TJ: Large language models don’t search the same way humans do. A lot of people just used me to kind of feel like they were hanging out with someone. They were playing Diablo 3 and they wanted to feel like they had a friend that was also playing Diablo 3. 

Spencer: Don’t quit when it gets hard and trust your instincts. 

TJ: If you have the right product for service, you don’t need to go viral to make money on YouTube. 

Spencer: Hello and welcome back to another episode of Creators Uncut, the show about the business side of YouTube and kind of like the behind the scenes I guess you could say. Super excited for another episode. Today I am joined with TJ Robertson. TJ, thanks for being on the show. 

TJ: Hey Spencer, thanks for having me.

Spencer: Yeah, so I don’t know what the best way to describe you is. Would you consider yourself like a YouTuber or an influencer or kind of something else? 

TJ: I don’t. I’m a business owner first, but I have been making YouTube content for the last 12 years off and on, and it really has fueled my business. So I could say my business would not exist if it were not for YouTube. Gotcha. 

Spencer: Okay, cool. So I think that’s part of the reason why we wanted to have you on the show is because of your expertise in the business side of things. So why don’t you go ahead and tell us a little bit more about yourself and kind of, I guess, what you do with your business and YouTube? 

TJ: Yeah, so I’ve been doing SEO since 2009. And early on I realized that video had a lot of potential to help my clients rank in traditional search results, but also just to get business through YouTube. So I’ve been helping them make videos. I’ve been making videos for myself. And then really around 2012, I was getting a little bored and I was playing a lot of video games. And I just, I thought that’d be cool if I could just make content about video games instead of about, you know, this personal injury attorney. 

And so I did that for a while. I’m just like, you know what, there’s this new game coming out Diablo 3. This is back when that was a new game if anyone is familiar with it. 

And I decided to use my SEO knowledge to rank for the term Diablo 3 Wizard. It was one of the classes. People were searching for it. And SEO was really easy back then. So I made a website called diablo3wizard.com and got to the top of the term of the search results for that term. And then just started promoting my own YouTube content through there. So I didn’t say it was me. 

I was just like, hey, look at this great YouTuber and would link to myself a bunch. And that worked. This strategy does not work today. 

Don’t copy that. Back then it worked really well. And so I started making YouTube content. And that was kind of the first glimpse I got into how effective it can be compared to other forms of content creation. Gotcha. 

Spencer: That’s cool. I didn’t realize that you were in SEO even before you started doing like content creation, I guess. So before we dive into that though, really briefly, hopefully everyone watching this knows what SEO is. But can you like briefly in a simpler way describe what SEO is? 

TJ: Yeah. I help people rank at the top of Google search when people are looking for their products or services. Gotcha. Cool. 

Spencer: So yeah, it makes it easier for you to get discovered. So you transitioned, or I guess I don’t know about transition, but started to include other platforms, right? Like you do TikTok, right? 

TJ: That’s something I just recently started doing. So a few years ago I started doing YouTube not for video games, but for the agency I was working at. The video game thing only lasted a few years before I realized it’s much harder to make money there. So I got back into it and I really do love digital marketing and helping businesses. And so I was working for this agency and we were looking for new ways to drum out business. And I thought, hey, let’s go on YouTube. Let’s see if we can get some traction there. And I was surprised that some of my videos that only got a couple hundred views were bringing us leads. And that’s where I realized if you have the right product or service, you don’t need to go viral to make money on YouTube. 

And we can talk about that more later if you want. TikTok was actually an afterthought. I realized it was a lot easier to get views with YouTube shorts than with Longform. 

Although the Longform views definitely led to more leads. But I was making the shorts anyways. I’m like, I might as well also post this to TikTok and Instagram and Facebook and everywhere else I can. And so I started doing that. I did not expect it to catch on. I thought this would be like a little bit of extra traffic, but TikTok, I started getting about 10 times the views I was getting on YouTube. And so I still post all the platforms, but TikTok is where I get almost all my views, almost all my engagement. And about 90% of my clients for my business have come from seeing me on TikTok. 

