AI Is Replacing Content Creators (Here’s How to Profit) w/ Sam | Ep. #172

Did you know that people are making money with 11 labs pretty passively?

I have zero doubt. Tell me about it. I have videos right now that are indistinguishable from human. You might as well monetize it.

Facebook is probably paying them six figures.

I need an AI army of faces that I trust to like actually go and say exactly what I need them to say when I need them to say it. So I can go make these ads generating leads without my face. Or without a real human needing to shoot content. It's a no brainer.

Listen, get rich. We all want Silver bullet.

Yeah, everyone wants to get rich, dude. Let's get rich together.

Welcome to the Kerner office. I'm back with Sam today. My good friend. He's been on about 12 or 15 times and we caught up today about this hummingbird nectar e commerce business that he launched in public. He's got thousands of dollars in sales. The first few weeks we didn't get to that to the end but it's really, really interesting. We talked about video generation, we talked about hey gen content creation, making money by having your Voice cloned with 11 labs and then licensing out your voice. Of course. We talked about Facebook ads, billboard marketing, comparing the two lots of gold nuggets in here. It's going to be a two parter. So come back a week from today for part two and share with a friend. Did you know that people are making money with 11 labs pretty passively?

I have zero doubt. Tell me about it.

They're like basically licensing out the rights to their voice. Sick. They're like uploading their voice and saying this can be one of the voices that 11 Labs users choose from. They pay out like a proportional rev.

Share of that actually.

Isn't that cool?

Yeah. So I actually noticed when I was picking voices yesterday I was like these feel user generated. Like they didn't all follow like the 11 Labs default voice vibe. It was like it felt like actual everyday people that were like uploading it. I didn't know they had a monetization model on the back. Yeah, that's awesome. Honestly like and you don't have to have a killer voice to do that. I spend so much time on 11 labs trying to find non professional sounding voices because like I don't always want studio quality, like deep. I don't want Morgan Freeman reading my ads. Like I want Sally from North Carolina in there like to run.

One of the commenters was saying like I have a really stupid voice or I have a thick accent. Like I'm an Indian American with a thick accent. Wouldn't work for me. And the original poster was like, this is the best for you. Like, you have like an outlier voice.

Yeah. A unique voice. Like, that's. People want it. It's a no brainer.

So for those listening, you want to make a few hundred bucks a month, go clone your voice on 11 labs.

I might do that. Like, get this voice all over the Internet.

I want to do it, but I'm kind of scared, like, are people going to copy my content even more perfectly?

True. If they had my voice true, I would. I'd be over here with an AI version of you in two minutes. Dude, to be honest, it's not going to stop them, dude. They just download all your podcasts, upload them all into 11 labs and clone anyway.

Make a few box prizes.

Yeah, you know what I mean?

That's the frustrating thing about that guy on Instagram that copies me is I can't file any copyright strikes because it's a different language. Like the algorithm doesn't care. It doesn't see it as the same thing at all.

Yeah, that's fair. But if you post it first in that language, then you can absolutely. I see what you're going. That's easy.

Have you heard of cling AI K L I N G?

That's the backend video engine. The 2.0 version, most expensive, takes the longest but by far the best. You can get similar results from 1.6, but you have to be way better at prompting versus 2.0. You can just like, let it ride, dude, it's so nice.

Okay, so, like, I still want to get to hummingbirds, but there's like three things I want to talk about before that.

That's fine, dude. It's fine. It's all. It's all about YouTube retention, baby. It's cool. We're going to talk about hummingbirds at the end.

Everyone's dying to hear about hummingbirds.

They're like, dude, what are they saying? Dude, Hummingbirds.

The weakest retention hack ever.

Hummingbirds.

It's this Facebook account that posts 100% AI generated videos. They post like 5 to 10 a day and it's. I swear, it's just some click farm out of China. You could tell they're trying to look westernized. All right, Jacob and friends, this popped up in my feed and I'm not trying to offend any of our older listeners, but this is Boomer bait. I'm calling this Boomer. Okay? So look at these. Look at these videos. You can tell before even clicking that they're.

I know what's good. Yep.

Millions of views. I'm going to click on this one. That has. Okay, here's one. Oh, geez. The tiger got hit in the face by the chainsaw like four times.

But it's totally fine.

Oh, my gosh. Like, okay, it's AI right. It is AI, but, like, right? It's so realistic. Is it? Or am I just.

There's another one. Oh, he's got one with the snow leopard.

Which one?

Underneath the log, dude.

Right here.

