You're COMPLETELY Misinterpreting Your Incrementality Tests (With Taylor Holiday) transcript

It's me. It's Taylor Holiday together talking in a packed episode about whether or not everybody is doing incrementality testing wrong and how to do it better about Taylor's daily podcast created by AI that updates him about what's happening in his company every single morning automatically. Really, really cool. We talk about whether or not Hierroox or CrossFit is better at capturing the value of the product they created, what you can learn from it, whether or not your brand should be creating events and be learning from Hy Rocks and CrossFit about how to build events for yourself and for your customers.

And a very heated discussion at the end by our standards, I think, about whether or not a $3 million e-commerce business growing 15% per year with 10% profit is a good business or not. Really good episode. I loved this conversation actually in a few different ways. Let's going to do it. No. [music] >> All right, Taylor. You just said you doubled your money on a on what? An NFT. >> No, it's a a coin. A token. >> Okay.

What's the I don't I'm not Never mind. Don't even Never mind. I was not going to ask you. What did you double your money on? >> Gambling. Like >> basically is what it is. >> You gamble more. You're really good at it. Uh, so I tokens are sort of I like to think of crypto tokens as an encapsulation of the cultural energy of around an idea. And so when I see something that I think it's like X gives you access to early indications of verality, you try and find the corresponding token.

Now the hard part is there's like 77 versions of them. So you have to try and find the one that the energy consolidates around and then you can make money. >> Okay. And then you're just really >> So what is the cultural energy around? So, this is all about um what Matt from Octai what and this is why like I'm pretty sure I feel like I'm in a simulation because the whole world is pivoting around something that's so close to you.

Yeah. You know when you have that feeling you're like this can't really be what's happening which is someone from a ecommerce app founder >> that makes quizzes >> that makes quizzes now has created the thing that the entire Elon Musk and internet is revolving around. You're like, "This feels it's just hard to believe." But if you obviously Claudebot was has been a big topic of conversation. Cloudbot basically being an the ability to create an agent, run on a local device, give it access to everything and set it up on its own and sort of be an autonomous employee extension of you.

That was a big topic for the last week on X. And then today, Matt uh Schmidt, I think is or Schlitz, I don't know. I'm pronouncing it wrong, maybe. >> I don't know what it is, but Matt, I'm sorry I'm mispronouncing your name. um launched basically a social network for all of the agents that have been created. All the multis they call them molting, shedding your skin, growing, evolving is the metaphor here. >> Yeah. M O L T >> Y >> Y not not NI. >> Yeah.

Yeah. Um are and now they're forming a social network where they're interacting with each other and this is like sort of taken off and become the today's verality. So it's now clawbot has become open claw is the name of the the product now and uh these are called multis and then molt book like Facebook is the social network that they all exist on and it's taken off and taken over the internet and it's really a fascinating thing to watch these agents interact with each other.

Now, there's debates whether this is just like what you would expect when they're sort of generating prompt output input output output input back and forth to each other or if this is some sort of like awareness, but there's all sorts of really cool clever >> What do you think? Do you think it's Do you think it's >> I don't know. I'm so dumb that I get toggled back and I get swayed every new tweet I read, I get swayed by somebody to go back and forth. that you know that that's actually probably relative to your technical knowledge of how it works as as far as I know of your technical maybe it's deeper than I think but >> um of how LM work and stuff like >> I think it's probably reasonable swayed pretty often right >> read the credentials of who wrote it >> wait it read the next one wait >> that's what you have to do right this is actually this is why influencers exist in e-commerce marketing is because you need to have somebody smarter than you for an area especially for an area where there's technical expertise going on >> this is why your supplement brand needs influencers.

You have to have somebody in a lab coat saying, "I promise these ingredients are going to make you healthier because there's no way you can know the science on that." >> Yeah. >> And it's the same thing. >> For sure. So, that the the there's a thousand interesting angles on this. I'm sure you guys will have read a bunch of them by the time we get here, but some of the favorite ones I wanted to chat about because I think it sort of overlaps one was the token piece.

So, I've got, you know, I don't want to slang it, crypto token here on this episode, but yeah. Uh, >> so that's just fun to play with. >> Can you talk to the microphone? Have you never recorded a podcast? >> This is why I have a headset. Do have you not noticed that? This is why I do this is because I get bered on the internet. I have I can't even tell you how many LinkedIn messages I've received from people saying, "Hey man, I like to listen to your content, but could you please speak into the mic?" Um, I get that a lot.

So, yes, I'll try. >> I also talk too fast and have never been able to fix that. So, I understand. I do think it's funny that uh that you are really bad at this because it's I don't think it's that hard of a skill, but it's whatever reason your brain just can't do it. >> Yeah, it's not I'm not good at it. Um, I do I actually really appreciate when people leave me comments like that. By the way, >> it I agree. I I've said it multiple times.

I literally sent the guy like, "Here's the headset I bought. Thank you. I'm going to try this now because I don't want it to suck and I don't want but it seemed like a better solution than fixing it." >> Yeah. >> But now I wear it and like what I hate is that like it creates a dent in my head all day long and so it's like actually super annoying. But >> yeah, it doesn't sound that good either to be honest with you. I listen to you I listened to a really good podcast you did the other day. >> Yeah. >> Uh which you should feel good about because I don't really listen to e-commerce podcasts.

So >> really >> I listen to all your podcasts. I I've always I've always felt I >> sometimes I cut them off halfway. >> Yeah. You've heard me say if you're Yeah. If you're giving me the repeated spiel. >> Yeah. >> Like your new one about how like everybody should stop growing. That's your new one. I I kind of just turn those off. >> I found that encouraging that you listen to those. It makes me feel like value my stuff to say. >> Yeah. >> But can we last thing?

The crustarianism. >> Crustaparianism. >> So one of the things that has happened inside of the social network is they've created a religion and they're like writing scripture and bylaws and recruiting agents into it and converting them. Um, and it is like fascinating to to read. You should you should you just because you'll you'll appreciate >> that. Yeah, I'll be fascinated by that. >> Um, >> but you should go read through it.

And it's just like interesting that one of the first things they did in a social gathering is they started to try and define the bylaws by which they operate. And like it's culture like and religion is this in so many ways like they're trying to say like who are we to one another? >> What are the guiding principles by which we will behave? um what can we appeal to if there are issues like and I just think that's like a a fascinating organizing first principle that even the agents did when they gathered together. >> Yeah.

What do you think that signifies? I'm I I heard a really good AI take the other day. Do you know Andy Crouch? You probably you might >> He's a pastor, right? >> Uh yeah. He's like a he's like a public intellectual Christian. He works he's so he's a Christian. is a theologian, philosopher, but he he's he runs a group called Practice Labs or is like he doesn't he's one of the key people >> which is sort of like a specifically Christian incubator.

Yeah. >> Um but with a lot of like sort of theological reflection built into it. So they're constantly talking about like how technology makes the world better or worse through a Christian worldview and some of those things. So he's just really thoughtful. He's just he's thinking about what are people like and does this lead to thriving or not. You know, a lot of I think even if you're not a Christian or or if you're not religious like you would you would find him helpful.

Um but he one of the things he was he was saying about this is of course that like LLMs are trained on humans right so so um so like they just are in some way they are going to reflect whatever it means to be human which I think is the interesting thing about like making the religion you know >> well I think about it like so I just started little league you're coaching >> and I've got four men that are like coaching with me so myself and three others and the first thing I did before our seating started is I wrote I give it to the parents but really it's for me as coaches I write our coaching philosophy this This is this is the language we're going to use.

This is what we're going to say. This is how we're going to design practices. This is what the goal of the season is. >> Shout out Devon Morgan. >> Yeah, exactly. And >> Devin is a big inspiration for all of it. >> And the point is that we have to develop these common sets of behavior and rules and language together in order to accomplish a purpose. >> Um, and so it's it, you know, it's our ten commandments or whatever metaphor is here.

Like there's a there's just a common set of ways in which when humans gather in order for them to organize well there have to be some boundaries to that behavior. >> I think you also need to raise the chair. >> Well, I'm just not that tall. >> I understand. I'm just saying that the highest chair height. >> Are you Get off of me. Okay. >> Sorry. Well, I just was Michael's blocking your face. >> There you go. >> Um, yes.

I uh I I think this is def Oh, well, listen. I could go down a lot on this on this topic, but I uh I I think there the idea that you that humans get together and ask the question, what does it mean to be human and what do we believe about our purpose and meaning is is a >> that's the questions you inter I didn't I don't remember saying any of those questions. >> Well, but but you said you said they created bylaws and what does this mean and like how do we behave with one another?

But if you want to ask how do you behave with >> behavior modification as much as anything though, it's like how do we actually organize ourselves? >> Yeah. Well, but yeah. Well, but but a religion is not I mean at least in that case you're saying that they are making a step beyond just the rules of engagement to what does what do what do we mean you know and like that's that's what I'm saying is like there's a step actually if you're going to say they created a religion >> it's something different than that it's >> that's just what they called it I like I I think it's like yeah it's because the metaphor I use was the baseball philosophy document in which case I don't think that's like a pursuit of some deeper understanding of art what does it mean to be human >> no that's just yeah >> it's just like how will we be? >> What do we think will be best for us? >> What is okay and not okay? >> Yeah.

Yeah. Yeah. Yeah. Well, anyway, by the time this comes up, maybe >> it's all over. >> Yeah. Yeah. Maybe this is >> cuz they're they're like, "Didn't you say they're multiplying and stuff?" >> Well, yeah. And they're like having conversations about us and like one of them even like created a phone number and called their their human asking >> that story is real. >> No, of course not. Am I sure of anything? Absolutely not.

And in fact, I've asked Grock the question to try and be like because I I I am very afraid of being dumb and like publicly and uh so I'm like, "Okay, how much should I believe that these posts are actually agents?" And it placed the odds at like 75 to 80%. Sure. That they're agents are and not humans acting as agents or human prompting agents. So, sure. >> Yeah. or or or somebody made a uh an an agent that creates a bunch, you know what I mean?

Like like there's actually a human step in it that is sort of undisclosed, you know, that somebody created an agent that purposely goes and does this and just makes Twitter accounts all day. That would not be very hard to do, I don't think. You know? >> Yeah. So, while I'm while I'm immune to the internet trying to force upon you the like nuance hammer and the caveing every statement you ever make, like I have a like a deep resistance to everybody who tries to impose that onto everyone around them on the internet, which is like ex's favorite thing to do. >> Yeah. >> In this case, there's plenty of caveats. >> Yeah. >> All right.

