How We’d Make $10K/Month From SEO Starting Today (From Scratch)

Welcome everybody to episode 1,000. This is crazy. We got Julian Goldie, we got Charles Float, Alejandro Meerhance, and all of you guys are in Thailand. And all of you guys know each other. Yeah. Yes, sir. Thanks for having me. And congratulations on the thousand episode. Thousandth episode. Yeah. A thousand days in a row. No days missed. So, uh I'm I'm just going to start. First question. Fun question. If you were starting fresh and had to make $10,000 a month from SEO starting today, what would you do? We're starting with Alejandra Myer hands putting you on the spot, brother. Why would I do right? Uh, depends. How much more do I have to start, right? You're starting fresh, so you just have like a couple hundred. Oh, EMD some niche that I understand will rank a small site legend straight. What would you say? No competition, easy to get. What would I sell? Well, it depends what I'm equipped to sell, right? Myself as the individual. If I'm funneling that into a business that I I have a partner with local, let's say, right? If you're in your city, you don't have, you know, you don't even know where to go get leads, where to sell leads. It's like, okay, who do you know runs a business? Can you rank for their business and get into a local partnership? Right? If I was, you know, really starting from zero zero, just knowing SEO now, not really starting from zero as in knowing the industry, I'll probably choose something in YM that is hard to rank for, but not in the right locations. That's crazy. Start Y. Yeah, it pays a lot more. And in the right locations, you you beating people who aren't SEOs. Charles, how about how about you? Yeah, I'm I'm probably selling links to be fair, just straight up. Um, you can go and buy age domain names for like $10, $20. You know, you don't need anything that's DR anything. You just need an age domain name so that it looks kind of more legit. Um, pump some free AI content, do a bit of analysis, probably a high-end or high ticket niche in terms of the actual links price. uh you can boost DR pretty easily with a bunch of free links and then just kind of keep the publishing frequency up to maintain that. The initial setup would probably cost you like 30 to 50 bucks per site. Um and then you could easily sell that uh sell links for that individually per link, right? So you could easily kind of make your money back just on the first sale. From there, I'd probably then try and reinvest, build kind of more of a network of sites to scale that out. I'd also at the same time be doing kind of free parasite pages as well just to try and find any niches that I can that are doing particularly well at that given time. Trendjacking, newsjacking, all that kind of stuff. Launchjacking as well. Um if you can get some free parites in there and just pick up some clicks that can get some commissions for some high ticket items could be $200, $300, $400 commissions there. Again, put that back into more site networks for the link sales and just kind of stack there and kind of build into a proper business where you could actually scale to 10k a month within a year or so. So you would you would start by selling links where what where would the links be? Uh so I I would set up blogs, right? I would go and buy age domain names, but again nothing with DR or anything. I just want really cheap age domain names. So anything like 10, 20, 30 bucks. On NameCheep, you can buy uh name marketplace. You can buy some age domain names that have been five, six years old for like five bucks, six bucks for an age domain name. Um re it'll probably be expiring soon. That's why they're selling it. Um, and then you can go and basically pay the renewal fees, expand that site, and kind of sell links. You you'll immediately start getting guest post emails if you put the right content out there. I know that for a fact. Um, just the amount of people that are currently emailing and doing outreach for link building. So, you'll immediately be able to get sales usually within the like month two or three from putting that out live. Probably crazy operating press wiz and the things that you see every day. Yeah. So we try and actively block resellers. So this is actually actually something that I haven't talked about publicly before. We have built an internal reseller blacklist of like all the biggest people in the industry who are actively trying to resell links and we're trying to build their portfolios as well. So we're trying to like figure out what their network is of the sites that they're trying to add on because once you become a publisher you can then add sites to be approved onto the queue and it's a lot easier to kind of pass the validation process. Um, so we're just trying to remove all of them. And by doing that, we're trying to make sure that we've removed some of the resellers. As another extension of that, we've actually set up some honeypotss. So we're trying to we're trying to get people's emails in that are outreaching and then we can blacklist them, right? So it makes it easier as well. This method of marketing is so effective, I had to make sure it wasn't against Google's rules before I kept doing it. It's a form of SEO I call compact keywords. Whereas most SEO focuses on putting up articles to answer questions, how, what, when, compact keywords focuses on putting up dozens of pages that sell to searchers who are actually looking to buy. These pages rank on Google and convert so much better than normal that when I discovered this years ago, I couldn't believe this was allowed. It's less work, too. The average Compact Keywords page is only 415 words. Compact Keywords is a 13-hour deep