The Exchanges

Q So Jeff, I'm a marketplace founder in this hypothetical scenario. We had a great first meeting and you were like, yeah, I'd really like to see your cohorts. And I go, great, great. I'm going to send them over to you. I send over my cohorts to you. Where do you zoom in? What do you want to see? What excites you when you look at those cohorts?

A The steepness of the decline, or is there a decline? And then the, um, uh, is, how are they trending over time? Um, so in the case of Instacart, um, the number of users declined pretty precipitously from, you know, first purchase to, you know, 10th purchase. Yeah. A year out, you know, the, the number of users declined something like three quarters, um, But the users that stayed spent, uh, you know, three or four times more money. So their cohorts were actually on revenue basis were, you know, close to staying at 100%. You know, you look at that and you're like, okay, that's pretty good. And then when, you know, are, are new cohorts declining or improving relative to older cohorts? I, I put a lot of, uh, uh, interest in that. If they're constantly improving, it could be a sign of a network effect. You know, just like, okay, you know, this is, you know, this business is improving because, you know, the definition of network effect is, you know, the more people use it, the more valuable it is. So if more people are using it, it's become more valued, it's going to show up in the cohort curve.

AI assessment note: “The steepness of the decline, or is there a decline?”

Q You said that there were two things, really, that you look for in terms of characteristics that make a great marketplace. Number one was fragmentation of supply side. What did you mean by this, and why is this so important to you?

A So I, I had a very, uh, interesting conversation on this with a, uh, with another entrepreneur at one point. I, I always contrasted OpenTable with, um, Fandango. And so, Fandango, you know, there, there, what are there, six or seven major movie chains in the United States? Aggregating that supply was very straightforward. It's like, okay, that's great. The challenge is if there are only six or seven major suppliers, each of those suppliers has supplier power. So I remember I used to go to AMC and one day they weren't in Fandango. So what did I do? I went to amc.com and made the reservation open table by contrast had, you know, the target market was 30 or 40 or 50,000 restaurants Almost all of which were the, had one, the owner only managed one restaurant. And so it was brutally hard to aggregate, but once you're aggregated, you know, it's, it's brutally hard to compete with it. And no supplier has supplier power. I think the biggest chain had like a hundred units. And so that's it. When I was telling this to that entrepreneur, he goes, um, uh, one thing you might not know about me. I was the founder of Fandango. It's like, oh crap. It sounded so good coming out.

AI assessment note: “no supplier has supplier power”

Q Yeah. That was, that was brilliant from back. A lot, a lot of times, you know, we look at like CACs, especially for a lot of these businesses and they're pretty, pretty volatile. How much weight do you place on CACs given how much they change over time?

A The concern I have on CACs is they typically go up in my experience as you try to scale the acquisition. Um, and, you know, part of that could be competitive dynamics. If you're paying, you know, your, the, the, the, the 10 mattress in a box, uh, customers trying to buy the, find the same, you know, consumer. And, uh, and so for me, I, I, I look at a CAC and I assume it's only going to go up. So if, you know, the LTV to CAC is tight, it's kind of, it's definitely a cause for significant concern. If it's enormous, like, okay, I get, you know, I have a 20 to one LTV to CAC. It's like, okay, I can stomach that. If it's two, you know, that's a big problem. If it's four, I, you know, I'm still leaning way back just because they erode and, and it's not in your power.

AI assessment note: “I look at a CAC and I assume it's only going to go up.”

Q is, like, be so specific. Like, choose, you know, people who love fishing and also, um, love protein snacks at the same time. Like, so specific that it seems outrageous, and then slowly expand, expand, expand. Do you agree with that extreme specificity? And how would you respond then to founders who say, yeah, but then I go to VCs and they say, oh, it's not a big enough market.