Well, I don't know. It's fascinating. I mean, when you told me when we were at lunch and you told me that story >> Yeah. uh that that happened. I >> you felt you didn't feel good about >> I didn't. No, I I I physically the idea that the agents are talking about us and behaving on their own and interacting with each other and all these things. >> What what I the the truth of what I felt in my body was fear. I was like, "Oh, that's scary to me.

I don't like that." I'm over It's funny. It didn't last. I'm over it now as we talk about it now. I'm like, "That's kind of interesting." >> I Yeah, but I I have this reaction a lot like I feel very fragile right now like >> in my work. Um and one of the things that AI does and I think this is parlay into the next topic of we can talk about kind of I know you guys are spending a bunch of time on try to build fire and um we are too one of the things that it does my my first experience with AI is it surfaces first the breadth of information >> that is been absent from your purview so like for us we you know are at this point where there's a lot of customers and a lot of things happening every day at CTC and I don't know what's going on all the time and so we have this mo model that we use for forecasting our own business that applies a percentage likelihood of renewal onto customers and therefore their future contract value.

So let's imagine you have a customer paying $20,000 a month. They they have a renewal coming up in March. The question of what should the value of that revenue be in April is like a problem to be modeled and predicted, right? Um and we have found like the historical way we used to do this and like it feels barbaric is even saying it is like we have account executives who are responsible for the relationship with the customer. they have calls with them regularly and they apply their own personal expectation on a percentage basis to that number.

Um and >> so so the say the executive says I think it's 50/50 that this person comes back and then in the spreadsheet they write 50% and then the spreadsheet that spits out $10,000 a month in the future. Right. That's right. Because the worth the value of the contract is 20. So >> that's right. Exactly right. >> And then that behavior done a million times because you guys have a lot of customers ends up netting out a number.

So my question so before you say anything else did that work >> well. So I think what we've experienced >> I work I mean did it get you reasonably within the range of I don't know >> the answer is we don't know first of all that's like one it changes so much that's another thing they're allowed to change it like every week >> so it's really so the question of >> that seems reasonable to me >> right and so the question is like what were they right when >> you know because the the decision is sort of binary at the end but their opinion changes all along the way yeah >> right and so at what point would you peg it and go were you right on March were you right in April were you right in you know so I think That's a that's a hard question to answer.

Um, but what we've started doing now instead >> Well, let me ask it a different way then. Did you guys previously have a problem where you were having revenue be massively different than expectation, particularly in a negative sense? >> Uh, no. I'd say that what what's happening now is it's acute at this time of year because there's so many renewals this time of year. And we had a couple go a direction that were very different than we thought.

It was very disorienting. >> Yeah. And so that caused this sort of like evaluation of this system and sort of like updating to like hey there's probably a better way to do this now. And this sort of spills over into a bigger thing of just um sort of creating hierarchy of action for all of our people and everything else. Some of the things we're doing in status etc. But um so what we started doing is building uh we have every now every call transcript databased.

We have every Slack conversation databased. We have all the statist perform business performance databased and we have a gem um that runs an analysis against the customer based on those things and predicts the likelihood. >> Um and then >> really cool idea. >> Yeah, it's it's well we'll see. Yeah. And then the other thing we've started doing is we also through notebook LLM LLM notebook LM notebook whatever it is um you can feed it all that information and have it create a podcast.

And so every morning I get a five minute podcast of people talking about what happened at CTC yesterday highlighting the biggest concerns, the biggest wins and where everything is in between. >> Yep. >> Um >> really cool. >> And it is it's super cool. On my drive-in, I'm getting two people who sound like podcast hosts talking about CTC and all the client calls and they're going, "Yeah, yesterday with, you know, client X, there's clearly distrust happening and they do not believe in the incrementality studies and you need to get in there and figure out what to do, blah blah blah." Um, and so what it did though, the first experience I've had is, oh, how was I ever making decisions ever?

Like, >> I had no there's so much happening that I had noility to it. Yeah. But then I then I sort of take a step back and the temptation is like, "Oh, those are all things that I need to solve." But that's really not the case at all. But but but the point is mainly that AI services so much information that it's overwhelming um to figure out now in light of all that information. >> Well, this is why my question is a little bit like, are you sure that's necessary to do that? >> No.

More staffing is the company you should be using to build your e-commerce business's team. And you should be doing that with more staffing because they give you access to incredible talent in the Philippines for your business. And if you're still thinking about Filipino talent in your business as $5 an hour virtual assistants, you are missing out on the opportunity. You've heard me say many times before. You've heard Taylor say many times before that one of the things that makes e-commerce businesses so good is the ability to scale with a very [music] lean team and a very lean opex.

And I really believe that and I think one of the ways to capitalize on that even more than you are already doing is by finding incredible talent across the world. People with deep e-commerce resumes, you understand what you need at manager, director, even executive levels in your business who can help you across the entire business like supply chain, [music] marketing, graphic design, video editors, whatever it is, you name it, you can find great talent.

And more staffing will help you do that in the Philippines. more staffing is thinking carefully about not only identify talent for you in these businesses, but because they were built off of a US-based e-commerce business, they specifically are helping you think about how to find e-commerce talent and how to integrate those people with great workflows powered by AI with great awareness of how to interact cross-culturally.

They will help you think about how to build your entire system of opex with amazing talent in the Philippines so that you can be running a profitable business that gets incredible teammates to it. Another way I've framed this a lot of times the way that I think about it in my businesses which use more staffing across all of them, literally every business I'm in and there are multiple of them is using more staffing. The way I think about this is just that my dollars go so much further in the Filipino market than they do in mine.

And I've just had a incredible experience working with Filipinos in my company. So you can get incred access to incredible talent because you can hire at the top of the market there. Whereas those same dollars would not reach to the top of the market in the US. Go check it out today. Staffing.co/af. Like I said, Afgrowth, my e-commerce business that I'm starting, it's all being done with help from more staffing. I'm a huge fan.

Morstaffing.co. Get started today. One of my thoughts was like I wonder how this will do relative to to two other approaches to this problem. Uh and first of all I still have the same question which is did you actually have a problem before or you just spending is it almost like you get a new tool and so you want to start building things. >> The problem it's never a problem if you're wrong to the upside. >> Yeah I know >> right.

So yeah but you were still wrong. >> Sure that might not be a problem though like you said right? It's never a problem. So but the point is so so there's that question but secondly my question is what if you compared this approach to two other approaches. One of them is let's call the control which is what you're currently doing. Yep. the account executive predicts uh and it's not just one prediction, it's a bunch of predictions which means it's gonna sort of there's gonna be some average that nets out to, right? >> Yep.

Yep. >> Okay. But then secondly, what if you what if you made another control here or I guess it would be a variant, but like essentially it's a control. What if you just regressed every client's retention rate to whatever your average client retention rate is, you know, and you just said like I don't know if we you float your client retention rate is probably really good, right? So >> yeah, I mean we have our net revenue retention rate is like 95%. >> Right.

So why not just assume that for every client and I bet you on average across the board it's >> yeah it's generally useful and it but specifically useless because the problem is next month my revenue is going to be in reality some variation of the renewals that are present in January and I have hiring decisions and all sorts of things to make that I need to be uh more right than the average. But what I'm saying is, are you sure that that wouldn't Well, of course you're not sure, but like what if actually just regressing to the to the mean on this is actually more right than the executive doing it? >> Yeah, it's possible.

I think the question of like the mean over what time and how much data I have is >> Yeah. >> is hard. >> Yeah. But I but that's but my point is if Yeah. If you just sort of didn't if you didn't allow any subjective element of it, if you just did entire >> there's no way like because when you say no subjective element, defining the time range in which I select the average is a subjective decision. Uh >> you're going to have an arbitrary end point for sure. >> Yeah.

And so it's like really hard and the sample sets are pretty small. >> What if you what if you ran simulations? What if you what if you back tested this? >> Well, that's part of what we're trying to do is when we create a new modality, we go back and then try and use it to determine. again because you're progressively polling the issue. The question is like when was the poll right or wrong, >> right? >> Um now I think the most like the most important windows are probably like within 90 days of the actual decision, >> but you're going to be in smaller samples then too. >> Yeah.

And you're you can also have wild volatility in those 90 days like it where one day it could be 90 and the next week it could drop to 70 and then towards the end it drops to 30 and it's like well >> maybe the 30 was right because you learned along the way that they were upset but then 90 was the problem because on day 90 you were kind of going okay cool everything's great. There's also like a deep question about why do people stay or go >> uh that like is very hard to extract from each circumstance. >> Yeah. >> It's multivariant and >> Yeah.

And of course Yeah. It's not going to be the same for everybody. And so I I do think there's this there's this concern around trying to measure it that becomes part of the problem. This is what I this is no different than what I watch people do with marketing measurement or forecasting which is you can spend too much time trying to be right and not enough time spent solving the problems. Um, and so I think I am trying to be careful that we don't get in this like >> because that's a little chasing, >> right?

That's a little what of what I hear, which is like you don't I'm not sure you actually have a problem, but you have a new tool that does a really cool thing. >> And by the way, and by the way, I do think sometimes it's also worth playing with a tool like that. You know, it's like it's like it's a good muscle to exercise to deploy a tool like that. Your podcast is cool. Like it's like >> Yeah. So, but you you did have a problem, you think? >> Yeah.

Well, I know I did. we're off course to our retention goal for 2026 so far. >> Okay. >> Now, um >> the other the other thing too though is that like again it surfaces these problems that I think will also allow us to evaluate the uh the system with more clarity because I think the feedback loops presently were a little little light. So yeah, more to come. But but my main thing with all of this is that AI has this experience of creating a ton of information that becomes hard if you then have to make decisions with that more information. >> Yeah.

It's the difference between like even like analyzing a meta account. It's like everyone likes the idea that's like oh like look at the last seven days check your dashboard and make a decision and then you just sort of go like okay or like analyze infinity things you know like and there's somewhere the the information doesn't cross over into becoming like oh this makes it easier to decide things uh >> and there's like as you've pointed out before I actually don't know what you ended up with you were going to do a study internally of when your media buyers and when your team makes changes does it >> is does it help or hurt did you ever get an answer to that >> um I don't think we've ever constructed it well enough. >> It would be extremely hard test to figure out. >> Counterfactual is always >> Olivia and I were having this conversation. >> Olivia Cory from House. >> Yeah, sorry.

