Pod Insights a16z Podcast a16z Podcast | Money, Risk, and Software About this episode Financial services are overdue for an overhaul. With a16z's newest general partner, Alex Rampell (who just officially started), this segment of the podcast explores the world of fintech... How software backed up by data is being brought to bear on lending, insurance, and the science (oftentimes art) of underwriting risk.
We also get a taste of what life was like for Rampell running a successful internet business out of his bedroom -- an experience that would lead him toward the world of monetization/ payments and eventually co-founding numerous startups.
Transcript 0:00 Speaker 1 Hi everyone, welcome to the A6 and Z podcast. I'm Sonal, and today Michael and I are doing an interview with Alex Rampel, who's our newest GP. And among the things he'll be focusing on is FinTech. And it's about time we had a FinTech pod, so we've been excited to get him on the podcast to talk about this. Welcome, Alex.
0:18 Speaker 0 Good to be here.
0:18 Speaker 1 So Alex, let's actually just start by talking about your background. I think you break the record among all our GPs for being a serial entrepreneur in terms of how many companies you've started.
0:28 Speaker 0 Well, more is not better. It's it's quality, not quantity. So uh so it's actually kind it's kind of funny. I mean, I fell into all of this by accident, and the way that I started was I got a Mac as a computer. I got the Mac 2, which was the first color Macintosh. Everything up until that point had been black and white, and I was a very ungrateful child. So because all my friends had PCs, all the games were written for PC. So I, hey dad, you know what, why'd you get me this Mac? I'm so angry. Like you got me this like massively expensive. Actually, it was my grandmother who bought it for me. Um, but my dad helped her pick it up. Um there's no games for this, I can't do this, because you literally you go to like some computer store and 99% of all titles were for PC, and one there'd be like this little anemic Mac section, which just made you feel bad about yourself if you owned a Mac in the in the 80s.
1:14 Speaker 0 Um and Apple, of course, learned from that mistake many years later, and now they they really prioritize developers. So my dad said said, well, why don't you learn how to make your own programs? So he got me a book on C. And um I also uh there was a program that some people might remember called HyperCard, but it had it introduced this concept of the hyperlink. So I started writing programs in HyperCard and also in C, and then there was this movement called Shareware where it was kind of the counterculture of commercially shrink-wrapped software that would get sold in Comp USA or Best Buy. This was something where you were encouraged to share it, and if you liked it, you were supposed to pay for it, and that was via checks in the mail. This is actually how I got into payments.
1:55 Speaker 1 So people actually paid you for your stuff. I mean, what were they paying for? Was it a game?
1:59 Speaker 0 So um the first program that I I think I got anybody to pay for was a screensaver. And uh I wrote the screensaver called Mini Screensaver. It drew ovals on your screen. I uploaded it to AOL, CompuServe, a bunch of other bulletin boards, and uh I said, if you like this, uh send five dollars to me in Palm Beach, Florida, and if a couple days later I got a five dollar check in the mail. And then a couple days after that, I got a $20 check in the mail because some company bought four copies. Actually, I still have I think I still have um the photocopies of all these different checks in the mail that I got. Um so I was making like a hundred dollars a week as uh as a 10, 11 year old, and I just I I really focused on these utilities, like what was an annoying I mean, screensavers were not a good example of this, but uh I remember on the Mac, it was very hard to change the sound volume. Not very hard, but you'd have to like go to the Apple menu, go to control panels, and then find the sound one, and then like click somewhere else and then like change it from zero to seven.
2:51 Speaker 0 So I wrote a little control panel, uh, if you remember that term, that you hit control and then a number, and then it would change the sound to that level. That was it. And that was like eight dollars. It was called volume quick change. This was not like complicated code, but uh it solved an annoying problem. So like a lot of people bought that and just kept doing things like that.
3:10 Speaker 1 So how did you go from these kinds of products to payments?
3:13 Speaker 0 I was running my business off of AOL because in the the early mid 90s, AOL was kind of the de facto internet for most people that didn't have, you know, weren't at a university campus and didn't have access to like a T1 line or something. In the I think it was the winter of 1996, AOL used to charge uh three bucks an hour. They used to charge six dollars an hour when I started using it. They cut it to three, then they went to twenty dollars a month flat rate. And understanding the laws of economics, if you lower the price of something, typically the quantity demanded goes up. If you lower the price of something to zero, the quantity demanded it goes up a lot, but they kept the number of modems the same. Consequently, it became impossible to connect to AOL, and AOL would dial each number one time and then give up. So I wrote this program for me that was called Always Online. It would dial each number a thousand times in a row. So it would guarantee that you connected, and then once you connected, it would keep you online forever by emulating modem activity and dismissing these dialogue boxes that AOL would pop up to kick you offline.
4:07 Speaker 0 Always online sold tens of thousands of copies, which was kind of crazy. I was 15 years old, and I might get like a thousand ten dollar checks in the mail.