A You know, the, um, we, we strive, uh, as a firm not to base investment decisions on current market size, because a lot of the best businesses created new markets to create new markets. And so, um, I, I, I looked at eBay when he, the, Jeff Skoll wrote the business plan for eBay with Pierre. Um, and, and I got a copy of it when I joined the business and they were focused entirely on the collectibles market. Now their argument was collectibles market in, uh, in the U S I think was a, was a hundred, you know, it was a, you know, a ten billion or a hundred billion dollar market. Um, and they were going to own that market. And so when Benchmark wrote the joint, Bob Cagle and Benchmark wrote the check, he was investing in a, in a collectibles marketplace. Now what happened is users found the utility and said, That's a great utility. And they started listing things out. You know, we started in collectibles, the collectible messages got home and then people said, Oh, this has utility beyond collectibles. And they started using it that way. And the, you know, the best motion we had to, um, grow the business was watch what the users were doing and, and double down on it. So, you know, when Simon Rothman was there, he was an early employee and he loved collectible cars. And I think he searched for Lamborghini and two real ones showed up. That was the beginning of eBay motors. It's like, wh…

AI assessment note: “we strive, uh, as a firm not to base investment decisions on current market size”

Q Do you agree with that extreme specificity? And how would you respond then to founders who say, yeah, but then I go to VCs and they say, oh, it's not a big enough market.

A You know, the, um, we, we strive, uh, as a firm not to base investment decisions on current market size, because a lot of the best businesses created new markets to create new markets. And so, um, I, I, I looked at eBay when he, the, Jeff Skoll wrote the business plan for eBay with Pierre. Um, and, and I got a copy of it when I joined the business and they were focused entirely on the collectibles market. Now their argument was collectibles market in, uh, in the U S I think was a, was a hundred, you know, it was a, you know, a ten billion or a hundred billion dollar market. Um, and they were going to own that market. And so when Benchmark wrote the joint, Bob Cagle and Benchmark wrote the check, he was investing in a, in a collectibles marketplace. Now what happened is users found the utility and said, That's a great utility. And they started listing things out. You know, we started in collectibles, the collectible messages got home and then people said, Oh, this has utility beyond collectibles. And they started using it that way. And the, you know, the best motion we had to, um, grow the business was watch what the users were doing and, and double down on it. So, you know, when Simon Rothman was there, he was an early employee and he loved collectible cars. And I think he searched for Lamborghini and two real ones showed up. That was the beginning of eBay motors. It's like, wh…

AI assessment note: “we strive as a firm not to base investment decisions on current market size”

Q So Jeff, I'm a marketplace founder in this hypothetical scenario. We had a great first meeting and you were like, yeah, I'd really like to see your cohorts. And I go, great, great. I'm going to send them over to you. I send over my cohorts to you. Where do you zoom in? What do you want to see? What excites you when you look at those cohorts?

A The steepness of the decline, or is there a decline? And then the, um, uh, how are they trending over time? Um, so in the case of Instacart, um, the number of users declined pretty precipitously from, you know, first purchase to, you know, 10 purchase a year out, you know, they, the number of users declined something like three quarters. Um, but the users that stayed spent, uh, you know, three or four times more money. So their cohorts were actually on revenue basis were, you know, close to staying at 100%. You know, you look at that and you're like, okay, that's pretty good. And then when You know, are, are new cohorts declining or improving relative to older cohorts? I, I put a lot of, uh, uh, interest in that. If they're constantly improving, it could be a sign of a network effect, you know, just like, okay, you know, this is, you know, this business is improving because, you know, in that definition of network effect is, you know, the more people use it, the more valuable it is. So if more people are using it has become more value is going to show up in the cohort curve.

AI assessment note: “The steepness of the decline, or is there a decline?”

Q Yeah, that was, that was brilliant from back. Can I say, a lot, a lot of times, you know, we look at, like, CACs, especially for a lot of these businesses, and they're pretty, pretty volatile. How much weight do you place on CACs, given how much they change over time?

A The concern I have on CACs is they typically go up, in my experience, as you try to scale the acquisition, um, and, you know, part of that could be competitive dynamics. If you're paying, you know, your, the, the, the 10 mattress-in-a-box customers trying to buy the, find the same, you know, consumer. And, uh, and so for me, I, I look at a CAC and I assume it's only going to go up. So if, you know, the LTV to CAC is tight, it's kind of, it's definitely a cause for significant concern. If it's enormous, like, okay, I get, you know, I have a 20 to one LTV to CAC. It's like, okay, I can stomach that. If it's two, you know, that's a big problem. If it's four, I, you know, I'm still leaning way back just because they erode. And, and it's not in your power.