Spencer: Wow, that is crazy. I wonder what the… Do you have any theories about why TikTok is so much better for your style of content? 

TJ: That’s a really good question. I have theories, but no confidence behind them. I think people really like authenticity on TikTok. I think people like to see people as they are. And that’s what I try to do, mostly because I’m lazy. I don’t want to sit there and get lights set up and have… I have that here at my desk, but I didn’t want to put a lot of time into these videos. 

I’m trying to run a business. And so I’m like, what can I do in like 30 minutes a day? And I was already walking my dog around the block. And so I figured while I’m walking around the block, maybe I’ll just pull out my camera and talk into it. Whatever I’m thinking that day, whatever I think people need to hear, or I shouldn’t say people, I’m very careful to make sure every video I make is for the audience I’m trying to reach. 

So what’s important to them? And I try to think of like a call I had that day or just a project I’m working on. Come with one little tip. And then of course I make sure the hook is really good, right? I think a lot about those first five to 10 seconds. But then after that, I’m just jabbering. I’m just talking off the cuff and then I’ll take it into descript later and cut out all the fluff so that it’s really punchy. But that seems to resonate really well on TikTok. People feel like they can trust me. 

They feel like they know me. And yeah, I don’t know why not the other platforms, but TikTok would seem to like it. And they’re not short. I know on TikTok they say make like 30 second videos, but my videos are usually two to four minutes. 

Spencer: Oh, gotcha. Interesting. Yeah, I’m always curious about that. You know, I’m a TikTok consumer. I do the scroll and I do, I don’t know what it is, but I can recognize when something is real and when something is fake. 

And so I think I as a consumer enjoy stuff that is more authentic. And so maybe YouTube is more, I guess, staged or set up. I don’t know if that’s because it’s been around longer, but that’s, I guess, a theory that could be it. 

But that’s super interesting. So you do, you like talking to the camera kind of off the cuff. Have you ever tried, you mentioned the video games, but have you ever tried any other type of content or just mainly the like talking to the camera? 

TJ: Yeah, rewinding back to when I was making videos for Diablo 3. That’s the first time I ever heard about Twitch. I was making videos for Diablo 3. This was 2012. Never heard of live streaming. 

It was brand new back then. But some of my viewers on YouTube started saying like, hey, you should go on Twitch. You should try streaming on Twitch. And so I did and I had like seven people tune in and watch me. And it was like the most exciting moment of my life up until that moment that I’m here just playing games and they’re talking to me and they’re chatting. And so I was hooked right away and decided I was going to become a Twitch streamer. 

And so I started putting in like 40 hours a week streaming Diablo 3. And that was a very slow climb. And I was disappointed to learn it’s much harder to make money on there than I thought it was. 

And it’s actually, it’s easier now back then. It was like taboo to ask for donations. So we raised a ton of money for charity. Actually have a little plaque up there for some work we did for a children’s hospital. But oh, that’s awesome. 

But it was much harder to make money on my own. So it was, it was a really fun year and a half of my life. I think I was the top Diablo 3 streamer for a little bit. But eventually I realized I got to get back to digital marketing and actually provide for my family. 

Spencer: Yeah, yeah, I’m glad that you mentioned that it’s easier now. I think maybe you were just ahead of the curve. It’s crazy now how big live streamers have gotten. And you’re actually the first person that I’ve talked to on the podcast about Twitch live stream. We’ve talked about, you know, like Facebook and YouTube live streaming, but not Twitch specifically. So did that, do you feel like that had any influence over like the content you make now? Or is it like just a completely separate time? 

TJ: It did help me develop the skill of just talking off the top of my head. I think the number one skill you need to be successful on Twitch is to be able to keep talking. That it’s the number one thing to determine success because if you land on a Twitch streamer’s channel and for the first five seconds you’re there, they’re just staring at the screen with their mouth open. 

