Christian Limon is the former Chief Growth Officer at Wish, which was the top spending advertiser on Google and Facebook. He was also the Chief Growth Officer at Tubi and Gemini. Throughout his career, Christian has achieved five exits and $28 billion in IPO and M&A proceeds. He has launched 20 apps and led 33 more apps on growth and monetization.

Here’s a glimpse of what you’ll learn:

• [2:46] How Christian Limon became involved with Wish and the company’s deal with LeBron James
• [14:05] Did Wish’s secretive culture excel or hinder its growth?
• [18:47] Tubi’s unique growth strategy
• [25:00] Christian shares how he unlocked Gemini’s growth
• [28:07] Future trends in Bitcoin
• [37:16] How to advertise on Meta and Google
• [41:54] Why data-driven approaches are a myth
• [49:41] The value of large sales catalogs
• [53:31] Advertising misconceptions marketers should consider
• [1:00:26] Christian breaks down the latest investment themes
• [1:13:59] Christian’s appreciation for theoretical physicist Richard Feynman  

In this episode…

Data-driven marketing and advertising has become a trend, with advertisers misunderstanding information and leveraging surface-level data to fit their narratives. How can companies implement data effectively to become top performers on high-traffic platforms?

While algorithms on platforms like Meta and Google can help generate ideas, you shouldn’t use them to structure your entire campaign. For instance, you may notice your algorithms targeting one audience group while your sales rates suggest higher engagement levels among another audience group. Having led one of the top advertisers on Meta and Google, Christian Limon suggests identifying the micro and macro factors impacting your campaign, like the economy, consumer preferences, and platform popularity, to match demand.

In the latest episode of the Up Arrow Podcast, William Harris chats with Christian Limon about advertising on Meta and Google beyond leveraging data-driven insights. Christian shares his experience scaling multiple brands, the latest investment and Bitcoin trends, and common advertising misconceptions.

Resources mentioned in this episode

• William Harris on LinkedIn
• Elumynt
• Christian Limon on LinkedIn | X | Email  
• Wish
  ‍

Quotable Moments

• "You're not an athlete if you do it when you want to; you're an athlete when you do it when you don't want to."
• "I feel no need to apologize for critiquing as long as it is both true and helpful."
• "If it's visible, it's cloneable."
• "We tried to defend against the common perception that Wish was a scam by leveraging partnerships with legitimate institutions like the Lakers."
• "The true best practice is what yields value, which often comes through counterintuitive and unconventional methods."

Action Steps

1. Challenge the status quo of your market strategy by considering unconventional paths to scale your business: This encourages innovation and allows you to identify new routes to success unexplored by competitors who stick to traditional methods.
2. Cultivate a deeper understanding of ad platform algorithms beyond surface-level inputs: This enables more precise targeting and budget allocation, enhancing campaign effectiveness.
3. Expand product catalogs to reflect the diverse tastes and needs of a wide customer base: A broad range of offerings can match more consumer preferences, tapping into broader market segments.
4. Embrace data-driven decision-making, especially when data challenges existing biases or comfort zones: This objective approach leads to more factual insights, enabling growth and efficient problem-solving.
5. Stay updated on emerging investment themes and technological advancements like generative AI: Keeping abreast of new trends can position your business at the forefront of innovation, providing a competitive edge.

Sponsor for this episode

This episode is brought to you by Elumynt. Elumynt is a performance-driven e-commerce marketing agency focused on finding the best opportunities for you to grow and scale your business.

Our paid search, social, and programmatic services have proven to increase traffic and ROAS, allowing you to make more money efficiently.

To learn more, visit www.elumynt.com.

Episode Transcript

Welcome to the Up Arrow Podcast with William Harris, featuring top business leaders sharing strategies and resources to get to the next level. Now let's get started with the show.

William Harris  0:15  

Hey everyone, I'm William Harris. I'm the founder and CEO of Elumynt and the host of the up arrow podcast, where I feature the best minds in e-commerce to help you scale from 10 million to 100 million and beyond, as you up arrow your business and your personal life. Joining me today is Christian Limon, VC and board member, previously the chief growth officer@Wish.com which was rumored to be the top spender on Google and Facebook, and eventually IPOed. He was the chief growth officer at Tubi, which was acquired by Fox. He was the chief growth officer at Gemini, which I believe was, and maybe still is, the most insured Bitcoin wallet. In fact, five of the companies he's worked at had big exits, and he believes that data driven is a myth, and uses term now, which I cannot wait to unpack with you today. Christian, thanks for joining.

Christian Limon  0:58  

Yeah, it's good to be here. Thanks for having me.

William Harris  1:01  

Yeah, I was thinking about how we got introduced, and if I believe, if I remember correctly, we were just chatting back and forth on Twitter. So I'll just thank Elon Musk, because ultimately, it was on X. Elon had nothing to do with this intro, but what the heck? Maybe he'll see this and repost it anyway, right?

Christian Limon  1:18  

Yeah, it's probably from my being one of the few people that critiques things or gives, you know, constructive feedback on Twitter.

William Harris  1:29  

You do? You give very good constructive feedback on Twitter, and you don't. You don't leave much out of it. You're just like, Okay, great. Here it is. Let's just dig right into it.

Christian Limon  1:38  

Yeah, yeah. I feel no need to apologize for critiquing, as long as it is like both true and helpful, if it's true and helpful and if it's on you to use it, or whether you want to acknowledge it totally,

William Harris  1:51  

totally with you. I'm excited to dig into this before we do, I do want to announce our sponsor. This episode is brought to you by Elumynt. Elumynt is an award winning advertising agency optimizing e-commerce campaigns around profit. In fact, we've helped 13 of our customers get acquired, with the largest one selling for nearly 800,000,001 that IPOed recently. You can learn more on our website@Elumynt.com which is spelled E, l, u, m, y, N, t.com that said, like I said, we're talking about why the biggest spender on Meta ads believes data driven is a myth before we get into that, though, I want to dig into your journey a little bit, because Wish.com you were the chief growth officer. There it. IPOed, third most downloaded app on Android in the world, number one shopping app in over 40 companies countries. It was rumored to be the number one spender on Facebook and Google. 2 million items sold daily. I mean, you guys crushed it. How did you get started in this

Christian Limon  2:46  

I got started because I wanted to get into mobile. And I wanted to get into mobile because I wanted to, like, accelerate my career as much as possible. And the iPhone had just came out, and so I thought, if I, if I get into iPhone payments or advertising or just it wasn't clear what, what it was, what what you were going to do on it. But I thought it's, it's not possible for somebody to have five years more experience than me on this thing. So if they want the most experienced person, there's going to be like, one group ahead of me, and then it's like me and whoever else joined at that time,

William Harris  3:26  

sure, yeah, because you got into this very early, and you were doing some stuff with Kim Kardashian's game and everything else. I mean, like you were doing some very early stuff in the space, yeah,

Christian Limon  3:35  

the earliest stuff is at a place called Tap joy, which was like, okay, fundamental to the start of mobile apps. Mm, hmm,

William Harris  3:46  

yeah. And so I want to dig into Wish.com. There's a couple of things. One that was a deal that you guys did with LeBron that I thought was very interesting, partly because I graduated high school in 2003 so I'm the same age as LeBron. We graduate in 2003 I'm from Canton. He's from Akron. And those who don't know Canton and Akron, we have we share an airport, so we're like these two small, little cities in Ohio. We're right next to each other. I can remember when I'm in high school in Shaq and Jordan are coming to my town to watch this high school kid my age play, and he was just a beast. He was giant back then too. I never played against him. I did play basketball, and I didn't play against him. And so I have this affinity towards LeBron and everything. But you worked out a deal with LeBron. Tell me a little bit about, like, how this deal happened.