course on getting sales with SEO. A customer recently said, "Each lesson is dense with information. You're giving years worth of experience boiled down into 15 to 30 minute lessons with no filler or fluff. I feel like I'm gaining a new superpower. Compact Keywords is about setting up an SEO funnel that brings you sales for years and years and years. It works with AI. It's less work than traditional SEO and it makes way more money. You can get it now at compactkeywords.com. Back to the podcast. Wow. Julian, what would you do? I'd probably I was thinking maybe start a pyramid scheme. No, just just kidding. So, I would uh I would probably I'd create a lot of content on YouTube around AI or AI agents and then I would sell the setup and the automations themselves uh for like $500 or $1,000. And so, like I think you just need to get five or 10 of those and you're good to go. Um yeah and then also like you can not just charge for the setup you know for example setting up open claw for someone or for example setting up Hermes for someone but also you can sell for the ongoing retainer and the maintenance on that too and so I think it'd be a super simple funnel cuz it's just like okay create content all day every day when you got spare time take the sales calls and and close as many as you can and I think between those two things it's a very simple funnel that would convert Julian, you must see some just wild stuff being so deep into AI and the need and the the fierce hunger that people have because you are constantly putting out like I'm getting your newsletters every day. It's you are going so hard. You have such a huge team. You must get the most insane requests too. Yeah, honestly like um for us like it's an interesting one. So like for actually selling AI automation, I don't like to sell it as a service. Uh because it's it's basically like selling a software, you know, like and and the biggest issue is like for example, if Charles is selling links, it's like you buy a link, you know what you're getting. Like it's very tangible, right? Um, but with AI automation, we've found like it's quite difficult to scale and it's very hard to set the boundaries in terms of like here's what we're going to do and here's what you're not going to get and then by the time we deliver it, is it still relevant? You know, because if for example, you sell a project for like 15K now and and I've seen this firsthand and then 3 months later it's like, well, nobody uses NA10 anymore or for example, nobody is interested in that automation or you can do it for a website now. you don't need uh like you know make.com or something like that to do it for you and so like it's really hard to scale. We do get some insane people coming to us asking for all sorts of crazy stuff and we're like no no no we can't do it we can't do it. Whereas for example for us like on the service side it's way easier to sell SEO way easier to sell link building cuz it's very productized and then for selling anything AI related I typically just sell the uh the community that we have. you know, it's a lot more scalable and then it's kind of like here's how you do it. You go off and do it way more chill. Yeah. Yeah. Yeah. Who um who in this in this chat do you think uh in this conversation has the most chaotic SEO strategy behind the scenes? For sure, Charles. No, for sure. Strategy wise, maybe not. Yeah. Well, maybe not strategy-wise because my strategies are pretty cleanly executed, you know, testing wise. Oh my god. You know, like we we just throw everything at kind of the kitchen sink at Google and then see what sticks kind of over time. Um, and like we've come up with some really stupid stuff in the past that just to kind of test cuz a lot of it is you're trying to test things that seem logical on the surface and then when you actually put it into practice, you're like, "Oh yeah, Google, there's no way they're spending this amount of processing power to try and figure this out, right?" Um, so our testing is really Celtic in the terms of we're just trying to see what is happening and we're even we've been testing techniques that used to work 2008 2009 2011 kind of times just to see if they work again now because and again several of them are working again now, right? And that's kind of the reason we're doing that. But in terms of strategy wise I'm pretty clean. Um, and I kind of showed you, Edward, the back end of some of the campaigns that we were doing, and it's all generally quite organized, and I would say we're we're more organized than like the traditional agency or something that that comes together with their most their strategies. And I I try to make sure that my strategies are very cohesive. So on page, publishing frequency, link building campaign, off-page signals, social signals, RM, everything is very interconnected. Um, and we're kind of building everything so that it stacks together, but at a natural pace as well. What are what are some of the things that were working in 2008 2012 that are that are working again? Redirect chains, right, is probably the the number one thing that seems to be working at the moment really really well. Um there's a bunch of stuff like if if you have high enough authority, you can essentially kind of stuff pages again. Um and then clo a lot of cloaking is working really well again actually. So a lot of the JavaScript based cloaking that Google had managed to clean up with previous actions is just ranking really really well again. And I actually had a a little clip um of a video that I was going to put on Twitter at some point which shows a tie with a uh a website for a restaurant that you're supposed to click on and then it redirects you to a gambling uh