Yeah, yeah, yeah. Oh, that one too.

Oh, this one has blood on its mouth, too. That guy looks real. His movements look real. Like this right here. This looks 100 real to me.

I know it does.

And then this, the one that. That popped up in my feed. It's always babies. There's always a baby in danger because they know that no one's going to swipe. Look at this. Look at all these tree on animal videos.

Tree on animal videos work, dude.

They're in tree on lion helicopter picking up giant turtle. Oh, my gosh. Okay, here we go. This is a baby floating in the water in a flood. And a German shepherd is in a car pulling the baby up out of the water. Okay, so. So here's how this works. I make you free videos. I actually know what I'm talking about. I have no greasy sales pitch at the end. And if you implement what I talk about, you'll make a lot more money and have a better life. And all I ask for in return is that you hit the subscribe button and maybe even the notification bell. Just like that. Thank you. What do we do with it? Is this just interesting? Can this be monetized? Like, these guys are monetized. Facebook is probably paying them 6.

And that's what it is. It's. I think that you're. You're limited in your monetization. Right. Because you're just getting paid for impressions, really, like, at the end of the day. But, like, that's not a bad business model. Like, if you can get to the scale of content, like let it ride.

You know how to make videos like these that are this long because.

Yeah, you just. You put multiple scenes playing video. Yeah. Scenes together.

Okay. You just keep adding on to it.

Yeah.

I just love that they found a format like, oh, yeah, a tiger under a tree is going to work. So let's do a lion under tree. Oh, a baby flood's going to work. Let's do a baby in a house fire, you know? Yeah, it's Just wild.

It works, dude. There's a bunch of people that are doing similar, like, micro movies. I'm trying to get to that spot with ads as well. Yeah, it's an interesting use case. Like, these just look like they're like impression farming. Right. Like they're just trying to get as many views as humanly possible. All of these boomers are sending it to their friend. Like, oh, my gosh, look at how real this AI looks. Right? Like, yeah, the whole thing. And they're getting paid for that. And that's sick.

Yeah.

That's all.

You see the comments. Oh, amazing. Wow. Thank you, dog, for saving the baby. Great work. Well done. Yeah, seriously, Someone stood by and filmed this. Oh, my gosh.

So funny, dude.

So what is this cloning business? How are you cloning people?

Basically, the problem that we've been running into is at scale, content creation, specifically for ads. Because, like, when I'm testing ads, I want 15 variations of hooks, of headlines, of all of these different pieces, which gets hard when you're dealing with real human beings. People are busy and it's just not on their priority list. And it slows me down from being able to generate ad creatives. And the ones that perform the best, like, we use statics, don't get me wrong. But, like, the ones that perform the best generally have some degree of a video element going on. And so I was like, all right, like, I need an AI army of faces that I trust to, like, actually go and say exactly what I need them to say when I need them to say it so I can go make these ads. And so this week I just decided that I'm just going to solve it. I was like, I'm just going to figure this thing out.

And when you say solve it, you mean get it from good to great.

Yeah. Like, actually where the hit rates. That's the issue. When you go on a. Hey, Gen out of the box and you, like, type in your script and whatever your hit rate on. I don't have a term for it, but it's like, there's like when you watch an AI generated video, there's things that, like, show you that it's AI generated. Right. Like there's little tunnels. Yeah. Like there's little pieces. Yeah. That just, like, they don't get the inflection on. On sentences. Right. Or like, one sentence doesn't flow into the next in the same way because it's all this AI stuff. And so it was just like, unreliable. And I decided that we're going to make it reliable at this point. And so we're getting pretty close. The stack is ChatGPT. To generate the image asset that you're going to automate. It also can do like, the script writing and stuff. Like that's its own piece. But like, we generate the image that we're going to use, we take that image, we pop it into photoshop. Because chat GPT likes to. For the, for the YouTube crowd that can see my face. ChatGPT likes to give you like this much of the frame. But on, on an ad creative, you want to be able to layer in text and other elements. And so you go to Photoshop and use generative AI, expand to create.

Fill in the gaps.

To create framing. Yeah. So that you have way more space. I found this other tool that does AI face swapping. I could take a photo of you and put it into ChatGPT and say, hey, you know, describe this person well enough so that you can replicate them in a photo, but they never get the face right. Like, it'll get close enough. Then you go into an AI face swapping tool that actually like mimics your face directly on that image. And it's just, it's way better.

We're just talking about still images right now.