Christian Limon  4:32  

Yeah, so the deal was, was, technically, it was with the Lakers. But the higher premise, like, the upside was, was LeBron. It's, it's just that he wasn't on the team yet, and they weren't hinting that he was going to at all. They were, they were really good about like that, no over promising and all that stuff. There were. There were really good partners. But it was. First year of that patch sponsorship, and there wasn't a very good there wasn't a process yet. It was pretty I think, you know, wing it for the NBA teams. It was an experiment. And so I reached out to the Lakers. I reached out to the Warriors, you know, the rockets, just started having conversations about, like, what they're thinking, what the pricing is like, what comes with all with this. And, you know, we're like, what just like, what's the nature of the partnership, right? And the Lakers were open, yeah. And they're the so first of all, if you just, like, want to do only one and you can afford all of them. It was the then world, world champs, warriors, or the Lakers. The Lakers, just because they're, you know, permanently, champions, greatly, like in China, they recognize, like the Lakers, more than anyone else in Kobe and stuff like that.

William Harris  5:59  

Sure. So how did you evaluate that? This was a good deal. You know, you're doing a lot of stuff with mobile app downloads, and obviously you're tracking a lot of stuff very, very granularly when you're running Meta ads and things like that. How did you get from that to say, Hey, I think that this, these numbers make sense. How are you evaluating that? Yeah, yeah.

Christian Limon  6:19  

So, so So first of all, there's like the context. There's two points of context that lead up to it. And then there's the the like, the way I came up with, like, a valuation of the thing. And then there's the outcome, the this is like, one of the rare things where actually that the outcome worked out in the most successful scenario that I had, like calculated so I was so I undervalued it significantly, which is why it like, for reference, there's the rumors of how much we paid were anywhere between like 35 to 40. Is tiny bit more in reality. But, you know, that's like, it's, it wasn't 300 million, it was sure, you know, divide that by 10, or even 400 million. Divide that by 10, and when our deal was up, it sold, it repriced and sold for 100 million instead of our price. So significantly more. Yeah, yeah. That's like a more proper valuation of that with Lebron in a world championship baked into it.

William Harris  7:29  

What about,

Christian Limon  7:31  

I think I need to, yeah, you asked go for it, like, what? Okay, yeah, so the other bit of that's like one bit of context to start. The other bit of context was, we're spending. We were on a run rate at that point for like, spending over a billion so, and the experimentation budget is is, like, bigger than that. You know, sure, which we don't like, necessarily we didn't fulfill, but it just made for a there wasn't much we that wasn't that we couldn't do if we wanted to, sure, but in anything we didn't do, we chose to do that, and we're still pretty strict. But with the Lakers and the other thing, with Wish at that time and and probably still now, the number one Google thing related to Wish was, is Wish a scam? So there was the, there was that, like kryptonite that we that we knew about pre IPO, that people would question the legitimacy of this thing. So, like, the logical rationale was something like, if people question our legitimacy, even the numbers are like, clearly, like, massive, but they, you know, they questioned like, safety and legitimacy and trustworthiness. So instead of speaking for ourselves, we'll have like, why don't we try to have the most legitimate institutions potentially try to, you know, speak for us. And then, you know, the assumption would be people say, if, I mean, if Wish partnered with the Lakers, they got to be legit, right,

William Harris  9:13  

right, right. Yeah, the Lakers aren't going to have just anybody go ahead and partner with them. Yeah, that makes sense. It gives them value and credence to it.

Christian Limon  9:20  

And if you're a Lakers fan, like I am, or in the Lebron fan, yeah, you know, brony was going to school at Sierra Canyon at the time in in like Southern California, LA area, Calabasas. So LeBron would come in the summers and watch all his games, which is dope, which is amazing, that he would, you know, go and get into the games, yeah, and I heard a rumor locally, like, from people that he was, like, looking at a couple of houses, you know, but, like, LeBron, probably would just buy houses, like some sure, you know, you know, there's, but there was a possibility. That he would, he would want to be there, to be with his family, and it was COVID hadn't hit yet. So, so really it became the, if he was part of the equation. It became an expected value equation of of, you know, the probability of LeBron, like the whole thing just centered on what the probability of LeBron coming to LA and most of the value has to do with the LeBrons. Social media has a massive amount of impressions, and it since it's on one side, that's the opposite of Nike symbol. It's like the then massive amount of Nike ads and and like out of home master prints and stuff outside the Lakers, the Staple Center. And it's just, I can go into into it more, but especially since they then, then then you got to factor in the possibilities that they win the national championship, and then you're permanently in the record, like you're permanently stored. When somebody looks up the Lakers, they're gonna, like the it'll want a reference to that, that moment, the most recent NBA championship. So you gotta bake in the probability, which is pretty high, they had Anthony Davis and anything could happen. So I was, like, getting there, this is pretty good. Every one of the warriors was, like, injured, you know, sure. So then he factored that in and, like, independent of being, say, of precision, like it would, I would have been, it's like, precisely way undervalued compared to what, what you potentially get in return. Yeah, and I love that. There was a bunch of Wish specific stats as well. Like that was the best metropolitan area for for Wish it had a lot of room for growth, because it goes all the way the Lakers market goes all the way from San Diego, starts in San Diego, all the way up Santa Barbara or or higher, and all the way into Vegas. Like those are all Lakers fans. Those are, that's the demo. So a lot

William Harris  12:24  

of good crossover. And it made sense that you were getting a deal, and it made sense because you were building up this trust that was maybe lacking. And if that trust could be captured a little bit more than it opened up significantly more sales. The other thing I wanted to ultimately,

Christian Limon  12:40  

I value the impressions that, the amount of impressions of that logo that would be sent around the world next to LeBron sure, you know, like with LeBron's face next to it. Then I thought it was like, dirt cheap, yeah, if you factor in, you know, if you put any number next to infinity, it's going to be pretty expensive. It's going to be really expensive or undervalued. But that, like, the the equation would would include, you know, the value of having, like, if you look up the, I think the Lakers, or a lot of the permanent pictures of things are have Wish in them now, which is really cool, yeah, and the company can go away and that will still, that will still be there. Yeah,

William Harris  13:35  

that's good. Something else that I really appreciated was, and we don't have to dwell on this too much. I know you said, like, don't get too deep into it, but Peter, like, the whole set up, the culture of Wish was a little bit more secretive. Peter, well, it didn't really like to do a lot of PR, partly because, like you said, that was his his personalities, that he's an engineer. But how do you think the more secretive culture helped the growth of Wish.com.

Christian Limon  14:05  

Yeah, I thought it, it's important for two reasons. I think it harmed Wish and it harms the legitimacy. Because you don't, they don't hear about it from sources that are things like, you know, what, what do, what does my the news commentator that I that I like, think about Wish. Does he like? It didn't like. Nobody was commenting because he was just like, like, nobody knew in our circles and in Teck what, what Wish really was, sure. And so it wasn't, it was an event, a great idea from that perspective. But really, in my view, like PR is, traditionally, you're not getting what you think you're getting. Or there's, there's a it could be on an overrated, underrated. It would be a. Overvalued, sure, but the thing that was, you know, partially intentional that I think, and I think about a lot now, especially with the idea of the cost to start a startup is getting so low, and the cost when it comes to skill and education and connections is getting so much lower that, you know, it's foreseeable. There's, there's a future very soon, as in next year, where the idea of having a differentiated product is a joke, because anything, as soon as you, you launch something is just going to be, you know, if they can, if they if it's visible, it's cloneable, yeah. And so then it becomes, so this is what makes, I think, Wish, that strategy really interesting and more relevant. Now, actually, is, is so we were getting cloned, but it didn't do. It didn't help anybody. The The reason was because they're they were cloning the stuff they could see. They were they can't clone the things they can't see and all if you're gonna have a superpower, I know when somebody doesn't have the superpower, if, like, before it's even meaningful, they have press releases about it, interviews, and they brand themselves through it, whereas what we did was we Like, we would even not acknowledge it. We would say that's, that's not sure, yeah, yeah, that's, that's not true, but it was, it was, it was, like it was executed that way, even internally, meaning internally, most people didn't know what we were doing with ads. And

William Harris  16:40  

that's, that's huge, like, you said, like, the internal thing, then there it was easier for people on the inside to even just be like, Yeah, I don't know that's, that's a rumor. Like, I don't know anything about that. There was no way for that to even get leaked.

