Alex Rampell on why the best companies have hostages, not customers

Alex Rampell is a general partner at Andreessen Horowitz, where he leads the firm’s $2.7 billion Apps practice and has driven much of the firm's fintech investing, including Plaid, Mercury, and Opendoor. Before a16z, Alex co-founded multiple companies including Affirm (NASDAQ:AFRM), FraudEliminator (acquired by McAfee in 2006), Point, TrialPay (acquired by Visa in 2015), TXN (acquired by Envestnet in 2019), and Yub (acquired by Coupons.com in 2013). In this episode of Summation, Alex and Aure...

Speaker names marked ? are inferred from speaking time and the show’s known hosts — they are not voice-verified.

• 0:00Speaker 00

  Every human in the workplace wants two things. They want to be lazier and richer.

• 0:03Auren Hoffman?

  My guest is Alex Rampell. Alex is a general partner at Andrisian Horowitz, where he leads the apps fund and he has driven a lot of the firms fintech investing like Plaid, Mercury, Open Door.

• 0:14Speaker 00

  Hey, do you want to go home at like 11 o'clock tonight or 5 o'clock tonight? For the next like five weeks, everything else stays the same. It's like, I think I'd rather go home at 5 o'clock. Yeah. Would you rather make double the amount of money or half as much money? Right? You just gotta choose, right? It's like, I think I want to make double as much money.

• 0:32Auren Hoffman?

  Hofer Deter, my guest is Alex Rampell. Alex is a general partner at Andrisian Horowitz, where he leads the apps fund and he has driven a lot of the firms fintech investing like Plaid, Mercury, Open Door. Before A16Z, he was a serial founder. I met him a long time ago when he was co-founded Trialpay, which Visa acquired. He also co-founded a firm with Max Levchin. He's even started in building software companies like when he was like a kid, I think when you were like 13 or 14. Alex, welcome to Summation.

• 1:03Speaker 00

  Yeah, it's great to be here. Good to see you across country.

• 1:06Auren Hoffman?

  The first course I've been like meaning to ask you for like a really long time, I've never actually asked you this is like, why are Visa and Mastercard so enduring?

• 1:14Speaker 00

  So there's a great term that was coined by this guy, DeHock. And he was the founder of Visa and it's called chaotic. And it's like chaotic. What is that even a word? He made it up. It's a portmanteau of chaos and order. And the origin of, and Visa, like he wrote a book, I think it's called The Chaotic Organization. So if you really want to dig into this.

• 1:35Auren Hoffman?

  Yeah.

• 1:36Speaker 00

  Oh, you did? Okay, great.

• 1:36Auren Hoffman?

  Oh, yeah. Yeah, I actually talked to him a few times back in the day.

• 1:40Speaker 00

  Oh, so did I. That's great. Yeah.

• 1:41Auren Hoffman?

  He's a true legend. Yeah.

• 1:44Speaker 00

  Okay, so here's why it's so impregnable is that normally it's almost like a decentralized thing before crypto. So the way that I describe it is first, it's helpful to understand what Visa and Mastercard are. They are the intermediaries between really four parties. So there is, I'm going to use the example of you, Orin, have a Visa card that is issued by a bank. So a bank gives you a card. So you're party one. Party two is your bank, we'll call it Bank of America. It's that, that runs on the Visa networks. So Visa's party three.

• 2:19Speaker 00

  You go to Chipotle, that's party five, but who's party four? I skipped party four, but I'll get to that. Party four is, call it Chase, who signed up Chipotle to go process cards for them. So there are five parties. There's like, you really think it's like, well, I'm just buying a burrito at Chipotle. There's Chipotle, there's me, but no, no, like Chipotle has a bank, you have a bank. And then in the middle is this company called Visa. And what Visa and Mastercard really do is they compete to get issuers to issue their cards, but they don't set prices, right? They don't, they don't take, like, most people say, oh, you know, I'm paying 2% or 2.9% for credit card processing. Visa is so greedy. Most of that doesn't go to Visa. Visa makes a very, very small amount per transaction.

• 3:04Speaker 00

  Most of it goes to the issuing bank, but Visa constructs the rules. They handle the, the kind of settlement of funds. They're just kind of this like permanent intermediary and originally it was owned by all of the banks. So the origin of Visa is that Bank of America, which originally, if you go way, way back in the, in the eons, it was called the Bank of Italy. It was in San Francisco. There was a major earthquake. There's this famous story of like the proprietor, I forgot his name, Italian guy, Italian American, was giving out loans after, in 1905, after the mega earthquake. You know, Bank of America now is, is based obviously on the East Coast after many, many, many mergers. Kind of changed the name for Bank of Italy to Bank of America when it was a West Coast institution.

• 3:46Auren Hoffman?

  It was like Giannini, right? Like that's like, cause all the, there's so many things in San Francisco are named after him, right? Yeah.

• 3:52Speaker 00

  And because he famously like helped the, the city rebuild when, you know, why would I give you a loan? You're not, you have nothing. Your entire thing burned down in the fire after the earthquake. Bank of America released something called the Bank of America Card. And this started in Fresno, California. They kind of basically just gave everybody in Fresno a card.

• 4:10Auren Hoffman?

  This is like in like the 50s, right? Where you like the famous drop or something like that?

• 4:14Speaker 00

  Yeah, it's either late 50s or early 60s. I forgot which, because the first credit card was Diners Club. It was really called a charge card. They weren't giving you credit. It was just like, hey, remember me? I'll pay you back next time. And then it kind of codified that in a card. So Diners Club was the first one to do that, because otherwise you'd have to take wads of cash to, you know, fancy restaurants in New York City. And they just kind of created this cooperative where every restaurant would just write down your number. And Diners Club facilitated that. American Express, which originally was like, if you look at their logo, it was like the Pony Express. They had, there are horses on the original American Express logo or something, just like Wells Fargo. They were like doing transcontinental transport or something like that. They had travelers checks. That was their first foray, I think, into money. And then they copied Diners Club.

• 5:00Speaker 00

  But then Bank of America and this guy, DeHock, came up with the Bank of America Card, dropped it to everybody in Fresno, but they kind of realized, and this goes for your question on impregnable, it's a long, many way of getting there, but kind of realized that this is a good idea. I'm not going to be able to get every other bank in the country to go issue a Bank of America Card. Like, that's not going to make sense. So let's come up with a consortium. We'll let other banks own it. We're going to call it Visa. And there was a competing thing called Master Charge, which became Mastercard, a different consortium of banks. I think that one came out of Chemical Bank or I forgot the background there. So all the banks owned this thing, because they all owned this thing, like, you know, they, they all had a account. There was like a council of elders that would basically set what pricing was.

• 5:45Speaker 00

  But it was a nonprofit, like the, the objective of Visa was not to make money for Visa. It was kind of like this holding apparatus for all of the banks to coordinate on pricing.

• 5:54Auren Hoffman?

  And there's some common tech stack too, right? Or something like that?

• 5:57Speaker 00

  Yeah. I mean, it was just, it was basically a backbone that connected, again, the issuer and the acquirer. That's the, the issuer is the, going back to my five party example, the issuer issues the card. The acquirer acquires the merchant, hence the names. And all that Visa did was connect the issuer to the acquirer. And because it's a very, there, there is a, I mean, it's not heterogeneous. It's heterogeneous in terms of banks, but they're all like, every issuer kind of looks the same. They all agreed to use the same card. It all looks the same. It's always 16 digits. It always starts with a four if it's a Visa card, you know, it's a set of standards. But consequently, like, if you get all of the merchants and if you get all of the consumers, like, that's the network effect. It's very, very hard to dislodge that. But the real power that Visa has, and it is a duopoly, it's not a monopoly,

• 6:42Speaker 00

  is they compete for Chase to issue their card type. And this is actually why Visa bought my company, is that it's kind of classic enterprise sales where imagine, I mean, you've been in this space, like, imagine that I say, hey, here's my product. You've seen this, right? It's like, there's a, there are like checkboxes and X's, right? So it's like, here's my product, here's the competitor product. Those are the two columns. And then it's like feature, feature, feature, feature, I, you know, my product, check, check, check. There, X, X, X, X. So I, I go to like Barclays or I go to Chase or I go to Citi. And it's like, hey, you should issue Visa cards. Don't issue Mastercards. Here are the superior things that you have by running our network.