4:15 Speaker 2 How long till Steve Case came knocking on your door?
4:18 Speaker 0 So it's it's funny you mentioned that. So that was actually the best thing that happened was because I was 15 and AOL was like this enormous hundred billion or whatever, how large company, very large company at the time, they banned the product. And um Forbes did an article about like little tiny 15 year old gets bullied by big bad AOL and here's his product and then sells another 10,000 cops. So anyway, I I needed to figure out how do I get people to pay me. How
4:43 Speaker 2 were people paying until then?
4:45 Speaker 0 The biggest innovation that I had was a uh check endorsement stamp. So I I'd open my mail, um, get a lot of mail, open it, I'd endorse every check, and then I'd go to the bank, and then I'd like to write down like the total deposit. I mean, it was fun. Um, but uh I did two things. One is I bought a credit card terminal on it was called auction web, it became eBay. This was probably in 1996. So actually I started processing credit cards and had to learn how to do that. And what I did was I I with Always Online and all my other products, I'd bundle a little application where you would enter your your name, your address, your credit card number, it would encrypt all of that information, and then it would send it to me via email. This was kind of like the early web order form. And then it would get processed on my little tiny credit card gateway, which was actually a modem.
5:33 Speaker 0 It would dial up to the bank and then like authorize the card. Um and then I'd just all
5:38 Speaker 1 this at age fifteen.
5:39 Speaker 0 Uh yeah, yeah. So that was You
5:40 Speaker 2 probably couldn't even get a credit card, but
5:42 Speaker 0 I could not get a credit card, but I could charge a credit card. Um actually that was very complicated. I I had a lot of respect when Square came out. I only got the credit card because my dad um runs a small accounting firm and it was actually in in his firm's name was how I initially got the credit card. I think I transferred like the credit card processing relationship to my own company eventually. Because the idea of a virtual company in 1996 was not really well known. So I mean it was it was an incorporated entity, but like, well, where's your principal place of business? Well, it's in my room. Like that doesn't make sense.
6:13 Speaker 1 So that was your first foray into payments.
6:15 Speaker 0 Actually, like the the real way that I got into uh trial pay, which is kind of my my largest company that I've founded, was most people that were using Always Online and all of the other products that I came up with, they didn't pay. Because a lot of this is like here I am holding a vitamin water, which probably costs five dollars at Whole Foods. Oh yeah, I'll pay five dollars for vitamin water. Oh, I'd I'd never pay five dollars for a song on iTunes. That's a ripoff. That's outrage. So people like to pay for tangible goods much more than they like to pay for intangible goods. So
6:44 Speaker 1 even true today.
6:46 Speaker 0 Oh, I think more so true today than than ever before, because there's so many different types of intangible goods. You're kind of hit across like if you were to pay full price for every iPhone app that you ever touched, it's like you'd be paying five dollars every day for something, and paying for coffee, that makes sense, paying for an iPhone app, like not so much, at least not as often. Like even the top games on the on the app store, like they might only get two percent of people to pay. Uh so always online uh varied in price from eight dollars to to twenty dollars, and a lot of different changes in between because you kind of get a feel for the elasticity of that demand. Um so I tried coupons, I tried discounts, I tried this, there's the pro version, there's the light version.
7:21 Speaker 1 So, where did the tangible goods part come in?
7:23 Speaker 0 Um so the idea was wouldn't it be great if I could let you pay for something tangible, and by doing so, I get paid from the tangible goods company, and then I'm able to give you my product for free. So I started doing this um kind of in the early 2000s, which was this was the the harbinger of of trial pay, which is okay, you don't want to pay for my product, I'll give it to you for free if you go shop at gap.com. It turns out gap is a 50% gross margin or more. You go spend 200 bucks a gap, that's a hundred dollars of gross margin that they have to play with. They almost every major retailer has something called an affiliate program where they will pay for the referred traffic. So that's what I started doing, and that seemed like a major problem as well, because I wasn't the only one out there. Little shareware Alex wasn't the only one that had a problem of most people not paying me. Every major software developer, every major kind of intangible goods maker had this exact same problem where a hundred percent of the revenue comes from one percent of their users.
8:17 Speaker 0 If you can get one percent of the other 99% to do something, um whether switching to Geika, which I try to do, but it turns out you can't incent people to switch their insurance, um, or shopping at Gap or signing up for Netflix. So this actually worked really well and started branching out beyond that. Aaron Powell
8:33 Speaker 2 So that's the story of trial pay. Now you're here at Andrews and Horowitz, so what are you gonna be focused on?
8:40 Speaker 0 So I think the number one thing is is financial services, uh, which is now commonly called fintech for financial technology, which really can mean anything. It's kind of it's reinventing the banks because the notion of banks haven't really changed that much in well, probably thousands of years, but they haven't really met the 21st century very lately. Um brokerage, uh insurance is another category, payments, all of these things um have a long way to go and are ripe for disruption.