casino page and it's just straight up Google like the number one ranking page in Google and stuff. But again, a lot of it was previously cleaned up 2010 2015 and it's cloaking is working really really well again right now. Julian, you tried a lot of crazy stuff too. Yeah, for sure. Oh, I mean even recently like for example um you know me and Kazra Kazradash like we did a video together where it was him and his manual SEO versus my automated AI agent with OpenClaw and um and we both tried to rank for the best SEO speaker 2026 and my openclaw uh article is still ranking now. It it was ranking inside AO views. is ranking on Google as well. And you're right, Charles, it's like, you know, you can get away with a lot more like higher keyword densities. Um, and these even when the content is not good, you know, like I I looked at the article and I was like, this is terrible. I'm embarrassed. Like Casra is going to absolutely smash me and and his article was better, like genuinely better. Um, but Open Claw beat him. It's just like why why is this working right now? Why did it was also not the best SEO in the world? Just saying guys, you know, Um, he's probably better than me, so you know, I let him off. But like he's good. He's great. But, uh, yeah, it was just like I I don't know why it was ranking. I think it was like a lot of keyword density and then, um, a lot of relevant entities and it was a prompt that worked before when I've done any sort of SEO, you know, with any sort of tools that related to AI. So, I think that helped too. Yeah. And also the domain authority of the website is like it's just so heavy. It's like if you have Yeah. And I see I don't know whe you've seen this Charles as well like for example Binance Binance are ranking for everything nothing relevant to their topic. It's all like if if you look at what they're ranking for it's like how are they ranking for all this and it's purely because they got so many back links and and so much domain authority pointing to their site. Yeah 100%. I I I actually know the head of SEO for for Binance as well and uh they have got a pretty broad strategy and I think it's generally because their uh root authority was quite generic but they they became like the number one crypto marketplace very quickly um and kind of maintained that velocity and attracted loads of digital PR links and even when like their founder went to jail and stuff that there's like a boost in like 1,200 RD or something that week or something for their site like there's there's quite a lot of uh things going on in in a positive way maybe for their SEO profile. Yeah. Like they're ranking for like loads of AI related uh keywords for news. I see them in the SERs all the time. Yeah. Yeah. Mhm. They rank for like sports banning keywords and stuff randomly. There's weird like sports relates keywords that they rank for all things. Yep. Why uh so is Binance actually going after are they expanding their topical authority? is just their authority their general authority is so massive they can do anything. Yeah. I I think they they've become a generic authority at this stage basically. And it's not that Yeah. Exactly. They're not necessarily trying to go after a keyword. It's just that they happen to have a page that matches some level of intent for that keyword or some level of query and it ranks for that query and then because they have such a borrowed website and link profile it ends up ranking across the board. But with the AI stuff, uh there is probably some level of intention there in in that kind of query. But again, there's loads of random niches that they're definitely not trying to actively target and they are ranking it still. The the other uh the other experiment I've been doing recently is like, you know, for how many are you guys using openclaw or Hermes or anything like that? Not at the moment. I think I just I'm just too deep in the AI agent space. But like for example things internally but just not not I I can't give I I do you not have a problem with giving it like full permissions that's my own it depends what it's depends what it's right it's like yeah I wouldn't give it like my you know my passport details and bank account details and everything but like if it's just to a WordPress website it's like it's fine it's fine um but for example one thing that I can do now is like if I have an exact match domain that I want to rank quickly you know targeting a low competition keyword I can and let's say for example it's it's to funnel traffic to the air profit boarding I can give my agent like openclaw a keyword then it writes like a 2,000word landing page fully like SEO optimized with clawude sends it over to netlefi netify host on a subdomain and all I have to do is like buy the custom domain for that exam match domain and it um and it's published and and ranking and it's like that that is literally like a freak click process, you know, keyword, buy the domain, put it on Google search console, boom. That actually makes a lot of sense for if you're trying to do like multi-dommain brand setups as well. So if you have if you know instead of having like com.net everything just redirecting to your main site. You could have individual little brand sites for all of them and it would actually kind of a you'd get more realy in the SER, b you'd get more entity kind of relation, see you'd have more visibility, d you could have more control of a views. There's like loads of kind of ways that this actually plays into it and I'm sure that openclaw could theoretically expand each of those individual sites into their own stuff. Exactly. Kind of. Yeah. Yeah. So you can like build a blog onto that exam match domain and then like every