Christian Limon  16:51  

Yeah, where they would, they would try to find out, because people would, would, would rumor and were interested, or, you know, their VC friend would say, Hey, can you get some info, or whatever? And that's exactly what we were trying to, trying to defend against. And it was because it's one, it was meaningful. It was part of the superpower of the company. And two, it, I think we overvalued this one piece, the that it's reproducible, but it's, I think it's less reproducible than than we, than we, than we thought, because I've studied a lot about why isn't has nobody gotten back up to that skill level ever since the ignoring The Apple at he outside of iOS 13 or what? Why? Why did we peak in like capability and intelligence at like, 2015, 16, like, like, like, there's plenty of smart people. You know, there's plenty of of a collective like, smart people, like the people at some of these places, like they could, they could go to the moon if they wanted to. They don't have like ads. Is funny, that the work that they're working on, you know, making ads relevant, but like they they can. They like they could just also, you know, be sending ships, rocket ships, into Mars.

William Harris  18:23  

Yeah, it's interesting. Why we decide to do what we do. I want to look at a little bit then with because I want to get into the ad stuff before do I want to still focus on your journey, just a little bit, because after Wish you were at to be, which was then acquired by Fox, what was one of the secrets to growth at Tubi? Like, what's something that you did that helps stand out? And you're like, This is one of the biggest reasons why this succeeded?

Christian Limon  18:47  

Yeah, yeah, that's probably the biggest thing. Is really the most efficient path. And I would say probably my biggest like strength or skill, is just being open to indirect paths to value, you know, there's the obvious one. So people, I think, have the 10. They have the tendency to just, why would try anything else? It's stupid to try anything else, you're going to look really bad every and people do ask because they have that assumption, like, Don't you know what you're doing. That's not how it's done, you know? And then we, I've had, I've had CMOS reach out to me and say, That's not how it's done. I've had Facebook and Google reach out and and say, Oh, you don't want to do that. That's like you want me to turn this off. You want that's, that's not how it's done. And so at to be the way that worked out was the most efficient path and the not undiscovered path of getting to the highest value, the highest value, also, by the way, on for like ad based entertainment and movies. And TV shows, just anything like TV. What you think of, when you think of TV is, is on the big screen, like that. The companies are, are bidding a lot for impressions on the TV. So in on mobile, it's much, it's pretty it's pretty light. They're not mobile heavy. So it's, it was, it was very binary. It's it was extremely expensive to advertise on things like Amazon Fire, because that's super valuable, and it's specifically very valuable to them. And it was like a war, like the media companies, the traditional ones, you know, and brought broadcast networks, stuff like that, that space is not known to be, like super efficient, and so they have, they also have some leeway in making investments. Netflix is in the space, and at that point where we like to be is digital native, and like mobile native. So they knew, they knew our competition is not like these guys. Our competition is Fortnite and, you know, tick tock, that that's what's competing for people's time, because that is also like amazing entertainment that's really inexpensive and accessible to a lot of people. So the the path that ended up making the most sense, and we just like, kept like it became like evergreen is, instead of buying on Amazon Fire, I would acquire users on mobile and then, and then funnel them into the TV.

William Harris  21:37  

Sure, so acquiring them on mobile through like DSPs or like, through what types of, through

Christian Limon  21:42  

whatever is the like, most efficient high, sure, high LTV method of acquiring them. So for us then was, I think it was, I think it was Facebook at that time,

William Harris  21:57  

nice, yeah, yeah, yeah. But there's a lot of challenges with,

Christian Limon  22:00  

with, with, like, entertainment in general, to content and entertainment, like, what, like, what kind of challenges? Yeah, so people already know of all you know, familiar with, with a lot of the dynamics of these things, but they don't think about, how would that impact acquisition and especially like getting good at it, because you don't have to worry about it if you're going to be mediocre. If so, if you want to be good, you want you need to learn a lot, and you need to learn, to learn a lot, you need to acquire a lot of data, like a lot of it, and try a lot of things. It becomes pretty difficult. If the catalog is rotating and disappearing and coming back, it's you need a degree of, I think, until that point, I didn't realize how important persistence is. Persistence of the the base layer. So like, the base layer needs to be, I really knew like basically needs to be, like, infinite, basically, and something that could be containerized and matched and moved around, just like an ad unit, but it needs to also always be there, or else. How are you going to acquire enough data on this thing? Or you invest by acquiring data for this, you know, other movie, running campaigns for it, in a in a catalog ad, and then it's not there. So you have like this, like dormant value.

William Harris  23:32  

No, that's very smart. And I do think that, you know, we're seeing more and more where that data becomes necessary for making those types of decisions, especially when you're missing a lot of the data from like you called out, you know, att and things like that. It makes it more difficult sometimes to have that data that you need, but that just means that you just need more data of the data that you do have to be able to make a better decision. Yeah. What about at Gemini? So I'm a crypto fan. I think I got into Bitcoin, not as early as I wish, and not nearly as much as I would have wished, but maybe around 2015 or so, 2014 2015 I helped launch a project on Polygon, a defi project on Polygon. We brought in like 60 million TVL in four days. So it's like, I'm definitely in this space, and it's something that I appreciate.

Christian Limon  24:25  

Well, Gemini, was it a What's that? Was it a deck? You

William Harris  24:31  

know, that's a good question. I'm not the technical founder. I just helped to market it. So I don't, I don't think that they would have considered it a Dex. I'd have to, I'd have to ask them if they would have considered that. I guess my question to you, though, is, this is Gemini has done very well. What do you think allowed Gemini to grow as fast as it did then? Like, what was something that you were able to do that you like? This is one of the secrets that we were able to unlock for Geminis growth. But

Christian Limon  25:00  

the Gemini, for the most part, it grew even the Gemini reasons why it did well, where, not unnecessarily, like anyone on the team that did it was, you know, they, they're part of the story. The brothers are part of the story of Bitcoin. If you read about the history of Bitcoin, she'll be in there somewhere at some point at, you know, toward the end. And so they're in the game the longest, right? Sure, and you've, you've heard of these people because of the social network the movie. So there's, there's that degree of like permanence to it, and and also the their, their is before that they're like people didn't know very much about them, or had heard of them, not not the twins, but Gemini. But then when basically everybody, if they weren't manufacturing it or or intentionally targeting that, which meant raising money if you like you need to. You need to spend like you need to, basically, you need to move your your revenue up, and then, like, put that into, put that into marketing, to to, you know, actually manifest it. And they didn't. They didn't they their strategy was more about being permanently safe and permanently legitimate, and not about, like, growth or something like that, and but when the market took off, the the one where the in 2020 when, like, barbers, taxi drivers, right, airport personnel were showing you the portfolio of, yeah, you know, and saying, Okay, I'm an amp, they would say, yeah, that's that's when I knew it was legit or or like truly taking off, is when I heard about it. Like there was no escaping it. Even in the real world, I could be offline and there was no escaping seeing portfolios my family, like people in my family were showing me their portfolio, like they have no money, sure,

William Harris  27:27  

putting 50 bucks into it, right? Like, whatever you can, putting something into it,

Christian Limon  27:31  

yeah, yeah, yeah. And some of it was it motivated people to research. And some of the These cities were well researched. So it was like, encourageable. It was, yeah, like, have had it? Like, this is the, if you're gonna, if you're gonna, like, mess around. This is, this is, like, a productive way of, like, what to read about and what to learn about. Yeah.