AI assessment note: “I look at a CAC and I assume it's only going to go up.”

Q How do you think about introducing kind of artificial marketplaces? And what I mean by that is where you suddenly become Netflix and you start creating supply yourself, or when you are Airbnb and you say, Hey, Joe, let's use your new startups inventory and add that to Airbnbs, and then we own supply as well. How do you think about when you can own both in that manner?

A You know, I think it's pretty hard to do that on, you know, on, on two-sided marketplaces, just, just because of channel conflict. Take the case of Instacart, you know, they partner with physical grocers to, you know, you know, to give them incremental business, uh, through the digital mechanism. Now it would be very easy for Instacart to say, okay, we, we're going to do, you know, Drop a warehouse into each city and do, you know, ABC grocery. Um, but at that point you're competing with your customer and, you know, that channel conflict is something that Instacart has no desire to do. They want to be the go to partner for the, uh, for the gross, the physical grocer. So you get some very interesting choices like that. When Netflix made that choice, it was, you know, a one way door. It just worked really well.

AI assessment note: “I think it's pretty hard to do that on, you know, on, on two-sided marketplaces”

Q So Jeff, I'm a marketplace founder in this hypothetical scenario. We had a great first meeting and you were like, yeah, I'd really like to see your cohorts. And I go, great, great. I'm going to send them over to you. I send over my cohorts to you. Where do you zoom in? What do you want to see? What excites you when you look at those cohorts?

A The steepness of the decline, or is there a decline? And then the, um, uh, how are they trending over time? Um, so in the case of Instacart, um, the number of users declined pretty precipitously from, you know, first purchase to, you know, 10 purchase a year out, you know, they, the number of users declined something like three quarters. Um, but the users that stayed spent, uh, you know, three or four times more money. So their cohorts were actually on revenue basis were, you know, close to staying at 100%. You know, you look at that and you're like, okay, that's pretty good. And then when You know, are, are new cohorts declining or improving relative to older cohorts? I, I put a lot of, uh, uh, interest in that. If they're constantly improving, it could be a sign of a network effect, you know, just like, okay, you know, this is, you know, this business is improving because, you know, in that definition of network effect is, you know, the more people use it, the more valuable it is. So if more people are using it has become more value is going to show up in the cohort curve.

AI assessment note: “The steepness of the decline, or is there a decline? And then the, how are they trending”

Q Yeah, that was, that was brilliant from back. Can I say, a lot, a lot of times, you know, we look at, like, CACs, especially for a lot of these businesses, and they're pretty, pretty volatile. How much weight do you place on CACs, given how much they change over time?

A The concern I have on CACs is they typically go up, in my experience, as you try to scale the acquisition, um, and, you know, part of that could be competitive dynamics. If you're paying, you know, your, the, the, the 10 mattress-in-a-box customers trying to buy the, find the same, you know, consumer. And, uh, and so for me, I, I look at a CAC and I assume it's only going to go up. So if, you know, the LTV to CAC is tight, it's kind of, it's definitely a cause for significant concern. If it's enormous, like, okay, I get, you know, I have a 20 to one LTV to CAC. It's like, okay, I can stomach that. If it's two, you know, that's a big problem. If it's four, I, you know, I'm still leaning way back just because they erode. And, and it's not in your power.

AI assessment note: “I look at a CAC and I assume it's only going to go up.”

Q Can I ask, how do you think, I, I totally get you in terms of not wanting to rely on paid. How do you think about when is the right time to really double down on paid, and actually just spend a lot?

A I mean, it's, it's when you know your unit economics and, you know, you're comfortable that it works. Early on, I, I was a big proponent of that e-commerce was going to clean the clock of department stores. I got the thesis half right. Share did shift. Um, no one could make money other than Amazon because of, um, there's just no defensibility, no barriers to entry. So, you know, with the, you know, look at that mattress in a box company got to a hundred million in eight minutes. Well, so did the other nine. And so then they're all fighting for the same customer on the same channels with the same product. And as a result, you know, the cost of acquisition went up and prices capped because you can, you know, you typically are trying to compete in price and the movie ends badly. So, you know, that's the concern.

AI assessment note: “it's when you know your unit economics and, you know, you're comfortable that it works.”

