of our users having one very, very quickly because it's just a really, really good product.

Again, our user, their mindset has long been changed from revolving is okay, I don't understand it, but I don't care to. I don't like revolving. I don't want to pay late fees. I don't want to pay excessive interest. This affirm thing is neat. The card is a logical next step. It's how much more can I have of this type of financial life?

How much insight do you have into, how much people are borrowing from other Bnpl lenders? So the sort of stacking issue, which has been discussed multiple times now, it's a great question.

We have a decent insight in the sense that we understand pretty well the overlap between user basis. So there's two prongs to his answer, and this one, it's worth delving into pretty deeply, because this is where we differ from the rest of the industry pretty significantly. So before I get into the stacking part, I'll address the, the overlap.

So we do a lot of studying just externally through surveys. Also because when we log into your bank account or we ask a provider to log into your bank account and pull in some of your transactional data, we get a glimpse into what else you're spending money on. So we have a decent sense for where else you might be using buy now, pay later. And to date, the user base overlap with other providers is quite minimal, so there's not a whole lot of stacking possibility simply because people who use affirm tend to state with affirm they sometimes straight and use other providers, but it's quite minimal for a meaningfully lower transactional amount. I think we average a significantly higher average transaction relative to the rest of the industry.

So far, that's not been a problem in the long term. You, of course, can make the argument like, sure, right now you guys are all still tiny. You're growing really quickly. Everybody's taken a bite out of the credit card industry. What happens five years from now when everyone has signed up for their favorite deal, and they're now signing up for a second one? The right way to address this problem, and this is the right way. And it's it's traditionalist of me, but I really do believe is the right way is for everyone to report to the credit bureaus.

Today, when we tap into one of the three major credit bureaus, we see what you're borrowing and relatively current state of your borrowing from other traditional sources of credit. And it is a variable that we use inside of our own underwriting. A firm today is the only one that furnishes. That's another fancy industry term, delivers the data of both positive and negative as you're on time, you're not on time to the credit bureaus. The rest of our competitive set, generally speaking, does not. Although everyone has experimented a little bit from the very beginning, we saw the idea of building credit history and improving credit scores for people who are on time as a key value proposition for our borrower,

like we are not going to be the only lender for our consumers that each bar for a card they need to borrow for a house one day. That means whatever they're doing with us, especially if they're doing well, which 97% of our consumers are doing really, really well. They're on time all the time. It has to reflect in their permanent record somehow. So we've been furnishing basically from the very beginning. It took a long time to persuade not just the credit bureaus, but also the companies that provide traditional credit scores to understand the data we furnish and treat it in the right way, so that when you are on time, the scores go up, when you're not on time, the scores go down, etc..

And so we did all that work basically on a volunteer basis over the last decade. And today, right now, the state of the industry is if you are a Bnpl provider and you're not furnishing data, you kind of have no excuse. Like, we pioneered this, we did all the work, we worked with the score builders and the credit bureaus, and we're now sending all of our data. And so we want everyone to join us. Because most people do pay on time, their credit histories should reflect their good repayment, but also people who should not be borrowing should be reflected. And I think that's just really, really important.

If I can, if I can do one bit of advertising on your show, it won't be for a firm. It will be all of you out there. If you are in a buy now, pay later industry, furnish your damn data. It will help consumers and it will eventually agreed to your brand too. But for now, we are the only major one.

Wait, what's the resistance then? Because it seems kind of obvious. If everyone cares about the underwriting, you would want more data out there. So why not provide it?

So I don't want to pretend that I understand all the motivations, but I think a basic inferential thought might occur that if you're making a lot of money from late fees, it's a cool thing to tell your consumers, don't worry, you can just pay me off and it won't go on your permanent record. And if you don't make any money from late fees, you have no incentive to tell your consumers, oh, it's okay, just pay me some sum and, we won't tell on you. The connection between late fees and lack of reporting is unfortunately, the, the underlying fact that because we don't make any money from late fees, because we don't charge any, we're very pro reporting just out of curiosity.

So if you're talking about late fees and credit and all that stuff here we are recording this on December 2nd, 2025. You have any, you take the temperature of the consumer for us. I don't think you're doing seeing any signs of growing missed payments or anything like that.

Not as of December 2nd, 2025. The important thing to understand I I'm asked this question fairly often. I am a great I am a firm is a great lens into the financial hearts and minds of North American consumer. Operate in the US, Canada and now UK. It is a specific set of people. These people are more financially responsible. They made a conscious choice. They picked us not just because it's installment pay. And yeah, it's convenient. It's also because we told them we don't charge late fees. It's also because we told them he'd be asterisk next to the 0%. We don't have any asterisks. And one of our core values is no fine print. And we mean it. And on and on and on.