Yeah, we'll get to the video part. So then you face swap. And so now you have an expanded frame with their face of them in. We've been doing a lot of like, podcast style. And so it's very much like them sitting in front of a microphone because all of the AI voices are studio quality. And so like when you're listening to like, what is a fake selfie video and it's podcast quality audio, it's a tell. It's one of those tells. And I'm like, all right, we can't mess with that. So it's all podcast right now. I'm gonna get to the spot where it's better and I can do the other one. I know how to do it. I just haven't gotten it yet. So then you take all of that, you go to one of our favorites, 11 labs, and what I realized is everybody's just being lazy with 11 labs because they are doing text to speech. Everyone's doing text to speech. Yeah, they're typing in their prompt or copy and pasting from ChatGPT and they're going text to speech, which is what causes those weird inflections and the unsmooth dialogue. And so what I'm doing is the voice cloning feature in 11 labs. I record myself saying the script.

Do you have to record yourself. Or can you just upload like a podcast?

If you talk, you could upload a podcast for sure. But for. For something as specific as an ad, where I have a specific pauses and like how I want that, right? Like you want your hook to be, like, very emphasized. And so like when you record that, you can control that better than in AI Uploading the podcast and stuff is going to be the voice cloning of myself, right? And so I'll have whoever is going to be the face of each of the things give me as much audio as possible so I can create the voice clone, but then also pop in there and record what I want them to say. So then I take the voice clone, which matches my inflection, so it's voice to voice, take that output, then you put that image, the face swapped expanded image into. Hey, Jin. With their new Avatar 4 model, which is.

Oh, is it a big leap?

Oh, it is way better.

Give me a scale. If 100% is a human talking Avatar 3, what percentage was that? What percentage?

Avatar 4, I relate that very much to like hit rate, right? And so like on Avatar 3, maybe 60, 60%. So 6 out of 10, Avatar 4, I have videos right now that are indistinguishable from human.

Okay, what are we doing? Show me. Come on. Okay. You think I can wait for this?

Dude, I'm so down, bro. I'm so down. So this is Ethan, our head of sales.

Okay.

At Service Scalers, but I'm gonna show you this one. Unlimited content rights and publishes daily SEO blog posts for your business. 100% done for you and optimized to rank for only $399 a month. We'll handle everything from keywords to publishing so you can grow traffic, get more leads, and build authority without lifting a finger.

Oh, I don't know if I'm being too critical. It feels like it went from not opening your mouth enough to opening the mouth too much.

That one, though, I've sent to people and they were like, wait, is this real? This one? And this is like, we're getting there. This is the Ethan one. Are you tired of sitting by the phone waiting for leads that never come? You work hard, you deliver a great service, but if no one's calling, your business can't grow, and referrals alone don't anymore. That's where we come in. And this is of a real person. Like, that's above Ethan. This is a fake human that just. Okay, ChatGPT came up with.

I feel like Ethan's Better. Is it because it's a real person or is it just a. I don't know.

Yours get consistent there. His messes up on this piece right here where it says high quality. Yeah, like yours get consistent, high quality leads.

Every single opens really wide.

Really wide. But it's like, I actually started running these ones as ads, okay. One of them actually, this one specifically, because I do a bunch of headline testing, right? And so I'm testing all of them. This one specifically is getting me $10 booked calls, which is unheard of.

What's the show rate?

I launched them yesterday, so I don't know yet. But generally our show rate's been like 60% across all of the ads that are both of me and others crazy. And like, this is 100% AI. This is a fake human being. The voice is fake. Like, the video is like the whole system.

I wonder if. If you had something. I don't. I can't tell if I'm staring at the mouth because that's where the tell is or if I normally do that. So.

So yeah, the other piece on this one, which might be helping is the subtitles because it pulls your. It pulls your attention away from the mouth a little bit. I found that to work even on Avatar 3. And so, like, that generally helps. It's getting there. Here's my barometer. They are converting on ads, right? Like, are they good enough to actually support the use case that I want to, which is generating leads without my face or without a real human needing to shoot content? And that answer is yes. And so, like, that's what's exciting to me. There's going to be more iterations of it for sure. But yeah, it's getting super close.

That's exciting. Like, oh, geez, it's frightening. It's exciting. I want to share my screen. There's like five or six guys that copy me on Instagram.

Yeah, I remember. I remember you telling me about it.

Yeah. One of them is, like, not good at all. One of them is. I'm going to show. Let's see. I'm just going to pick one at random. In Dubai, mini burgers in stylish little boxes with different sauces. It's convenient, delicious, and visually striking. This idea can easily be implemented in existence. His mouth is hardly open. Delivery services, food trucks or food court stalls, many. What do you think? Is that obvious to you?