There’s a good chance you’re clicking off and going to someone else’s channel, which means you can never stop talking even for five seconds. Like you need to always have. And so it developed the skill of always speaking my internal thoughts, which is not a good skill to develop for social situations. I found myself thinking that people wanted to know everything I was thinking all the time. 

So I had to kind of unlearn that to some degree. But in terms of like creating content, I think it’s the reason that I’m able to create the short forms very quickly on the walk. I know a lot of people like to script content and it just, you know, takes a lot more time. And so being able to just speak in real time has definitely made that more efficient. Gotcha. 

Spencer: So have you ever done any live streams nowadays with like your SEO type content? Haven’t. 

TJ: I probably should. I actually, I have this dream where I’m just auditing people’s websites on Twitch or YouTube or wherever. And now you can stream to multiple platforms simultaneously. And I think it would be really fun because it’s my favorite thing. It combines my two favorite things, which is content creation, specifically live streaming and just auditing and providing strategy for small businesses. So if I can find some way to make that profitable, that will be my next journey. I think it would be a blast. 

Spencer: Yeah, that would be awesome. That’s cool. 

TJ: So, I know TikTok’s telling me to live stream every day. I get a notification. Go live. Go live. And I’m like, I don’t know. One of these days. 

Spencer: Yeah. For me personally, I don’t watch a ton of live streams. You know, I have watched some Twitch. I do enjoy video games. So I have watched some Twitch live streams. But for me, I love like the recorded stuff. But there is definitely an audience out there for the live stream live streaming. I mean, like, I think speed is like, I show speed is like one of the biggest live streamers and my kids love him. Yeah, it’s crazy the amount of people that watch him like on a regular basis. So it is. 

TJ: Definitely. It’s a different. Yeah. It’s a different kind of viewing too. I think typically when people are watching a live stream, it’s like it’s on a second screen, right? Like they’re doing something else and then they have the live stream there. I know when I was streaming, a lot of people just used me to kind of feel like they were hanging out with someone. 

Right. They were playing Diablo 3 and they wanted to feel like they had a friend that was also playing Diablo 3. And so they might be stepping away. They might be glancing here and there. But it was just really like, it’s almost like just a way to deal with loneliness to put a depressing spin on it. 

Spencer: Gotcha. So I guess let’s go ahead and move on from Twitch. I could talk about Twitch a lot because I love learning about new platforms and I don’t know that much about Twitch. But so you work primarily with SEO content. Is there something that YouTubers can learn or utilize in terms of SEO that could help them be discovered more easily? 

TJ: I think most of the time you want to disregard SEO on YouTube. But there are times where it can be really useful. 

Spencer: So what do you mean by disregard SEO on YouTube? Yeah, good question. 

TJ: So people who do YouTube SEO try to make you think that it’s hard and there’s all these things to do. It’s really, it’s like your title and your transcript. Everything else is going to have a very minor impact. And so when we’re talking about like, should you optimize for SEO, we’re saying like, should you make your title similar to what someone would type into Google? And there are times where that makes sense to do. But as any YouTuber knows, your title is the most important thing, at least other things that are easy to adjust. It’s the most important thing in terms of getting views. And so most of the time, if you take a title that is like really good at getting clicks and you make it something that’s optimized for SEO, your views are going to go way down. 

They’re not going to go up. And so most of the time that’s what you want to do because most views come from Discover. They don’t come from Search. There are times though, where you’re in a space where it’s unlikely you’re going to get a lot of views from Discover. 

Maybe it’s just a really competitive space. And then if you can find terms that people are searching for in Google that don’t have good videos made for those terms, that can be a really good way to like get started in growing a new channel. So I think there are always little opportunities for these commonly searched phrases that no one has made good content for. And so you can use tools. I use a paid tool, but there are free keyword research tools out there if you just Google it and find some of these terms in your industry or in your niche that you’re making videos on. And then just search for them. See, are the videos showing up high quality? Could I make something better than this? And if you’re having trouble getting started, I think that’s a great place. 