And so it stands to reason that affirm consumers will be probably much later into any kind of a macro economic earthquake territory. If there was one to come. But as of right now, we're we feel quite good about our book and also about the U.S. consumer.

I'm going to ask you a question and, I'm going to mention a competitor. I won't make you name the competitor, but I'll just say it my question and then I'll ask you about. So one of your competitors, Klarna, recently announced a new dollar pegged stablecoin. And I'm very cynical. And so when I see, like, random companies announcing some crypto thing, like sometimes it makes sense, but, but often it's like, well, maybe is this just a good press release or something like that? And I haven't really looked at the Klarna one. The but for you affirm right now, when you look at and you as someone who has been in the internet payment trenches for literally decades right now, does do stablecoins or anything crypto unlock anything for you right now? When you think about the future and you think about opportunities that you can't do with traditional rails,

the short answer is I don't think so. Now, I'm also very cynical, so I think you're unfortunately tapping into the exact same mind flow that that you have.

This is our way of asking them to want to come on the podcast and have argue against it, but keep going, keep going.

I, you know, I, I am confident that there is some thought that went to the press release. So I don't think it's a pure press release. So I will, I will I get I try to tell us about I won't make you. We're not going to make you speculate on a competitor, but just tell us about from different perspective.

Sure I will strawman the reasons why. Yes. So why not is obvious. I think you probably can do a much better job than I can. The why? Yes. So there are a couple of reasons. One, in general, if you have a lot of cross-border commerce, foreign exchange, the complexity of just diligence around, you know, everything from sanctions to reserve accounts, etc., you can sort of squint a little and say, you know what? If we just lived in dollar pegged stablecoins everywhere all at once, it would be a little bit easier. And I think that's true.

E-commerce is not a very significantly cross-border discipline. And so before you sort of say, oh, wouldn't it be amazing if I bought lots of things from Zanzibar and just paid them with a firm, but they got stablecoin as a settlement currency, they'll be very happy because they want dollars in this thing is essentially dollar equivalent. It's a good story. Just most people in the US don't buy a lot of things from Zanzibar. They buy some things in other countries, but most of the time, because we all want things shipped to us very, very quickly, we're probably buying it from a US subsidiary that's warehouses somewhere pretty close to the US, and so it's not that big of a market just yet. But in like five years, if you see me putting out a press release, it's not because I finally realized that I'm missing out. It's because there's enough cross-border commerce in our scale where it starts to make a difference. That's the best I got in terms of why would you want to do this?

The obvious of why nots. It's like, well, maybe you want to move money around person to person, but that's not our business. Maybe you want to believe that one day Joe will have a wallet full of stablecoins. I really have a hard time imagining you supporting a whole bunch of different kinds of brand. A dollar pegged, effectively equivalent stablecoins like a coupon book of I got my affirm stablecoin and my brand XYZ stablecoin, and I'm going to choose between the two like none of that.

Maybe some of the like. What about for like rewards? Like, you know, that's one thing that credit cards currently have that like, oh, I get miles. Like, could there be something with if I use the stablecoin, I get remitted, but I don't know. I'm just I'm just throwing stuff out.

So for what it's worth, I spend a lot of my time contemplating the word ecosystem and mostly raging against it because it is one of the less documented wealth transfers. And by the way, the credit card industry in general is a regressive wealth transfer where people who never evolve transactions, like probably both of you and me, when I use credit cards, which is really rare these days, we're paying our bills. At the end of the month, we're not paying a penny of interest. I think Joe on the record saying you've never paid any interest. And so I think that, that, that, that that's a nice place to be, if you can afford it, telling people that they should just not borrow and pay it off at the end of the month is an incredibly let them eat cake equivalent for the 21st century, payments. Lots of people in America revolve half the country involved in a $10,000, and so the idea of those people caring about rewards is sort of preposterous. They do not. They care about making their minimum payment.

People who are figuring out which reward scheme they want to be a part of. I, I'm just not so convinced that they're going to be expanding their horizons to have more stable coins, because ultimately they're they're getting dollars in the world of a firm people that are people that we serve every single day, 24 million of them in the last 12 months. They really, really care about the interest rates they pay, because a lot of them understand that the alternative is revolving. And so for them being told, hey, your reward is a 0% loan is incredibly powerful. And so before we get into using our hard earned revenue into stablecoins, we will just give it back to the consumers in the form of no, no interest.

So we talked about how you make money, but we should talk a little bit about your cost side as well. And you mentioned funding earlier. So you have to get money from somewhere and you're paying for that money. What levers do you have to pull if the cost of money goes up? And again, I'm aware that we're recording this in early December and everyone expects an imminent rate cut. But in theory, if interest rates were to go up, what could you do to offset that additional cost?