• 7:26Auren Hoffman?

  I have one more check now.

• 7:27Speaker 00

  Exactly. I have a check and it's called the Commerce Network. And that's basically what they turned Trialpay into. So, so the reason why it's just so impregnable is it's like, they are a network of networks.

• 7:40Auren Hoffman?

  It's like a true network effects business, like it's a true.

• 7:44Speaker 00

  It is the purest network effect business in the world. But it's also strange because you'd think if it's the purest network effect business in the world, you wouldn't have this duopoly setup, right? Like there isn't like an eBay Junior to eBay. I mean, there are a lot of like specialized Ebays. There's Copart for, you know, used in total cars. Totaled cars, that's Copart. Copart, I think, has a bigger market cap than eBay. There are a lot of auctions, but if you're selling like, you know, a home run baseball, you're going to sell that on eBay because it has all the demand, because it has a lot of demand and has all the supply. You kind of need to issue a card as a bank that you know consumers will want because they can use it everywhere. And this is why, you know, Discover Card had a hard time. There were these kind of like smaller networks, but they didn't get used everywhere.

• 8:28Speaker 00

  So you really have to have a Visa card or a Mastercard and then you get this ubiquitous presence.

• 8:33Auren Hoffman?

  Why do you think there is a duopoly?

• 8:34Speaker 00

  Early on in the origin of this whole thing in the United States, there used to be, and this is where the name Master Charge came from and then later Mastercard, is that like every department store would give you their own card. Like there were, there were these private, they were called PLCCs, private label credit cards. And I remember my mom had this purse and she had like 80 of them. Because it's like, oh, we're going to Sears, we got to use the Sears card.

• 8:55Auren Hoffman?

  The Sears card, right? Back in the day. Yeah, I remember that.

• 8:57Speaker 00

  The JCPenney card, like everybody had, and these were not like co-branded Visa cards. These were like, this is a card that only works at Sears. You, you had this like, this giant, this wallet that had like 50 cards and then that didn't make sense. So you kind of simplified down and then it's like, well, you know what? I'm just going to get my, and then a lot of these private label credit card companies, like the Sears of the merchants of the world, just kind of aligned with one of these card networks that were emerging. But merchants just wanted to make more money. So getting Visa didn't get you everything because Master Charge, Mastercard was too big. Getting Mastercard, Master Charge didn't get you everything, but if you get the two of them.

• 9:40Speaker 00

  You just, you kind of got both because they both were, it's like, it would be one thing if Visa just operated, if Bank of America to Visa operated unopposed for a long time, but there were plenty of banks that didn't enter the Visa consortium, they entered the Mastercard consortium. Because Mastercard was also a nonprofit. And the interesting thing is that both of them went public in the late aughts. I think Visa, I believe, went public in 2008, maybe 2007, something around that timeframe. And it was the biggest act of all time when it happened, but it was all secondary. There was no primary that was raised because normally, why do you go public? You go public to both be liquid and, but also you go public to raise capital. It's like a venture capital fundraise, but very big.

• 10:20Auren Hoffman?

  But it's so profitable, they don't need to raise capital.

• 10:23Speaker 00

  It was so profitable and they had FTC issues because again, once a year, more than once a year, but all of the banks owned this thing. It's like, what if all of the airlines owned Kayak and decided how much to charge for airline tickets? It's like, they'd go to jail. So there were all of these antitrust investigations and whatnot. And the way to solve that problem was to have banks divest their ownership of these institutions and make them independent. Whereas once upon a time, they were, they were both nonprofits. They were very profitable nonprofits. And the same with the NFL is a nonprofit. It's not a nonprofit like a 501(c)(3), I forgot what, what term it is. But it wasn't, it was just like a, an entity that was constructed for coordinative, coordination purposes.

• 11:07Speaker 00

  That was, that was the role of Visa and Mastercard.

• 11:10Auren Hoffman?

  There's a lot of great businesses that started as these kind of nonprofit co-ops and then like Verisk is like that, where they started that way and then they later became like a for-profit company.

• 11:18Speaker 00

  But this is the tricky thing is that I believe that JP Morgan has a higher market cap than Visa, but not by much. It's kind of weird where like this thing that was owned by all of these big banks became more valuable than all but one of these big banks. It was meant to never happen. It was meant to be like, hey, we're going to create this, we're Bank of America. You don't trust us? Okay, we're going to, we're going to make it called Visa. We own a little bit, you own a little bit, nobody owns any of it. So we're all good here. And we're never going to have a situation where the child bests the parent, both from a market cap perspective, but also it's like, what if I change the rules to benefit my acquired banks or benefit merchants or benefit consumers to the detriment of you

• 11:59Speaker 00

  who once owned me and created me? Don't worry, that won't happen because you own me. Right? That, that was kind of the, the covenant. But the covenant was kind of broken, or it had to be broken. And it went public and now, you know, Visa is a much more powerful institution than any single bank.

• 12:15Auren Hoffman?

  One of the things I, I heard from the outside, I'd be interested in your thoughts, is that Visa actually could be way more profitable. They're actually worried to be too profitable. They don't want to look too profitable to the government. So they literally just like spend money on like random things, like they sponsor the Olympics, they do all this, they're just literally trying to like shove the money out the door. So they don't look too profitable to people.

• 12:36Speaker 00

  I think a lot of large companies work that way. I don't know if there's a conspiracy behind it per se, but it's, it's called managing earnings. In this case, it might also be like managing optics, because yeah, you could probably run the entire company with less than a hundred people. Like if Elon were in charge of Visa, you'd have like 50 people working there. And it would be the most profitable thing ever. But it still has, I think it has better net margins than, than Microsoft. I mean, it, it still is from a margin perspective. It could be a lot more.

• 13:01Auren Hoffman?

  It's one of those companies that's such a good company, that it's kind of the Warren Buffett ham sandwich type company that like, like if you fired the top 100 people and replaced them with my dead grandma, like the company would still be amazing. Like it's, it's kind of hard to kill this company.

• 13:17Speaker 00

  Oh, for sure. I mean, I, I think they have so much potential for growth in ancillary places. And that's the part that would, would hurt. But like, it's, yeah, it is, it is very, very hard to, to, to challenge what they have. And actually, like, maybe another way, I'm surprised I didn't say this when you asked me the question, which was a very good one, like why is Visa so profitable? It's this weird thing where higher prices actually yield more profits and more, more market share, which I know sounds really weird, but think about it, like there is a Visa card. I wish I had my wallet with me, I could pull these cards on and then you could blur the, blur the numbers for me, for your, for your audience that clearly won't steal my, my card numbers. But Visa came out with a product called the Visa Signature Card. Like what is that? Like, how is it different than a Visa card? Basically, the only difference, I mean, I, you get like free wine tastings

• 14:02Speaker 00

  if you go to Napa Valley. They throw in like more checkboxes of like what this thing does. But the net result is that it, it charges more money to the merchant and gives more rewards to the consumer.

• 14:13Auren Hoffman?

  It's just the extra X-bips to the merchant, essentially.

• 14:16Speaker 00

  Well, from the merchant, not, not to the merchant.

• 14:19Auren Hoffman?

  From the merchant.

• 14:19Speaker 00

  Right? So like, like, wait a minute, like, how did Visa gain share by inventing a product that costs more money? Well, think about how this works. I am Barclays and I'm deciding, do I issue a Visa card or a Mastercard? I'm the issuer. Well, wait a minute, if I issue this Visa Signature Card, I get 300 basis points. If I issue this Mastercard, I get 200 basis points. The merchant doesn't know the difference. And actually Visa has, it's called the interchange table. Interchange is the kind of interconnect between the merchant and the acquirer. And it's like, if you have, you know, a small business card, that's 2.74%. If it's a debit card, it's this amount, like, it's like five pages long, maybe 10 pages long.

• 14:58Auren Hoffman?

  It's very confusing also because like, if you're a merchant, you think you're just accepting a Visa card, they should all be the same. But like, certain Visa cards, you're paying more than others.