Q Yeah. That was, that was brilliant from back. A lot, a lot of times, you know, we look at like CACs, especially for a lot of these businesses and they're pretty, pretty volatile. How much weight do you place on CACs given how much they change over time?

A The concern I have on CACs is they typically go up in my experience as you try to scale the acquisition. Um, and, you know, part of that could be competitive dynamics. If you're paying, you know, your, the, the, the, the 10 mattress in a box, uh, customers trying to buy the, find the same, you know, consumer. And, uh, and so for me, I, I, I look at a CAC and I assume it's only going to go up. So if, you know, the LTV to CAC is tight, it's kind of, it's definitely a cause for significant concern. If it's enormous, like, okay, I get, you know, I have a 20 to one LTV to CAC. It's like, okay, I can stomach that. If it's two, you know, that's a big problem. If it's four, I, you know, I'm still leaning way back just because they erode and, and it's not in your power.

AI assessment note: “I look at a CAC and I assume it's only going to go up.”

Q which is great fun. Uh, Uber by Instacart's Deliveroo's. If you look at the time between payment and execution or action is very quick. It's, To 10, 20 minutes, and then Airbnb, Incredible Health, it's much longer. So essentially, I'm just being blunt, you're the master here. Are these businesses much better, where there's a much greater payment time between, because of the liquidity that one gets and can invest?

A You know, there, there are tons of features that we look for in business. That, that, the get paid in advance is very rare, but when you see it, I've come to value it heavily, just because it, it changes the capital vision. Iman and incredible had the, um, just had this dream of, she called it A to IPO. She arguably could have never raised a B round and gone public. Um, the, the, the business has that kind of momentum. Um, and because she was quote unquote accidentally profitable, the bookings came in way faster than, um, particularly coming out of COVID than, uh, any of us anticipated, you know, she, she literally could have gone A to IPO. That is a pretty rare characteristic.

AI assessment note: “the get paid in advance is very rare, but when you see it, I've come to value it heavily”

Q which is great fun. Uh, Uber by Instacart's Deliverooze. If you look at the time between payment and execution or action is very quick. It's To 10, 20 minutes. And then Airbnb, Incredible Health, it's much longer. So essentially, I'm just being blunt, you're the master here. Are these businesses much better, where there's a much greater payment time between, because of the liquidity that one gets and can invest?

A You know, there, there are tons of features that we look for in business. That, that, the get paid in advance is very rare, but when you see it, I've come to value it heavily, just because it, it changes the capital vision. Iman at Incredible had the, um, just had this dream of, she called it A to IPO. She arguably could have never raised a B round and gone public. Um, the, the, the business has that kind of momentum. Um, and because she was quote unquote accidentally profitable, the bookings came in way faster than, um, particularly coming out of COVID than, uh, any of us anticipated. You know, she, she literally could have gone A to IPO. That is a pretty rare characteristic.

AI assessment note: “when you see it, I've come to value it heavily”

Q Do you agree with that extreme specificity? And how would you respond then to founders who say, yeah, but then I go to VCs and they say, oh, it's not a big enough market.

A You know, the, um, we, we strive, uh, as a firm not to base investment decisions on current market size, because a lot of the best businesses created new markets to create new markets. And so, um, I, I, I looked at eBay when he, the, Jeff Skoll wrote the business plan for eBay with Pierre. Um, and, and I got a copy of it when I joined the business and they were focused entirely on the collectibles market. Now their argument was collectibles market in, uh, in the U S I think was a, was a hundred, you know, it was a, you know, a ten billion or a hundred billion dollar market. Um, and they were going to own that market. And so when Benchmark wrote the joint, Bob Cagle and Benchmark wrote the check, he was investing in a, in a collectibles marketplace. Now what happened is users found the utility and said, That's a great utility. And they started listing things out. You know, we started in collectibles, the collectible messages got home and then people said, Oh, this has utility beyond collectibles. And they started using it that way. And the, you know, the best motion we had to, um, grow the business was watch what the users were doing and, and double down on it. So, you know, when Simon Rothman was there, he was an early employee and he loved collectible cars. And I think he searched for Lamborghini and two real ones showed up. That was the beginning of eBay motors. It's like, wh…

AI assessment note: “we strive, uh, as a firm not to base investment decisions on current market size”

Q work out, um, not to the surprise of some people. My question to you is, how do you retain mental plasticity when you see something from your operating career that didn't work, or you have an investment that doesn't work, and then like Mike, you know, he did Webvan, and then he did Instacart. How do you retain mental plasticity as an investor to not be jaded by prior experience?