So the most important thing to understand about our business, from the capital sourcing part of the game, everyone understands that rates are not super volatile. But, you know, if you look back a couple years, they were very volatile. If for for a brief, not so shining moment. And so we all knew it was possible. So all of our contracts with various lenders of money to us, but also people who will buy our loans, the all adjust fairly gently over a course of a fairly long period of time. So it's not about can we deal with rising costs of credit for us, it's how quickly do they change for us? And of course, we understand that they might change quickly. Therefore, a lot of our contracts stipulate that, hey, if they do move up, the adjustments to our cost will go fairly slowly over a period of time.

So the more likely outcome in December is probably a downward motion of the current rates. Or perhaps they stay steady, but just the same as I described, we're not going to wake up to all 25 three basis points of incremental revenue. It will eventually come to us in a form of accrual margin, but it'll take its time because these contracts adjust both up and down quite slowly. And we're in no way unique in the industry. Everybody does that.

The levers we have to deal with such cost changes are exactly what you would expect. So we will either pass it through to the consumers and moving up in the increments or down of 25 basis points is not that significant. Then most people, I think, generally don't care that much. In some cases, we actually choose to tell, hey merchant X, you are providing these rates for nothing. You should continue doing so, but the cost to you will go up a tiny bit because the rate of change now, because of time value of money, the merchants pay us in real time, essentially as a transaction is consummated, the true cost to them is truly de minimis. And so in either of these two cases, so long as the movements are not very violent, it's not that major a component of our business.

I just have one last question. I so setting aside the technology for the underwriting, which I'm sure is very high tech, data intensive application and setting aside that, probably many of your engineers on staff are using code generating for that. Just putting your tech hat on. As someone who oversees a large organization, people looking for how large language models in particular are being deployed in productive capacity at a firm today, is any of this technology in use in production and saving human hours in some respect or another?

Yes. Emphatically so. So we just wrapped up the Cyber Weekend, Turkey five, Black Friday, Cyber Monday, whatever, whatever you want to call. I don't actually have the stat yet because I will get the report immediately after this podcast is recorded, but I'm constantly, well, I probably can't, pronounce the results we report early next year, but we will have handled tens of thousands of consumer contacts, people saying everything from, hey, I just borrowed money from you and I don't understand when my first payment is due. And, you know, maybe it's because they hadn't read the emails clearly or everything all the way down to I just borrowed money and I realized that I need a refund because I actually have no intention of buying this thing and all this stuff, and a lot of it.

We try to serve them in real time as they're transacting, but plenty of people contact us right after or maybe sometime, you know, after the purchase. A huge percentage of that is handled entirely by AI. Now, now, that has not caused us to lay off our wonderful customer service staff at all. What it has allowed us to do is to go really deep into specialization. AI is very, very good today at handling basic questions. It can do cool things like look up your account and say, no, you're not late, you're mistaken. Don't worry about it. Or yes, you're late, but we don't charge late fees. So just please make yourself current and we'll move on. All of that can be handled by AI wonderfully.

If it's something like I changed my name or, you know, I changed where I live, right as I was consummating this transaction, there's all sorts of crazy things that come up out of that is something that a human can handle. AI models are not smart enough yet to handle some of these things, and also frequently, if we wouldn't want the possibility of hallucination to derail what is a good customer relationship. So we've been able to move our human helpers into a much more sophisticated, much more specific role. So the theme in our customer service for last year has been specialization, specialization, specialization where we train people now to serve very, very specific subset of problems because they can be effective and move faster. And so these tools are actually making them more efficient by letting them focus on just a very specific thing that they're good at.

So that's one example. We track all sorts of interesting metrics about AI tool usage internally. The interesting or sort of random factoid, our engineering group is not the single largest consumer of AI tools. I will not reveal exactly who it is. And by the way, the vast majority or cannot reveal who it is.

Give us a hint or something.

It's actually finance. Our, finite finance group uses these tools obsessively. I'm cheating a little bit here. So for a very long time it's not the case anymore, but for a very long time, because we are so AI and ML heavy company. One of the core requirements for any employee, but certainly in finance, was you have to know how to read code and probably write code. And so huge percentage of what we do in our finance group actually looks a lot like coding and a lot of software engineering. And so but there's some how maybe because the smaller group on average. But, they're great adopters of these tools. Our legal team uses it all the time. We have literally hundreds of thousands of very custom contracts. Every time a merchant pays your interest for you, they signed up for it. Contractually, that's a custom contract. Imagine managing half a million of these things. They're all carefully written and spoken, so I can read and find errors and corrections, etcetera, etcetera.

We use. So one of the things we're responsible for as a regulated business is we cannot allow merchants to advertise our service incorrectly. We're actually on the hook when a merchant says something like hooray! Affirm interest free for everyone. It's not true. Some people will actually pay some interest sometimes. And so we have to. We are responsible for finding that and telling the merchant, please stop saying things that aren't. Strictly speaking, true. AI tools are amazing at reading boatloads of advertising copy and saying hey, this second, that is inaccurate. We got a fire off an email and tell this merchant to fix it so you can sort of imagine a flood of ideas that we had with the ChatGPT moment happen, and put it to work for the last few years.