• 15:07Speaker 00

  So this is why, the more money that I charge, and this is like product at Visa is not like the way that you think about product at a tech company. Product at Visa is I'm going to design, it's like a CPG product, right? It's like, I'm going to design a new type of card with new benefits, new SLAs, new, you know, chargeback fraud, insurance, like whatever. Like this is the, the finished product that I'm delivering to issuing base. And if this product is more financially compelling, because again, merchants are accepting both types of cards. So it's this weird thing where the more money that I charge, the more market share that I gain. If I said, hey, I'm going to design a new card that charges merchants nothing, right? I'm F'd because I go to all my issuers, like I don't want

• 15:48Speaker 00

  to issue that card. Because you know, what issue, and by the way, most of the interchange goes to the issuer, the vast, vast majority goes to the issuer, very little goes to the acquiring bank. Because the acquiring bank, it's like, they kind of have to accept all cards because if they don't, their merchant will leave them. Right? Like if, if I'm like, hey, I'm Chase Pay and Tech, the biggest online acquirer, and I am in a big spat with Visa, I won't be accepting Visa anymore. Walmart.com is like, well, screw you, I'm switching all my processing volume.

• 16:15Auren Hoffman?

  They're also software, so they have other ways to make it like, there's like a toast and they're like, we've got other things in there that just things.

• 16:22Speaker 00

  Some of them, yeah. Like this has become like more of a thing where it's like software and credit card processing have gotten blended. And toast would be a good example, or mind body would be a good example. There are a lot of, you know, niche examples like that, or not even niche, like very large, but also very small ones like PoolTrack that does it for pool maintenance companies. It's software plus credit card processing. And, and that's not like an acquirer in the traditional sense. It is tied to a bank on the backend. Whereas like most acquirers were meant to be banks. It's, it's shorthand for acquiring bank. It's the bank that acquires the merchant. But the point is, the 300 basis points of interchange that I get to charge for the XYZ blah, blah, blah, new product Visa card, you know, let's just say 270 of that goes to the issuing bank. And now the issuing bank can be like,

• 17:07Speaker 00

  hey, take out advertisements on, you know, every single subway car in New York, say 2% cash back with the fidelity card.

• 17:15Auren Hoffman?

  Yeah, because CACs are, CACs are high to get new, new customers, right?

• 17:19Speaker 00

  If you kind of flip that around, I mean, like one of the ways that I've heard it is it's like this reverse Robin Hood product, where it's like, if you're really rich and you can get a credit card and you get rewards, like that's awesome, but who really pays for that? Well, like, well, first of all, you do, because everything is marked up in the store. It's like, there's no free lunch. The hardest businesses to change, I love this saying, are the ones that have concentrated benefit and diffuse harm.

• 17:43Auren Hoffman?

  There's a small number of families in the US that own like the sugar plantations and stuff like that. So they, like, you know, those type of things.

• 17:49Speaker 00

  Or like, you know, there's a saying, it's in a play by this playwright, George Bernard Shaw, every profession is a conspiracy against the laity. And that's another kind of a play on this. But, but what, what happens here is that all those little pennies or fractions of a cent are worth a lot of money to Walmart. So they really, really hate Visa, they hate Mastercard, they hate this entire like complex. But do you really care if you're paying 98 cents or 99 cents for Coca-Cola? You don't even notice, but they really notice, like, you know? So in this case, they have kind of concentrated harm, but it would be concentrated benefit if it flipped the other way. People that are wealthier get the rewards. People that are less wealthy, they are using debit cards that have much lower interchange, or at least most of them do. That's a whole other complicated subtopic on the Durbin

• 18:33Speaker 00

  Amendment of Dodd-Frank. But most of them get no rewards, but Walmart doesn't price discriminate for that Coca-Cola.

• 18:41Auren Hoffman?

  It's the same, same price. In fact, in some cases, they're not allowed to. They're not allowed to charge a different for cash or for this card or that card.

• 18:48Speaker 00

  That actually has changed. So now they can. So before, it's not, it's not illegal. It was just like you would violate your acceptance rules.

• 18:55Auren Hoffman?

  Ah, got it.

• 18:56Speaker 00

  Because that was part of what Visa and Mastercard were meant to do.

• 18:58Auren Hoffman?

  So now potentially online, you could have like, you could, they could say, hey, you know, you can put in your credit cards, like, great, here's your price. By the way, do you happen to have another card? Because if you do, you might be getting one or 2% off if you add this other card in or something.

• 19:11Speaker 00

  It's just, it's too complicated. This is the problem is that like this interchange table, before you couldn't even say, it's cheaper if you pay with cash. That was, you know, not illegal, but it violated card network, card association rules. So you'd have merchants like Blue Nile that would give you a cash discount. Right. So they were, which is kind of the same thing, but it, you know, it didn't violate that particular rule. Now you can do it all you want. Like you see some restaurants where it's like, hey, here's the, there's a card surcharge.

• 19:36Auren Hoffman?

  I think in Australia or something like that, like they cap the interchange fee.

• 19:40Speaker 00

  But this is the bizarre thing. This, this goes back to their very good opening question of like, why are these things so unpredictable? You take the example of Australia and all that has happened there, I would posit is a giant transfer of wealth from consumers to merchants.

• 19:54Auren Hoffman?

  Cause they haven't actually lowered the price.

• 19:56Speaker 00

  They haven't lowered the price and now rewards are gone because there's no mechanism to fund rewards because interchange is capped at like, I think it's like 50 basis points in Australia. That's the cap. So I can't get 200, you know, I can't get two cents back on every dollar spent like with my fidelity cash back card or I use my, I love my Apple card. It's, I think it's the greatest product ever. I tap to pay, I get 2% back. It goes into a cash account. I love that. How do you do that? If the aggregate fees are 50 basis points, but that 2% cash back that I get from Apple, well, where does it come from? It kind of comes from me, but actually it doesn't come from me. It's like, that's why I mentioned it's like this well, it's this reverse Robin Hood effect. It's like the 70% of people that are paying with debit

• 20:39Speaker 00

  cards or, you know, lower interchange cards or paying with cash, they're not getting 2% back. I am, right? And disproportionately it's wealthier people that can get these, you know, postpaid or credit cards, even if they don't have a revolving balance, but you know, it's called a credit card. It is extending you credit. Those are the ones that have high interchange. The other types of tender have low or no interchange, like pay with a check. That has no interchange. Pay with cash. That has no interchange. So I am kind of stealing from them because the merchant just has one price for everybody. And that's kind of not fair, but it's why like you don't have, like you would think that somebody would come up with a lower price mechanism and the government of Australia did, all it did was make the Walmarts

• 21:24Speaker 00

  of the world richer and the consumers of the world poorer because of this concentrated benefit diffuse harm, right? Where it's like, I don't know the difference between a 98 cent Coke and a 99 cent Coke. It's like, it's the same to me. There's no differential, you know, purchasing behavior. It just means that Walmart keeps a couple of extra pennies, but times, you know, billions of transactions. That's, that's a lot of money.

• 21:45Auren Hoffman?

  In some ways Visa has like two things that like almost guaranteed, like there's the, the GDP of the world, which, you know, goes up X percent every year. And then they have the GDP, let's say of the internet, cause like more people are now buying things with cards, less things are happening with cash, or even when you're in person, you're just like, you're, you're using cards more than you use cash all over the world. So you have like both of these things times each other, which like increased Visa's itself.

• 22:12Speaker 00

  Well, I'd actually argue there, there are three, there are three things. So one of them is GDP. That's just a secular tailwind. If GDP shrunk, then that would be bad. But GDP, if it just grows by 2% every year, Visa will grow by 2% per, every year. That's one. The other is the destruction of cash and kind of the transmutation of like cash transactions. Like, you know, how would you pay for, I didn't say there, there are four things. I'm going to go back. So first is GDP growth. Second is like new types of things that go onto card that used to be cash. So like parking meters or, um, you know, vending machines, like all these things that were like cash, cash, cash. And, and maybe that's not, that's not new growth. Maybe it is. That's going to go category three. I'm just coming up with this on the spot. Category

• 22:54Speaker 00

  three is it's like, how many people took taxis before Uber? Not that many. I mean, people did, but like Uber has like ballooned the taxi industry by like, I don't know how many orders of magnitude, but I think a lot. And that wouldn't have been possible without cars. That wasn't like destroying cash from an existing transaction type. It just created a new transaction type.