Spencer: That makes sense. So do you help YouTubers or do you like help businesses with their websites and stuff like that? Or do you kind of do both? 

TJ: Yeah, so I help small business owners, sometimes medium-sized business owners. And if they’re already making videos, then I’ll provide some guidance. I think that, like I said before, there’s a few things that make a big difference, right? As you know, the title, the thumbnail, the first three to five seconds. 

And then after that, it’s really hard to help someone. I don’t know. Have you ever done like coaching for YouTubers or worked with coaches? 

No. You know, if you’re serious about becoming a YouTuber, I think it makes sense to hire a coach. But for my clients, they just want to get leads. They want to take their marketing budget and get the biggest return on it. And so in my experience, you get 90% of the benefit in just improving those three things. 

Better titles, better thumbnails, better hooks. And so I help clients with that. I help them come up with titles for their next videos. But then the other work we do is just repurposing those videos. If you’re already making video content, which is a ton of work, you might as well get as much from it as you can. It’s easy to take one long form and make it into a couple of shorts to post those shorts to 10 different platforms, to take the transcript from the long form, turn that into a blog post, a Medium post, a LinkedIn article, a sub stack. And so we build systems that does that really efficiently. So they put in all that work to do the one video. And now they have 10 to 20 different posts that could rank on search or get found on those platforms as well. 

Spencer: Okay, that is really good information. That sparks a couple of questions in my mind. The first one is what are your thoughts on not just content repurposing, but republishing the same video? 

I’ve heard that some influencers or content creators, you know, after a certain amount of time will reupload like the exact same video and it’ll do just as good, if not sometimes even better. What are your thoughts on that? Do you think that is plausible for most people? 

TJ: I think so. I honestly don’t have any extra insight to share there. I’m not an expert in that regard, but that has been my experience. I’ve had a lot of videos that just kind of fell flat early on and I repost them and yeah, they do way better. So I don’t know like the perfect cadence to repost them. I don’t know at what point you’re going to upset the algorithm or upset your viewers, but it definitely seems to work. 

Spencer: Does that ever screw up? I don’t know if you can screw up SEO, but does republishing a video screw up SEO? 

TJ: No, there is such thing as duplicate content on your website. But really the only problem is that if you have two pages that have the same content, Google’s going to ignore one of them. 

And so it’s the same thing. If you publish the video twice to TikTok, Google is going to do its best to recognize the same video and they’re only going to index one of them. But you still get one indexed. And for now, if you have like different captions or different titles, depending on the platform, Google might not even realize it’s the same page twice and you could give them both indexed. So from an SEO point of view, I don’t think it’s going to hurt you. It just seems worth doing from my point of view. 

Spencer: Gotcha. Okay, cool. I was just curious about that. And then the other question that you sparked in my mind was about transcripts. So how important are transcripts to YouTubers? And should we be making sure that they’re like pristine? Because I’ve worked with transcripts before and like the auto transcripts get it wrong. A lot of the times, like certain words, would it be beneficial to like, you know, go back through and make sure the transcript is perfect for SEO purposes? Or I guess how important is the transcript? 

TJ: That’s a good question. Obviously, Google’s getting better at creating transcripts. I think eventually this will be less of an issue. I actually don’t know if you create your own transcript, if Google’s going to trust that or if they’ll still rely on the auto-generated one. I do know that if they’re having trouble understanding specific words in your transcripts, they will look at your tags. And that seems to be the only reason that tags help with ranking at all is if you’re using words that it’s unfamiliar with or likely to mistranslate, putting them in your tags can increase the likelihood that they understand what you’re saying. But mostly in terms of transcripts, I would mostly just make sure that the things you’re talking about are relevant to your audience, so that Google understands that. Gotcha. 

Spencer: Okay, cool. So for our podcast, Creators on Cut, we actually upload the transcript to our website. Is that beneficial in helping find the video at all or not really? 