Yeah, I think the other piece of it, though is like, we stare at AI generated content all day, trying to. And so, like, we pay attention to it. But I Don't think, obviously, like, it's working. Sure, you can nitpick it all we want and like, yeah, it will always continue to get better, but like, if people are liking it and commenting on it as if it's real, then I think it's passable.

Dude. This freaking guy right here, he takes. I'm just going to show you. You know what? Hold on one sec. Let me show you this. I go here, let me show you this. Here it is. All right. You spend half a million bucks on an ugly house that looks like this. When you can spend half a million bucks on. On a 3D printer that prints houses that look like this and then build all your neighbors houses as well. The future belongs to the owners. All right, so there's that. And then this dude, he's in Canada, but he speaks in Hindi. He takes 100% of my videos, translates them verbatim to Hindi, changes not a word, and recycles all of them.

Wow, that's sick, dude.

I hate it so much.

Someone like, you need to go launch so you can do that in 11 labs. The overdub. You should just launch. So. So Mr. Beast does that.

That's where I'm going with this. Just to wrap it up, there's this guy that's copying me in Hindi and I'm going to beat him to the bunch and publish everything in Hindi before he can. Yeah.

Using 11 labs.

I'm going to be big in Delhi, dude.

That's it.

My next T shirt is going to say the career office Big in Delhi. Big. And Hyderabad.

It's like that, you know that meme of the guy standing in like the corner of the room and he's like, none of these people know that I'm big on Twitter. It's like, none of know I'm famous in Hindi, dude.

I felt like that in Nashville. We went to Nashville as a family and this guy in a random boot shop was like, oh, my gosh, I follow you on Instagram. And I was like, super embarrassed, but flattered at the same time. And then for the rest of the trip, I was like, you guys don't even know this guy recognized me in a boot shop. Whatever.

That's so funny, dude.

The whole dream is to use this to just more video ads at scale.

Yeah, yeah. And I think there'll be a world where I'm using it specifically for the Facebook ads. Use case. Right. I think that there's a world where I can go and get to like, AI influencer level in a very real way. I Think a lot of people talk about it. I'm. Dude, I'm saying like I've been deep this week in like 15 different discords with all these nerds using all these different tools to achieve all sorts of stuff. I mean there's, there's even a few like I know we want to, we want to talk E comm a little bit. I'm getting there on UGC Creatives as well.

User generated content.

Yeah. For physical products I'm getting there. And so very similar flow but using a different tool.

So video is cool. But you know what's better? Long form audio via podcast and my newsletter, tkopod.com go there to subscribe for free to my newsletter. It's one email a week, very tactical. And then go to my audio podcast, three episodes a week. Stuff like this. You're going to love it. All free. No sleazy sales pitch. Tkopod. Com. Did you see what icon, the company icon that launched. They're all over Twitter. Icon. Com.

Yeah.

What do you think about that? So for those listening, Icon.com is this new startup. They raised $24 million. They spent half of it on the domain name Icon.com. half of people say their landing pages ugly and terrible and poorly optimized. The other half people say it's genius. The point of the matter is people are talking about them anyway. What icon does is it. They're just putting AI in the meta ads library. Like so you basically let AI run all of your ad campaigns so you can a B test thousands of times. You can iterate the moment that it sees something that needs to be iterated. What do you think about all this?

It's very interesting. I mean obviously like with creative os, like I'm in that world, I haven't played with them as much lately just because I've been trying to figure out more of the backend. It to me comes down to hit rate again, right? When I think about AI generated content, that is the metric that I'm paying attention to is you create 10 versions of something, what percentage of them is good enough where you trust that that output on a consistent basis is going to work? A lot of the out of the box tools, there's a bunch of them like I think make UGC AI. There's like a handful of people that are stepping into this space and trying to move it. And I think that the hit rate is the piece that causes pause for me on a bunch of those out of the box tools, which is why I've stayed away from them and gone to trying to build my own. I think that I need more control over it to limit the f ups. But for people that aren't as technical or just like, want to move faster and are okay with spending the money for 10 videos, knowing only two of them are going to be good, and knowing that you're getting charged a premium on all the models that they're using, all 90% of, you know, these tools are using the same video generation models on the back. And so for me, I just kind of decided to go straight to the source. And I was like, I just trust it direct. Which is the tool that I've actually been using more is called Flora, Fauna, AI.

Ooh.