• 23:12Auren Hoffman?

  More, more type of spending.

• 23:14Speaker 00

  Right. And then kind of category four, which is a super exciting one, just similar to category three, is everything that AI is doing. And why do I put AI in its own category? Because I'm saying what was once payroll is now payments. Like I, I did this tweet about this a little while ago and I know, I know the CEO of Visa liked it. I said it to him. But what I mean by that is like, I invested in a company that is, it's like, uh, it's called revenue cycle management for dentistry. It's like, you know, imagine you're a dental clinic, half of your patients pay with insurance. You got to get paid back by the insurance company. You have to fill out all this paperwork. It's so much work. You normally hire a person to do that. The person that you hired to do that, she or he just quit. You got to hire a new one.

• 23:57Speaker 00

  And then this company called Lassie shows up and they're like, hey, we'll do that for you. It's $15,000 a year. Dentists normally don't spend $15,000 a year on software. In fact, they never do. But they need to hire somebody to process their, you know, $500,000 backlog of insurance claims. They need to pay that person like $80,000 a year. If they paid that person $80,000 a year, that would go to ADP or paychecks or a payroll company. But now it goes to Stripe. It doesn't go to Stripe. It goes to Lassie. But how does Lassie charge them with a credit card? So it's like this net new category of like what was in the labor market is now kind of, you know, it's this, it's moving to like pure payments because payroll has becoming payments. Does that, if that makes sense? Like, and most of it is not destructive

• 24:41Speaker 00

  of labor at all. Like I haven't seen it. Like we back tons of, of like amazing AI apps companies. None of them are like, hey, fire everybody and like use us. It's like, you couldn't process this before. Like it was just too expensive or you can't find somebody or like who's willing to speak, you know, who, who is competent and willing to speak Mandarin, Portuguese, Spanish and Mongolian Saturday at 4:00 AM and they can answer the phone for a, for a transaction that has a lifetime value of a dollar. It's like, well, no, nobody. So like, that's not displacing human labor. That's like, wow, I can hire an AI that does that. It was a human like role before, except the financials obviously didn't make sense, which is why there was no human role, but that doesn't run on ADP or paychecks. That's not a payroll transaction.

• 25:25Speaker 00

  It would have been, like if the economics worked, that would have totally been, that went to ADP, but it doesn't work. It goes to AI. It goes to tokens and Stripe is the backend of that. And that's how I make money.

• 25:35Auren Hoffman?

  Have you ever read the book, A Piece of the Action? This is like one of my favorite books of all time.

• 25:39Speaker 00

  Yeah, of course. Joe, Jonas Sarah. Yeah. I've, I've recommend, I, I probably, so he likes me and he follows me on Twitter. So his book sales picked up because he wrote that book like 30 years ago.

• 25:49Auren Hoffman?

  That book is so good. I like, it's one of my favorite, but I give that book to so many people.

• 25:54Speaker 00

  I, I, so I've recommended this to so many people, but like one time I, like I wrote this, uh, this kind of like Twitter thread on like how Visa works and it got lots of, it got lots of action on Twitter, should we say. And I wish I had set up an affiliate link. I just wasn't smart enough. And at the end I was like, for more, read this book. And then like he got like a ton of sales. He was like, thank you. Because it's like this, this book was kind of like, you know, puttering along. And then it's like, I don't know. A lot of people went to Amazon and then bought the thing.

• 26:19Auren Hoffman?

  I hope after this podcast, he gets a ton of sales too, because it's, it is such a great book. It's definitely, it's a book that I think anyone who's interested in this should read. Um, no. Okay. Now let's talk, we, we touched on AI. I sometimes go back and forth. Like, do you think AI gives more of an advantage to the startup or more of the advantage to the incumbent?

• 26:37Speaker 00

  I think it's tricky. Like there's a saying that I use a lot, which is the best companies have hostages and not customers. And one of the ways to tell the difference is like, are you, are you a hostage company or are you like a beloved product company? And one of the ways to think about that is, is the customer, the user, right? The user's the person that has to use the product. The customer is the person that pays for it. And then at the shorthand for all of this, I think I tweeted about this with respect to Workday. If when you're signing up for a product, if in the dropdown menu, the United States of America as a country is listed right before Uzbekistan, you're a hostage company, right? Because you know, why, why do I say this? It's because if you were trying to optimize product experience.

• 27:19Auren Hoffman?

  Exactly. You would never do that. You would just like, you know, exactly, you know where the IP address of the person, you know, you know, all these other things like.

• 27:26Speaker 00

  Or you just put like Canada, US, right? It's just like at the top. I know that 99% of my customers come from here. If you don't give a shit, you're like, I got you captive because your CFO cut me a $9 million check. Ha, ha. You have to use Workday. It's like.

• 27:37Auren Hoffman?

  By the way, I, I freaking hate that. Whenever I see like the dropdown and they know I'm coming from the U. Even if it's a foreign, you know, thing, like they know I'm coming from the US. Like why have it alphabet? Like why not just have like the first choice be the, it's so obvious.

• 27:51Speaker 00

  It's exactly why I said, I love it because it makes my shorthand heuristic, uh, like, you know, evidently true. So I think for companies like that, they're not going to lose customers, provided that their pricing model is not challenged by AI. So like Zendesk, they have a major challenge, their pricing model, because they charge per seat per month and seats are, are like part of the process, the mechanism of delivering outcomes. And now you don't need seats to deliver outcomes. So Zendesk might do great where they change their whole business model of instead of paying us a million dollars a year for seats and a hundred million dollars a year for people that use those seats, pay us five, pay us $5 million a year. But if they don't change anything to their business, that company will go to zero

• 28:32Speaker 00

  in a year, maybe two years. I mean, it's, it's like, but it's a private company. They got, it was a two billion ARR business that was taken private by Hellman and Freeman and Primera. Those guys are really smart. They know that they have to do this. So it might actually be an amazing app. I just don't know. I hope it works out well. But then you have things like Workday. Workday charges per employee per month. It's not seats used to deliver outcomes. It's like, okay, well, your GE, you have 289,000 employees. We charge $10 per employee per month. That's $2.8 million a month per software and 36 million or whatever, million dollars a year. Done. Like, and by the way, it's a five year contract. And then why are they so well positioned? Because they don't have to play defense for that. They have long

• 29:17Speaker 00

  term contracts. You know, the price is not like, they do charge per seat, but it's not per seat in a way like Zendesk charges per seat, or it's not per seat in the way that Salesforce charges per seat. It's like, I just multiply your number of employees. It's kind of crazy. It's like Excel charging for the number of rows in the Excel spreadsheet. I don't know how Workday got away with this, but they did. It's like the agreed upon pricing for like that ecosystem. And now they are going to be the point of diffusion for AI into HR. And the reason why I say that is, think about it, like what, what happens when you hire an employee and onboard them? Well, I know at our firm and all of my companies, it's like, well, we want to do a background check on you, or we want to make sure that all of the five places that you put on. On your resume, we want to make sure that you worked at those five places.

• 30:01Speaker 00

  So that means that somebody on the HR team.

• 30:03Auren Hoffman?

  Or your continuous monitoring, like we want to make sure you didn't have a DUI, even if you've been an employee for the last three years, right? And, you know, just to learn that.

• 30:11Speaker 00

  You worked at LiveRamp from 2013 to 2016? Okay, that's on your resume. Let's go validate that and we'll call their HR department. You know, that's going to be a button in Workday that says, before you hire Orin, click here, it's $500, we'll make 19 phone calls for you and validate his employment. And that's not the, you know, $10 a month per employee. That's like a new revenue stream. They are the system of record. So a lot of people are saying, you have system of record, system of action. Like the systems of record can become systems of action. And I've, I did a whole presentation on this a little while ago where I go back into the early origins of software. And my theory is that software basically transformed a filing cabinet, like the old

• 30:55Speaker 00

  fashioned filing cabinet with lots of files in it, into a database. And originally that was an on-prem database. Eventually it became a cloud database. But humans, this is the controversial part of my statement that was meant, I can't prove it, but it's meant to be intentionally confrontational, if you will, or controversial, I should say. Humans did not get more efficient with software for a lot of these things. And that's probably not true, but like, think about it this way. Workday, what happened before Workday? Did we have HR people? Yes. Where were the file, where was the Orin Hoffman file stored? It wasn't in PeopleSoft, that was before Workday. But let's go back to the, you know, 1950s. There was a filing room.