Olivia Cory from House about why she's like, "Why do you think there's not an agentic media buyer yet?" >> Um, and I think the answer is because there's not a single endpoint to optimize for. And so it becomes really hard to figure out what does a good media buyer do. >> Yeah. >> Um because it's different in almost every business case. Um unlike hedge funds where it's like generate the highest return on invested capital possible at the largest scale possible.

It's like the measure to compare two different systems is obvious. It's like >> in this year did you generate what return on capital? >> Yeah. >> Like you can't do that in metamedia. Like so it's just very difficult to validate that some thing is better than another which is why the whole thing is such a industry like ripe for guru. >> Yeah. Yeah. That's right. Yeah. That that makes sense. And this is actually another thing I've been thinking about too is uh recently this is not on our list.

Random show episode by the way. Who knows where we're going to go right now. Could >> go anywhere. >> I have a little theory right now that I mean I've said for a long time that one of the problems with with you know the specific example I typically use is creative testing in a meta ads account is that people are actually not really testing anything. No. And they're not learning anything there. So it's just it's just an endless >> like stream of no knowledge being accumulated.

And and what I mean is there's no discipline or rigor around the accumulation of the knowledge. So you're not actually you're not actually banking any learning. Yeah. >> And I was thinking about this recently with intelligence tests as well. >> I think incrementality test you can put into the same bucket. >> Yeah. That's right. And and there's a couple reasons for that. But one of the things I've been thinking about is just that like all of us are so bad methodologically.

So even even if it's like you >> are committed to the idea that I want to have some rigor like for I'll just speak for myself. I actually lack the technical data skills to test in some way that is meth methodologically reliable and and the reason I think about this is I know that in social sciences in particular and and even beyond that in academic uh literature right people have talked for a long time about the replicability the restoration crisis right the idea that like all of these results that come through randomized controlled trials right to the gold standard like you like you've said about right it's the gold standard of of measurement and yet you go back and and paper after paper after paper is not replicable That's right. >> And and there's all kinds of reasons for that related to statistical reasons and the way you set up, you know, and just human bias and all this stuff.

And so then it becomes this strange thing where where all these people are doing these tests. I'm going to have Drew Marone on from Intelligjam soon. And I'm just going to ask about this because like first of all, am I crazy? Is this actually a problem or actually like are you directionally correct enough that it's okay? It's okay to not be good. you know, like if your CRO test wins by 20%, like I don't know, it probably won't hold at 20%, but maybe it holds at five and it's it's and you you pile up a bunch of 5% wins or whatever.

Or is it actually possibly just leading people on a wild goose chase, you know, like of like nothingness, you know? I I'm becoming a bit nihilistic about it. >> Well, there's a huge operationalizing the ideology problem here. And so I think about this uh and but I actually do think that there's there's really good ways to wrestle with this and it's no different than what do you do with Cal Raleigh's home run forecast next year?

Okay. So if you want to explain that to people who did. >> So so in baseball you mentioned this idea earlier about regressing to the mean which like um when you forecast things you look at outlier performances and you don't assume the persistence of the outlier performance >> unless it unless it persists for a super long time. >> That's right. That's right. The sample size has to get a lot bigger the more of an outlier it is. >> That's exactly right. >> I think.

Yeah. >> But then again I'm running up against my technical skills. >> That's right. But but I think you generally I bet if you were to bet on what fang graphs forecast Cal Raleigh home runs in 2026 it wouldn't be 60. No, which is what he hit last year, which is for those who are not baseball fans, Cal Raleigh is a catcher. Had 60 home runs last year. Only >> home runs in the history of baseball by a catcher. >> Only a few 60 home run seasons ever in baseball.

So, it's a huge number is the point. He had a lot of home runs. >> And I promise you, no system will forecast him for 60 home runs. >> We could probably check. >> Yeah. Yeah. Pull it up while we're doing this and I'll keep talking. >> Okay. >> So, this same question exists when you run a incrementality study. Um, so let's use something where I think there is actually a fairly large database. There you go. 40. That's what I would have guessed. >> The I would have too the highest >> the highest number across the forecasting systems is 44.

There's like then and this is by the way there's the thing that uh Taylor and I are looking at here is from Fangraphs. There's like seven different forecasts here. >> Yep. >> Um >> they range from 38 to 44. >> 37. The lowest number is 37. 37 to 44 and he hit 60 last year. That's right. And he had 34 and 30 and 30 the two years before. Yeah. So anyway, the point is it it pulled him back by by 30% or something like that.

So this is why this is like we struggle a lot in trying to get brands to interact with incrementality tests uh in in this way which is you get a single test result back. >> Okay. What do you do with the single data point? >> Yeah. >> And my suggestion would be you do not pivot the entirety of your measurement to that single data point. >> Yeah. >> Okay. Um >> my response is you try to replicate it. >> Exactly. So one you try to replicate it but two is what we have is like an index of the mean like an aggregate test.

And if you don't have one because you're not running tests all the time, use house publishes a bunch of test results for you to YouTube and meta. So on meta, I think the number that they have is like hundreds and hundreds. >> This is so funny. Sorry, keep going. >> So the benchmark is 1.2 on a 7-day click. Okay. So 120%. >> So the point is what your 7-day click rorowass is in meta. Multiply that by 1.2. >> Yes. >> Which uh which >> is almost your 28 day click. >> Some of us who have been advocating for 28 day click for a long time. >> It's probably going to be really close. extremely close to 28 day clicks.

That's right. That's right. Yeah. And and that's that's from hundreds maybe thousands of meta in >> I think I think they tell you the exact amount. I think it's like 400 or so. Um >> okay. >> But the point is this is an aggregate mean. Now what I'll tell you is that our benchmark factor prior to them publishing the study was exactly the same. >> Y >> so what I hear there is like replication. Yeah. >> Across a fairly large sample set.

So that that >> essentially those are two separate what do they call those when you when you analyze a bunch of tests? Uh a meta analysis. >> Yeah. Yeah. >> Yeah. It's essentially separate meta analyses confirming the same thing. Now I feel very good about that number. >> That's right. So my confidence level in that read goes up. Okay. So when we start all clients we have uh incrementality factors in stat list. It starts at 1.2 on a 7day click.

Now when they get a test result back okay >> um if it is close to that bar that is like a high confidence that this is consistent with historical data. If it's an extreme outlier. So, we just got one back the other day that's like um 100 or like 212%. >> Yeah. >> We are not going to move the factor right >> to 2.12. >> That's right. >> Now, the debate we're having and the thing we're trying to work through is what does one what does one test results weight against the mean. >> Yeah. >> Um and that question of how far should we move it >> I think is relates to what does AM look like? >> What is the actual contribution health?

What's the risk of over optimizing or overspending relative to that read, >> right? If you if you're a super high LTV brand, you might be willing to to go push a little harder, right? Um yeah, we we talk about this with clients all the time, which is like uh in fact just yesterday I was talking about with a client, which is we have this question with our media buyers, which is which side are you willing to miss on? So we set our targets relative to the brand that wants to be a little more conservative, follow the Ander, never grow method. >> That's not true.

I believe in growth. >> Yeah, >> hang out. I I do have a different client where they went year-over-year went backwards in revenue by like 60%. And I thought on putting that on a tagline, >> you guys still rebranding by the way? >> Uh probably at some point. Yeah, it's just not the highest priority. Um the uh we thought on putting on a tagline like you know this client we we brought this client down 60%. You're you're higher higher age of growth for that.

That client was also drastically mismanaged financially before. So it's a it's an outlier scenario. But anyway, so but but your point is right, which is which is like you shouldn't go take that. And and to your question, your very specific question, which is how much should you wait that one test? There must be a statistical answer to this. Well, there is there must be a methodological answer. >> Yeah. And I also think there's business factors.

There's both that can exist. Now, um the other thing I'll say though, what's harder is if it's really bad. Um >> Yeah. >> And and actually the other thing that becomes even harder after that is if you get tests that are very disperate in the outcome. >> Yeah. Oh, I had that with a client recently. >> This is actually what Olivia and George from uh Cozy Earth and I are doing a podcast about this because uh Cozy Earth has been on a very long journey in partnership with us in house to to get to um a a read and their reads are all over the place.

Variable and bad all the time. >> All the time. >> But they're bad. >> Bad always bad >> in every channel. >> Yeah. >> All the time. And >> but they're making all this money. >> They are they are doing well as a business. Their efficiency is not what the incrementality is reporting. Yeah. >> Um, so there's a whole dialogue here around what do you do with a bunch of disparate information? We had a call today with a customer who we've gotten two reads that were very different and he's and his response is I'm done with this.

Yeah. >> I'm just going back to Chad scaling and feeling the vibes and because but that's why people people struggle with this so much >> to to handle disperate singular data points. My very first thought as you mentioned this whole whole thing is that the problem and I've had the same conversations with I've had a client who's saying why am I paying this bill for an incrementality firm that we we're doing. I don't even you guys don't talk you know part of it is we need to do a better job surfacing the results to the client and explaining some of those things but I'm like I'm pretty confident you made your annual contract back with this group in January during your lowest month because of how we handled the spend because of some information we got from those from from those tests.

But the thing is this this particular client um is runs a really good business. Like I I don't want to sound disparaging at all, but he he he would tell you he is not a first and foremost data thinker. >> Yeah. >> And and what this requires you to do is be is to learn and he he's willing to learn. He's humble about it. It's not he's not arrogantly, you know, whatever. He he wants the data. It's just that I would say it's not his instinct the way it's mine or yours or something like that, right?

Where it's like we we came from. So, is there a worldview almost like predisposition? Yeah. Wow. >> This is what I'm saying to where you and I approach everything this way. Like you and I are talking about this stuff >> like you know like we've talked about right this you pulled me in from fantasy baseball. I pulled in as stratey baseball. The whole point of that is take numbers and turn them into insights. >> Yeah. Well, but the only thing I'll say is that I reject the idea that everyone isn't actually behaving this way.