9:07 Speaker 1 So fintech and disruption all sounds really interesting in theory, but how does this actually play out in p practice? I mean, as a consumer, how would something like this affect me personally?
9:17 Speaker 0 So I'll give you an example. If you borrow money, there's a company called Lending Club. And what is Lending Club do? They kind of they help l most of what they do is actually refinance credit card debt. So if you're a bank like JP Morgan Chase, you have 250,000 people that work there, you have thousands of branches across the country, uh, you have millions and millions of people that have deposits, and you pay those people one basis point, you know, 0.01% interest, if even on their on their deposits, even if they're quite large. Then you have all these credit cards that you issue, and when people rack up debt on their credit cards, you charge 18%. And it's kind of a good business because you're paying one side 0.01%, you're charging the other side 18%. And you have a company like lending clubs That actually a lot of what they do is they show up and say, hey, if you have lots of debt on your credit card, why don't why don't we're going to pay it off for you?
10:06 Speaker 0 You're going to owe us 12%, but where do we get the money from? We're going to take one of those little poor guys that's getting one basis point on their on their savings account deposit and we're going to give them 8%. So in a small way, actually kind of a meaningful way, because it's a multi-billion dollar company now, lending club is helping to disrupt finance. Trevor Burrus, Jr.
10:26 Speaker 1 But don't the incumbents have a structural advantage here? Aaron Ross Powell
10:30 Speaker 0 The structural advantage that Chase has is that they have lots and lots of depositors, they have a very well-known brand, they have all these different systems in place for managing risk. And also the other structural advantage is they can borrow from the Fed for zero percent, which is kind of nice right now. The structural disadvantage is they have 250,000 people. They have thousands of branches. These are all things that were super relevant 25 years ago. That's how you grew as a bank. But if you were going to develop a bank from scratch today, chances are very, very high you wouldn't say my my first order of business is to go sign 9,000 leases at different uh commercial banking centers across the country, you'd say, I'm gonna have a mobile app and that will do it for me. Aaron Powell
11:09 Speaker 2 Alex, are you suggesting we're going to start to see new FinTech companies directly take over Chase's business? Aaron Powell
11:15 Speaker 0 You're probably not going to see somebody who's going to take over and take on Chase directly. It's more of you'll see minor things like lending club that attack little bits and pieces of of its business. And there was um there's something I looked at a while ago where it showed a picture of Craigslist and all the major categories on Craigslist and had little arrows pointing in each one of those categories. Unbundling graphic of credit. The unbundling, and you're going to find that with banking as well. So I think
11:39 Speaker 1 someone already made one actually, CB Insights did.
11:41 Speaker 0 Yeah, they're a bit of RoboAdvisors as an example. So this is like companies like Wealth Front and Betterment, which are doing a better job and um arguably than some financial advisors and charging a lot less because it turns out that software can do a better job of that. So stuff like that. Aaron Powell
11:57 Speaker 1 One thing is you mentioned that you know if you had reinvented banks today in the world we live in. Can you break that down for our audience a little bit more? What does that mean to reinvent the financial services industry? I know we're talking that's a big catch-all category that describes a lot of things from a mobile native, like a software-centric world. Like what does that actually mean? Aaron Powell And
12:17 Speaker 2 also you know, you you talk about lending club, but they didn't reinvent the business at all. They're in the same business. They're marginalized. Exactly. Their margins are just different, right? Um how do you reinvent the banking industry and how do you reinvent the business? Aaron Powell
12:30 Speaker 0 So at the end of the day, um you don't need to reinvent it. Um if I'm a bank and I can line up if if I can source loans in a meaningful way, I can control my risk in a meaningful way, and I don't have all these what I would call structural disadvantages like thousands of leases, hundreds of thousands of tellers, and you know, no disrespect to the profession of the bank teller, but it's probably not one that's going to be around in 50 years, much less 15. If I don't have to pay all these people, and uh I can lower that spread, lending club is kind of a bank. It's not it's not a bank for a variety of
13:04 Speaker 2 reasons, and they can't call themselves a bank. Officially, right. Yeah.
13:07 Speaker 0 Officially. But it really is a bank because they're accepting deposits on one side. Um I mean, that's the thing that they're not a bank, and that most people are not going to want to give their money to lending club because they don't want to say, okay, here are the 19,000 loans that I own a small piece of, and here's how much interest I earn today, and oh, this one defaulted and that one didn't, and I can go sell this loan. No. They want a safe place that they can keep their money. They want to know that it's insured and it's protected. Um but really lending club behind the scenes, exactly as you said, is doing the same thing as Chase, but lending club has, I don't know, 1,500 people and not 250,000 people. Um what Chase, what the new version of Chase for the 21st century, and who knows, maybe it will be Chase. So a lot of times companies can uh figure out how to regrow themselves or become reborn in a way that's more technologically meaningful, but it's going to be the same thing.