• 31:33Auren Hoffman?

  Yep.

• 31:34Speaker 00

  For the HR department, guarded by a guy named Phil with a key. And like, I would just go to the H section of the filing cabinet, pull out the Orin file, oh, you know, we need to do this, or we need to do that, or like, we got called to validate that he works here. Yeah, he did.

• 31:50Auren Hoffman?

  Yep.

• 31:50Speaker 00

  It's in, so what happened with PeopleSoft and Workday? Well, you got rid of that room, but now you have to have a CISO. Now you have to have like 50 IT people. And the same person that would operate on the physical file in the H section of the filing room, like, is now operating on the record from Workday. It was the same person that was operating on the record from PeopleSoft. Like, what, what got more efficient here? Well, what got more efficient is, I guess now we can collaborate across different offices. We don't have to use FedEx as much, you know, fewer phone calls, like, sure. But like, how many IT people did you have in the 1950s? How many CISOs and less security people? And how much did you pay for Palo Alto Networks to make sure that nobody hacks into your instance and, you know, somehow accesses your Workday and everything else? Like, zero.

• 32:34Speaker 00

  Whereas with AI, now you have an ability to use the digital filing cabinet to its fullest extent. Like, you don't need to have the human that operates on the file, because that's the thing. It's like 1950, you had the HR person operated on the file. Or like, take this to accounts receivable land. NetSuite, what was, what was around before NetSuite? Yeah, I don't know. Great Plains Software, what was around before that? You know, filing cabinets. So the concept of an AR, you know, accounts receivable aging summary, that's been around since like the East India Sea Company.

• 33:08Auren Hoffman?

  Let me push back on you just a tiny bit. One of the reasons that these software companies have their customers as hostages is that just like moving off of them to something new is very, very difficult. And sometimes there's like a multi-year process. It's like, we're going to move from here to here and then we'd go higher, like Accenture or back in the day, Anderson Consulting, right? And it would be like a, just the move itself would be multimillion dollar to go move it. Today, that, that, that price of move, both in terms of dollars and in terms of time, has come down greatly. And we should assume it will continue to come down as AI gets better. Again, if a product is great, if Workday is truly the best product for my, my thing,

• 33:53Auren Hoffman?

  I'm going to stick with it even if it's a higher price. But if there is a product that I presume will be better for my organization and the main gate before was this migration, well, now I'm much more likely to do it. Or what do you think about that?

• 34:08Speaker 00

  Well, so here, here's why I disagree. I mean, I agree conceptually, but like, you know, the story of Goldilocks, right? And so this one's too hot, this one's too cold, this one's just right. Like, if you have the right Goldilocks pricing, like, imagine that you take over GE. You took over GE after, who was the disaster guy after Jack Welch? I forgot. But like, you take over GE, it's like falling apart, ML, right? So like, you're ML and like, you know, oh my God, like, you know, our GE finance was the propping up the whole business and now this is falling apart, you know, our, like, we might go bankrupt in 2008, blah, blah, blah, right? And then I show up, you're Jeff Immelt, right? It's like, hey dude, I can save you 12% on your HRAS. What do you think, man? I can get it done tomorrow. And you're like, I don't even care. I don't even care enough to find a person here who does care.

• 34:53Auren Hoffman?

  I mean, it's not about the price though. It's more about like, okay, if something else truly is better, like, if it truly was, HR is a very, is an important piece of an organization. It's strategic potentially, right?

• 35:06Speaker 00

  But I, I think about this as not really being an AI problem. I mean, AI reduces the activation energy a little bit, but you still need somebody. I, I used to tell my team at TrialPay, we have a janitorial services problem, right? And I use, not the HRAS, but it's like, if I went to you and said, I'm going to get your toilets 12% cleaner for 7% less per month, and you're the CEO of this big company, not only do you not care, but you don't even care enough to find the person who does care, even though I guarantee there is somebody who does care within the organization. So you kind of have, if you're doing a top-down sale, you have to find the person that does care, you have to save them enough. I just think it's really hard for a lot of these things because they have, some of them have really effed up their pricing and it's way too high.

• 35:49Speaker 00

  And like, you know, what did, this is why I think a lot of the private equity portfolios are a little bit screwed, because it's like, they layer it on all this debt, you know, Vista and Tomo Bravo were geniuses. Everything they bought, like, was, you know, it was just like this incredible coupon.

• 36:01Auren Hoffman?

  And then, and part of, the part of the thing was, hey, we can increase price.

• 36:04Speaker 00

  Right, or Constellation Software, or Tyler Technologies, like all of these things. And if my only plan was to raise price and I have like, I have hostages, but I have like perfect negative 100 NPS, which I'm pretty sure some of these products have, like, it's like, I don't know why they're going to make this the number one priority, but like, if you piss off Jeff Immelt enough, like, he will be like, fuck it, we're going to drop everything and get rid of Workday. Like, if you get that level of hatred, then sure. But I think a lot of these companies, they operate in the Goldilocks zone. They're smart enough to operate in the Goldilocks zone and you still need, and then, you know, think about the principal agent problem, which is a whole other topic that I thought about a lot with, with respect to AI. But like, what is good for the principal, the ethereal, what is a principal?

• 36:48Speaker 00

  It's, it's the ethereal being of the corporation. And I used to, I was, I was so naive back in the day. It's like, I'd go, it's like, wow, I can make you like $100 million a year. But I'd go to somebody who used to go home at 5 o'clock every day and they'd look at me being like, now I have to go home at 6 o'clock and none of this $100 million accrues to my benefit? Go away. And I could not understand this when I was 23, but now it's like the principal agent problem, it's not a problem, it's just a reality.

• 37:12Auren Hoffman?

  By the way, it's one of my favorite things, like when I'm talking to a company and they're like, go meet with Apple and they're like, Apple told me this. I'm like, Apple didn't tell you anything. Like, you didn't talk to the, yeah, you talked to some like random person who has like some group who has an opinion, who's obviously trying to do something in the best interest of the company and they're not, you know, they're not a, probably a bad person, but it's just their opinion of whatever they decide. It's not Apple, you know, these are not things that like rise up to like the board decision.

• 37:40Speaker 00

  That's the power, is you have to figure out how to appeal to the agent. Like, how do you make this the number one compelling thing? And sometimes it's like, wow, I want to get, I mean, a lot of it is like, I want to get, I, I, so I wrote this, this, this little blog post. It's like, somewhat crude, but humans only have, every human in the workplace wants two things. They want to be lazier and richer. And you might push back on that, but like, what do I mean? It's like, hey, do you want to go home at like 11 o'clock tonight or 5 o'clock tonight? For the next like five weeks? Everything else stays the same. It's like, I think I'd rather go home at 5 o'clock. Yeah, would you rather make double the amount of money or half as much money? Right? You just got to choose, right? It's like, I think I want to make double as much money. So like, that's what every agent wants.

• 38:24Speaker 00

  And it's actually really interesting watching the diffusion of AI into organizations where I, so this is, this is part of what I wrote about in the context of the principal agent problem. It's like, I call it the principal agent problem versus agents, is that shadow AI exists at all of these companies, even if these companies do not approve AI. And they are, they are kind of benefiting from a product perspective, but not economically. Because think about it this way, like, you're a Goldman analyst, this is the example that I use. It's 11 o'clock and you get a note from your managing director boss and says, hey, please fix this presentation. Right? So you are probably looking at an all-nighter and you're going to finish this at 6 AM before he says, this is not good enough, please fix it again. Or you find this amazing tool like Gamma or something and then it's done at 11:01 PM, right?

• 39:09Speaker 00

  So you benefit from that as the, as the agent, right? The principle of the ethereal being of Goldman Sachs does not benefit at all, because that presentation was going to get done by 7 AM, one way or the other, either by the human doing it or by the AI doing it.

• 39:25Auren Hoffman?