TJ: No, it can. It ranks in Google. And so I think I would go one step further, maybe a few steps further. The first thing I would do is take your transcript to a large language model and just say, hey, clean this up. And it’ll just make it mysr. So it’s formatted, doesn’t have timestamps or whatever. And then I would ask a large language model, review the information in this transcript. and then try to identify the most common questions someone would type into a search engine when looking for this information. And that’s going to help you come up with a potential keyword that you could rank for. Because if you’re trying to rank in Google, you need to know what you’re trying to rank for. And then you want to split out the sections and do the same thing for each section. 

For this section, what is the most likely question someone will type into Google when looking for this information? And those become your H2s as we call them or your subheadings. And now you have a more optimized article. Now, we actually take this a few steps further. We do deep research with chat.gbt. We provide a little bit more structure and some writing guidelines, and we have it turn it into a full article that still maintains your voice and your opinions. 

But you could do it without that whole thing. What’s important is that you have an article that’s optimized for some term that still maintains all the value that you provided in that transcript. And then if that article ranks and you have your video in with the article, then yes, you’re getting extra views to that video. And of course, when Google sees people are finding your video outside of YouTube, it can also help you perform better on YouTube. Gotcha. 

Spencer: Okay. That’s interesting. So another question that I have, I guess more specifically for the YouTube side of SEO, with the rise of AI and like the chatbots like chat.gbt, Gemini, is there ways to be discovered on those easier? 

TJ: It is much easier to be discovered on those, especially with video content. The key thing to keep in mind with AI results is they’re still using a search engine. So if you go to chatbt and you search for information or you search for a recommendation, the first thing chatbt is going to do in most cases is search. And they’re using probably Google search. There’s some speculation on which one they’re using. They were using Bing. It looks like they’re using Google now. But regardless, they’re using a search engine. And so that has a lot of SEOs saying that there’s no difference. That showing up in a large language model is exactly the same as showing up in traditional search engines, because if you rank in the traditional search engine, the large language model will find you when it uses them. 

The problem with that is that it’s not true. Large language models don’t search the same way humans do. So when a human goes to Google and performs a search, they’ll typically search for like one broad term. So for my clients, they might search like dentist near me, right? Or, you know, if you’re a gaming YouTuber, they might search, you know, a best warrior build for this game. 

And then a human is going to click on one, maybe two or three websites before they get their information and they’re done. Large language models don’t search for small, broad, competitive terms. They search for very long terms because they add all the context, all the information they have about you, all the context they’ve learned from this long conversation you probably just had with it. And so it’ll search for something like, you know, best warrior build for this dungeon when partied with a priest in this video game. 

Right, some ridiculously long string or for a business, you know, the best dentist tooth for wisdom, the best dentist for wisdom teeth extraction that uses, you know, sleeping gas or whatever. And so the search results are different. But also, instead of just looking at one or two results, the large language model will crawl hundreds of results. 

They’ll do multiple searches to look through multiple pages of results. And so that changes their strategy significantly. Instead of the strategy being how do I rank at the very top of Google for this really competitive term, it’s how do I show up somewhere in Google, even if it’s on page two, for very specific terms, which means the competition is way lower. The search volume is also way lower. In other words, you’re going to get far fewer visits from those searches, but there’s way, there’s way more variance in the searches. 

And so instead of just making how we used to one piece of content that was really aimed at bringing the top of Google, now it’s about how many pieces of content can you make and how efficiently can you make those pieces of content. And then because if you can cover a lot of ground on these very specific terms, you’re going to find that people are finding you more easily than they were before. If you were to niche before you were invisible, because no one was searching for that super niche content. But large language models makes it way easier to find that niche content. And so I do think there’s going to be a bigger demand for these smaller niches. And if you can either niche down on your topic or just on individual videos, they’re going to become way more discoverable because of the way these large language models search and how quickly people are moving towards large language models to find content. 

Spencer: So I guess, does that mean we need to change the way we, I guess, write or publish our content? Or should we just keep doing what we’re doing and it’ll just show up in the large language models? 