13:55 Speaker 0 Aaron Powell
13:55 Speaker 1 Which kind of sucks because people really seem to hate the banks.
13:59 Speaker 0 People hate their banks because it's really not transparent at all in terms of what you're going to be charged. And there's also there's a major market failure when it comes to lending in general. Aaron Powell What do you mean by that? A lot of people have this this mental image of they don't want to see their bank account balance drop below a certain amount. And maybe it's irrational, maybe it's not, because they can always borrow money or maybe they can't, but uh they'll say, okay, I always want to have $5,000 in my checking account. I don't want to see it drop below that. So they'd actually feel more comfortable taking out a loan to buy something than watch that mental barrier get breached. Um the market failure that you have happen is I go to Walmart, I want to buy something, it stretches my budget a little bit, and I know that if I charge it on my credit card, I'm kind of going to get ripped off.
14:45 Speaker 0 Or maybe maybe I'm not, I just don't even know. I know I'm going to get charged a lot of money, and if I pay it late, I'm going to get charged a lot more money. It's not transparent. And the thing that the credit cards, and that's it's not a credit card companies, because I I, you know, I think highly of visa and I've I've worked there for a while. Um it's not just the credit card companies, it's lending in general. It's not really based on the item that you're purchasing. So the market failure happens where I would have bought this if I had been offered a loan at a reasonable rate. Instead, I'm offered a loan where all loans are the same size. And you know that they're different, like there are car loans, there are hour there are housing loans, mortgages, and there is kind of like general purpose credit, and those are kind of the three major categories. But why does a home have a different interest rate than a car than what you're paying on your credit card?
15:34 Speaker 0 In some cases it's it's a secured asset, so that's one reason. But it just people behave differently. Like they they're a little bit different when they when they think about taking out a loan to buy a property versus a car versus some general purpose credit card item. So the market failure happens where like Walmart wants to sell more stuff, they would sell more stuff, they can't sell more stuff because people are hesitant to buy it because they don't know what going into debt for this particular item has or what what that means for them. So that's one area of just lending in general can be made so much more efficient. Aaron Powell
16:06 Speaker 1 Right. Well, I do want to talk about um the financial inclusion aspect of this as well. Another aspect of lending that's inefficient in a way that's regressive to people who maybe don't have the best credit history, but actually it could be very credible. Trevor Burrus, Jr. Totally. This is a
16:21 Speaker 0 huge problem. Like FICO, you know, the Fair is Fair Isaac Corporation has been around for a long time, and that's what FICO stands for. Um a lot of this is based on I I would call it the mathematical concept of induction. If you paid your bills on time, you will pay your bills on time in the future. And there are two flaws with this. One, what if you've never paid a bill on time before because you're 25 years old and you've never had a bill? Well, yet guess what? You're a terrible credit risk, which is not true. Aaron Powell
16:45 Speaker 1 Right. Or what if you pay your bill in cash all the time because you're just not in credit? I remember
16:50 Speaker 0 I I could not get a credit card in college, and the reason why I kept getting rejected is because I had never had a credit card before. And this is the definition of a catch 22. If you look at most people between 20 and 30 years old, they have a debit card and not a credit card. But the point there is that they they are known as thin file customers. They're not bad credit. The term in the industry is thin file. And how do you underwrite credit for somebody that has no file or a thin file? And the answer is you can't using kind of traditional FICO scores. Right. The other thing is that you have this whole category of people that went through some kind of hardship in the past. So 2008, 2009 might seem like a distant memory, but it's not for your credit report. So you might have been doing everything right and you had this double whammy of you lose your job and uh your house is foreclosed, and it doesn't mean that you were being reckless with your spending.
17:39 Speaker 0 You actually you are predisposed to always paying your bills on time. You're a very honest ethical person, and now you have this this black spot on your credit rating or your credit score. You actually might have a lot of cash now. Maybe you won the lottery. Um and all these good things happen to you. People should
17:55 Speaker 1 bank on you, no pen intended, but exactly.
17:57 Speaker 0 People should bank on you, but they don't because they're looking at this one thing that hasn't really changed. So actually you have a whole category of consumers that aren't even thin file, they're thick file, but the file itself is misleading. So this is by the way,
18:10 Speaker 1 where software does have an opportunity to change the FICO definition of how the signals for FICO scores are sourced. Aaron Powell
18:16 Speaker 0 Yeah, it's already changing. I mean, like right now you rely on kind of there, there are the three big credit bureaus and they have to kind of redo their own stuff. And like one thing that's actually happened is there's a change in that algorithm for what
18:27 Speaker 2 if you go to debt settlement.