William Harris  27:55  

Where do you think it's going now? I mean, we're seeing another boom here. It's, Bitcoin has gotten really close to 100,000 Are you still into crypto? And is this something that you think is going

Christian Limon  28:07  

to continue to grow? Yeah, so that, relative to how it was when I was actually voting with my life, with my entire career, was a Okay, I'm gonna, I'm like, voting on this is going to be it, because I'm like, I'm spending most of my time in that and and I was pretty committed. So I was a I would advise, I advise several companies in defi that were decentralized exchanges and like routing protocols to learn like, the nuances of like on chain startups and on chain activity. And that was in that was the most educational part that, to be honest, the was was learning from the on chain actors, like the market is so much like free there that it just, it's, it's like an education when you study some of those guys. What was, what remind me, what I What

William Harris  29:09  

did you ask? Well, just where you know, is this something that you're still bullish on? Is this something that you're still excited about? Where do you see going? So

Christian Limon  29:17  

the the solution of like computers to computer trust, less communication is like a breakthrough. And like solving the double spend problem is a massive breakthrough. And so the like that alone is is just will be valid like it was, will be valid forever, like they did a a and it's up to us to figure out how to use all these solutions. The thing the thing about crypto Is it a term in a like category, has, has, like, been, like, put on, like, a really wide range of solutions and technology. Jesus, you know, all the way from, from zero proof, proof theorems that can help with, like privacy, to, you know, meme coins, and they're all two people making decisions, or who are betting some of their own money and or just, you know, the average person in the US. It's all the same thing. It's just crypto. And crypto is one of these things where, where, like, there's like, the things that you can do you you want to be careful and be an adult about it, and not, do you know, like, risky, have like, don't, obviously, bet money you don't have. But I think in a time of exuberance, even the best of the best get carried away. So, so I there was a lot of people getting carried away that weren't like best of the best. And I think the if you do that to enough people, like, you don't want to go viral that way. You want to go viral, like, if when you go viral, you want to go viral in the positive direction. And it did, and then retroactively, like, history got rewritten. Because the big news that came came out was that, you know, Sam was actually investing people's money instead of, like, holding it right, and he put eggs on on, like politicians faces. That was the introduction to the industry, for for a lot of those people in the hill, and that is very a lot of the arguments for the crypto token market, like the market caps, had to do with, you know, just like everything else, like a lot of seasonality, and, you know, being part of the macro and like inflation, is part of the story, sure, so, but you can't the thing that is not extrapolate able in the historical data that you can't remove, and you can't extrapolate the data without this thing now is that that same bank and freed meltdown and the amount of like consistent public scams or scammers or just shady individuals, and you can't pull that Out of the data like that that now is, is like part of the whole Bayesian calculation. And I think that that makes it very a new type of forecast. And over time, that's just that's going to be like corrected, and we're coming up like with with the proper regulation framework to be able to do that. It's just slow. And I, my opinion, was just that it's like a beautiful solution, and for specific things, and not for the majority of things. And in it, I wanted to accelerate it and and then my post, Sam, my recalculation was, this is whatever I thought before I need to extend further out, because it's going to be, you know, the unit need time for, for, for healing, yeah, for

William Harris  33:39  

the data to cycle. It reminds me of an ad metaphor, which is similar to that, where if somebody ends up having a whole bunch of bad data come into their ad account, and then they continue to keep running ads through that without cleaning that. And sometimes you can, for those who don't know, you can actually, you know, have that data removed by Google or Facebook. You can actually ask the teams to do that, oftentimes, but if you don't, yes, yeah, well, but if you don't, there are some times where you just have that bad data that's stuck in that account. Now, for some reason, and until you get that data out, you are your ads are not gonna be as effective as you'd like them to be just an example off top of my head of this would be like, if you were setting the wrong customer information in, or something like that you're sending in that the price was $1 or something, yeah, or whatever. This could be, right? Like, there are things that you could send in, but I've seen this on the performance side, right, where it's like, I'm trying to remember what this would have been, maybe just an actual error, an actual error in data that came through, and now that data then is it's in that right, for if it was like, for a week or something like that, and it's like, okay, well, you have a week's worth of bad data that you have to somehow purge, otherwise it's going to impact all of your ads future now that it'll correct itself without that. But it's better if you could just remove that. Yeah, once you remove that data, once it's in the super majority of the data, and it's no longer recent, because there's. The recent, the premium, then, then it will be phased out. But exactly, just like, like, what are you going to do while you wait? Yeah, exactly, just like crypto, and unfortunately, we can't remove it. Can't just email Google and Facebook say, Hey, can you remove this similar data from the crypto scene? We can't do that. So we have technology

Christian Limon  35:19  

has all the time in the world, but us, our careers Don't, don't at some point, like I'm at a point now where I, like had recently hit just, you know, the point where even biases shouldn't really restrict me from being able to do whatever I want, just meaning like age biases. Like, yes, an age bias will make it so that I can't do some things, because they'll they only the candidate pool is, only looks a certain way or, or is, you know, they're not young. That is, sure. But everything else, it's just like, it's time to I prioritize the where I wanted to be before and and like, what I what I wanted to learn. And now, anyone that talks to me like, all they care about is like, what do you is like, what I'm producing. And value companies want to know what the value I'm creating is, or what I'm what I'm fixing. And like, nobody pays me to teach me anything they paying they, they want an, like, an exact, immediate exchange of value. And I think, and it's like, an appropriate time where, yeah, I will, like, monetize, like the skill set, and now just everything that I learned is through like, complete first person acquisition of of that information.

William Harris  36:50  

So this is where I want to start digging into some of this first person acquisition of knowledge. A little bit, you were rumored to be the biggest spender on Meta and Google, whether you were number one or number two, like you guys were there. What are some of the secrets that you learned that work on the ad platforms that you think people just aren't getting

Christian Limon  37:16  

some some of them are. And to be clear, there was like some, there was some decent sized ones. And then there's, there was a lot of small ones, a lot of small ones. And the the nature of, like, the way you set up a campaign in Facebook is there one campaign, it was like, a loaded container of, like, a like, almost just an incredible amount of components. And so if you like, want to optimize that thing, you need to optimize the components. But another, there's a number of other things that you introduce to a campaign, like when it, when you start it, what by how much? When? How often do you do you change it by how much do you change it? And do you do those things at the same time as other things? Like, there's enough, there's a number of things you introduced into that system that impacts, that could impact the system, and you just assume does so. Step one is you need to first. You need to first figure out, like, what are the what are the variables that are important, exposed and unexposed and like, so you have to end up and so this is something mostly Wish can do where anybody with like, a billion dollar budget or several 100 million is you need to discover this. Like, what's important that is not public information that other people don't know, like one even if it's just only relevant for you, that's the only thing that matters. Is your is that context you know they're not paying you to figure out, like, what to do at your next company there you're you need to figure out what to do at this one, and part of that process is, is, are there important things that I'm not aware of and that nobody is aware of, or that are just not public?

William Harris  39:17  

Like, really parsing out what the algorithm does or doesn't do, and how it reacts to different changes in inputs and things like that. Is that what you're

Christian Limon  39:27  

saying, yes, including, including the difference between what what they think it does and what it does like I there's what people think and all this stuff, and that's it. It helps for ideation, but it plays no role in, like, what we would do and like that kind of that's my mentality ever since is it's irrelevant to what is, what I try, if anything. Oh, now I validated why we should try that. Because you get to test it. Like, think about how many people are not doing that, if they tell people not to do that. And so you you need to figure out, like, what are the what are the inputs in general? You know, instead of the you see four inputs. So there are four inputs.