Yeah, Very interesting. More control. Really like what they're doing. Following a very similar flow that what I was doing for the podcast clips.

Let me show you one more thing. So did I tell you my billboard test to grow the podcast in Houston? So I spent like all morning trying to see if that has actually been working or not, because it's been one month to the day that my billboard has been there. And so I'm going to share my screen and I'm going to show you these numbers and I want you to break it down for me. So.

All right.

Okay. So what you're looking at here is basically my last 30 episodes, not on a per day basis, but on a per episode basis. So episode 140, 141, yada yada, the blue line. This is basically when the billboard started. Okay. Now this y these yellow rows is when I was running Facebook ads with pictures of 30 different variations of the billboard to find the winning one in the same Houston area to people that did not follow me because the people that drive by aren't going to follow me.

Right, right, right.

And so here's what's very interesting about this. This number right here in column I. This is the percentage of people in the Houston area, the greater Houston area, that listen to my podcast before any Facebook ads, before anything, it was consistently one and a half percent. While I was running Facebook ads, it was about two and a half percent. Yeah, okay. But here's the kicker. I only ran ads for two days. The 27th, which was like right here and the 28th. And so most of the downloads came after those two days because people subscribe, but then they.

And then saw the next one. Yeah, right.

And so, so here's the thing. I spent $260 a day for two days to get these downloads. So basically I increased my downloads. These are the raw numbers an average of 74.

I mean, you almost. I was going to say you almost doubled. Not quite, but yeah.

So.

Oh, average was 74. I got you. I mean, that three seven one was. If you take out the outlier on three seven.

Yep. So call it like 50% growth in the Houston area by spending $250 a day.

Yeah.

Okay, now for the billboard in the Houston. It's just one billboard in like, kind of rural outskirts of Houston, but it gets 20,000 cars driving by a day. My average is now granted, my podcast as a whole has grown 30% and my podcast in the area have grown 60%. So basically double the growth rate. Double the growth rate. And so here I'm spending $20 a day for the billboard and getting almost as many new downloads.

That's sick.

So on one hand, like, a lot of the people on Twitter when I post about this were like, dude, just. Just use Facebook ads. It's like, people don't use Facebook ads to grow podcasts either. That's because no one's ever tried, which I doubt it. Or because they don't work. And so I used them like I had a link to the podcast. So I wasn't just getting data for the billboard, but I was leading people to the podcast. And it worked. But it worked 10x worse than the billboard does.

Right.

What do you think about that? Break it down for me.

Yeah, I vibe on it. I think the big piece on it, like, on the. Even the comment where, like, Facebook ads don't work for growing podcasts, I think that's more of like, a cash conversion cycle issue for most podcast creators than it is a. Like doesn't work. I hear that all the time. People love to say Facebook ads don't work. And it's like, no, no, no. Like, your cash conversion cycle on that media spend is actually what's making it not work. You can't afford to have a thousand dollars a day of cash sitting out there for 30 days since you don't have direct monetization on that subscriber. Like, you're paying, like, CAC to LTV payback, period. Game two. I think that that's awesome. The most interesting to me is like, okay, how do you scale that? How do you identify where those billboards need to be? Have you ever looked@blip billboards.com?

Yeah. Yep. Yeah, for like a day or like, yeah, yeah, for every hour.

Yeah, it's all the digital billboards. I haven't had a use case for it, but it's one of those like, you know me, I'm a big fan of obviously Facebook ads, but like non traditional marketing. Yeah, like non traditional advertising. Which, like, hilarious that we're saying that about billboards. That was like, historically one of the most traditional ways to advertise. I think that most people in the digital age, like, aren't thinking about it like that. And it again just comes down to, like, potentially slower growth, but also like, your lift per dollar spend is significantly better because you went through all of the steps to go and do that.

Here's what's really interesting is the billboard actually went live on the 14th right here. This is kind of an outlier right here. This green, but really no bump for the first month. Then my account manager slash friend David, he drove by it in person. He lived like an hour away. And he's like, we actually have a billboard 300 yards away. Same side of the street. Same street. But like your billboard, it appears like as you come around a corner and this other one, you've got like a full frontal of it for like 30 seconds. As you approach it, he's like, this one's a little more expensive, but if, you know, we'll move it for free and I'll keep the price the same because I think you're going to have a better experience. And I was like, do I have anything to lose? He's like, not really. Like, this is objectively a better billboard.

Yeah.

Like, okay, so on the 21st, right.

Here is when he switched to that one.

And then look at that. So for those just listening, like it went from the ratio, which higher is better in this case, went from 65% to 95 to 120%.