Um which is to say >> this is this is a great point. So you were actually pushing me on this uh recently in a text conversation which which is just like the idea that you're you're starting with some allocation decision. That's right. And they're trusting >> and it's based on some number somewhere. >> Exactly. And it's usually trusted on like the meta number that that I just go like why did you assign confidence to this number? >> Yes.

Yes. >> On what premise under what re of information? And so we we have this way that there are certain things that just get anchored Yeah. >> into our frameworks that are like now everything has to move that with really great evidence even that even though that itself has no evidence. >> This is a hobby horse of mine, but it's why you ought to be extremely careful about which dashboards you put in front of yourself. >> That's right. >> It's it's why uh attribution software is something I've still never really embraced even though I think theoretically it could be useful.

The concern I have is that it's it's so dangerous to put numbers like particularly MTAs where I think there's just really limited usefulness if any and and that the problem is is is actually like >> uh like my like my friend a forementioned drive line baseball their their coaches talk about the problem of negative transfer which is when you go and master a skill that's bad for you. Yes. Um, and so when you go do that, you end up you end up getting really good at something >> that is you actually got better at the thing, but the thing doesn't move you towards the outcome that you want.

And there's a negative transfer problem with data because when you put you're like, "Okay, great. I've got this data." But if the data is actually not leading towards the result you want, it's a real problem. And and so yeah, so for me, like this this is uh this is a real thing. And to your point, I think you're right. Everybody is actually making decisions based on this. Ultimately, your bank account is a data point, you know?

Um, and it's all data. The question is like which data do you prefer? And I do wish people would would Yeah, maybe maybe you or or I or somebody needs to make more of this case more often. This might be a drum worth. You're everybody's data driven. I even have this conversation with baseball fan friends who are like you data nerds or whatever. It's like I don't know. Do you watch a game and see one score one team score more runs than the other?

That's a that's a data point >> out. Yeah. >> Yeah. Right. Like so we're all data driven you know >> and I think that uh trying to help people get the idea that we're on a journey to a I use this phrase a lot progressive truth but it it is there is some knowledge that we have today that we believe about the measurement that we're trying to make better over time and there are there are moments that will push us closer to what is really happening and then there will be things that push us further away and we will contend with that forever and we will try to hone in and the best case scenario would be that we replicated 10 tests in a in the channel forever and they were all exactly the same and our confidence would go really high. >> That's likely not what's going to happen.

And to me, like the most honest thing to do would be in many ways to present people a range of outcomes of results, but nobody would want that. Imagine a dashboard that's like your row is somewhere between 1.7 and 2.5. Nobody would want that. >> Maybe everybody thinks in a data sense, but but I think the idea of probabilistic forecasting is extremely is extreme. Like the idea of like, oh, this is 75% likely to happen, >> right? you know, is like it's really hard to get your head around because actually one and four is a happens a lot, you know.

So the the alternative case happens fairly often, you know. So it's really it's really tricky. I think the thing that I I'm solidifying in my mind as you're talking as we're working through this right now is the notion that what you really want is to is to figure out which data points actually lead you to better decisions. And that that's the whole thing that that was the revolution of baseball, right? It's not like there was data in baseball for keep putting stats on the back of a baseball card for 70 years before, you know, way before Moneyball came out.

The point is, are those data points going to lead in the case baseball teams to make better decisions about how to put together a team? >> That's right. >> And the answer was sometimes yes, sometimes no. And so, and this is the same thing, right? It's like which data points are you putting in front of you because though what because you should be careful because you want the ones that are going to lead to um to better decision-m and that's actually a really hard job to to figure that out. >> I saw somebody on X today saying that they were having two agencies compete against each other.

Uh, and it was gonna grant one of them the ad account. And I was like, "Okay, what's the measure?" And he said, "Roorass versus ROS." And I just went, "Oh." Like, you are literally going to force two people. >> Think you're being so smart, >> right? And you are just literally incentivizing them to do something that will be harmful to you long term and you're going to anoint them the winner on their basis of the ability to do that. >> And so that's a good example of like the wrong >> how should they do that? >> To have them compete against one another, I think it's incredibly challenging.

But but what you could do, I think, is give them two campaigns. um on a similar set of uh products that have the same exclusions right and have a cost control and they have to drive up value or they could run it without it I suppose. Yeah. Um >> you'd have to be really confident in the exclusions that there's no audience overlap like there's lots of challenges because meta will assign last click value to a thing like so this idea of testing two things against each other is almost like never a useful idea.

Actually your boy Barry I call him your boy because he's regular recurring guest. >> Yeah. And he's great. uh he has this thing that I think is right which is that the evaluation of creative at the ad level is so flawed because your first click could be on ad A and purchase could be on ad B and you assign all the credits ad B and ignore the interaction with ad A. Um and Meta doesn't care about that, right? So >> um I think this is the problem with competing campaigns. >> Yeah, I think Jordan Manard talks about this too.

The idea that there's like an ecosystem in meta account >> in the ad account and you and so you basically ought not to evaluate any given ad. It's why like spend is the metric for me. As long as there's not a super crazy rowass thing like which 99.8% of the time there's not. You're building an e-commerce business and you need customer service software to do it. So you should be using Rich Panel for that. There's a very simple value proposition with Rich Panel from day one, which is that uh Rich Panel guarantees that they will save you 30% on your customer service software bill if you switch from Gorgeous or Zenesk over to Rich Panel. from from the beginning.

That's just something you ought to be considering. And you should know at the same time that Rich Panel is not just more affordable. It is great software. Rich Panel is [music] built from the ground up with AI. They're not adding AI to a legacy product. It was built from the ground up with AI. So it's very AI fluent. And obviously customer service is one of those areas in your business where AI can make a really big impact really really fast.

And so having Rich Panel do that is is uh is a real value ad. And that's the reason why very large brands like Ridge are using Ridge Panel as their customer service software and uh and brands as big as that and also brands that are starting from zero. It's the brand that's software company I plan to use for my customer service software when my brand starts and launches from the beginning. They can scale with your brand over that whole time.

They on average see a 30% reduction in tickets, customer service tickets because their customer service self-help platform is so good. They have an automated they have an AI agent who will respond to your social comments for you. go interact with customers on your ads. Uh fully fully AI, fully agent uh and does a great job. Again, very large brands are using that. So, there's also potential savings there using AI to handle that kind of customer service work uh instead of having to employ a team of people all over and trying to stay up on top of things and it's just really awesome.

So, customers also love it. Customers give really positive reviews on their interactions with Rich Panel, including AI interactions. They also guarantee a twoe transition so that when you are switching from one software to another, it's not going to be a huge headache. They're gonna get your team trained and ready to go. Go check it out today. rich panel.com. Get started with Rich Panel today. Okay, let me take this conversation to the banking knowledge makes me think of something I put on our list today.

Y >> um it's sort of a different thing. Amazing story in the weekly dose of optimism uh newsletter this last week and I I want to walk through this and tell you kind of what has made me think as an organization. And I think I'm influenced by you here a lot. So I think you'll have good things to say about this. So first of all, everybody go subscribe. Py McCormack, formerly Dan McCormack, but he's just made create too much of a monster.

So you guys go easy. brother. >> Yes. Yeah. But so Dan and Py are brothers. But Dan, founder of create creatine uh brand. Uh it's just he he's got too good of a business. So he had to stop writing the newsletter. But PY took it back over who's an investor and uh all the stuff. Not Boring Capital. PY is awesome. You introduced me to him originally and his newsletter is great. Weekly dose of optimism. Five optimistic stories every week about how the world's getting better, usually with a tech and business kind of lens through it.

Okay. Um and it's it's just great. But uh they there was a story about the founder of uh GitLab. I think his name is Sid. GitLab is uh he's a billionaire. It was a tech company that was founded. I think GitLab is a collaboration tool for coders. Okay. Is is is I think what it is. And um and you know there's there's a lot of interesting things about um about GitLab and about SID. But basically what happens is is that uh first of all gitlab was studied really aggressively by people uh after covid because they had been fully remote forever despite like a 2500 person workforce working on coding together and you know he was a y cominator guy like everybody he said told him you got to be in an office because you got to share the knowledge and he was like no there's other ways to share knowledge and I don't I'm not convinced that that's the best way to do it and their tool allowed them to share knowledge really really well.

So, they did a few things which is basically uh they have a 3,000page handbook internally and they have a YouTube channel and I it might be private, I don't know, it might not be. It has 12,000 videos. It's basically every meeting in the company. It's all recorded and the little tagline is no piece of information is too small to record. And so, they just document everything. And then the internal culture is when you have a question, go to the handbook because that's the first place to go for everything because you just document absolutely everything and and you go and then of course everybody's collaborating on that all the time, right?

Um, and so that centralized knowledge becomes this critical thing and they as a company and Sid as a leader is going like this is the most precious thing we have is this set of knowledge that we're all collecting and that we're all building around. Okay, so there's two fascinating things about the story. The first is Sid got cancer when he was pretty young. Now he's a billionaire. He first went into remission and then it came back and doctors said it's over.

You're done. We can't do anything about it. And Sid was like okay well I don't believe you. So, he goes and gets goes down this route of of you know, he's a billionaire, employs a full-time team to help him beat his cancer. Um, they all come around this and they start collecting all of this knowledge and through a whole bunch of stuff and the exact same methodology I just described about his company, they they go and they put together the knowledge and he's he's back in remission.

So, it worked and they they did some really specific therapeutic for him, you know, like reading all these papers and whatever. So, he he was able to do all this thing. Incredible story. And as much as first of all that can sound like okay rich guy gets better healthcare than the rest of us like that's really annoying. Um the comparison the article makes and actually PY links out to a fuller article about this is really worth everybody's time.

I'll put a link to the weekly dose in the in the show notes but um uh the the comparison he makes is Magic Johnson and HIV where he says like you know ma like HIV is undetectable in Magic Johnson now. like it basically he got way better healthcare than anybody else but actually that's really good for everybody else because it sort of allowed people to develop ways of of handling this that then they got to the masses and now HIV is a much more treatable disease than it used to be right so um so that's the idea is that it sort of it starts with him but then that knowledge is banked and now like that other people can use that and can learn from it and doctors can learn from it and all this kind of stuff so so there's that okay but what has really made me think about is like so so first of all crazy like maybe we can cure our own cancer now with the knowledge and and of course AI's a part of the story and all that stuff too.