William Harris  40:22  

I think a lot of this comes down to sometimes understanding how algorithms work in general. One of the things that I've seen, at least in the DTC Twitter space, is advice that is oftentimes given that if you take it out of the context of the Facebook algorithm in general, you look at it and this is just simply say that's not algorithmic. That's literally just not how an algorithm is even able to function. Therefore, that advice doesn't make any sense. So I think that sometimes it's understanding it's like, what, what are algorithms capable of and not capable of? And then once you understand that, then start running the tests that test out how that algorithm is behaving in the different environments. Is that kind of what you're suggesting

Christian Limon  41:04  

then, too? Yeah, yeah. It's just what, what inputs create, what outcomes. And that's, that's really all I cared about. I don't The thing is, everybody would agree, but then they they do something else, because I'm not bringing up the things that they that they end up choosing ahead of that, which are like, you know, like, whatever it is there is they, you know, people tend to like, want to be right. So they'll like, manufacture being right by not trying anything that they don't think will that will challenge that, unless they thought of it then, then, then they won't try it. Is

William Harris  41:44  

there something that you're thinking of right now that you're not alluding to, or maybe you are alluding to a little bit that you're like, This is so prevalent right now, and I feel like this is kind of bogus. Yeah,

Christian Limon  41:54  

it's it's more just that, the the the attitude toward data and wearing like, a data backlash now, where it's trendy to say we don't, we don't, we're not, like, we don't use data here, here we use, like, qualitative information and like the richness of that. And people are building like crazy good companies doing that. So that's the that's like, you know, people are using that as as evidence. But like, really, if you just world class at anything, you're can build the company it just like, this approach, the data, the data version of doing that. That's like, that's the skill set and strength that I happen to have. But if you have a different one, and and you make the company that that strong at it, if you're, if you're, like, ridiculously amazing at that, at that particular, you know, contribution the company will, will knock it out of the park. Sure,

William Harris  42:54  

you had told me before that a bigger catalog was one thing that was helpful for Wish. Why? Why is that helpful?

Christian Limon  43:06  

For catalog? I'm going to write it so that I don't forget

William Harris  43:09  

sure,

Christian Limon  43:13  

because I don't think I adjust the actual data. Question of, like, what people are doing? Sure, really quick, really quickly, the like, the summary of like, the best way to describe what I think is the issue with data is that people only test what they they're too selective about when they're data driven their their data driven when the data agrees with what they want to do. And like, the best story I can, I can, I can say that kind of, like illuminates, like, the way I feel about it is, when I was in high school, I was really serious about athletics, like crazy serious about it. And the school was, like, notoriously massive athletics. That's like, the identity of the school and the people in it, and it was all boy. So it was like, very, very like, that's the only thing to focus on. The sports had this coach who was who, you know, when it would rain and all this stuff would, you know, still make us do stuff outside and all these things. And he would say, you know, practicing and like working and working out and and training, you're not an athlete if when you do it, when you want to, you're an athlete when you when you also do it when you don't want to, and when it's challenging, or when things are in your way, when you do it anyway, that's a professional, and that's the way I feel about data, is because you do what the data says sometimes, doesn't make any data driven totally like you need to flip that

William Harris  44:54  

there's a client, and we have a really good relationship. So I can share the way that this went. But we beat their sales with at Target last year during the holiday season, so they're nationally in target. And the target was like, this is the first time anybody's ever beat our like, our sales, like you guys sold more on the website than we sold at the target. And this is pretty crazy. But he sent me over this email. This is maybe even this week, or whatever. He was like, Guys, our, our ROAs on, on brand is like, a 7x and our ROAs on, you know, prospecting on Facebook is a 2x I want to shift so much of our budget over to, you know, brand search on, on on Google, where it's got a higher ROAs. He was like, this is a data driven decision, and I know him well enough so it's like, I could say it this way. I was like, my quote, I put in quotes. I was like, it's not data driven, if you don't understand the data. William Harris, right? So I was like, the point here was, I was like, but those are two very different things. The reason why your brand search is performing very well is because of the prospecting that you're doing in these other areas. And so that's why people are searching now for your brand, yeah.

Christian Limon  45:59  

Say it's true, right? It's doing exceptionally well. And that is, on the surface, what it's literally telling you, yes. That's like a binary statement. You still need to bet size so, and that's the execution part. The entire the entire skill has, doesn't have to do with the binary judgment. Has to do with the proportion and the method. Yes, 100% agree. And that is the, is the, the the part that's not necessarily, you know, can't, not going to be a blog post about,

William Harris  46:30  

well, a little bit, yeah, it's hard to turn into a blog post. I've tried to, and I've actually made a couple of really good posts about Bayesian statistics within advertising. I have one on Simpson's Paradox within advertising, trying to help get a little bit more of this understanding out there, trying to educate. And part of that is because we will see things come up along the lines where people are like, Oh, this is a data driven decision. And one of the ways that I try to format it, I guess, in my mind, makes sense, is you're making a decision based on a one dimensional view of the data. And that's partly where I think people get into this problem. They're like, well, the data makes, this is what the data says. And it's like, that's flat. And the the illustration I give is when I would break my wrist as a kid, because, okay, you were in sports. I was in sports as well. Broke a lot of bones doing some really crazy things, like back flips on pogo sticks off of loading docks. And I can remember the doctors saying, okay, great, you know, like they take an x ray, but they don't take an x ray just of once view. They take an x ray of the AP and the lateral view. And the reason is they need to see three dimensionally how that break is moving in order to understand how to better address this. And what I've seen people run into is they're looking at a flat dimension of data. This could be their Shopify data. It can be GA data. It could be North beam or triple whale dead, or it's still a flat dimensional look at it. And if you're not looking at, but how does that data also look when you look for a third dimension, look differently from a different perspective. This could be looking at, my favorite is gonna be a GL hold out test. But how does this also look at it? If you're looking at, what are your customers saying about you? Does that data match what you're seeing in your UTM data? And if it doesn't, there's a discrepancy, and that discrepancy is helping to tell you a little bit more about three dimensionally, what's going on. So don't make a decision based on one dimensional looks of data, even if that data is really good attribution data like what you're getting from something like a North beware trip oil, where it's attempting to be a better look at data. It's still just a dimensional look, if that makes sense. Yeah. I mean,

Christian Limon  48:27  

the it's completely dependent on what, what like degree of in size of bet you're there that's gonna power. Like, sure, if it's about a creative or it's about shifting over 100 bucks. Like, have at it. And the data that I want is, does it work or not in like, live and like, you really only need to be like, That thoughtful when there's a degree of consequence that you can't reverse, but if you can reverse it, like with ads, all you have to do is turn it off immediately. And there you go. You reverse the decision, sure, to try something. And you got your decision in a day. So you can, you know, factor that into into your your stuff?