18:30 Speaker 0 This is actually interesting. I've seen a number of companies in this space where they'll say um that the highest quality consumers are the superprime or ultra prime. Like that's the category where you have 800 plus FICO and these guys always pay their bills on time and until they don't, but uh but they they look very good for the time being in their induction-wise, they've done very well in the past. There's another category where imagine you did fall into hardship. It turns out the people that raise their hands that say, I want to get out of debt and I want to do the right thing, I want to be honest, they're actually very, very reliable people. The signal there is actually quite good because what you want to find is you want to find groups of consumers that uh at least at first glance, you know, judging the book by its cover, the cover doesn't look so good, but inside the book is actually quite good and it's going to perform well.
19:19 Speaker 0 So there's a lot of stuff that is just going to be done much, much better from underwriting, but you do run into you do run to issues with the uh with regulation and whatnot by law, you have to explain why you're rejecting somebody for a loan and come up with I think it's I don't remember off the top of my head, but I think you have to provide the three reasons for like why you're dinging them for the particular loan. If you're using complicated machine learning and you're looking at 9,000 different inputs, that's that's challenging. Um but the other market failure just kind of in lending in general is that the reason why interest rates tend to be very, very high for people that are deemed a high credit risk is because you're counting on half of those people, or maybe 90% of those people, or a huge number of those people to never pay you back. I actually I I take issue with people that think that payday lending um or just high.
20:08 Speaker 0 High interest rate lending is is this evil, terrible profession. Because imagine that I'm trying to just make ends meet. And I I need to pay my landlord on Friday. And my my paycheck is coming the following Friday. I owe the landlord I don't know, a hundred dollars. I have zero right now. I'm getting two hundred dollars Friday. So if somebody showed up and said, okay, uh I'm going to loan you the $100, but you owe me $150 the following week. So that's 50% interest effectively in one week, that's considered like super illegal usury. But the problem is that if you were to charge like a normal interest rate of the max of 30 six percent in the state of Utah or whatever it is, um, you'd get a couple dollars for making a pretty risky loan of 100 under the hope it's completely unsecured. It's like, okay, I'm gonna count on you of your own recognizance to come back in one week and pay me 102 dollars when I'm loaning you a hundred dollars, like nobody would be in that business, and therefore the person that actually really needs the money would not get it.
21:02 Speaker 0 Aaron Powell
21:03 Speaker 1 I think the problem that people have with that though is that it they feel it punishes people already poor. There's a huge power differential in that system. So I kind of I see your point, but I also think there's a point to be addressed on the other side of it, which is this sort of you seriousness that takes advantage of that situation as well. Aaron Powell
21:19 Speaker 0 I I totally agree. I mean, I think this John Oliver did a really good segment on this, which is a lot of the payday lenders are kind of trained to get people to just come back and get more loans, and this is where you run into this problem where the the interest dwarfs the principal, which is bad, but one way of dealing with so one way of dealing with this is regulation. Um, but that can have some unintended side effects, such as nobody will loan to these people because they can't make money. Right. You can have regulation where nobody can loan to these people, but somebody sees an opportunity through software to kind of separate the wheat from the chaff so that by charging $2 on a hundred dollar loan when it's completely unsecured, they actually could make money on that. Right. That's another way of looking at it. But then the other is if you have lots and lots of competition, then that actually takes care of itself.
22:05 Speaker 0 Um so actually I'm in favor of regulation just being like requiring absolute clarity on what you're being charged, because a lot of people don't realize this. This is actually what, in my opinion, uh contributes to that cycle of poverty because you go take out a loan, you don't know what you're getting charged, there's a hidden late fee. So the $100, oh, well, if you don't pay it by, it's actually due at 11 a.m. You paid it at 2 p.m. Aha, $50 fee, and then that's what really kills people. If you make it super clear and super transparent, you're not going to run into that problem. And if you have the law being around uh clearness and transparency, um, and you have lots of competition, that's what will bring the rates down. Aaron Powell
22:41 Speaker 2 What you describe um uh is it sort of just a long tail scenario where you know companies now, whether it's lending tree or or some or lending club or something else. Is it a question of now finding the right risk and all that and doing making all those relationships, you know, person to interest rate um to risk more efficient and and uncovering the ones that actually work and then finding the the right amount to charge for a week where you can make money and they still will be a customer. But how does technology enable all this and how is it starting to change things? Because I hear you describe the sort of landscape. How do we then come in with new ideas that help change what you describe? Aaron Powell
23:26 Speaker 0 So I think software is a huge part of it because this is exactly like the reason why you charge very high interest rates is you just it it is this regression to the mean of, or not even regression to the mean, it's like I don't know. So I feel like I'm protected with a margin of error if I just charge you a lot. Part of this is a regression to or just going back in a time machine to an earlier place. So I find this very fascinating. Like a hundred years ago, you go to the general store and you left your wallet with your gold ingots and your silver back in your home, and you say, hey, general store owner Joe, you know me, I'm Alex. Uh I don't have the money right now, but I live like right near the saloon, like I'll I'll go pay you back next week. They
24:06 Speaker 1 would have trusted you.