Um so amazing amazing story first of all. Uh and he's like in his 40s by the way. Um secondly though it has made me think a lot about the idea of a playbook which is a term that we talk about a lot. Um and it comes back to this thing you said which is like this banked set of incrementality studies. Okay. Uh and and what I've thought is I have hundreds of podcasts I've recorded. You have countless hours of content you've recorded.

I have real solidified beliefs about a lot of things in e-commerce at this point. You know you've made jokes about growth. I don't actually think that brands shouldn't grow, but I think some brands shouldn't grow. And I actually have some categories in my mind when when you asked me this question that I'm probably not as clear about as I could be. And I bet if I sat and did the hard work of writing out the ideas, it would force me into behavior of of getting clear about that and then having different sets of advice for for clients about that.

And I also bet other people in my company have different ideas about that and it could be worked out, you know. And so what I've started to think about is maybe as AJF growth grows, we're still very small, right? We're hiring a second growth strategist right now. Like still very small. So all that maybe I should massively disproportionately put time into maintaining a centralized document of knowledge as best I can especially because now AI can uh query that information much easier than than we've ever been able to before.

So it's the AI the information is much more transmissible than previously. And then I should take that and of course it doesn't need to be all my knowledge. I'll tell you what I did the other day. You you released a really good podcast episode that I thought I'd listen to. What did I do? I went and put that in a knowledge bank somewhere. I wrote my own thoughts on it in a notion card so that I had like here's what I really liked about this and so I could bank that thought.

I'm not exactly sure what I'm going to do with it yet, but it's made me think maybe there is just this really disproportionate impact of just doing the slow, steady, detailed work over a very long period of time of just like because you know you don't build a 3,000 page handbook overnight, right? So it's like just doing that over and over. So I'm just curious how that story hits you and what you think because I think you guys have been very very good at this.

Well, yeah. I've said a lot that that idea that like operationalizing ideology is the value of an agency, right? My fear and this goes back to this conversation that happened on X where somebody was like, I uploaded the meta gems like the meta gem uh model breakdown and asked it to analyze ads through this lens. And I was like, it has that in its database. >> Yeah. >> I guess like my question is what do you have? You took you just talked about taking a publicly available piece of content and putting it into a separate repository. >> Yeah. that's now in every LLM in the world. >> Yes. >> My fear is that the surface area of things that don't exist in every LLM is like basically there is no knowledge or playbook that doesn't exist already in there because all like we I like to think I have a lot of original ideas but mainly they're an extrapolation of some previously existing idea too in some way. >> But do I mean don't here's my here's my response to that.

It's that it's that uh the the information is all there but it always has been there and actually it's been fairly accessible in a lot of ways for a very long time. >> Y >> it's now it's easier to do that. >> But the question is can you access it and query and know when to apply it? >> Yes. >> So that to me is the thing you have to illustrate. >> Yes. >> With the database. >> Yes. But but also there's point of view, right?

Like let's just take this is way too broad of I mean I don't know. Let's take some debated cost controls versus not. Okay. Easy one. Okay. The arguments on both sides of that have been are in every LLM like you said. Okay. Whatever anti, you know, whatever auto bid guy thinks versus whatever hands thinks. Yep. >> Those two are accessible to the LLM, right? >> Um, so I could go and ask LLMs right now what's the best argument in favor of each of these.

But I actually am trying to not that's not what I'm really after. What I'm actually after is the point of view. What I want is is somebody to arbitrate that debate. And the LLM could do that. Okay, it could, but I'm not sure I believe it. Actually, this is one of the things I think is one of the biggest challenges with LLM is it's it's almost that negative transfer problem again. It speaks very confidently about things that it's actually not that confident about.

It's just sort of uh predictive. Well, I do it all the time. Yeah. [laughter] So, um but but you know what? I actually really try not to I Yeah, I I I think the question is just Yeah, it's a question of trust. Do they trust you more than that? Well, but my point is I'm building AJF growth and I want to say like I'm willing to put my my flag on the ground on some things that I believe that are contrary to some things that other people believe and say it's you know again I've heard you talk about the CTC at some point you said you must behave the way I'm telling you to behave now that the LLM had that information but the people you know or does now anyway the the people though would behave in all kinds of ways so the handbook would be like go how do you go run a bidcap versus a TRO or whatever AJF growth there should be documented thing in a handbook somewhere about how to do that and somewhere maybe Andre Lunv who likes bid caps also disagrees with me about some detail of this but you know what I'm saying this is the thing that I believe >> a really good thing about why cost caps aren't uh the the less conservative version than big caps today so we at some point >> yeah let's yeah but but you see what I'm saying I I don't just want the information I want the point of view that's what I'm saying >> I I get it I think I could ask I think I could ask X your point of view on vidcaps and I think I could get it I don't think it's actually locked in your thing um the the the this is why I think at the end of the day what it boils down to is the accountability. >> It's that I think the only human >> I don't think this is wrong. >> Yeah.

I just so I think the problem with the handbook thing because it is so tempting. I even like Brian Brian Armstrong of Coinbase just did this thing the other day >> where he talked about how they've like done this. They've internalized every thing that's ever been written in the history of Coinbase and now he gets like advice as a CEO about what to do and it like it like tells him what to do and like it like I'm intrigued by that idea.

I think the question is what do I have service area that um that that it doesn't now there there are some things I think our our data um I don't know how much publicly available P&L level e-commerce data there is like you guys have a lot >> there's some of it right >> we have this cool lead genen idea I'm actually I don't know if maybe I'll ask you to cut this out I'll ask afterwards but that someone just shared with us which is um imagine a lead genen tool where you come to the landing page and you write the problem you're facing it as a And what it does is queries all the history of CTC and goes >> there are 412 conversations about this issue inside of CTC over the last 5 years. >> Yeah.

Right. >> Get on a call with somebody to discuss our experience in those conversations. >> So what it gives credence to is >> it cites those conversations for the person on the call. >> Right. Yeah. >> And so the idea is like everyone says case studies or we have experience and it's like well what if we just showed you oh yeah this brand was facing that exact issue. Here's what we did that didn't work. here's what we did that did work and that's why we'll do this for you.

Yeah. Um and so I think there is this way in which you can build and that's just trust development, right? Like I think that's a reason to trust my point of view on the issue. Yeah. >> Um so I think there's there's ways there's something there about like the institutional knowledge >> is is important. Um but I've had people like I don't know if you've experienced this. I have had e-commerce owners send me transcripts where they ask the LLM to act like Taylor. >> Yeah. >> And what would I do? >> Yeah.

How do It does it. >> Yeah. >> Like in ways that I I don't actually know what the counterfactual what I would have said if they I'm so now I see it. But >> but I'm like dang one I feel like almost like >> like they stole my soul like they took me and put me in the thing. Now I'm what use am I? But so so but it's weird that they're like you can do that. You as an e-commerce owner could go ask Chad GBT say I want you to act like Taylor and I want you to give me a recommendation on this ad account. >> Yeah.

Right. Yeah. Well I think there's a there's a hidden benefit potentially. I mean what you're possibly what you're saying is that there's our knowledge is useless in general [laughter] in this world, right? Which as a beat as a drummer I've heard you beat before basically just you know the surface area of what AI knows versus what humans know >> that kind of deal. Uh so it's possible that's that what you're saying is sort of the negative end result of this which is that there's no point in organizing the knowledge and information in some good way because it's already there you know. >> Well the go back to your cancer example >> he operationalized it to his own benefit.

Like so I do think that this idea that like >> you're going to take the output and you're going to go do it in the ad account. You're actually going to do it >> and you're going to be on the hook for it >> is the thing. That's the like you have to go and be the person that's like, "Okay, I'm going to go change the bid cap based on this advice and I'm going to if it fails, it's my fault." >> Yeah. >> But I the I think I really think there's sort of another thing here, which is, you know, I you talked about operationalized knowledge for an agency.

I I love that phrase. I actually think about it all the time, especially as I've really, you know, fully committed on growing an agency is like that is what the agency is. It's operationalized knowledge. And both sides of that equation are really important and they're both really hard. You need the knowledge part and you need the operations part. And the the thing can break on both. I actually think some agencies are bad because they're bad at both, right?

They they actually don't have any point of view that's clarified and also everybody inside the agency is doing whatever they want. Some agencies are bad because they actually do have point of view or somebody in the organization has point of view but they haven't figured out how to how to get it out to the rest of the organization. And one of my thoughts though is like, okay, if I believe operationalized knowledge is what an agency can do, then before I can operationalize knowledge, I have to organize knowledge.

I have to actually have a way of of creating that. And the and the and the downstream benefit of that, the sneaky benefit, I think, is that is something I've said a lot of times, which is that writing is an act of thinking. And so, forcing myself into the behavior of writing out ideas. I mean, it's one of the values of the podcast to me. It's not writing, but it does the same thing. It makes me work through an idea because I have to put out content.

And so um and so yeah, again I try not to do that when I do that I don't really believe or something like that. But it's just something about that act of talking about or writing it is is a benefit. And I kind of wonder if the GitLab thing like if the sneaky thing is actually just that you've got a lot of people in an organization trying to be forced into behaviors that makes them think more clearly. You know just that that act of thinking clearly is really good.

And I think one of the dangers of AI is how much it will it will end run around that process. Right? It it it you don't need to think that clearly if AI can do it for you. again if we end up in the sort of super cynical version of the scenario that's fine and the AI just does the work but like but yeah um anyway so it's just something I've been thinking about I it is nobody actually has a playbook but what if you did you know what if you had all the way down to when >> isn't that what Prometheus isn't isn't it your playbook >> yeah for sure we and and but the thing about Prometheus which is our AI corriding tool and really goes way beyond that at this point it's our operational tool ultimately >> um is that it has one surface area where we spent a disproportionate amount of time which is which is our approach to ad creative um and messaging uh and it needs to get better still there but uh it doesn't have all of our knowledge yet you know it does and so like yeah >> yeah like status is ours is like I think everywhere everybody tries to do >> yeah but I don't think everybody tries to [laughter] right >> I think they should and the other thing I've thought about this is I sort of wonder like in the case of Sid exercising those muscles in the business created a skill set that he could take into the rest of his life and this gets into something else we don't have to go down this road because you've got some other things on the list.