A Well, the good news on that is you fail a whole lot, and that's pretty humbling. And then, you know, you know, I spend a lot of time going back and looking, you know, trying to, you know, do the Annie Duke thing of, you know, did, you know, it's not the outcome, it's the quality of the decision. And then, you know, why did I make this decision? What did I get right? What did I get wrong? I mentioned early on, I was doing DTC e-commerce before it was even called DTC. I mean, it was just, I'm, I'm, It's differentiated from Amazon, you know, cause they have their own brand and it's not sold on Amazon and, and all that stuff. And then, you know, that, that basket guy, you know, is, is resulted in an uninspiring set of outcomes. And I actually got two of the best e-commerce companies out there, fanatics and Zulily, and, and it still is an uninspiring outcome. And you're kind of like, okay, why is that uninspiring? And that's where I kind of doubled down on the, um, the LTV to CAC, you know, if it's, If it's not incredible, the need to buy users at scale, typically the, the economics erode. And so that, that, that developed the thesis on, okay, I am not, I am leaning out on, you know, heavy paid acquisition businesses. And, you know, it was through just pattern recognition, you know, trying to assess, okay, why did this decision not work like I wanted it to? And, you know, what can I…

AI assessment note: “trying to assess, okay, why did this decision not work like I wanted it to?”

Q How did you determine a critical mass of supply side within a specific region?

A Well, the good news is we, we didn't have to, the diner did, you know, they just, it was, you know, they, you know, they would discover the service, and if they searched, you know, for Saturday night, and there are five restaurants, they're like, ugh, yawn. If they search on, there's a, you know, interesting assortment of restaurants, including some of the ones they know, we typically try to go to the very best restaurants and work our way down, so if French laundry's on there. You can't book it, but it's on there. Um, you know, that, that gave some credibility and it just, there was, there was a tipping point in each market. It's something like 10% of the target market supply where diners finally said, okay, this is so much more convenient than calling a restaurant one at a time, getting put on hold, getting voicemail, getting all this and that. So I, I would, I'm going to forego the complete selection in, in, in lieu of convenience.

AI assessment note: “It's something like 10% of the target market supply where diners finally said”

Q tough, no turning point, uh, as Jeff Bezos says, you know, one way door decision. You mentioned channel there, and that was the other element that you said was a crucial element in your kind of desire to see in marketplace businesses, which is like really intelligent lead gen in terms of how they acquire their customers. What do you want to see in how marketplaces acquire their customers first?

A Let me, I mean, contrast two things. Yeah. When I got to open table, it was such a good idea. I said, okay, what other industries Can I do this in, you know, just, and, and owned a hundred percent at the beginning, not just, not what, you know, the, the measly CEO, uh, package I got. And, um, you know, the lead gen was really important. If so, um, say, you, you know, your car repair or hair salon, they typically have a customer that is a regular that, you know, I, I go to the same place to get my hair cut each time. If, if you go and say, I want you to pay me a slice of each transaction. Yeah. They're going to just say, no, there's free scheduling tools online. Why would I pay you for, you know, for access to my customer? And so what was wonderful about dining is people crave diversity, the new, new restaurant, the new different cuisine, this, that. And so because they crave diversity, we often were introducing a diner to a new restaurant and the restaurant finds high value in that. So if we were just giving, helping the restaurant, you know, uh, enable digitally enable their own customer. We, we couldn't have charged what we charged. It was that new customer that they valued so much.

AI assessment note: “we often were introducing a diner to a new restaurant and the restaurant finds high value”

Q The element that we were talking about demand and supply, bringing them together, what have been your biggest lessons in terms of maintaining equilibrium between the two varying levels of demand and supply?

