Do you think AI avatars are good for blue-collar work advertising, such as top of funnel, middle of funnel, and product reviews, or will they hurt credibility?
Very interesting question. I think your issue is you are attempting to find leverage where it probably doesn't make sense. Why not just record your ads yourself? If you're going to go through the process of scripting them and stuff like that, just record a video of you saying it. You're trying to make your team seem larger. Actually, there's a tremendous amount of value in consolidating that team into one public-facing person. If they like, you know, a piece of content that they watch with you, then when they see an ad of you in their feed, you'll get significantly lower CPMs. So, I think you're just trying to solve a problem that doesn't exist. What are you planning on doing? Spinning up 4,000 AI avatars talking about all different types of of services and whatnot? No, I mean, if you're here on my channel asking me that question, you probably don't have the ad budget. So, don't worry about that at all. Just do it yourself. I don't know if you've noticed, but I do all my videos entirely myself. I also do all my ads entirely myself. And the reason why is because this is the most important part of my business. It is the top of my funnel. It is the step that every single person that ever gives me any money needs to go through before they eventually get to the point where they transform into a paying customer or client. If you think about it that way, you'll realize that there is no point in trying to save money at this stage. You should focus entirely on growing, not saving. And all AI agents or AI ad avatars and UGCs things can do for you at that stage is typically save.
You find clients in fashion much like in any other niche. Assuming you're going DTC or retail brands, you could target online businesses with terms like fashion on their Shopify stores, or scrape Shopify DNS ranges. You could also look at all the retailers in a certain geography, not the Macy's but smaller retailers, find out what brands they stock, and reach out directly to the smaller brands offering fully automated AI-synthesized UGC-style creator videos for their ads. Another way is to go bottom up from the ads: see which companies are running non-AI UGC ads right now on the Meta ads library, Google Ads Transparency Center, or even Pinterest since it's so visual. Make a giant list of advertisers, apply your human judgment or standardize it into a skill, and have AI pull out the ones likely performing well. Then enrich them, reach out directly, and offer a free AI UGC video, which costs you somewhere around 10 to 20 cents. Your total addressable market isn't very big, so giving value up front makes sense. Once you've made it, maximize the view rate: reach out over email, LinkedIn to everyone in the company, X, Instagram and so on, because most emails don't get viewed and you've already sunk the cost. As for packaging, you could sell a minimum number of videos per month, but videos per month don't really matter to them; what matters is the money they make on the ads. A simple way to align incentives is to focus on revenue, for instance: "I'll guarantee you an additional $20,000 a month, and if I don't hit it I'll keep working for free until I do." If a company already makes $150,000 to $160,000 a month through that funnel, adding $20,000 over two months is a tiny proportional improvement. If you can't improve a business's ad effectiveness by that much after 60 days of pouring over it with AI skills, you're probably not good enough for this game; a 16-year-old with no experience could find a 4 or 5% improvement by staring at it long enough. Then just rinse and repeat.
My AI and automation agency makes significantly less money than my coaching. Coaching is extraordinarily scalable—when you run a YouTube channel with distribution like I do, the optimal next choice is to sell something hyperscalable with zero delivery or logistics, which is what my coaching stuff is. If you're a total beginner, your optimal choice changes depending on whether you have distribution: with distribution, you'd go for e‑commerce or info‑product with a templated service; without distribution, you'd run an agency because agencies let you get your first 10, 20, $50k in revenue super quickly. Agencies allow you to make a lot of money very quickly, but you tend to plateau relatively fast, with further gains being incremental. Contrast that with an information product business (like a community or SaaS), which has zero marginal cost after a fixed development cost, so you make little money at first but way more later. Does this make sense? From my perspective, when I started my agency I was up here, made money quickly, hit resistance, wanted to scale with as little headcount as possible, moved over here, and now I'm significantly higher. Would I recommend doing what I'm doing now to total beginners? No—many people throw away their businesses after starting coaching because they stop doing the main service, become disconnected from implementation, lose skill, and start teaching the wrong way. That's why this YouTube channel is valuable: I'm showing you how I actually generate leads for my agency. I want to change the fraction of my total revenue that comes from this channel; I don't think it's the most effective use of my time per dollar, but I believe it will indirectly lead to more money through authenticity, so I'm getting back to selling outbound for the agency, setting up more inbound funnels, adding calls‑to‑action at the end of videos, and going to hire a team eventually.
Interesting that you point that out. That's probably a measure of buyer intent—if you're searching for the term, you already know what a website is, so it's a more mature market; you don't need to explain the value as much. You can just say your websites are better because of X, Y, Z, since people already know what the product is. You could do something similar with automations, but with automation services you need to be specific—like lead-gen automations, proposal generators, CRM automations, etc. If you're asking whether I've ever thought about paid ads for my own automation agency: I did experiment with paid ads when I first started, but I didn't see many results because I'm not very good at paid ads. That's why I've focused mostly on outbound work initially, and now that I'm doing more inbound, I don't really need paid ads. However, I am considering a retargeting campaign—just for transparency—where I retarget people who visited the Maker School 'About' page with YouTube and Facebook ads. As for the $20 a day to get good results: that's pretty great. I'm not sure about the exact customer lifetime value, but $20 a day is about $600 a month. Assuming you advertise on weekends, that's roughly one website per month. I don't know your COGS, but they're probably high unless you have a team. In Australian dollars, $600 a month is even less, which is great. I'm actually hiring someone right now to handle retargeting ads, looking to spend about $1,500 a month, and we'll see how it goes.
You clearly have an extremely valuable set of skills; your question is how to take this valuable set of skills to market. Very few people can get 300 million views in two years—you could genuinely change the course of the global economy with this skill, and millions of politicians would be interested in achieving 300 million views in two years organically on a super broke budget. So this sounds pretty good. The thing is you need distribution to begin with; you could show your current audience how to do that. That’s reasonable. But you should also leverage the systems you already have that enable you to do it—this is just your bottleneck. Six to eight hours to make a video is tough; you could bring that down to three or four hours, allowing you to make two videos for two approaches. In six months you’d have a solid base to sell—maybe a fitness channel program or something similar. That’s probably what I’d do.