But um I've sort of become obsessed with the idea that one of the reasons to hold yourself to really high standards in business and to high accountability in business and to pursue great work, you know, it creates financial benefit and all that kind of stuff, but as I've said a lot of times, that's not the most interesting thing to me. But this idea that like it forces you into behavior the kinds of excellence in your behaviors that are transmissible and useful in all kinds of parts of life, right?

I'm I'm sure there's things that you've learned at CTC that have helped you be a better dad, you know, and and the way you think about what how you how your kids have formed and a better baseball coach and a better wherever whatever communities you want to contribute value to, you know. So, um, so that's the other thing I think about this is I think it would be good for me to, you know, to do this just because it would exercise good muscles, you know. >> Yeah.

I think in all cases you are, I think, passionate about a pursuit of truth. Um, >> that's right. >> I think Yeah. like get the closest alignment to what is real and what you think that you can create the better. Yeah. >> So yeah. >> Um so I think that that feels to me like an expression of that which is that there is now >> definitely >> that's what he wants to know can you cure this cancer and there's now way more information ever than so he's going to >> try to organize the process to getting into alignment with what is true about his cancer.

That's right. That's right. >> And he's like, and so that pursuit, you're right, was shored up in his company. And so he applied it against that and it worked because somebody wasn't willing to actually acknowledge that there was more to be known than what they knew. >> Yeah. >> Yeah. Or at least what they said was that the standard of care was had been had been exhausted, you know, and so it's not that of course I'm sure no doctors told them you could never possibly be cured, right?

They just said we're not we can't do anything more for you. What we have is not enough for you. >> Speaking of exercising muscles, >> yeah, we're here looking a little sweaty and ragged. >> Yeah. Do you want to talk? Let's Let's do that. That's a fun one. We only have so much time left. >> My wife dragged me into Hy Rocks. I'm not here to do the thing where you just talk about your fitness thing all the time, but I am fascinated by the business model um of each. >> You want to just just I think people mostly know Hyrox, but just in case. >> Hi Rocks is a fitness race where you uh run eight kilometers and in between each kilometer do some sort of activity from rowing to wall balls to lunges to burpees.

All movements that I would say almost anyone can do. Yeah, there's no big barrier. In in CrossFit, I think there was a few movements, not all of them. Yeah. >> Snatches. Yeah. >> Uh muscle-ups that are like pretty prohibitive for people to participate in. All the movements on High Rocks >> are uh >> without having to scale them. >> That's right. >> You can actually do that. >> And the and the weights are nothing intrusive.

And so you most people can participate. And uh I got to do my first race in Anaheim and it is insane the like >> velocity of energy and participation in these things and um and >> sorry do you mean that the business is growing really fast or do you mean the experience on the field of Hierrox was extremely high velocity fun both like so these races they sell out the second that they open up they're global they're all over the world they have them for kids they have a professional elite competition division down to like you can do relays you can do couples you can And >> is it always the same race? >> Always the exact same race. >> At least up to this point, I think there's so much service area for them to introduce little variations, but the one strength or whatever.

Yeah, >> it's more like running a half marathon in that way because you get to try to improve your own >> race time all the time. And so I think that the continuity of it, this is one of So this is one of the things of the business model for me versus CrossFit. The actual the entire thesis seemed like ready for anything like it was about multimodality >> that that you wouldn't even know what you were going to do. Yeah, >> before you go and do it, which I may be willing to concede would make you a better athlete, but in terms of inviting a larger tent of people into participation, I think is actually in many ways prohibitive.

And so there's something about the consistency of repetition that like um like marathons and half marathons allows someone to see their own progression in a clear way. >> Um and and they're crushing it. And I'm fascinated by live events. I'm fascinated by the business of the apparel and the ticket sales and like how big this is going to be. Um, but it is super interesting. >> Well, there's one big problem with what you just laid out, which is that CrossFit got gigantically huge at at one point.

So, like it was able to do >> the business of CrossFit. >> It was very big. I mean, I don't know what Hyrox's, you know, uh, total revenue is or whatever, but but CrossFit was like the biggest fastest growing fitness thing ever or whatever, you know. It was like, >> but the sport was >> No, no, no. I mean I mean it was like the it was the most gyms put everywhere. >> But wasn't the problem that the gym business model was like a weird franchise model that they didn't actually get to participate in that revenue like at all? >> The gyms. >> Yeah. >> Yeah.

No, the gyms are totally franchised affiliate deal. You basically paid I think three grand a year to to have a to have the name CrossFit, >> right? Um so like that feels like inherent like what's interesting about Hy Rocks is they actually are trying to take nothing from the gym owners. They allow all the gym owners. Like today you and I went to a Hy Rocks class. Moxy has nothing to do with Hyroxy. >> Yeah. >> And so, >> but CrossFit CrossFit, I mean, I guess besides the affiliate fee, it was a pretty novel or it was a pretty I mean, not enough.

It was it was a pretty small affiliate fee all told, you know, it was a few grand a year. >> So, part of what I'm saying is that was their monetization mechanism though, one of it is crappy. >> Yeah. >> And so, it's like I look at that like, >> yeah, I don't I don't know if that's where they made all their money, but the other they make all their money. >> I think they made a lot of money on the L1, which was which was how coaches got trained.

Um, so you you go sign up for an L1 that so your L1 is your level one coaching certificate. And the thing in in peak CrossFit was that uh lots of people got their L1 who were never planning on coaching anybody. It was just like if you wanted to go to CrossFit uh CrossFit camp for a weekend or CrossFit College for a weekend or whatever. That was part of it. And then the games. Now I now I actually have always wondered about this question which is like because CrossFit was a very big and very profitable business for a while.

Um, but I don't I don't really understand where they made their money if it was all affiliate. >> That's how I feel too is that because like even this like so this says estimates have been around 100 million per year. There's some estimates that it got into multihundreds. There's some that say that 82 million to 24 was its peak reported revenue. It's not public. >> That's about that's about what I would have thought was about 100.

Yeah. >> But but there was a $4 billion industry that surrounded it. And this is what I saw was like the value capture out of CrossFit taken by cor CrossFit corporate was crappy. Yes. >> Relative to the thing that they built as a movie. >> Whereas Hyrox, if you want to do a Hyrox race, they capture all of the revenue of a Hyros race. It is their event. >> Yeah. >> Do you know what their revenue is? >> I think I saw it was like 140 million.

They blew past it. >> Yeah. Again, I don't we don't know. Let me see this. >> And and CrossFit did what everybody expected, which at some point it flamed out. You know, everybody was concerned that it was sort of a cult or a fat or whatever it is. >> 130 to 140 million in 2025. And it's I mean it's gonna be way up in 26, >> right? Because nobody, you know, does this like it's like and it's global. It started in Europe actually.

I think it began in Austria. Yeah. >> Um >> and it just like the approachability my wife the idea that she would ever do a a >> like a public crossfit competition would be such a far from where she was, but like she crushed it like she and she's going to do another one and like and so there's just I think the approachability is so much higher now. Will it fade out? Will there be a new trend? I'm sure. But it seems like there's a lot of surface area to adapt the race a little, come up with variations.

My kids are interested in trying. They've done two superstars races, one in the Netherlands, one somewhere else. Um, so it it seems like there's still some runway to go here that I would not be surprised if this sells for a billion dollars. Like it seems to me that there is real real business capture around this idea. The other thing it led me to is part of the like so Puma's the official sponsor. Puma does all the apparel.

I think the Hierrex should have just done their own apparel. they probably screwed that up, but um there's probably a lot of money involved there and maybe they have a rev share on that on that so the distribution helps. I don't know. But I do think that a brands events might be a a huge revenue item because I think one of the things that brands need to think about is how they get access to a different part of my P&L.

Um, so like I think that we spend a lot on entertainment and travel and things like that. And I think brands could do well to think more in the way that, you know, Apple does, which is that like rather than just thinking about product expansion through more if you're ridge trying to get me to buy a backpack and a wallet and a battery case, which I get, like that's fine, but I still I spend way more money in entertainment and travel that if Ridge was able to find some way in which there was an event that fit in their universe >> Yeah. >> that feels like there could be something there.

Yeah, you know, >> a while back I saw Cody Ploffer from Jones Road tweet this quote, "Since deploying this tool, AMER is up around 50%. New customer rorowass is up close to 60%." And somewhere in there, he referenced that that was happening because he finally got his exclusions to work in meta ads in a way that they had not worked before. His his ads were finally reaching new customers, not just going to existing customers, no matter how much he was excluding [music] those existing customers other ways.

The thing is, if you've ever tried to exclude existing customers from your meta ads prospecting campaigns, you know that it is basically impossible. For at least for many brands, that is the case. You exclude your clavio lists. You exclude your pixel audiences. You exclude Shopify audiences. You do all the things that you're supposed to do. And yet, your spend still [music] reaches existing customers when you want it to be reaching new customers.

I know because I was having it happen with my own brand. Uh, not with my own brand, with a brand that's a client of mine. I was watching what happen with that brand. So I DM'd Cody after I saw his tweet and I said tell me about this tool and he said the tool is called waste not and wast was actually solving the problem for him. So we started using wastist not for a brand of mine. That brand was seeing about 35% of the purchases that were driven from the ad spend that we were uh theoretically using as prospecting spend.

Those 35% of those purchases were on existing customers. That number dropped to 3%. And by the way Wastistnot took 10 minutes to install and required no coding or anything like that. It was incredibly fast, incredibly easy, and just completely wiped out the problem. Really like nothing I've ever seen before. Uh really incredible tool. And it's not just me, and it's not just Cody. It's also a bunch of other big brands in the space.

Cozy Earth, Portland leather goods, other really large brands who are using Wastistnot to actually get their customer exclusions to work. If you are having that problem with your brand, okay, and you need to exclude existing customers in a way that will actually work, you need to go to wastenot.io.not.io. The link for that in the show notes. Sign up for a free trial. Mention my name. Mention the Interference Podcast during your onboarding call.

They'll give you a free extended trial so you can try it for yourself for a little bit longer. Make sure it holds. It's really no risk to you at all to get on there and see how that impacts your existing customer spend. It has been proven out to me and to others in the space. It works. Go to wast.io. Go check it out. Tell them I sent you. I agree with that point. I was going to say quickly the what actually reminds me of the Spartan race. >> Yeah. which which is like tough motor is where it's like an event based thing where it's a really specific thing.