24:06 Speaker 0 And they'd say, or maybe they wouldn't. Right. Because they owe you,
24:09 Speaker 1 but they know you're not going to be able to do that. You're the one
24:10 Speaker 0 that got drunk last week and ripped off the barbershop. Um but they Joe can make a judgment and say, okay, I'm gonna just pay me back next week. I'm gonna keep that on my ledger. Can you imagine doing that at Walmart? You go to Walmart and you say, hey, remember I come here every day and I buy four bananas and and uh a rubber ball. Uh I'm not gonna I don't have the money right now, but I'll pay back tomorrow. Then like they escort you out of the store. And the same thing goes for like banking and lending.
24:38 Speaker 1 It's kind of like time traveling through software, like back in time, going back in time because what we're able to do today. Aaron Powell
24:45 Speaker 0 It's really interesting. If you look at a lot of the micro lending operations that are happening in a lot of the developing world, some of it is it's like somebody goes to a village and they kind of just they know people in the village really like they know their parents, like their aunts married somebody else and so-and-so was the officiant at the wedding. Like there's that kind of personal connection, which is very hard to quantify. Like putting that in software is actually like that's not software. That's kind of the human touch. And that's actually work quite well. Um you've lost that when you go to branch banking, you've lost that when you go to these like massive, massive stores where knowing if you're a credit risk or not, you can't say, aha, like this is my uncle's best friend's roommate's sister, therefore, like he's vouching for her, and I think it's gonna work out okay. You can't do that with software, but you kind of can because there are other signals out there.
25:33 Speaker 0 Like humans think that they're very smart, and in many cases they are, but there's a lot of other information out there that can be brought to bear. You mentioned lending trade. You meant to say lending club, but you mentioned lending tree, but actually it's good that you did that because lending tree, like what is lending tree do? It's like, okay, I have I I want to look at every mortgage available to me. I want a Google, but for like seeing all the terms, how much I'm going to have to pay, like you have little meta marketplaces like kayak for lending, which is lending tree, um, and applying that to more I've seen a number of companies in this space as well where they all say, like, we're, you know, the the the hip thing in startup land is I'm the X for Y. So it's the I am the kayak for payday loans, or I am the kayak for whatever you want to call it. And then being transparent around all the different terms there, that's another thing that's gonna push rates lower and be friendlier for the borrower.
26:22 Speaker 0 But the main thing that's gonna change this is the software to identify, okay, we think that you are a positive quality payer. We see what's going on in your bank account. We see things that are just much, much better than what FICO can provide to us. Aaron Powell
26:35 Speaker 2 We're at a phase in at least to financial services where it's it's kind of in the transparency phase, right? It's shining a light on you know terms, conditions, rates, um, and also people and behaviors. And then uh, you know, I eventually it moves on, or maybe not eventually, maybe right now it is into changing how we do things. Um but right now it sounds what you're describing is still this this phase very much of you know shining a light on things and and describing it and looking for inefficiencies. Aaron
27:07 Speaker 0 Powell Looking for inefficiencies and finding these examples where you truly do have a market failure where a transaction would happen or something that's very good for all parties involved would happen, but you either have a lack of communication, you have a lack of transparency, you have uh too much resilience or calcification in terms of why rates are set. It's like, well, why is the credit card rate one size fits all regardless of what item you're buying? That doesn't make sense, but it's too much work to make it dynamic. Aaron
27:34 Speaker 2 Powell When federal fund rates go up, does that change kind of what you might see in the marketplace in terms of like what people are interested in and where the opportunities lie? Aaron Powell
27:43 Speaker 0 So definitely the fact that we're in uh it's called ZERP. This is the zero interest rate policy, we're in whatever it is, year seven of ZERP. So I think we're gonna remain in go to year eight and year nine, that's that's my own personal opinion. But um that does change things a lot. So if you look at uh SoFi, SoFi is a great example. So SoFi is now a multi-billion dollar company and their their kind of big innovation. I mean, I uh not to oversimplify it, but um, if you go to Harvard Medical School, um after you graduate from from Harvard Medical School, or pick on any like high-end medical school, you're going to become a resident, you get paid very little, then you become a high-paid doctor, you get paid a lot more, and across this entire trajectory of your career, your chance of declaring personal bankruptcy is very, very low.
28:30 Speaker 0 The only thing that I need to look at to know that is you graduated from Harvard Medical School. Um Citigroup, on the other hand, so my wife went to actually she went to Harvard uh business school and had a bunch of loans from city, and they were they were pretty high priced because it's like, well, they're saying, well, all student debt is at the same relative waiting. Like they weren't looking at this one vector, which is like you went to a pretty good school where the average person makes so much money and this is the average default rate, and SoFi did. Because SoFi will go say, okay, your city loans are at 7% because they were struck in 2003 or 2004, or I don't know what the numbers are. But uh, this is what they were struck at. Now interest rates are zero. We're gonna come in and do it for three and a half percent, and you'll save fifty thousand dollars and you say, holy, holy moly, this is great.