A That's like the whole game, you know, you have to figure out, because, um, you know, I, I have a belief that, um, the demand side is, you know, kind of the, is the more, more strategic side, because suppliers typically will go wherever the demand is, and so for me, keeping a healthy demand Bucket, you know, is what, what, what your suppliers are looking for? Because if, you know, if I list my place on Airbnb and I get, you know, and it's 60% occupied and I make a bunch of money, I'm going to continue using Airbnb. If I go through that agony and I get a, you know, two orders a year, um, it's, it's, it's not worth it to me. So for me, Healthy demand results in healthy marketplace typically.

AI assessment note: “Healthy demand results in healthy marketplace typically.”

Q um, not to the surprise of some people. Uh-huh. My, my question to you is, how do you retain mental plasticity when you see something from your operating career that didn't work, or you have an investment that doesn't work, and then like Mike, you know, he did Webvan, and then he did Instacart. How do you retain mental plasticity as an investor to not be jaded by prior experience?

A Well, the good news on that is you fail a whole lot, and that's pretty humbling. And then, you know, you know, I spend a lot of time going back and looking, you know, trying to, you know, do the Annie Duke thing of, you know, did, you know, it's not the outcome, it's the quality of the decision. And then, you know, why did I make this decision? What did I get right? What did I get wrong? I mentioned early on, I was doing DTC e-commerce before it was even called DTC. I mean, it was just, I'm, I'm, It's differentiated from Amazon, you know, cause they have their own brand and it's not sold on Amazon and, and all that stuff. And then, you know, that, that basket guy, you know, is, is resulted in an uninspiring set of outcomes. And I actually got two of the best e-commerce companies out there, fanatics and Zulu and, and it still is an uninspiring outcome. And you're kind of like, okay, why is that uninspiring? And that's where I kind of doubled down on the, um, the LTV to CAC, you know, if it's, If it's not, the need to buy users at, at scale, typically the, the economics erode. And so that, that, that developed the thesis on, okay, I am not, I am leaning out on, you know, heavy paid acquisition businesses. And, you know, it was through just pattern recognition, you know, trying to assess, okay, why did this decision not work like I wanted it to? And, you know, what can I learn of …

AI assessment note: “it's not the outcome, it's the quality of the decision. And then, why did I make this decision?”

Q Uh, final one for you. A lot of young VCs listen to the show. What advice would you give to young board members who are assuming their first board roles?

A Watch what seasoned board members do. Is that person adding value? If so, how are they doing it, and, you know, and why? You know, just, just, you know, model what's on there. I remember, you know, I remember my first board, it was Hotwire, and, um, you know, David Bonderman's on it, of TPG, and, you know, uh, you know, the ex-head of, um, Sabre was on it. And, you know, just, so I just watched, you know, fascinating on, okay, what are these guys, very experienced people do on boards. And then there's the contrary example where someone's just driving you crazy because they're designing product in the board meeting and, you know, it's got it. So pay attention.

AI assessment note: “Watch what seasoned board members do. Is that person adding value?”

Q How did you determine a critical mass of supply side within a specific region?

A Well, the good news is we, we didn't have to, the diner did, you know, they just, it was, you know, they, you know, they would discover the service, and if they searched, you know, for Saturday night, and there are five restaurants, they're like, ugh, yawn. If they search on, there's a, you know, interesting assortment of restaurants, including some of the ones they know, we typically try to go to the very best restaurants and work our way down, so if French laundry's on there. You can't book it, but it's on there. Um, you know, that, that gave some credibility and it just, there was, there was a tipping point in each market. It's something like 10% of the target market supply where diners finally said, okay, this is so much more convenient than calling a restaurant one at a time, getting put on hold, getting voicemail, getting all this and that. So I, I would, I'm going to forego the complete selection in, in, in lieu of convenience.

AI assessment note: “It's something like 10% of the target market supply where diners finally said”

Q tough, no turning point, uh, as Jeff Bezos says, you know, one way door decision. You mentioned channel there, and that was the other element that you said was a crucial element in your kind of desire to see in marketplace businesses, which is like really intelligent lead gen in terms of how they acquire their customers. What do you want to see in how marketplaces acquire their customers first?