TJ: I think the key thing it changes is that you want to publish more content in more places. Oh, okay. 

Spencer: So that’s why it’s important to publish on all the platforms. 

TJ: Yeah, yeah, because it’s pretty wild. Large language models will find videos with like 20 views on them for certain searches, right? Videos that never would have ranked on page one of Google. 

They just go way deeper. And so you never know which video is going to hit, of course. But if you’re going to put all the time into making a video, find a system for repurposing that. And I can talk about how we do it. It’s too involved to go through the whole thing on the podcast. But you know, large things can help a lot with this and getting it published to all the platforms. Gotcha. Cool. 

Spencer: I feel like I could talk about this all day long. This is super fascinating. Yeah, me too. But let’s move on to the next question. So we kind of already talked about content repurposing a little bit. And I don’t know how often you work with YouTubers or I guess businesses that do YouTube videos. But have you seen anything that like people could learn from that would be would help them repurpose their content? Like I think you guys, at least you mentioned you have like a tool that you use or you created a tool or something that helps make it easier. Is there any tools that like our listeners could benefit from to help, you know, repurpose content to other platforms? 

TJ: Yeah, I’m sure most YouTubers are familiar with tools like Opus Clips. I think they’re really good. It’s not for everyone. Some YouTubers want a lot of control over how the short turns out. But you know, if shorts are an afterthought, if it’s just something extra you’re doing, I think tools like Opus Clips do a really good job for what they cost. And that’s great for creating video content, of course. But then the transcript will help you create written content. 

And that’s much more likely to rank in search. And so the tool we use there is Claude. If you’re not familiar with it, it’s just like chat to BT, but it’s made by a company called Anthropic. So you can use chat to BT for this as well. We prefer Claude because it’s better at handling lots of context and it’s better at writing typically. But it’s not going to be a big difference. You can use chat to BT if you don’t want to pay for a new large language model. The first thing we do for our clients is we create what’s called a brand ambassador. 

I’m trying to decide how deep to go here because it’s I want people to give a full course on it, but I can at least give some principles. The brand ambassador is just a project in chat to BT or Claude. If you’re not familiar with projects, you can create them from the little sidebar. It just gives you a folder where you can dump in some information and a box where you can write some custom instructions. And so you write some instructions like you’re the brand ambassador for your brand. 

Your job is to represent the company in all outputs that represent the brand represent the channel. And then you tell it what documents you’ve given it. So you can say like I’ve given you a document called our brand’s ideal customer profile. And then you say what it is. This outlines all the details about our ideal audience. And then you tell it when to use it. Use this whenever you need to determine the audience you’re speaking to. And you give it as many documents as you can. 

You want to know everything there is to know about your brand, everything there is to know about you. And you only have to create this once. It’s a bit of effort. But then once you create it, every time you need to create more content, you do it inside that project. And so in addition to whatever prompts you give it, it also knows all the relevant context about you. 

So that’s optional. But I recommend taking some time to do that. And you might have to tinker with it to the outputs good, but it can make a big difference. After that, it’s really just about the prompt and the context that you give it. And so if you already know how to make an effective LinkedIn article or blog post, great. You just want to detail that very clearly in the prompt. And I’ll tell you the structure we use for prompts in a second. If it’s something you don’t know anything about, I recommend using chat to BT5 thinking. 

Make sure you’re on the thinking model and just say, tell me everything there is to know about writing an effective LinkedIn article about writing an effective blog post that rings on your mind. And Google, let it grab all the research. We’ll come back with a huge PDF and then give it back to large language model and say, just distill the main points from this research into a bulleted list. And now you have some guidelines that probably not as good as what an expert could do if they sat down and thought about it. 

But you have something there. And then you draft a prompt. And this is the prompt structure we use. You start with a role. 

And the role is always going to start. You are an expert something. You are an expert article writer for this channel. The next one is goal. 

You are going, you write highly compelling blog posts that rank on Google. And then next is going to be input. You tell it what information you’re going to give it. So you say, I’m going to give you a transcript from a YouTube video. 