29:20 Speaker 0 I'm gonna sign up for this. That's a lot harder to do um if rates go up more. You have less room for somebody to come in and start refinancing debt. So that's one thing that would certainly be cause a challenge. Aaron Powell
29:33 Speaker 2 Does behavior need to change, and is that what we're kind of fighting against why things don't happen as quickly? Aaron Powell
29:38 Speaker 0 So inertia is an incredibly powerful thing, especially for banking. So why don't people change their bank accounts to earn more money? So all you have to do is go to bankrate.com or any of these other kind of savings account, like how much do I earn on my money that I have in my bank account? And you'll see like I mean, probably 99.999% of people do not have the highest yielding bank account because it's probably the bank. Of internet or some like crazy name that's listed on on bankrate.com. But think about what goes on with your banking relationship. You have your salary direct deposited into there. So I want to change my bank. And that means I have to go to my employer, fill out all these forms, bring a voided check. I don't even have a like what is a check? I've never seen one before. Where do I get one? I never asked for one. Can my ATM card work?
30:23 Speaker 0 No, that won't. Then you have maybe you have a mortgage from your bank, and they'll give you a 25 basis point reduction on your overall rate, assuming that you have a direct deposit salary integration with them. Aaron Powell
30:36 Speaker 1 Why are they incentive to even do that for you if you're so captive as a customer?
30:40 Speaker 0 Actually, it's twofold. One is to keep your banking relationship. But actually, more importantly, it's to watch your cash flow. So they know what kind of risk you are. They know if you've lost your job because you no longer have deposits coming in from your employer. So they're able to watch that in real time. I mean, it turns out that uh with companies like Yodoli and Intuit, um, anybody can do this now. Uh you don't have to own the banking relationship. If you want to see what people are doing with their own personal funds, like non-cash, but like actual like in a bank account, you can watch that through one of these third parties. So my my main thing is that it's just inertia is very, very it's very powerful. And uh Chase has trillions of dollars of deposits and they offer the lowest rates, which seems like that doesn't make any sense. And any other industry, like if there are like two gas stations, one's over here, one's across the street, and one of them offers gas for three times as much, you'd say, well, geez, like it it's a little bit of a pain to take a U-turn.
31:34 Speaker 0 But I'm gonna take a U-turn because I'm gonna save 50 bucks on my gas, and all gas is gas and it made out of the same chemical compound, so it doesn't really matter. Banking is just you have to get people to care, which is a big problem. Uh there are all of these kind of interconnected relationships which you have to dislodge.
31:51 Speaker 1 Even if it's painful and inefficient.
31:54 Speaker 0 It's not kind of point-click efficiency. You can't just say I'm going to click two buttons and then I'm done. Um, you know, the US government intentionally, so you have all these uh KYC and AML, you know, know your customer and anti-money laundering laws that require disclosures from a consumer. That takes a little bit more time. So you have all of these things as well. Like insurance is another one where inertia is just, I mean, nobody cares who they use for their insurance. Um I know I'll be hated by the geicos and the state farms of the world for for uttering such a falsehood. But I don't think anybody cares at all who they use for insurance. At the same time, they don't even care about getting the lowest rate because once they have it, and Geico spends, I think $600 million a year on advertising in the US to convince you that all it takes is 15 minutes to change your insurance.
32:39 Speaker 0 But no one bothers to do it. There's a there's a quote by George Bernard Shaw, which I'm very fond of, which is um every profession is a conspiracy against the laity. And I I firmly believe in that because there's actually a law in most states, I know this because we tried getting into the insurance lead generation space at trial pay where you cannot reward people for switching their insurance. Um and this is bullshit. So imagine that I'm an upstart insurance company and I do everything better than State Farm. And I say, okay, I know that you don't care if you use State Farm or Allstate or whatever. I will pay you 100 right now if you switch to me. Against the law. Yeah. Against the law. And why can't you do that? Well, because every profession is a conspiracy against the lady. There's no good reason to do that. Um so that's another one where like that should change dramatically. Data would make a huge, huge difference in that space.
33:26 Speaker 0 Like we're still giving people rewards if they go install a low jack in their car. Um that's that's the biggest innovation that you've had in like car insurance, um, or like you get ADT and your rates are lower for your house. But I mean, just think how much better of a job we can do, but you have to get people to care. And this conspiracy against the lady makes people it it makes it harder for people to care because you can't reward them in a meaningful way for switching.
33:51 Speaker 1 Let's transition then to something you wrote about five years ago now, I think, which is this concept of online to offline commerce or O2O, not to add another coinage to our list of B2C, B2B, but that's the same thing. That was
34:02 Speaker 0 the idea you were doing. Exactly.