A Let me, I mean, contrast two things. Yeah. When I got to open table, it was such a good idea. I said, okay, what other industries Can I do this in, you know, just, and, and, and owned a hundred percent at the beginning, not just, not what, you know, the, the measly CEO, uh, package I got. And, um, you know, the lead gen was really important. If, if so, um, say, you, you know, your car repair or hair salon, they typically have a customer that, that is a regular that, you know, I, I go to the same place to get my hair cut each time. If, if you go and say, I want you to pay me a slice of each transaction. Yeah. They're going to just say, no, there's free scheduling tools online. Why would I pay you for, you know, for access to my customer? And so what was wonderful about dining is people crave diversity, the new, new restaurant, the new different cuisine, this, that. And so because they crave diversity, we often were introducing a diner to a new restaurant and the restaurant finds high value in that. So if we were just giving, helping the restaurant, you know, uh, enable digitally enable their own customer. We, we couldn't have charged what we charged. It was that new customer that they valued so much.

AI assessment note: “because they crave diversity, we often were introducing a diner to a new restaurant”

Q The element that we were talking about demand and supply, bringing them together, what have been your biggest lessons in terms of maintaining equilibrium between the two varying levels of demand and supply?

A That's like the whole game, you know, you have to figure out, because, um, you know, I, I have a belief that, um, the demand side is, you know, kind of the, is the more, more strategic side, because suppliers typically will go wherever the demand is, and so for me, keeping a healthy demand Bucket, you know, is what, what, what your suppliers are looking for? Because if, you know, if I list my place on Airbnb and I get, you know, and it's 60% occupied and I make a bunch of money, I'm going to continue using Airbnb. If I go through that agony and I get a, you know, two orders a year, um, it's, it's, it's not worth it to me. So for me, Healthy demand results in healthy marketplace typically.

AI assessment note: “the demand side is, you know, kind of the, is the more, more strategic side”

Q um, not to the surprise of some people. Uh-huh. My, my question to you is, how do you retain mental plasticity when you see something from your operating career that didn't work, or you have an investment that doesn't work, and then like Mike, you know, he did Webvan, and then he did Instacart. How do you retain mental plasticity as an investor to not be jaded by prior experience?

A Well, the good news on that is you fail a whole lot, and that's pretty humbling. And then, you know, you know, I spend a lot of time going back and looking, you know, trying to, you know, do the Annie Duke thing of, you know, did, you know, it's not the outcome, it's the quality of the decision. And then, you know, why did I make this decision? What did I get right? What did I get wrong? I mentioned early on, I was doing DTC e-commerce before it was even called DTC. I mean, it was just, I'm, I'm, It's differentiated from Amazon, you know, cause they have their own brand and it's not sold on Amazon and, and all that stuff. And then, you know, that, that basket guy, you know, is, is resulted in an uninspiring set of outcomes. And I actually got two of the best e-commerce companies out there, fanatics and Zulu and, and it still is an uninspiring outcome. And you're kind of like, okay, why is that uninspiring? And that's where I kind of doubled down on the, um, the LTV to CAC, you know, if it's, If it's not, the need to buy users at, at scale, typically the, the economics erode. And so that, that, that developed the thesis on, okay, I am not, I am leaning out on, you know, heavy paid acquisition businesses. And, you know, it was through just pattern recognition, you know, trying to assess, okay, why did this decision not work like I wanted it to? And, you know, what can I learn of …

AI assessment note: “it's not the outcome, it's the quality of the decision.”

Q How did you determine a critical mass of supply side within a specific region?

A Well, the good news is we, we didn't have to, the diner did, you know, they just, it was, you know, they, you know, they would discover the service, and if they searched, you know, for Saturday night, and there are five restaurants, they're like, ugh, yawn. If they search on, there's a, you know, interesting assortment of restaurants, including some of the ones they know, we typically try to go to the very best restaurants and work our way down, so if French laundry's on there. You can't book it, but it's on there. Um, you know, that, that gave some credibility and it just, there was, there was a tipping point in each market. It's something like 10% of the target market supply where diners finally said, okay, this is so much more convenient than calling a restaurant one at a time, getting put on hold, getting voicemail, getting all this and that. So I, I would, I'm going to forego the complete selection in, in, in lieu of convenience.