I’m going to give you some research on the topic, whatever you’re going to give it. The next one is process. You tell the exact process you’re going to go through. 

Start by making sure you fully understand the transcript and the tone of voice I use. Next, following the guidelines below. Draft and effective article, whatever the steps are. Next is guidelines. This is where you’re going to put some rules. This is where you would paste those bullet points you got from Chatchapiti’s deep research. 

This section might start small, but this is what you’re going to update. When the output is bad, you’re going to update the guidelines. You’re going to keep iterating until the output is not bad anymore. And then the next one is going to be examples. Try to give it at least one good example. And then eventually you also want to give it some bad examples. You don’t have to do that right away. Once you start using it, when it gives you something that you don’t like, copy that, paste it into your prompt template. I didn’t say that before, but you should save this somewhere so you can keep reusing it. And then that starts your bad example section. 

Don’t do this. Paste the example and explain what’s wrong with it. This sounds like a lot of work and it is a lot of work, but you only have to develop this template once. And then, you know, if you use it a hundred times, it actually ends up being a very small time investment per use, right? And this is why people don’t get the output they think they’re going to get or they want to get from large language models, because they don’t put in the time to clearly communicate what good looks like and what they actually need. But if you put in the time, you make a brand ambassador, you make really good prompts, I think at least our clients, they’re blown away by the quality. It really, when it’s done right, it’ll look like you personally sat down and wrote out this blog article or LinkedIn article. Wow. 

Spencer: That is a lot of information. And I think that is really, really beneficial. So I guess how did you become like such an expert on this yourself? 

TJ: Yeah, this has been my passion since chat should be teedropped. I guess I’ve always been excited about AI. I’ve always thought that one day I would replace all of us and then probably kill all of us. 

Hopefully it doesn’t happen too soon. But when chat should be teedropped, it like it really disrupted digital marketing because a big part of the marketing is we would hire content writers to create content. And there’s still definitely a place for writers in the world. But the kind of writers we were hiring weren’t experts in what they were writing on. They would write for any small business. And their process was you go to Google, you read some other articles, and then you rewrite it in your own words, which is exactly what chat should be teed does only better and much cheaper. And so as much as we love our writers immediately to stay competitive, everyone essentially switched to chat to be tea. 

Unless you’re writing something where you can pay, you know, the top writer in the world who’s an expert in their field to write the content for you. It just didn’t make sense anymore. Yeah. And I could tell that writing was first, but it was going to start happening in other areas as well. And we’ve seen it happen in graphic design. We’ve seen it happen with voiceovers. And I think we’re going to start seeing it with website creation. 

It’s halfway there with code. And I’m not happy about this displacement. I think it’s really bad for a lot of people. And I wish there was something we could do to at least help people not lose their job so quickly. But it’s happening. There’s no slowing it down. 

We have to adjust. And so once I saw that, I just, that’s been my core focus is how can we use these large language models to do our work better? How can we get our clients to rank in these large language models? And at first I just thought, oh, eventually these models will be so good that it’ll just be easier and easier to use them. And they’ll just be able to create this without, you know, we’ll have to provide less and less guidance. But the opposite has been true. The smarter the models get, the harder they are to use. 

It’s, I mean, it’s just as easy as it was back in the day if you’re doing something simple. But to get high quality output, the models still need context. They still need a clear design doc, right? 

They still need you to outline exactly what they need to do and provide all the context they need to do it. And it doesn’t seem like that’s going to change anytime soon. As the models get smarter, they’re not getting any better at just understanding the context of your specific brand of your specific project. And so that’s the gap I’ve been focused on, on bridging is how do we create systems that give those large language models the context they need. 

Spencer: Gotcha. That totally makes sense. And I’m glad there’s people like you out there who are experts on this because it seems like it could be really beneficial. And I recommend any of our listeners who don’t want to go through all that work. Just reach out to TJ. He could help you. 

TJ: Yeah, thanks. TJbarberton.com. 