34:05 Speaker 1 So what is it? And actually, um, given that you wrote that piece five years ago, how have things changed since then? Because some of it is maybe not as relevant today. Aaron Powell
34:12 Speaker 0 The reason why I wrote that and the reason why I was very interested in that is at trial pay we had two markets. We had publishers or app developers that would show offers. So you don't want to pay for Zinga Poker, you can get it for free if you sign up for Netflix or if you shop at gap.com or go to banana republic.com or shop at jCrew.com. Notice the prevalence of the dot com. Um the other side of our business was uh you know, signing up more guys like Zynga. And I liken this to a wine list at a restaurant. If you're a restaurant and you want to optimize your revenue, I mean, the most important thing is you get more people to sit at your restaurant. And if you only have two wines and both of them are red, the marginal benefit to adding a white wine, your your very first white wine is probably pretty high. But if you already have a wine list of 900 different wines, and you add your 91st, like you're better off getting more people to sit at your tables.
35:03 Speaker 0 So those are the two levers that we had at trial pay. We either get more publishers or app developers like Zynga because that that was more tables, or we get more wines on our wine list. Um it turned out we had kind of tapped out on the online wine list, and it turns out, not surprisingly, that 93% of all commerce happens offline. Um that's up from 0% 25 years ago. But still, the majority of commerce in the United States and the rest of the world happens offline. And wouldn't it be great if we could say for trial pay, get your Zynga coins for free if you go buy a coffee at Starbucks or get your Zynga coins for free if you go bowling today?
35:38 Speaker 1 That's what you mean by online to offline commerce. Yeah,
35:40 Speaker 0 so that this was the idea of it's it's kind of a third category of commerce because the U.S. Commerce Department broke this out a while ago, which is there are two forms of internet commerce. There's travel, which is very large, that's expedia.com, hotels.com, priceline.com, all these different uh OTAs. And then there's for lack of a better term, stuff in a box, where I go to Amazon, I buy something, it gets shipped to me in a box, and that's how I actually receive the goods. There is a third category, which means you buy something online, it doesn't get shipped to you, it's not travel related in the traditional sense. That's online to offline. Um it's a huge one because all of that 93% of commerce, sure, you might buy your coffee and redeem it, redeem it. Drink it offline redemption, right? I think it's online parlance. But you go to Starbucks, you go get that, or you go to Burger King to actually eat your big or whopper.
36:27 Speaker 0 Um but at the end of the day, these transactions should happen online. I mean, it should be on your mobile phone. Like you're paying for it online that's happening on an internet connected device, and that allows you to do something really cool, which is actually do performance marketing. Because the problem is if Burger King wants to go advertise on the internet, they don't know if any of those advertisements actually work. Priceline knows every single ad. Did it work? Did it not work? Because people click on the ad. Burger King can figure out if they clicked on the ad. They don't know if they bought the Whopper. Priceline knows that they clicked on the ad and then they went from that click to actually go buy something at Priceline.com because they can track that end-to-end on the internet, and you just don't have that for offline. So I thought this was going to be a huge category that would really redefine online commerce.
37:14 Speaker 0 Aaron Powell And did the term take off? It's funny. Um I I was the one that coined this term, and if you do a search for O2O on on Google, everything is in Chinese. I met these uh this group from China, um, it was like uh CEOs of like all these large Chinese corporations that came over a little while ago and they're like you invented Oto Oh no, you didn't. That's not possible. If you're like you like
37:34 Speaker 1 the Rockstar of China for coming over. Well, not though. It's like the term is the rock star of China. So because
37:39 Speaker 0 like if you if you look at Baidu, if you look at a lot of the big uh companies over there, like a lot of their focus is on enabling online to offline. That's actually that
37:46 Speaker 1 was a big theme of a WeChat primer that we put out. So how have things changed since when you first wrote that? I mean, that was five years ago. Aaron Powell
37:53 Speaker 0 The first category of online to offline was really for things that were easily measurable. So uh Uber and Lyft came about a little bit after I wrote that. Obviously, it wasn't in any way, shape, or form based on on my idea, or not it wasn't even my idea. It was just kind of trying to to uh come up with a canonical form for what was happening. Um but it was very easy to do, relatively speaking, to do Uber and Lyft because it's time and distance travel. Those are kind of the two variables that you're measuring on. And you're taking something that it was originally delivered offline, will always kind of be delivered offline, like moving you from point A to point B will be delivered offline. Um but you you start that process and you pay for that process online or on mobile.
38:32 Speaker 1 That's great, Alex. Well, thank you for your time on the AC and Z podcast. We I'm sure it's gonna be the first of many. Welcome, and we're excited to have you. Likewise.
38:41 Speaker 0 Thank you so much.
Previous Company Building in Crypto: A Conversation between Brian Armstrong and Chris Dixon Next What's Next for Education Startups Page updated August 25, 2026 · transcript and summary are machine-generated