AI assessment note: “It's something like 10% of the target market supply where diners finally said”

Q tough, no turning point, uh, as Jeff Bezos says, you know, one way door decision. You mentioned channel there, and that was the other element that you said was a crucial element in your kind of desire to see in marketplace businesses, which is like really intelligent lead gen in terms of how they acquire their customers. What do you want to see in how marketplaces acquire their customers first?

A Let me, I mean, contrast two things. Yeah. When I got to open table, it was such a good idea. I said, okay, what other industries Can I do this in, you know, just, and, and owned a hundred percent at the beginning, not just, not what, you know, the, the measly CEO, uh, package I got. And, um, you know, the lead gen was really important. If so, um, say, you, you know, your car repair or hair salon, they typically have a customer that is a regular that, you know, I, I go to the same place to get my hair cut each time. If, if you go and say, I want you to pay me a slice of each transaction. Yeah. They're going to just say, no, there's free scheduling tools online. Why would I pay you for, you know, for access to my customer? And so what was wonderful about dining is people crave diversity, the new, new restaurant, the new different cuisine, this, that. And so because they crave diversity, we often were introducing a diner to a new restaurant and the restaurant finds high value in that. So if we were just giving, helping the restaurant, you know, uh, enable digitally enable their own customer. We, we couldn't have charged what we charged. It was that new customer that they valued so much.

AI assessment note: “It was that new customer that they valued so much.”

Q Can I ask, the boards that you sit on, internally, are you saying, shit, we need to get on video, we need to get on TikTok? Like, what do those conversations look like? Because it's coming so thick and fast.

A I think everyone can You're, you're constantly looking for the new channel. I mean, there has been a duopoly for too long between Google and Facebook. Um, and, you know, you know, there's this meme in the venture business that, uh, you know, you don't want to fund companies where you're just basically, you know, you transfer money to the company and the company transfers money to Facebook and Google. So, uh, I, I am reluctant to invest in companies that, that require, um, heavy paid marketing to build their audience. Um, you know, the, the best, Concepts, uh, typically don't need, you know, they're, they're so compelling, they don't need to buy users. That's not, that's not a categorical rule, because TikTok buys users at monstrous scale. But, you know, Google really didn't buy many users, Facebook didn't buy many users, you know, tick, tick, tick, tick.

AI assessment note: “You're, you're constantly looking for the new channel.”

Q you do then if you say, you know what, Jeff, me and you're my board member, we're not gonna spend, we've got a great customer base that love it, we're gonna grow sustainably, but we've got a competitor who's raised from a big, big, big institution, and they are spending a ton of money, and they're gonna hoover up all the oxygen just by spending. What do we do then?

A You know, you're, you're faced with a super tough choice that's probably informed by the capital environment. In this environment, a, someone who, you know, raised a round and is spending it hard trying to drive growth that, that, you know, the sentiment is, uh, the pendulum swung a little bit away from that business. You know, we're today counseling our businesses, try to be, uh, try not to need to go to market for as long as you sustainably can without hurting the business. Um, uh, you know, uh, three years ago, it was spend, spend, spend, spend. You know, there, this is an environment where there's unlimited capital behind you. So, you know, you get the Uber, Lyft, uh, situation where, you know, billions of dollars went into the building of those two businesses, and they're in the United States, they compete with each other. So it, you know, it's, it's a bit of a, you know, a bit of a hamburger hill, I think is the phrase.

AI assessment note: “counseling our businesses, try to be, uh, try not to need to go to market”

Q Can I ask, how do you think about how like new customer acquisition has changed given the rise of TikTok, given the rise of short form video, given the changing landscape that we have today?

A The only constant is it's constantly changing. I guess, I mean, think of the advent of mobile. I mean, and, you know, it used to be Google had, you know, the google.com had that monopoly. And so, You just have to keep your finger on the pulse of, okay, what is new? My partner Connie Chen thinks TikTok has a chance, you know, in China is already replacing a lot of the Google use cases, you know, and one of the ones we argue about constantly is restaurant discovery. She says in China, people are discovering restaurants through TikTok. And it's just kind of, that's a different use case than I envisioned for that platform. But, you know, then again, if you're a restaurant, you got to stay current and try to figure out, okay, what's, what's the next thing? What's the new thing?

AI assessment note: “She says in China, people are discovering restaurants through TikTok.”