  And also potentially their data might be leaking, right? Or, you know, there might be other things that happen, right?

• 39:30Speaker 00

  This is one of the reasons why, like, in the diffusion cycle that we are in right now for AI, like, a lot of it is just benefiting the agent. I don't mean the agent as the AI agent. I mean, in the economic sense of the principal agent problem agent, it's not benefiting the principal, which is not, shouldn't be news to people.

• 39:47Auren Hoffman?

  Of course, and that, that, shoot for, even for when you're selling software and it's like, well, I like this guy because he takes me to the Yankee game or something, you know? It's like, wow, it's like, are you really buying the best software? Like, well, okay, you're not, you're not doing something like terrible for the organization, probably, right? You're not like a bad, these aren't bad people, but you know, there's other incentives that kind of play into somebody's head.

• 40:08Speaker 00

  And you have to figure out which they are, how to harness them. This is what being an enterprise-focused entrepreneur is. So, but, but your question was, is it going to benefit incumbents? And I think for certain ones that get to play, they don't have to focus on defense. Like, I don't have to completely blow up my pricing model. Oh, shoot. Now, the software that was occupied by a human seat, I was charging for a human seat. Now, the software could do that work. Uh-oh. And then, nobody wants to vote to be like, you know, they want to be lazier, but they don't want to be less rich. So, they're not going to like, vote to like, replace their job or do what, oh, yeah, it's going to be really hard. So, versus like Workday or NetSuite or like these things that are just really, really, really hard to rip out. And sometimes, like, the front end is the back end.

• 40:53Speaker 00

  That's another shorthand. Like, QuickBooks is an amazing part of Intuit where like, the front end is the UI. Maybe nobody uses that anymore, but it's like, this is one of the challenges with Salesforce, by the way. It's like, imagine you have a 500-person sales team and, you know, your, your boss discovers Claude code. And it's like, oh, wow, I built something that could log into the backend database and just summarize everything for everybody. Well, I don't need seats anymore. It doesn't mean that like, I don't want a CRM. It just means that I've figured out a way to hack the pricing. Whereas like, how would you do that for Workday? So, if you don't have to play defense on your pricing or on your seats or anything like that, and there's a credible, like, do work on the filing cabinet play? Then, I think, wow, like, incumbents will do great, at least for the, for the brown field, right?

• 41:39Speaker 00

  Like, whereas the green field is always an interesting opportunity because I am not hostage to anybody. Like, Orin, you and I, let's start a company tomorrow. I'm excited about this. I don't know what it's going to do. We're going to start something. Do we want to use piece of shit Workday with Uzbekistan listed right next to the United States? No. Well, we found some startup that has a better HRAS that doesn't charge like, you know, $120 per employee per year. We're going to use that thing because we're not hostage to anybody. And that strategy works pretty well, too. I mean, that was the Stripe strategy, right? Like, or I was the first investor in Mercury. That was the Mercury strategy. They didn't steal anybody from Silicon Valley Bank. That would have been impossible. Like, you're not going to switch your bank.

• 42:16Auren Hoffman?

  But they did, eventually. Like, the product got, Mercury got good enough at some point where people would switch off of like, and, and it's a very hard thing. Like, imagine you're using, I mean, obviously, we had like, we had some, we had some interesting things with Silicon Valley Bank, which helped people switch to Mercury. But even people who are using other banks, they're using JP Morgan or the other. Lots, so many people have moved on to Mercury because, okay, the old, you know, they were just logging into the UI and it was terrible. And now, they, you know, said that one of their friends sent them a Mercury thing. Or, you know, at some point, like, people do switch if one products aren't that good.

• 42:51Speaker 00

  Had SVB not failed, like, I, I just remember when SVB failed, it was like, that was the first time where it's like, we just got this giant influx of customers because they had no other choice. But otherwise, like, these were negative 100 NPS customers with SVB. But they just, it's like, it was the Goldilocks problem in, in reverse, right? It's just like, well, I guess it's not my, like, my startup sales are down by 50% per month. What's our number one plan? I know, let's switch banks.

• 43:15Auren Hoffman?

  We do like, 40 or 50 in new investments a year. And one of the nice things when you do a new investment is you know who, when you're wiring the money, you know who the, so we looked at the wire. And it's like, a long time ago, it was like, 80% of our wires were going to Silicon Valley Bank. Now, 80% of our wires are going to Mercury.

• 43:35Auren Hoffman?

  Like, it is, it is completely transformed. Like, almost all the YC company, you know, there's so many people are using, like, it, it, like, they have got into the zeitgeist. But in many ways, I do think that they're not, they don't have the hostage. Like, their product is clearly better, but I think it is easier to switch today than it was. Even though it is annoying to like, change your numbers everywhere, it's, it's easier to switch.

• 43:59Speaker 00

  It's, it's annoying. I mean, I, but, but I feel like this is where I'm torn when I, when I make that blanket statement of the best companies have hostages, not customers. Like, I hate those companies. Like, I don't really want to invest in a hostage company. I want to invest in a company that has the highest and fastest speed of execution to build the world's best product.

• 44:15Auren Hoffman?

  And one of the reasons I love Mercury is when I log in, it delights me that there's like, this cool new feature that I didn't know about. And I was like, oh, that's like, I'm, I'm probably still only using like, 10% of the features there. And I'm like, excited to try something new. Whereas, when I log into Chase, which I do have to do occasionally or, you know, some of these other things, it's like, it's literally, it's like, the worst experience.

• 44:37Speaker 00

  Well, then, it's like, call a 1-800 number if you want to do anything.

• 44:40Auren Hoffman?

  I can also email the CEO of Mercury, which is an extra like, benefit of having one of your friends.

• 44:45Speaker 00

  Yeah, yeah, yeah, that's a VIP tier. Yeah, I have that too, yeah.

• 44:49Auren Hoffman?

  We have all these smart people who are listening who want to start companies. Like, if you're like, hey, I, I want, I wish this company was started. Like, there's something going on in payments or banking or something like, what do you wish was started that someone like, has not yet figured out?

• 45:03Speaker 00

  Well, I, I think the mental model that I use a lot is, if you think about Econ 101, where you have a downward-facing demand curve and an upward-facing supply curve, where you have quantity and price, right? So, as price goes down, guess what? Quantity demanded is larger. And then, as price goes up, quantity supplied is larger, right? That's the upward-facing supply curve. And everything to the right of the equilibrium point is somewhat of a market failure. And a lot of the best companies that have popped up in AI land are addressing that market failure. Like, Sudo. How much supply is there for like, let's go hire Eminem for $20 million to write us a custom rap song about this, this, this whatever podcast that we're doing right now? Would he take 20 million? Yes, he would.

• 45:49Speaker 00

  He would probably take four weeks, $20 million, here's a custom rap that I made, it's awesome. What is the demand for that? It's just very, very small, right? Whereas, what does Sudo do? Sudo is, uh, yeah, I'm not an investor in Sudo, so this is not me talking. My book, or like, Sudo has like, $500 billion in revenue because they exploited the fact that like, it turns out, or like, Midjourney proved this for graphics design. Like, you could have hired a graphics designer on Upwork. If you go to the Wayback Machine and look at Upwork.com, which is a publicly traded marketplace for freelancers, the landing page of Upwork.com was graphics designers $15 to $35 an hour, right? So, that supply curve, right? It's like, over here, if you could actually price graphics, you know, you know, whatever artist for like, a penny per graphic, holy crap!

• 46:33Speaker 00

  Like, the market is like, a bajillion times larger and Midjourney has more revenue than Upwork has GMV for that entire category. So, that's just like, an interesting lens to think about company creation right now in general in the age of AI. And if you go back to now, how does this interface with payments? Sometimes, a market that seems very, very small can become, like, because the market for software, like, the actual unbundled opportunity for selling like, plain old software, like, replace the filing cabinet, is just small for some of these, like, you know, especially very, very tiny categories. So, I mentioned Pooltrack. I, I clean swimming pools. Oh, I'll use a filing cabinet or Microsoft Excel or whatever. Oh, no, this company Pooltrack came along. That's not an interesting software company. It became interesting because they bundled in payments.

• 47:18Auren Hoffman?