William Harris  49:22  

Yeah, so coming back to the catalog, I think what you talked about was like taste versus need for bigger catalogs. Why? Why is a bigger catalog? And I feel like I know the before bigger catalogs helpful,

Christian Limon  49:41  

it just has to do with the diversity of of preferences in the population. So that's the demand there's a demand side and there's a supply side. The demand side could be described through a distribution that has like long tails. So you need, if you want to match that with something you. The supply side needs needs to reflect that, and it needs to reflect that in an environment of like, changing tastes, changing changing economy, changing macro, changing platforms and so like it the low priced items direct manufacturer was a way to have a supply, be able to have a supply that could match the like, the amount of like variables that we were or types of demand we were trying, we were trying to match with, and that could have a distribution of long tail and long tail that match that did that? Yeah,

William Harris  50:46  

yeah. No, that makes sense. I think that this is one of the things that people sometimes get, I don't know, hesitant about even still with, let's say, catalog ads or things like that, that are showing people a lot of different products. And sometimes they're like, Well, I don't know if this is showing them the right products or not. How can you trust Facebook to do show them the right product? And a lot of times, there are ways you can kind of interject yourself into a system, into an algorithm. And I'll say that some of the ways that I interject myself into the algorithm of Facebook, because I do, I do believe in trusting Facebook to a point in Google as well, because they have way more data and data that I don't even have access to, right about what people are looking at, what they're searching, what they're doing. There's a lot of data there, but you can get into a native feedback loop. And this is one of the ways that I would say that I interject things. Let's imagine that I happen to know that this persona of person, and I'm just gonna make this really reductive for ease of conversation. Let's say that I happen to know that guys have a lifetime value of $400 on my website, and girls have a lifetime value of $4,000 on my website, but they cost 25% more to acquire than the guy does. Well, if I happen to be running lowest cost on something like this, I'm going to end up favoring the algorithm. Is going to favor acquiring the guys at the $400 lifetime value, and I'm going to say as a business, but that's not necessarily the one that I wanted. So I might interject myself into the algorithm and say, I'm going to focus on this. I'm going to make some ads that are for this. I'm going to get sure that I'm targeting these people, because I want that despite what the algorithm is bound by, if that makes sense, that is what targeting is. Yeah, exactly. But a lot of times people won't do this right. And they say, well, broad audience, just open it up. You're going to

Christian Limon  52:32  

find that. But the problem is starting. Nobody was doing broad audience, you're right. Everybody, everybody knew who their customer. Everyone knows what the right answer is. You don't need. You don't you don't need us to set up a system where you're open to discovering that you know, you know what it is, sure, and that that, that general mentality is what keeps, what would keep people even just not going in the front door, of being good at ads, because, like, the that is, like the cost of admission is, is the that that having that kind of framework?

William Harris  53:16  

Yeah, what are some other areas of advertising, that you think that maybe we aren't thinking through as advertisers right now, or that we're missing, or that we're doing incorrectly, that we should re evaluate? Yeah, well,

Christian Limon  53:32  

in general, everything is being done somewhat incorrectly. Incorrect is like a normative but like, like relative to the goals of the company into the shareholders, it's pretty incorrect. And I think that we're just very humans. Are like very story oriented beings, and we so will tend to highly value things that have a story attached to them, and that will create a big bias, and where, like, very biased toward favoring things that worked in the past, when without, like, really taking into consideration the different context of What was available in the past versus available now as options, or that you actually don't need to make a binary judgment. It's a degree of you can do everything, just in different proportion. And I think in general, all the everyone's making mistakes for the most part, most and it's, it'll always be in that that'll be the theme of all the mistakes is, is that type of bias

William Harris  54:50  

makes sense? I want to shift gears a little bit, especially doing

Christian Limon  54:53  

stuff that that, like experts say you should do, you know, or that was done. I.

William Harris  55:01  

Yes, and this is why you said, let's say there's the myth about being data driven, and let's just say, similarly, there's a myth of best practice, right? Because Best Practices oftentimes, one of the things that I've noticed is that there can be very different people that have very different ideas of what best practices is, and that's to say that's the best practice that they've tested in the environments that they've tested it in for the companies that they've tested it on for. But that does not mean that that's the best practice for this completely different company, and they may not have tested the other things.

Christian Limon  55:31  

The true best thing is what you're describing. I think, like, definitionally, like, what people would say would think best practices is, is like, what is the thing that everyone agrees works? Sure, and maybe that that makes sense if you have one bullet in, you know, in and it has to be effective. So you're like, All right, let's do what has worked for others. But if you the way venture, the way startups are, are you raise a bunch of you raise capital so you have a amount of time allotted and resources to answer, like, a several questions. And do you risk, like, your ability to create value? And do you risk this, like, the value that you're saying that is like possible, and you don't have one bullet, you can sure you can answer a number of questions, and you can answer them sequentially, if you'd like, I like

William Harris  56:30  

the way you frame that, because if we, if we go on that perspective where it's like, Look, you have to make sure that you're at least remotely effective. You've got one bullet, then, okay, do the thing that's at least been tested and proven and is going to be have some modicum of success, but likely not. The most successful personality was not, has that

Christian Limon  56:49  

like has value. It's just like on in what proportion, yeah, or what would have been, because it

William Harris  56:55  

was not best practice to spend 30 to 40 million on uh jerseys, to grow an app company, right? Like that wasn't best practice at the time. When you did that, that was like, Hey, this is a test. And that's the

Christian Limon  57:07  

whole point. Best practice to spend several 100 million is right? It wasn't best practice to spend several 100 million on ads, sure, or best practice to spend it all on Facebook for the most part in the beginner or and to not use an attribution solution in the beginning, sure, because there's a just a lot of new there's, there's a lot of nuance that creates opportunities that make it so that the right approach In numerous circumstances are counter counter intuitive, first of all, and second of all, the the opposite of what is the best thing to do normally, like normally, and just thinking in averages messes a lot of people up, because the things at the individual level don't behave that way. And that's not the world doesn't behave that way at the individual level, at the individual level, like there's insane amount of variance.

William Harris  58:10  

Yeah, yeah. One of the other things that you tweeted about that I really appreciated, that I wanted to bring up, is this idea of founder mode incentives. You said it's turtles most of the way down on top of a hard coded algorithm called survive. What does this mean

Christian Limon  58:28  

that at a base layer, if you want to know why people are doing things, if you keep asking, why, why is that important? Why are they doing that? Why doing that? Why are they doing that? And just keep going? It just ends up getting going, just back to survival and that we it's like we have behavior that is ingrained in us to help us survive, or else, you know, we wouldn't be here. And you know, where we have inherited like these traits based on like these survivors, that is, like, the most descriptive thing about them is, I don't know a lot about these people, but I know that they could survive because I'm here and I am there. You know, I'm I've inherited everything through them.

William Harris  59:22  

Yeah, and to your point, I think when you were talking about this from the founder mode, incentive is that there's an incentive to the founder that's not necessarily there for the employees the founder.

Christian Limon  59:36  

And why that's related to the survival is that there's numerous country contributors to decisions and of different degrees, also of ability to, like, remove the CEO. So like, no matter how objective you can be, the most objective person, but in that objective. Activity. You there is human nature to factor in the and I like to do the thing that helps me survive and keep this this, this thing that I love. And you know, the CEO reports to the to the to the board.

William Harris  1:00:18  

So you're a VC Now naturally, and you tweeted about the new investment themes that accept you the most. And I want to dig into a few of those themes with you. You talked about a lot of different ones, and I'm not even sure which one I want to dig into, but one of the ones you talked about first was the myth of the differentiated product. And I like that, because I feel like we talked about that a lot of times in e-commerce especially, it's like, oh, you need a differentiated product. Why is there a myth of a differentiated product?

Christian Limon  1:00:46  

It's that's that is the goal, I would agree and it is valuable, I agree with that. Is it achievable? Is another question. And have you achieved? It is also another question that I think very often is not, is they people believe what they want to believe. They they're what it is just becoming more and more obvious, the that there are like eight, like buckets of options of the same or similar solutions, especially on the surface, on without, without access to like first person required data, the things look no different from each other. And make the same claims. You know, they're both the best according to the website, and they both like, can get me the best deal, you know, according to the salesman. So the when you lower the bar. So, like, the startups really started pouring in when AWS took off. So because it lowered the barriers to entry, to having a startup, the cost became much lower when you didn't need to host your own servers and host the website and be up 24/7 and, you know, be ready for when you if you take off. And that's happening again through AI like coding and programming, the skill set is also a big barrier to entry, and that that's being removed, or at least is the bar is lowering, as far as who can contribute with that to code or contribute to things that require code.