I'm going to end with, a sort of theoretical question, big picture theoretical question. But if you could design the ideal payment system from scratch, complete scratch, you know, using today's technology, there's no legacy card networks, maybe not even legacy banks. What would it look like, and how much of a firm's current business model would sort of survive in that environment?

Well, I think you would look an awful lot like a firm. A firm is my personal attempt to build a system that I can be proud about. One of the core things we tell people who join us, one of the reasons we had this such a black and white, no late fees, no compounding, no deferred interest. None of the yuck is because I wanted the smartest people that would have otherwise gone to Wall Street and traded in quant funds to join us and build the system that they could be proud about. It's very hard to be proud about a thing that makes half its money and late fees, and that's why we don't do it. And so what we've built is biased as I am. It's pretty darn great. It really is something that I'm very proud about. And I think the totality of the team here takes an enormous amount of pride in how we went about building this.

It is dependent, intertwined with the legacy systems. But even as we go through guardrails to use another industry jargon or many other systems that have existed, when we tap into capital markets, which are certainly been here long before we came about, and everything else in between, we maintain the moral integrity of the product and its DNA all the way through. We might pay late fees if we borrow money to lend. We haven't, because obviously we're on a pretty tight ship here. It will not be pass through. We take an incredible amount of pride in the diligence we exercise within the system, all the way down to leaving a penny on the table for our partners to keep because we refuse to make more money than we said we would.

And so I would love to have a blank sheet design exercise in payments. But I think what would emerge is a system that looks a lot like a firm, just doesn't have some of the 1980s craft in it.

Max Levchin, founder and CEO of a firm. So great chatting with you. Very illuminating, helpful conversation, and I appreciate you coming on our lot. Tricia I thought that was great. I thought that was really interesting and just sort of understanding the real nuts and bolts about how the money is made and what's different about the Bnpl companies from the credit card companies, to my mind, very helpful conversation.

Yeah, I thought it was really interesting, the discussion about why other Bnpl companies don't want to report. Yeah, the credit bureaus and things like that. You know, I wonder if that'll change, anytime soon. Maybe politicians will start getting interested in it. Regulators as the space expands.

A thought that I had listening to Max was, he mentioned the sort of the regressive ness of the point system. Right. So there's people like us who rarely roll their credit card, but get frequent flier miles or other rewards, etc. but that's paid for by the people that roll, right? Or sorry, that's paid for by the people who roll. Now, if affirm is targeting, they're not they're clearly targeting people who are not thinking about points. Right. They're thinking about people who really do need to extend their payments or implicitly roll and say, we have a better option. So it's sort of cleaving off if, you know, if this grows, the firm is still just a $22 billion company, by market cap. Whatever this grows, you gotta wonder, like if in the maximal version where you've cleaved off a significant number of people who roll, what happens to the point ecosystem for the people like us?

Now, granted, I don't think like, oh, what about the poor freaking flier and all the great rewards you're getting, but you just someone think the point?

Well, someone please think about the points accumulators who never go into debt or who never, like, roll their debt. But it is interesting to think about, if you could, like, cleave off that, then a big part of the business model of credit card, the legacy card companies could potentially be unstable and part of the appeal of going to any credit card that offers a lot of rewards could decline if revolvers moved to this alternate payment model. Yeah, it's a really good point. Yeah, it's a really good point on points.

Okay. Thank you. That one was a little obvious I'm sorry.

No it's good I like that it was straightforward. I actually got this one. Okay.

Actually you know what? I really also appreciated that Max had substantive answers for things that they're doing with AI that aren't just code generation. Even if it sounds like the finance department is probably still using it for code generation, like that makes a lot of sense. Like being able to proactively scan websites to say who is sort of, claiming something that's not true about a firm, etc.. So maybe there are productive uses of all this technology or is sending out millions and millions of contracts automatically. Totally.

All right. Shall we leave it there?

Let's leave it there.

This has been another episode of the All Thoughts podcast. I'm Tracy Alloway. You can follow me @TracyAlloway and I'm Joe Weisenthal. You can follow me @thestalwart follow or guest Max Levchin. He's @MLevchin. Follow our producers Carmen Rodriguez @CarmenArmen, Dash Bennett, @dashbot and Cale Brooks @CaleBrooks. And for more all lots content, you should definitely check out our daily newsletter. You can find that @bloomberg.com/oddlots And you can join fellow listeners in conversation 24-7 in our discord discord.GG/oddlots and if you enjoyed this conversation then please like leave a comment or better yet, subscribe. Thanks for watching.