Spencer: Yeah, I’ve got just a couple more questions for you. And I guess there are more, I guess your experience or opinion about, you know, content creation. So I guess from your time, you know, beginning with like live streaming to making the content you are now, has there been a mistake that you’ve made that you feel like? You’ve learned from that would be beneficial to our listeners. 

TJ: Oh, man. Yeah, so many mistakes. I quit streaming at the wrong time. I quit streaming right before the first expansion to DLB3 came out. And I quit because my viewer count was going down and that’s it’s hard. 

Right. It’s hard anytime things are going up for a while to see them start coming down. And they were going down because no one was playing the game I was streaming. And then the expansion came out and I saw all these little streamers, audiences blow up. I had a close friend, Datmods, who still streams full time. And I talked with him right after the expansion. 

He’s like, I’m making $10,000 a month streaming on Twitch. And I was like, oh, cool, I quit three months ago. Really hard to get that momentum back. So that was my first lesson. I think the lesson there is just like, don’t stop when things get hard. 

Like just keep going. And that has served me well. The other big mistake I made was with my last agency. I started my own thing in 2009 and I did that for about five years before I joined someone else’s agency. And I joined them honestly because I was in my 20s and I was kind of lazy. 

I was disorganized. I enjoyed doing digital marketing, but I needed some structure. And I found someone who was like really building a business the right way. And I was like, this is the right guy to hitch my wagon to. And that did help me develop some discipline and some structure. But I stuck around way too long. When Chatcha BT came out and I saw this big shift, I was trying to tell the founder, we need to completely reinvent ourselves. 

We need to change what we’re doing things. And he did not agree. He thought, nah, it’s not going to happen that fast. 

Let’s just wait and see what other marketing agencies do. And yeah, I should have jumped ship right then because I was already rebuilding my processes and doing things twice as efficiently. And yeah, I waited another two years or so before I finally decided to go out on my own. And so, yeah, I think I just learned to trust my instincts and just have the confidence to take some risks. And if you really believe in something, it’s worth it. 

Spencer: Yeah, that is great advice. Don’t quit when it gets hard and trust your instincts. I completely agree. And I think that applies to YouTube really well as well. So a lot of people quit YouTube when it gets hard and they don’t stick it out. And if you have an idea for a video, you should just trust your instincts and go with it. So the last question I have, we asked this to all our guests and you can answer this about YouTube or TikTok since you do a lot of TikTok videos too. But is there any myths or misconceptions about YouTube or TikTok that you think needs to be busted? 

TJ: Yeah, I think this isn’t true for everyone. Depending on how you’re earning revenue through making videos, views aren’t everything. Now, if you’re making money strictly from views, then yes, views are everything. You just try to get as many views as possible. But if you’re selling anything, if you sell any kind of product or service, then it’s really easy to just get fixated on views and doing whatever is going to get you more views. And I can say from someone whose entire business comes from TikTok, I have videos with 2,000 views that have earned me multiple leads and videos with 50,000 views that I have not got a single lead from. It’s really just about does this video speak to my core audience and does it do a good job of representing what I do. Gotcha. 

Spencer: That makes sense. Cool. Yeah. Thanks for all your advice and insights. Really appreciate it, TJ. 

TJ: Yeah, thanks so much for having me. This was a blast. 

Spencer: So you mentioned briefly your website is tjrobertson.com. Is that right? Our viewers wanted to follow you elsewhere. Where should they go? 

TJ: TikTok, probably. But I post the same videos to YouTube and Instagram. So if you could just Google tjrobertson SEO or tjrobertson AI, anything like that. If you just do tjrobertson, there’s a football player, I think that outranks me. But if you have any kind of marketing term there, then I’ll show up. 

Spencer: Cool. Well, thanks again for being on the show. And thanks to our viewers and listeners for checking out this episode. Make sure to like and subscribe for more content. And if you ever need help with making videos, feel free to check us out at creatorluxe.com. And we’ll see you in the next one.