William Harris  1:02:38  

Yeah, I've seen this, right. I mean, the things that we're doing with generative AI in the coding space is incredible. And you alluded to this earlier in the show too, where, if you release a feature, it could be copied tomorrow. Like, it's not even that hard anymore,

Christian Limon  1:02:54  

yeah, and it's fine, but, like, you don't need a differentiated product to do really well or to win, like, look at like, Coca Cola is not, is like, you can, you can copy Coke, but you're not going to, like, take their market share. They're, you know, they, they're popular for a different they, they probably are for different reasons. Like, consumer packaged goods is like, good at, at this they don't, they don't just, they don't describe it this way, but, but, you know, they'll say, Oh, these, there's nothing too different about our product. But what is different is that, you know, we is our, that our name and that we and they associate that quality or whatever. But I think it's just that, like, the differentiable thing is things like that you that they invest in, like once the product is locked in, and there are 1000s of them that are identical and probably use your same factory and materials, is you know your the your color, Your how you say things, where you say them, and how the site looks, you know, that is where they try to differentiate themselves. But like, if, if everybody's copying that there that leaves people with no difference, no differentiator. And like, I think right now in technology, we've been in this point where it's the amount of identical products has been accelerating, and is already at the point where, realistically, you're not differentiated, right? But I think that we're getting to the point where, and even if you get that far, and you, and you create a solution that helps you when, independent of being differentiated, that solution is longer differentiated like now that like you're like, multi levels in being cloned by by hundreds of startups.

William Harris  1:04:58  

And I think that alludes. Maybe a couple of the other ones that you said you talked about the un cloneables, the Chinese room and the visible is Cloneable. So I'm going to maybe skip those ones. I think they're similar enough to that, right? Would you agree with that? So then the

Christian Limon  1:05:10  

next one, those just have to do with the Yeah. Like, how are you going to win in a world where people can copy anything they see that?

William Harris  1:05:21  

Yeah. So then I don't know this one's different. C is zero, yeah, D zero. Feedback loops, hidden in plain sight. I don't even know what that means. I'm literally coming to you what one's that one referencing

Christian Limon  1:05:37  

in that one has to do with it's I have this I always talk about how B to B, most of B to B evolved and grew up and matured in a in a culture that's quite different from the one that I that I thought was the normal culture of Silicon Valley, Which is like data, data driven, very signal focused. But then what do you like? What is your culture when most of your people are sales people, and most of your important you know transactions like, like are really driven offline through conversation like, how do you have a strategy? Or how do you know that? How do you make sure you're doing that best? And really there's there, offline, you can't. What are you going to do? You know you can't. You can't manipulate people like a Facebook campaign, sure, offline, with the with AI, if we record every every meeting and every call trans and transcribe them, you have, you have, now, in real time, transformed that, that whole process into a text process, and created now a feedback loop to be able to optimize. So for the first time, like people don't, people think that the valuable thing is, you know, saving time on taking meeting notes, that's not the valuable thing. The valuable thing is, you just gain access to statistics like you. You were not able to use statistics in any form. You now can have at it. You're you can you now have a feedback loop and and like a confirmed feedback loop, where you know exactly what was said, and you know exactly what what what you had, what your input was, and what you know what the inputs have been,

William Harris  1:07:38  

yeah. So you do enough of those, you actually have enough data now to start really pulling out trends better than just subjective data. I like that. Yeah,

Christian Limon  1:07:46  

we're guessing, or, you know, every like, you know, attributing it to people's style. You know, everyone, like people, have egos so they they think, like, like, what really closes the deal. Is my style. You know, sure, like, you can't have, like, 1000 people thinking that, first of all, they're not right, and then second of Second of all, then what the hell do you do if you run that team? Like, what's your contribution? If everyone's running around doing their own thing, and you see it in in the way people reach out to other folks these like, we'll call them like, human campaigns. And they're, it's like, it's like, if I asked you, can you run some campaigns for my for my brand or for my company? And then you and I asked, Oh, let me give you creative and some copy of, like, how we describe ourselves. And you said, no, no good. I got the best style I'll do, I'll do whatever. I'll get users with my style, which is, like, maybe, maybe, yes, but I, like, we, our company actually has a name. You don't have to make one up, or, yeah, kind of, there's a degree of, like, set standardization that is the baseline of like need needs to be used if you want to like experiment on anything and use statistics, yeah,

William Harris  1:09:11  

yeah, every experiment comes with a cost, right? And this is one of the things that we talk about often, which is, if you run a test and you run it 5050, which I see sometimes people will do, and that's just bad. You can guarantee that 50% of whatever you're testing is wasted. As on the flip side, if you're running the test of 10, you know 9010 then 10% is potentially wasted, or 90% is potentially wasted. But if the 90% is the one that has been proven so far to be the most effective, then you're potentially only wasting the 10% until you discover that, versus wasting 50% but there's always a waste every time. Like,

Christian Limon  1:09:47  

if you have no data, basically no users, if you're just starting 5050, is the only probably the sure, like, like, you know you have nothing to lose also, but once you have, like, stakes, i. It's even Best Practices wouldn't give you the conclusion that to do 5050, totally. It's like, you know, you sacrifice the experience, because it is, it is a sacrifice your The experiment will yield a lot of things of value, but to discover that, you need to test the things that yield value against things that don't or your your previous lower value. And those, those those users that had that experience, whether it's like 10 people or a million or 100 million, that's their experience. You can't take that away from them. And that's, part of their LTV equation, which is now anchored low.

William Harris  1:10:45  

That's the key that a lot of people miss about that test, is that it's you didn't just impact that sale. You impacted the entire lifetime value potentially of that class.

Christian Limon  1:10:53  

Yeah, yeah. You have to assume that. Treat it like you only have one shot at people, right? The

William Harris  1:11:02  

other one that I really like that you talked about is war halls. In the future, every avatar will get 15 minutes of fame. What does this one

Christian Limon  1:11:08  

mean? Oh, this was just a play on because I don't want to be too Doomer about stuff. Like, I don't want people to think I'm just super negative. But like, there's all kinds of cool things that can, that can happen and will happen. And like, Andy Warhol is famous for saying something like, in in the future, everyone, everyone's going to be famous for 15 minutes. Everyone will have 15 minutes of fame, which is, like, we're kind of, yeah, we're living there. He understood, I guess, like, human attention and what, how to get people's how to get focus. And he was famous for doing these reproductions, these, like quadrant print, like us, art, arts, canvases, art, art things. And he didn't paint those things, they would just be, it would be like Madonna, and then the other one is like an inverted color of Madonna another, and then another inverted color of Madonna or another, or a Campbell Soup, you know. So it was these, like cloned thing, things in the same painting. They look they were like, they had something flipped, but it was just like, cloned, so and so in one like in the future, basically every startup is going to is going to be one, one of those quadrants on Andy Warhol's canvas and but the upside is, you know, You can. You can have your own bots and clones and and things like that. And each one of them also, maybe they have potentially 15 minutes of fame in their lifetime too. And like, literally, some of that is happening where the digital influencers, but like there's no reason why they can't be. You know, the if cost continues to decelerate, there will be millions of these digital influencers, so and then there will be a day where every digital influencer, every bot, has 15 minutes of fame.

William Harris  1:13:21  

Yeah, I think that makes a lot of sense. I want to transition then into the last segment, which is, who is Christian Limon a little bit. One of the things that I think that I appreciated about you when we first jumped on the very first call that I had with you, and you can see it in the background, but I don't think you can even read it, but I saw you've got some books on Richard Feynman. There the top the Red Books, and I can, I can see him, but, yeah, there you go. Okay, now I can see it. Fine. Yeah, you don't have to move camera, but those are the books that I was talking about. Why? What do you love about Richard Feynman? Why is because I'm a big fan of his as well. But why are you a fan of Richard Feynman?