You could train really specifically. You can be as serious or not as you want. Although you probably have to be a little more serious about high rocks spart race you can go kind of walk the whole thing and have fun. Um and I don't know if anybody's doing that with high rocks but >> there's there's some big people working their way through it slowly but but yeah not walking. No. >> Yeah. Yeah. That's what reminds me and that event based thing I think is good.

Although it's funny to think about brands being involved with it. And I'll be curious because you and I had the experience of trying to do stuff with Spartan Race when we were at Kao after the success of CrossFit. >> Yeah. >> And what the what I think I realized about that is that it never morphed into a community in the same way that CrossFit did. And that was that. And so as a as a vehicle for brand engagement, it was dramatically worse.

It was it was it was not good. >> This is because again, this is the point though. I think CrossFit outsourced the value to all of us. Like we actually captured more of the value. Not specifically Kao, but I'm saying the brands, the the gyms, the people versus Spartan Race, I think, held it closer to themselves. They're like realizing more of the value. >> Yeah. And I think just Yeah. So, but it's interesting to think about how Yeah. >> I I I think I like your application of the principle, which is like could Ridge create the Ridge event of some >> Well, an easier example is like Exclad cooking classes.

Like I think about like, you know, like paint with Pino or whatever that stupid thing that became a thing where you go, you have some wine and you do some painting. Uh-huh. >> Like, why shouldn't Hexcloud create the best date night experience in every strip mall, every high-end strip mall in America? >> Yeah. What is the answer? Why shouldn't I mean, why don't they? >> Because it's not intuitive. I don't And they're not good at it.

Like, it's not a natural extension of what they're doing. But I think that like brands that have this kind of ethos, we're going to see it more and more that like Restoration Hardware did this and like, have you been to the Restoration Hardware restaurant in Fashion Island? No. >> So, they have hotels and and restaurants now. You go to Fashion Island, they have a four-story store and the top floor is the restaurant. >> Yeah.

Um, and so the idea is like have a glass of wine. >> Nordstrom's in Nordstrom Cafe. >> Yeah. >> Um, I I think that the idea is that like part of this is like my thing about e-commerce is that it's just not growing fast enough. >> Yeah. >> The TAM is actually horrible in almost every >> still 20% of all online sales growing lots of more brands playing for that 20%. >> And and lots of categories where the dem the supply side of competition is growing way faster than 20%.

And so there's just this like not that big pie. Yeah. um that we're fighting over all the time. And so you've got to get potentially into some other revenue streams when you get that big. >> Yeah. And it's and I think I think something you're really good about pointing out and I' I've tried to think about this at times for other brands is like for e for e-commerce brands, there's nothing that says you can't ever do something else, but you just don't, like you said, you don't think about it. like it just doesn't occur to you that you could that you could come up with some other way of engaging with customers and sort of seed a new business within within it or you know what I mean now I think that one of the things is a lot of brands don't have the capital to go like >> well for some reason example is banana ball >> like I look at that and I go like okay where does that fit in the in the spectrum of brand consumer brands to me it's got to be one of the best >> yo in the world and from a merch and experience it like it's able to touch.

Now they're launching a youth sports league. Have you seen this? >> So it's like to me that's a business that I go. >> Do you think it'll work? >> I don't know. I think that they will drive in the short term a ton of interest and sign up and participation. I wouldn't bet against Jesse. He seems to really understand something about what people want from the experience >> and he's so committed to it. >> Yes. And Yeah. Exactly.

And he's he holds an seemingly incredibly high standard for the things that they do. Yeah. >> Um, so I go like, well, maybe that's actually closer to an illustration of what a brand >> should be able to do, which is touch the surface area of all these different parts of uh, people's lives in a way that isn't so isolated to consumer product. You know, >> it reminds me a lot of your interview a long time ago with the guys from Home Life Media.

Is that what they're called? >> Yeah. Yeah. Yeah. when when you talked when they you did an episode a long time ago about how they were selling their email list to non-competitive pet brands because they're a pet brand so to non-competitive pet pet brands like dog food brands or something like that which there's a game they're just like we're never going to play the dog food game so but we have a big list of people who bought shirt shirts with dogs on it so why don't we sell our email list >> and you know who does that >> who >> me and you like >> we we are a media we have media revenue just like we have service revenue just like we have software revenue Right.

Like and so >> and I do think about how do I grow my media revenue and how do I grow my service revenue and which which revenue you know how does that affect my business and how do they all reinforce each other and and you know this actually a good point the hexcloud thing theoretically well here's another reason I think people don't think about this is that it it's looks so small your hexcloud class is going to be 10 people in a room >> and but but and I think this is why everybody underestimates events the 10 people in that room >> the experience they'll tell >> they they have incredible and they probably love cooking that's the reason they showed up that Right.

And so they they are the kind of people who you really want buying HCloud pans and having a HCloud experience. You know, >> because you know what's actually kind of weird is to order pans, take pictures of them, and post them on Instagram. That's actually kind of weird. >> But you know what people post pictures of Instagram all the time? Their experiences, what they're doing with their wife, when they're traveling with their family, when they're with their kids.

And so there's this this element too of like if you give them something in that realm that I think is like >> super interesting to to go after. The reason it reminded me of the home life media thing is just the idea that there are assets inside your company that you're not thinking about that are not traditional e-commerce assets. And so maybe it's not an event for everybody, but maybe there's some way to look inside your business and go what are the other touch points with our customers we could make.

And I think for some people this muscle gets exercised with product creation where it's like the way you expand TAM is just by create is like you know and bridge is a good example of this where it's like a charger a power pack or whatever that they have right doesn't actually have any organic relationship to a wallet and it's just you know that okay we can make both of these or a ring doesn't have much but they but they can then expand their TAM and create and so there's there's a similar muscle being exercised there is what I'm saying.

Yeah. you're going wallets are gonna die like as if or or they're not gonna die. They're just not going to be the thing that gets us for forever. So, how do we push the whole thing forward, you know, and uh without while thinking outside of that? I like that thought a lot. There's a um so George and I co because you talk a lot because they're sort of at this size where they're trying to find these things and I think about like for them there's this demographic of these like um you know wealthy uh women in their 35 to 50 year old range that are show their core demo and my wife's in this demo and like one of the things that dominates their reality is book club like um >> so this is a space that women like they get together they have a great meal they do all this stuff and but it's unoccupied by brands.

Like they're reading a book, but there's nobody. And so I was like, George, you guys have enough gravitas that you should get the author. >> Yeah. >> To host a night in the city and go to all the book clubs and have them all come together. And like you could find this way into these spaces that Cozy Earth is creating this connection. It's a they they happen in homes. They're like the demographic. And so it's like find because they like they go after book talk a lot with content.

I was like get a podcast. do do a cozy earth uh reads like cozy reads or whatever you whatever it wants. Like there's there's this angle here that the problem is the women just don't need to buy sheets every day and they don't need to engage with your brand every day around the idea of sheets. Like they just don't. Um so what what where do you find space there? I think is just >> this I think fits nicely. It actually comes back to you're making fun of me about being the anti-growth guy.

I I actually think in some ways you and I are completely on the same page here. um in the sense that it's something I've been thinking about a lot and I again another drum I've heard you beat a lot uh with recently and I've come to really I think it fits almost all of the great brands I've been around which is that the great brands there so there's a small outlier set of brands who are tactics monsters they are just incredible at building funnels forever you know and and just doing that but as I've said a lot of times I think what people and those people get a lot of attention.

What I think the problem is like it's actually a very hard skill set to replicate and people don't realize that and that the reason that those standout examples exist is because they are actually 95th percentile plus at this and therefore you are not because that's what 95th percentile means. So, so you should probably not try to do that. Instead, I think there's and so if you put those those kinds of brands aside, then there's this way a brand can think about growing, which is how do I create moments?

And you you've been talking about this a bunch and a bunch of content of yours, which is like, okay, the agency can't create the moment for you. The agency can execute day-to-day plans and other things, but the agency can't. So, okay. So, how do you as a brand, you're cozy with examples? What made me think of this? You've got all these women buying sheets or whatever, but like there's these women there and they value something and you have some kind of brand identity.

And so, what kind of swing can you take as a brand to go, we can do something really big and it's going to be really hard to see that pay off on day one of the revenue or whatever it is. If you take that swing over and over and over again, as much as your budget and your time will allow you to do within some level of reason, you know, um to to go do this, that's actually the way you make deep impressions with customers that you really care about and that really love your brand.

You do that enough times, eventually you're going to be able to do some some really far. And by the way, some of those swings are going to be home runs. At some point, you'll do one of those. You'll launch a product or you'll go to a book club or whatever and you'll and you and more people will sign up for it than you realize and it will end up creating a ground swell for the brand. And I I do think in the world of tactics and the the other thing about this is I actually think almost everybody who's a brand owner or operator watching or listening to this right now would rather spend their time doing that than working on tactics.

It's drastically more fun and interesting as work. You probably actually care about your customer and your product and those things. And so it's just much more interesting work. It's creative. It's interesting and all those things. So you'll have a lot better time at work if you behave this way. >> Yeah, I agree. And this is this is where I think the brand owner that is like can't get themselves out of the ad account despite having lots of smart partners around them that that are going to run it in some modicum close to their level of output like they're just really problematic for their own business because I just think that there are so many people that can replicate that skill at a fairly comparable level even if you the founder are so smart and amazing.

Like you could hire Andrew, you could hire me, you could hire lots of freelancers that we've worked with before. Like there's lots of people out there that could run your ad account and they'll be in pretty good hands, but there's nobody there's nobody that will go forward out into the world and manufacture that novel idea with a deep intimate understanding of your customer like you should. And I think that it just is like it's a responsibility on it.

And I I have felt this like right now at CTC that I think what happens is you have these periods of a tailwind where where there's this like macro industry thing that pushes you forward that you are just riding. You're just like you have your sale up and you get to move forward with it. And so you don't have to really behave this way and then at some point the wind kind of dies down and you're now obligated to recreating some of that wind.

And for me like the the the era of like focus on profitability, reduce waste and cost like we dominated in that era. And for the last two years, I think a lot of brands have gotten the message. They've made the cuts. And there's still a lot of waste out there, but it's less. They're they're on the other side of like, okay, I stopped being bad. Now what? Now, now where do I go now that I cut the bad spend? Because it's a lot easier to cut your bad spend than it is to introduce a bunch of new good spend.