  And so many companies end up like, somehow, like, payments become a piece of it somehow.

• 47:23Speaker 00

  Well, and this is why, like, why did Toast not exist in 1985? Like, the IBM PC had been out for four years. Like, you could buy an IBM PC. It's just like, I don't want to pay like, $5,000 for a computer and like, $10,000 for a software product. Like, you know, how many restaurants would have bought that? Like, zero, I, I'm 100% sure zero restaurants would have spent $15,000 on software for their restaurant. How did Toast, Toast has a higher ACV than $15,000. How did they pull that off, right? Like, same restaurants, people still go out, they eat food. We know how, because they bundled in payments. That's how. So, you, you kind of took a software opportunity that seemed small, and it still is, but once you bundle in everything, it actually is quite large. And AI is like, yet another arrow in the quiver there.

• 48:06Speaker 00

  So, like, why is there not a Toast of manicurists? So, I looked this up, you know, there's the Bureau of Labor Statistics, the BLS. And they were showing what are the fastest-growing jobs in America? So, one of them is compliance officer, which is a very sad thing. But, uh, don't tell my compliance, I love my compliance officer. But it's just like, you know, why, why are there so many compliance people? Because it's like, all these rules that have to be enforced and compliance people do that. The other is manicurist-pedicurist. That's actually number one. Number one is manicurist-pedicurist. So, it's a big market. Why, why is there no Toast for manicurists and pedicurists? Well, because it's like, it's the same thing with the restaurants in 1985. Or even if it's not the same thing as the restaurants in 1985, because we have Toast. It's like, it's a very hard CAC to LTV problem, right?

• 48:47Speaker 00

  Because I have to hire a sales team that goes and sells the manicurist and pedicurist, but they go out of business 50% of the time, just like restaurants. Can I charge enough to pay for my sales team? It's a hard problem. But at least payments makes it bigger. But now, I can say, hey, manicurist-pedicurist shop, you have somebody that answers the phone, right? You know, like, yeah, I was like, I will give you someone, I'll, I'll give you the, I'll, I'll bundle that in as well, right? So, I'm, but I'm giving you payments. I'm giving you a front office receptionist for $5,000 a year that speaks Korean and Tagalog and Mandarin and English and Spanish and every other language that's spoken in the United States by lots of people, right? I'm going to give you that. And that's, so now, this is both saving you money and making you money.

• 49:32Speaker 00

  And by the way, the sales team that I'm going to use to go sell this to every manicurist and pedicurist in America is an AI agent. So, that, that's actually interesting, where it's like, we already know that these verticals exist, but it was this problem of, you know, if your CAC plus your COGS is less than your lifetime value, you do not have a functioning business. And a lot of these categories, it's like, the CAC plus the COGS was greater than the LTV. And you have this opportunity of bringing down CAC with AI sales. I think that's an ephemeral thing, by the way, because like, I don't know about you, but the number of emails that I get that are seemingly customized is going up a lot, right? So, it's, uh, or the number of phone calls that I'm getting from AI agents that sound like, and I got one from Ben Horowitz yesterday. Alex, I would like you to send the wire.

• 50:16Speaker 00

  It was actually his real, I mean, I knew it was fake, so he doesn't call me, ask me to send wires. But, um, this is, this is, you know, it's, it's like the Yogi Bear expression, it's so crowded, nobody goes here anymore. Like, we're, we're nearing that point.

• 50:27Auren Hoffman?

  Yeah, back, back in your example before of like, where is the United States listed in the thing? Like, if you call up a place and there's a phone tree and it's like, press one for the, you know, and it's like, and, and it takes you like, eight. Things to get a human.

• 50:43Speaker 00

  I hang up.

• 50:44Auren Hoffman?

  Also, you know, they don't like.

• 50:45Speaker 00

  I go to the next restaurant.

• 50:45Auren Hoffman?

  That is just clear. Like, anyone who does that, like, does not care about you at all, could not care less about you, doesn't really want to help you. Like, you would want a friendly voice. Now, you obviously prefer a human, but if they have a friendly voice who's AI, who immediately picks up and is like, hey, Alex, hey, let me help you in the language that you speak. Okay, what do you want to get?

• 51:06Speaker 00

  At 2 o'clock in the morning.

• 51:07Auren Hoffman?

  Yeah, 2 o'clock in the morning. Okay.

• 51:08Speaker 00

  Exactly.

• 51:09Auren Hoffman?

  Oh, you're having a big thing tomorrow? You want to get your nails done? Or what kind of like, let me, what have you done in the past? Let's just talk about that for a second. Oh, you have these kind of nails or this thing? And oh, it's a big event or, you know, whatever? Let's do it.

• 51:21Speaker 00

  I help interview people. Like, I'll vet people's resumes or work product or whatever.

• 51:26Auren Hoffman?

  That's so easy to do today. Like, why have a phone tree? Like, it's just like, you know, it's like, yeah, so if you do that, you know you're like, you're gonna like your customer immediately.

• 51:33Speaker 00

  Yeah. So, but I just start off with this notion of it's like, focus on everything to the right of the supply-demand equilibrium point, because those are the businesses that are kind of most compelling from a why now perspective in 2026. And that's, that's the AI enablement. Don't lose sight of what makes the real margin work for a lot of these software businesses, which, I mean, now you have this like labor margin. It's not labor, right? It's like something that was once done by labor, and now it's software, it's AI tokens, and you're basically selling, you're transmuting the silicon and power into like, you know, work on the other end, which is amazing. But you still have all the oldies and goodies of it's like, but it's also payments, it's lending, it's insurance. It's like this

• 52:15Speaker 00

  kind of embedded set of solutions. And it's like, kind of crazy that like, why isn't there a toast to manicures, pedicures? Like,

• 52:21Auren Hoffman?

  Yeah.

• 52:22Speaker 00

  It's just two entrepreneurs that need to just go,

• 52:24Auren Hoffman?

  Make it happen.

• 52:24Speaker 00

  The guy that sells the thing and the person that builds the thing, they need to make it happen. And there are probably a lot of things that look like that. You know, I'm an investor in this company called Salient, which does like AI loan servicing. Was there a software company for loan servicing? Not really. Like, they have like giant call centers. They're like, you know, what if you're like, I'm leasing a car. What if I'm in a car accident? Right? And I'm not at fault. And like, the other person's State Farm is, needs to talk to like my insurance company, but actually it's not even my insurance company. It's actually BMW Leasing of America, because they own the car and I'm just less, I don't really care. Right? Not my problem. How do you handle all these situations with like boatloads of people that are accessing files? Like, just go and say, hey, I'll deliver

• 53:06Speaker 00

  a product that does this entire thing for you. And now you can afford to contest. I mean, actually, we talked about payments originally. What about chargebacks? This is a cool little, I don't want to like, tell people something that they can use to break the rules. But like, if you go buy something and it's under $10 and you put it on your American Express card, and then you're like, you know what? I wish I hadn't paid for that. I'm gonna do a chargeback. I'm gonna say, I never paid for that thing. Complete lie. I'm just gonna say, I never paid for that thing. Well, remember I said, CAC plus COGS has to be less than lifetime value, right? The amount of money that American Express has to spend to look into that $9 charge at McDonald's on my behalf is greater than the, the amount of money that, so they're just like, fuck it, you could keep the $9.

• 53:51Speaker 00

  But why, why should they not do that? Because of moral hazard.

• 53:54Auren Hoffman?

  Of course, if you do it more than a few times, they will, they'll definitely fly you.

• 53:59Speaker 00

  The only way to solve this, like, in economics terms, is to random sample. That's the only way to potentially solve this, because the problem is that even if they fight every one of these things, it's like, they're gonna spend $18 of human labor to get their like $9.

• 54:13Auren Hoffman?

  I mean, procurement in general always has these, a lot of companies like, they can't buy less than, you know, the $200,000 worth of software, because it costs like $150,000, it's just internal procurement to, you know, the Pentagon's like that sometimes, right? Like, it's like, well, we can't buy 20 grand software, because we have a procurement cycle that costs 150K.