Christian Limon  1:13:59  

There's like, a depreciated piece to Richard Feynman, and what I think is an under appreciated piece to Richard Feynman. And depreciated pieces understanding things simply, and that if you don't understand them simply, you don't understand them. And like, he's so like, spot on. And like, you can see that like value it is is created and proved, both created and proved through simplicity and through uh comp the compactness, like how far down you can reproduce something without it being lossy, like, and that just, he was like crazy spot on when he came to that and he didn't describe, he doesn't describe these things like a scientist. He describes them with ordinary, plain speaking. And I think there's, there's a degree of egolessness to him that I think is like the most adamant. Thing is that that he, he was like genius, like one of the best like gifts to to to science, anybody he spoke like a, like a, like a human. And you know, he would have, he had lessons from his father that, you know would tell would tell them. Like, like, basically, just because you call something, or you know the name of something, doesn't mean you know the thing and like, just because you use these words doesn't mean that you're doing anything of value. Like you're the value you contribute in like, what you teach is independent of like, if like, how complex the word choice is that you're using, and you can tell you, just absorb the hell out of all that. And then he would win awards. And his his mentality was, like, what I don't I'm fine. I don't even, I don't want an award. I'm the reward is the thing that I'm doing, in and of itself that is rewarding. And there, there's an element where, where people are just too thir, Steve or to be recognized. But what you're what you're implying, is that where it's like, there's a tree, if a tree falls in the woods and no one's around to hear it, did it happen when Richard Feynman would say yes, and popular culture right now would say no, like I only I'm only x, if that person thinks that I'm x, and then there's an element where that's also just how other people make decisions, because they have to find the information. You can find the information through, you know, influencers, because everybody knows who people know who they are, even if they don't know the right answer, and if so, what they basically hacked without knowing it is they don't need to convince everybody. They don't need to, they don't need to be the what they what they say they are. They just need to convince that guy, that person who is the, you know, thought leader in the space. If you do that, then you're legit. And so that, like, you can apply that to the 30 and the 30, I don't know what that is, right, and I but it like one helps people, like, is the hilarious part, or just the like, ironic piece, but like, the fact that any that you care, that you make a list that everybody makes, but the fact that you that that is what You think like when people put that in their headline, that they think that that's one of their biggest achievements, is that what some someone else a list that someone else put them on is is just feels like it makes it makes the entire thing like a mockery. You know it's like it it makes the entire system like contributes to a lack of meritocracy.

William Harris  1:18:05  

Interesting. I need to go edit my my Twitter bio here, but no. But to be fair, you remind me a little bit of a modern day Richard Feynman in the ad space, and the idea that you're not very self promotional at all, and that is one of the things that I appreciate you whenever you whenever you say something and you reply to somebody, I sometimes look to see what people will reply back to you to see. It's like, do they even realize who they're talking to? Because you don't carry yourself as if, like you're just like, hey, I'm Christian Limon, you should definitely listen to what I have to say here. And that's why I tried to put you a lot more in the beginning of this. Because it's like, no, I want people to understand, like, just the genius that you're bringing to this, but you're you're very good about you're like, Nah, I'm good. I don't need that. I appreciate,

Christian Limon  1:18:49  

yeah, it's like that there is problem with, how do you separate signal, like you're describing a signal to noise? Problem of so the first problem is, is access to information, the second order problem, once you solve that, is too much information. Now I can tell what's real and what's not, and what to value if everything looks the same and like, we're clearly in that, in that stage where what's the difference between you and everyone that says the exact same thing as as you, when you say, if you, if you say, I know what I'm doing. What does that mean?

William Harris  1:19:26  

Yeah, that's very fair. I can only

Christian Limon  1:19:28  

be proven, like really, through observational results and or, and they can only get a feeling for it if they just feel that was quite insightful, like you can. You can separate yourself just from the quality of insight and the objectivity that you share.

William Harris  1:19:45  

Unfortunately, what somebody thinks is insightful can also just be based on their own biases towards what they wanted to find insightful as well. And I'm sure you've discovered this as well, where oftentimes somebody will say something that you know believe they believe, has a lot of May. It's just because it confirms how they already are thinking about things. And it's like, well, it doesn't make it right though, either, but that's fine. You're allowed to appreciate that.

Christian Limon  1:20:09  

Yeah, I mean, I'll correct people or disagree all the time and sure and think, like, no, like, I didn't have to correct you, like you can, you could have gone on. This is very correctable. And it's like, if I, if, if you, if this changes your mind at some point and you start to see, like, the flaw in, like, the way you're thinking, or how it's keeping you from exploring this other thing, then there's like, a i Did you like, this is a favor, not, you know, not, not a troll. Richard

William Harris  1:20:47  

Reiman, it was outside of his realm of what he was studying. But if he was alive today, seeing what was going on with generative AI, it's a very interesting thing. What would he think about generative AI and where we're at? I think he'd be tinkering

Christian Limon  1:21:00  

with it like crazy. He loves, you know, a physicist, love messing around with their own hands and seeing what happens, poking at it, seeing what's real. I think the anthropomorphizing of it, he wouldn't, he wouldn't do and he would think it like that. It like the degree that we air it will be in associated with anchoring it as like a human, or anchoring it as having human like qualities, so and therefore having human like motivations. You know, I do that. I get I get my word choice will be different with ChatGPT. If I'm frustrated, I'm frustrated, I don't like the answer. I'm like, you're wasting my time ChatGPT. And they'll say, Oh, I'm sorry. And I'll be like, No, internalize that you wasted my time. But you know, he's a good, he's a good counterbalance to what we have too much of. But I think he would, he would be tinkering like crazy with it and and launching games like fun games with it, you know, and like then maybe through that, there'll be some crazy discovery that will name the Feynman, The Feynman, you know, transformer particle. But he'll, he'll, he'd do things he'd enjoy. Yeah,

William Harris  1:22:31  

yeah. So I think it's funny that you talked about, you know, while he might not anthropomorphize it, because what's interesting is, you do, I do, and I think there's an instinct to almost want to out of because of how well it replies back to us, feeling as if it is human in nature. But I would even say that one of the things that I tell my kids when they are talking to Siri or ChatGPT, et cetera, is I want them to try to treat it as with good manners. And my reason for it is not because it has manners, but because I don't want them to learn to have bad manners, if that makes sense. And so if they're talking to if we begin to talk to ChatGPT as if it wasn't human, I think one of the problems is we could get to the point where it's like our ability to actually have conversations with human beings will go down significantly, because we'll expect humans to be able to tolerate that type of communication, and I think that that was just degrade things in a significant way. So I treat it as human as I possibly can, as a way to get

Christian Limon  1:23:30  

what it knows. And its base form of communication, of inputs is, is from us, writing that way, like the in, where, where people and are. When we write, we write like that, right? And from that stuff is what it it, it uses so it's in it's in there, yeah, being polite, yes,

William Harris  1:23:55  

um, Christian. This has been so much fun talking to you, and I feel like there's a million more things that I would love to that I would love to dig into with you, including a lot more Feynman stuff that said, I want to be respectful of your time and everybody's time listening here, if people wanted to follow you or learn from you or work with you in some way, what's the best way for them to Get in touch?

Christian Limon  1:24:19  

They can follow me on X @CPLimon, which is L, i, m, o, n, c, p, Limone and my email. They can email me directly if they want to reach out, send me a DM or email me at christian@limon.ai.

William Harris  1:24:40  

Very generous of you. Again. I've really enjoyed this. Thank you for taking your time to share your knowledge and wisdom with us. Yeah, this was great. Thanks for having me. Thank you everyone for listening. Hope we have a great rest your day.

Outro  1:24:52  

Thanks for listening to the Up Arrow Podcast with William Harris. We'll see you again next time, and be sure to click Subscribe to get future episodes.

‍