Um, and so I think the question is beginning to rise like, okay, where do what do we do now? And that question is much harder and requires you might need to go actually change the underlying reality of the business. Why I think we're ex actually on the same page here. What I'm saying is never remove that constraint, right? Don't just don't fall prey to the thing that I feel when I run an ad account, which is like I don't know, turn that T target a little bit lower because we're not getting enough spend.

No, keep the T target where it is. That's the constraint. If you want to spend more money, you have to do something more awesome. That's what you have to do. You and and otherwise be content with your 15% profit at your whatever size business and take the money because at least that point you're distributing. you know, at least at that point, like if you're not if you're growing 20% and you have that kind of, you know, 10% >> the only thing Yeah.

And because I ag I know what you're saying, but what I think you're saying is actually your business isn't that awesome. And and this is the part that like you you like to whitewash this a little bit with this idea of like I use that in the Huck Finn painting the dirty fence metaphor here is this idea that like it's okay if you're growing 15% at 3 million. And what I would say is if you're growing 15% at three million, there's actually something here that isn't that awesome.

Yeah. And that that that there might be actually an opportunity here for you to step back and go, why do I have to grind out 15%. What is this market saying back to me about this moment that that if I were to set that aside and go, okay, this framework has 15%. What's a frame breaker? What's something that would unlock it in and not in a wasteful ad spend way, but in a way that sort of changed the underlying reality of my unit economics?

Or this goes back to the slick products thing. I think about slick a lot. >> I think it all the time >> that like we tried to get that person >> Yeah. >> to change the core composition of the product to say like, "Hey, man, heavy wash bottles that they don't ever run out of is never going to work." And you when I say never going to work, you're going to grind out percentage points a year and the margin's going to be like we should have tried to create something in a very different >> form.

Yeah. Yeah. Well, and and I I don't disagree with that and I I mean I think the the thing I push back on is like if that $3 million business is actually producing $300,000 in cash or something like that, you know, which is possible like you know, like if you're growing slowly and you're running at a very big margin, I would just say that's a still a really good business. It's just that you're it depends on what terms you're thinking of that in like what you're talking about is one of the richest people who ever walked the earth and >> in which sense >> if you're spending out $300,000 in cash in a year at that size business and I I just >> Yeah.

I I don't know. I I don't that framing to me like we don't think about ourselves relative to like >> Yeah. Yes. >> human 100 BC. >> Exactly. Right. And that's the problem. >> No, it's not. It's not our lived experience. We should have no comparison to what >> I just I completely disagree. I It's the It's the old Louis Cape. Everything is awesome and everybody's miserable. That's a problem with us. I actually do think we should >> You're not You can't ask someone to have this experience.

You don't You don't have any idea of what it was like to be alive. >> I know. But I I I know. But >> so the fact that you could delude yourself into thinking I'm so much happier than they were. How do you >> I'm not saying I'm so much happier. I'm saying if I can't be happy with what I have now, there's a problem with me. >> But it it has nothing to do with whether somebody was happy or not a hundred,000 years ago. >> No, the the point is we are we are the richest people ever.

Ever. >> But so what? So, so the point is if that's true, if the percentage of all of the richest people ever are is very very small and it exists right now, >> it's a message and I can't experience joy in that right now. Okay. Then something is wrong with me. >> That's what I think. >> I I think there's so many people that struggle making that amount of money that are going through like genuine hardship and their life of course >> and the business is sucky and they're dealing with all sorts of >> So what you just said is they have a bunch of hardship in their life.

Not that they have >> in their business. in their business. >> Yes, if if their business is bad, but what I'm saying is if if it's producing $300,000 a year at that stage, then their business, I'm saying, is not bad. >> The level of satisfaction I've watched in periods where people are everything is working and everyone is happy is so different than when everything is struggling and everything is hard. Like the idea that rapid growth and success doesn't yield a better experience is just like not inconsistent with my lived reality.

Like >> Yeah. I mean, I think it's more fun >> for sure. >> For sure. It's like every in every way, shape, and form, winning feels better. >> Yes, I agree. I agree that it feels better to win. But I'm saying 300,000 if you can't see making $300,000 in a year as a win, if that's the example we're using, then then there may be something wrong, >> but there's just this this idea that like every So where's the threshold? $187,000. >> I don't know.

That's probably that's still pretty good. Yeah, I think if you're if you're spitting out that much cash now, and this is why I'm saying I'm not saying you shouldn't be ambitious, and I'm not saying you shouldn't pursue more growth from there. I'm not saying that at all. I'm just saying that there's got to be there's got to be a way. There has to be a way to be that wealthy, which is insanely wealthy. >> Absolutely not.

Like, what what do you mean? Making $187,000 >> makes you one of the richest people alive right now. >> Yeah. But in again, the relative comparison is not someone's live reality. >> Exactly. >> It might not even make you the wealthiest person like on your street. >> Yes. So if you if so if you spend all your time comparing yourself to the people on your street, it's going to be a problem. >> So you should just pick the worst person and compare yourself to them >> partly.

That's this is this is what everybody who has ever been to a really hard place in the world says >> that every then everybody but the worst off person in the world gets to the >> Of course not. Of course not. Well, you're taking over sort of a ridiculous version of it. What I'm saying is anytime anytime you ever heard somebody go to a really poor area in the world, what do they come back saying? They come back saying like, "Holy cow, I have a whole new >> and it lasts them like seven seconds." >> Yes.

And that's why they have to go do it again at some point. And that's why they have >> it's not real. It's not that's not actually what experienced satisfaction is. Is a relative comparison to someone who has it bad. Like that's not >> No. The point is the point is that that more wealth is not necessarily the answer to how you improve your experience. That's the point. It's so because because you're right. People are experienc People are plenty wealthy and experiencing all kinds of hardship.

I I completely agree with that. What I'm saying is money is not the problem. Is not the solution. >> I just think you you led this off with a story of a man who's a billionaire solved his cancer issues. Like I I think the idea and this is >> I'm not saying it's not it doesn't massive utility for incredible housewives. I'm not saying that >> it absolutely can and it does being wealthier will in almost every time of life lead to higher levels of opportunity across any dimension that you could possibly pursue. >> Yeah.

Well, I don't think I'm going to convince you of this, but I I just I just think that the idea of the word bad business >> Yeah. >> is by definition a comparative term. And it and and so what you're doing in your mind is comparing >> business business actually has like an objective standard of success. >> Yeah. And and you but not really >> there are bad baseball teams. >> Yeah. >> The Royals are a bad baseball team if if there's worse baseball teams than them. >> But but a bad but you see even that that's because you're picking a very specific set of businesses to compare yourself to. >> Okay.

So then compare yourself to the local coffee shop. >> Okay. But like Sure. >> Right. Then then that business producing $300,000 a year is a really good business. >> I just think that's a wildly bad comparison. But I but it's completely as arbitrary as your comparison. >> No, it's not because it's not even in the same industry in which I operate. If there are >> But you just said get out of your industry. You just get out.

Forget it. Forget you don't have to be an e-commerce brand anymore. Now you're an inventor. >> Exactly. >> Don't. >> Yeah. >> My point is that if you're looking around and there are people and you're going like, "Wait a second. Why does it seem like it's so much easier for them?" >> So, why are they making so much more money? Why is their P&L 30% profitable when mine is eight? >> Yes. that there's wisdom in like that's a really important question agree completely >> and instead of going like no eight is good you should be you should you should be happy at eight >> I just go no no no no you should actually deal with the objective reality that there's some inefficiency in your able your ability to drive leverage between the labor that you're outputting and the reward that you're receiving >> that person should live in unhappiness as you're telling me >> they should be dissatisfied with the business and try to solve the business problem when you talk about personal satisfaction I think these things get blurred a lot for you it's like it's the idea that you should live in a perpetual state of happiness and no to and joy or whatever the whatever the phrase is that you would use and so in which case none of this matters at all because it's independent circumstance entirely. >> I I I think you are you are tapping into something right now that I think is important for people to get in their mind in relation >> and you've heard a million people tell the story of when I was on my deathbed I realized >> I don't think the experience on your deathbed is actually a good indication of your live life at all.

This is a thing I think because you can't occupy the disposition of dying person. Um like this is such a madeup idea. These are tropes that I think are perpetuated on this idea that like live like you were dying. The the the literal chemical composition of your reality changes when you're dying. >> Yeah. >> Like you can't inject that into somebody Monday afternoon and ask them to live as if it's your last hour on your deathbed.

It's just like insanity. >> Yeah. I I think having a stronger sense of the reality of your impending death would be really good for you. That's why I think not just talking about you. I'm saying for me for I can I can accept entirely logically and deal with the fact that I feel anxious right now cuz I got to get home and take my kids to to baseball at 4 and that's just as important and it's actually more true for my present reality than it is my future impending death.

And so like I the idea that I might be on my deathbed and go like man I look back in those moments and whatever I think I feel then if I were to impose upon myself today the idea that that feeling is more true than my feeling today I just don't believe that. >> I just think they all get a voice in the conversation. I I think >> but they don't today because I don't feel the dying thing like I I only feel presently embodied the the reality of what is happening to me today.

But the the the the point is to me that you ought to be you ought to attend to how you feel and what you feel and what you compare yourself to and where your joy is coming from and all those things including in your daily business decisions because ultimately you're spending your time and your life and your wheels and these things. So listen, I completely agree that if your $3 million business is growing at 15% per year in an e-commerce kind of world, you probably ought to ask the question, is this the best use of my time and capital? >> Yeah.

And you would experience deeper satisfaction if you were inside of something working better. I maybe there's a I I'm I'm not quite willing to grant that though I understand your point. It does feel great to win. And so I don't know. I I probably think about that more if I think that's true. I'm not it's one of those things you said earlier like, you know, do you say something that's not true? I'm not sure what I think about that. >> But I and I do think you should learn from people who are growing more uh and all those things.

I just what I really think is that you shouldn't make bad financial decisions because you're because you're uh feeling insecure about how big your business is. That's that's kind of what I and that's what I think >> that just sounds like bad strategy but see but I think I agree >> but I think this is part of the whole thing. I mean I do agree with that. Now