• 54:32Speaker 00

  With that as a mental model, I think it's a helpful starting point. And then you just want to do something that you know well. And, you know, I think the payments and financial services embeddings and substrate are, you know, there's a kind of time, those are timeless, if you will, but there are new wedges to attach and to bundle and to make this a very compelling why now case. Because the thing that you learn in venture capital school is why now is like one of the best questions to ask, because it's like, if this is such a great idea and markets are relatively efficient,

• 55:00Auren Hoffman?

  Why isn't someone else done this?

• 55:01Speaker 00

  And humans should not suffer from, you know, too much hubris. It's like, why, why didn't you do it five years ago? Right? Not just somebody else. It's like, if this is such a great idea, and it's like, you kind of, the best companies have a pretty distinct why now, or they evolve into one.

• 55:14Auren Hoffman?

  All right, two last questions we ask all of our guests. The first is, how are you personally like using AI in your day-to-day, either like at home or at work? Like, what are you doing? What are you building?

• 55:24Speaker 00

  I have a 14-year-old and a 17-year-old. And my 14-year-old has a very popular product called Gradesync, which logs into Schoology, which is like the learning management system that like half of all schools in America use. And then it kind of like gets your grades and assignments. And then it's also like got a social feature. So it's like, you got a 92 on the test, had it everybody else do if they want. It's like a gift to get, you'll, you'll love this from your background of data. So he is like 500 daily active users. It's a very popular product. And now he has assigning me work to do. But this thing, like, he's not technical at all. My 17-year-old is super technical, I guess. So, so one thing is that. So I just committed some code to get, to get home for him. And he doesn't know how any of this stuff works. And then the other is, I am currently suing Wells Fargo, just kind of for fun. So Wells Fargo, you'll appreciate this story. I had a bunch of treasury

• 56:09Speaker 00

  bills that I got for my bar mitzvah in 1994.

• 56:12Auren Hoffman?

  Yeah.

• 56:12Speaker 00

  Like, the paper treasury bills. And the reason why you give this for somebody's bar mitzvah is because it's like, you give it with a face value, but you pay half of the face value. So I can give you $500, but it's 250, because in 10 years it matures to that. So 30 years later, I'm like, okay, and they only accrue interest for 30 years. Like, my parents sold their house. I got these like, you know, a few thousand dollars of treasury bills. I can't do anything with them, because nobody takes, you know, paper treasury bills. Chase would take up to like $300, and then like Wells Fargo would take. So it's like, all right, I'll open a Wells Fargo account. They open the Wells Fargo account, deposit it, and then they mark my account inactive for inactivity, and they start charging me $15 a month as a minimum fee. And it's, this is like the apt event, like, all right, I'm gonna have to spend like two hours on hold to save like, you

• 56:56Speaker 00

  know, $500, and my time is kind of worth more than that, without bragging about it. It's just like, so it basically like moved the entire account balance. Like, it was like hundreds of dollars just like evaporated. Because when I opened it, they were like, there's no minimum, like, as long as you keep $500 in the account, there's no fee forever. And then they changed that. They stole like, I don't know, $500 from me or something.

• 57:17Auren Hoffman?

  I had a similar experience in Wells Fargo myself. Yeah.

• 57:21Speaker 00

  And, and then like, you know, I filed a complaint, like, oh, here's $15 back. And it's like, and by the way, they have spent more money, like, so you know what? You know, as unlimited time, my AI, I prepared a small, and I'm not spending time on this, but I prepared a whole small claims suit. So I'm excited mainly to just use the AI to see how far it could take my suit.

• 57:42Auren Hoffman?

  I'm very excited, like, like, to have a, like, essentially a consumer advocate, a true consumer advocate, where the AI always has your best interest, or Alex is the best interest, and does what, when you're not, not, you know, again, it's not, not someone else is paying for the product or some advertiser, but actually doing what's your best interest. And whatever it is, like cancelling things, or, you know, doing small claims court, or whatever it is, it's like going, it's doing stuff on your best interest. That is gonna be awesome. All right. And last question we ask all of our guests. What conventional wisdom or advice do you think is generally bad advice?

• 58:20Speaker 00

  One thing, not to be like too tactical, is that I think we talked a lot about like, what products can you build or what's happening here? A lot of it is like, you just can't really defeat the laws of labor economics. So like, most people, like, you almost want to start backwards. Like, the reason why I love Workday, and I've talked about Workday, like, as if I'm an advertisement for this company that I hate and whose product I hate using. But the reason why it's so magical is because the product is not HRAS. The product is a predictable sales cycle. And why is it a predictable sales cycle? Because like, here is the problem that I have with trial pay. I don't know if you had this with, with LiveRamp, but like, I would sign up a company like Skype, but I would have no idea from the outside in how much money I can make from them, because selling kind of began the sales process. Like, selling into Skype or selling, like, WinZip paid me millions

• 59:05Speaker 00

  of dollars. I made millions of dollars a year from WinZip. And then I signed Yahoo, back when Yahoo was really big, and I made like $10,000 from Yahoo. How would I know that? Yahoo is like a bajillion times bigger than WinZip. It's because WinZip used it correctly. And then, guess what? Yahoo didn't. Or at a firm, we had the same thing. We got every mattress company. We made more money for the mattress companies in our first year than we did from Walmart. Like, how would that be possible? So I think the conventional wisdom is like, you build the product and, you know, you just focus on product, product, product. If you think about it backwards, it's like, all right, how much do the best salespeople make per year? Right? How much? Okay, the best sales guys at Oracle make

• 59:44Speaker 00

  $2 million a year. Okay. I am going to figure out how to make that sales guy sell a product for me. That's the actual thing I need to figure out. I'm not starting with the customer. I'm not starting with the customer I need. I know that sounds like very, you know, very controversial or very whatever, however you phrase the question. I'm gonna start off with like, how do I get the 10 best sales guys to quit their job to make more money selling something for me? And the way that I'm gonna do that is by having a predictable sales cycle with clients that they already know. And if I do that, then I'm gonna have a kick-ass company. Whereas the other way around, I have to now convince or train other salespeople, or I'm gonna have, I call this the, a lot of companies have either stochastic

• 1:00:25Speaker 00

  gross margins or stochastic kind of revenue from customers, because it's like, they don't know from the outside in. Like, I don't know how much BNPL volume Walmart's gonna send me. I have no idea. Right? I think they can send me this much, but like, what if they don't, what if they put like the, an in-store BNPL like sign next to the bathroom and only in Kansas? Like, it's not gonna do anything for me. So, and then, one of the ways of dealing with this for transactional businesses is you have to say no to customers more often than you say yes. And one of the things that I did at trial pay is like, we built out this like little bingo board. And it's like, rather than letting customers choose how they use you, which sounds like choice is better than no choice, right? It's like, uh-uh. We're just gonna give them nine options. If they don't want one of those nine options,

• 1:01:09Speaker 00

  they can go work with somebody else. But like, we know that these nine options work. And if we leave it up to their creative imagination, they're gonna say, oh, let's do a test in like Afghanistan with expired, you know, subscriptions, because that's easy. And then they're like, oh, yeah, we did the test. It doesn't work. And I'm not gonna allow, and but, but most entrepreneurs, they start off so over eager to get the logos that they will do anything. They will build any product, any experience to get to yes. Whereas actually, what you want to do is you want to start backwards. It's like, what does it take to get highly predictable revenue from the outside in? I need a heuristic to know that I should prioritize mattress companies over Walmart. And I know that these best salespeople, I'm going to convince them to quit their jobs and work for me.

• 1:01:53Speaker 00

  And the only way that I'm gonna do that is by promising them more money. And like, if you kind of think about it a little bit backwards that way, you're gonna have a more successful company.

• 1:02:01Auren Hoffman?

  That's amazing. Thank you, Alex Rampell, for joining us on Summation. By the way, I follow you on X, @Arampell. Definitely encourage our listeners to engage you there. This has been like a ton of fun, super interesting. I knew it would be great. You and I have been friends for a long time. But this has been amazing. Thank you again for, for, for joining us. One more thing before we go. I read a blog called Summation. It's the same as this podcast. The blog is about non-obvious idea sharing on business, talent, data, longevity, and random contrarian takes. If you like the conversations on this show, you'll probably like the blog. It's free. New content comes out twice a month. You could subscribe at orin.substack.com. That's orin.substack.com.