Yeah.

After moving at 300 yards down the road, same side of the road, do.

You think that that tapers off in terms of performance? If. And I think it depends on placement. Right. Like, there's nuance there of like you said 20,000 cars a day sick. What percentage of those cars are the same cars? Because, like, when I think about it from a. From a Facebook ads perspective. Right. Just because there's like very quantifiable metrics across Facebook ads.

Your repeat rate. Right?

Yeah. Frequency. Right. Like, okay, cool. Does that taper off over time where you kind of max out that audience? If it. Let's just say that it's like a major commuting highway with 20,000 cars going, do you only spend that money for three months?

That's how long my contract is. It's funny, you would say three months.

Yeah. And so the three Month piece to me, I go like, you probably hit kind of your max on people that are actually going to download it, right? Like, yeah.

I think there's two things at play. I think with any marketing, you would agree you need repetition. You've got to see that thing five to 20 times before you click. Right.

No doubt.

So that works in my favor.

Yep.

The longer it's there, the more people are likely to click. Another thing that works in my favor is the fact that every day a unique personal drive by that will never drive by again. Right. That will never change. Maybe it's 1% of the drivers, maybe 50%, I don't know. But then what works against me is the fact that probably 70 to 80% of those drivers are going to see that thing hundreds of times, and if they didn't click at time 10 or 20, they're never going to click. And, yeah, it's wasting ad spend on.

Them, which you're also dealing with finite amount of ad spend. So it's like, it's okay. It all comes back to, like, what's the math of it all? Right. Where it's like, you sat there and you were like, 20,000 people. X percentage of them are probably interested in business. Cool. Three months, they probably drive by it X number of times. With this conversion rate, it's a little less. The reason I always talk about Facebook ads is because you get such quantifiable data from it to help you make decisions. Something like that. You had to go and build all these spreadsheets and do a bunch of backend math to work your way into that. But the math is still the same. It's impressions, times, viewers, times unique times conversion rate. Right. Like, it's all the same metrics, the same stuff.

Can you see that?

Yeah, that's awesome.

That's the billboard. So that 1 out of 20 different. Really? All of the pictures look the same, but the phrase. And for those listening, it was get rich. That one won, like, and I tested it a number of different ways. So it just speaks to what we all want, which is like a silver bullet.

Yeah, everyone wants to get rich, dude.

Yeah.

That's why they're listening to us right now.

That's right. That's why we get rich together.

I know.

We're no different, right?

No, I just want to get rich.

I think the billboard accounts for, like, 30 extra downloads per episode. The question is, what is that going to look like once the billboard's down? How many, like, loyal lifetime listeners am I going to.

That's all. And That's. That's cact. Ltv. Like, you know, all of those pieces, it's probably a little bit more challenging to track that with that channel, but it doesn't mean that that channel doesn't work. It's similar. I mean, when we run ads, it's the same. Like, there's downriver stuff of. Of ltv. We close a client, what's the close rate? How long are they a client? What's their average spend per month? What's our, like, all of the numbers behind it all. I've just been telling people a lot more like, entrepreneurship is two things. To build, like, a very solid business, you have to have a really good understanding of people. Like, you have to be super empathetic both to, like, your customers, your employees. Right. Like, your vendors. Like, you need to understand people really well, and you have to understand math, tracking metrics, tracking the right metrics, understanding what levers you can pull to create results. Like, that's been the thing that I've been super focused on this year is like, all right, this is a math, like, business is a math equation. At the end of the day, like, how do we get the equation to work the way we need it to? It'll be interesting on that to see, like, how many loyal people come out of it for what price point, how you monetize them at. On an average of, like, was that ROI positive at the end of the day?

Yep.

Not to mention, I've noticed this too. Like, you get. Because of the. The media side, you get secondary ROI from doing that. Because your tweet about doing that also attracted people on. Right.

It's a flywheel.

Yeah. And so it's like, for some people, I think it can happen very much in a. In a vacuum. But, like, I've noticed that, like, even with the community, like, me tweeting about Hummingbirds, like, has caused people to join the community, which I'm monetizing. And so it's like, even though Hummingbirds isn't necessarily, like, super profitable, I'm still picking up secondary benefit on the other side. And so it's like, it's interesting when you play that game, especially with the media piece of, like, the personal brand is like, just do interesting things because you'll get the secondary benefit even if the thing independently isn't profitable.

Yeah, agreed. All right, what'd you think? Come back a week from today for part two conversation with Sam, and we'll see you next time.