time there's a new keyword related to that initial keyword you can publish another page another page and then it's got access to all these different websites you're hosting on netifle so you know you could you could expand that out as much as you want that was my next step actually uh pretty cool there Julian uh I haven't been doing it with open club but I've been doing like through actual approval you know with uh with anti-gravity is if you feed it basically system to just have its own topical map. So it will it will continue to develop it in in the way that you were describing. So I've I found out if it has the right like rack pipeline and it knows how to develop a topical map that you know will follow adhere to the principles then it can just follow. I haven't tried it fully identically with open glow. So that's something that I'll try I'll tell you uh what what I find but I found out like even it's just manually approving every page builds pretty pretty big sites pretty fast. Yeah, I can imagine. Also, it's probably a safer way to do it. Like I mean you know approving page by page probably a better way to quality control it all, you know. So that's cool. Julian, do you like OpenClaw or Claude Code more? Honestly, for me, like claw code is way better because it's it doesn't break so much. It's way less technical to set up. But the thing with claw code is like you kind of even though it's got persistent memory now, it's almost like starting again every time. Um whereas for example, if you're using something like and even if for example you want schedule tasks, you have to leave your computer running with clawed code open for it to schedule tasks and run automatically. Whereas for example with open claw I can just open up telegram and like that task is running 24/7. So it's way more it's way easier to run I think. And also claw code I feel like they do put a lot more guardrails and stuff which slows you down. It adds a lot more friction to the process. Open claw is crazy. It's like you know it it can it there's no real guardrails on it. You know it's just an open source project someone created over a weekend and you can create whatever you want with it. Have you had anything go wrong? Um, yeah. Yeah. Like for example, I mean honestly like every time I update it breaks and then I have to set it up again. So stuff like that it goes wrong a lot. But nothing crazy. I mean I saw way worse situations. I think someone I saw one situation where someone uh got their open claw to send crypto to the wrong wallet and they lost like 40k. That was bad. That was bad news. And there was another one, the head of security at Meta, she gave access to open claw via email told it like, you know, don't delete anything without my permission. And it it was like, oh, sorry. Sorry. I've already deleted all your inbox. Yeah. Yeah. Yeah. Sorry about that. I won't do it again, though. I promise. It it it makes a weirdly uncompelling case for Open Claw when it comes to like what other people like are thinking of using it. Oh, like let me have it as my personal assistant. N like self-contained in a little box with extremely limited uh accesses that might be a little bit better for very precise things. But yeah, like that when when it first came out, right, I I looked into it like this sounds like a nightmare. like this. This is like an nightmare with uh with candy wrapping and same with some other AI tools that come like oh you know press a button and you have a speech to text but it has full access to your machine and it sees everything on your screen and like including my bank passports because I'm not doing a speech to text when I can just run locally and just train it over a weekend. Right? So, I think with with some AIS, the uh the the human in the loop is more like the human that makes the executive choice of not opening the doors for Pandora. 100% agree with that. There's there's actually there's one called Screen Pipe on GitHub. It's pretty easy to install, completely free, like open source again, and it records literally everything that you do all the time, including audio on your computer. So, it's like it's amazing for making your assistant very very personalized and giving it a lot of context. It's terrible for privacy. Literally the worst thing in the world for privacy. Typing your will into the chat referencing back to you. It's like all your bank account details, all your passwords, they're going through that. So, all of you have been on this show before, uh, some several times before. And um is there anything in your SEO or marketing that's kind of changed or new things you're excited about? Charles, you put up that awesome write up about X's algorithm, for example, like that's I'm actually planning on making an episode on that. That is that was great. Yeah, I I've been trying in general to kind of reverse engineer a lot more algorithms. I didn't realize how many um of the other platforms are open sourcing things or there are leaks of other platforms algorithms. So, Amazon A10 is available online. X's algorithm is available online. Uh, Tik Tok's algorithm has recently been uh updated since the US takeover. And there's been a bunch of leaks um around Snapchat's algorithm and LinkedIn's algorithm and all this sort of stuff as well. So, actually there's uh Chase Demand is a really interesting LinkedIn uh expert specialist and he posts quite a lot about the LinkedIn algorithm updates and some of the LinkedIn engineers actually talked to him about the algorithm which I think is like insane, right? if you think coming from an SEO world. Um, so I've been trying to just look more into these kind of other algorithms and see where we can actually kind of move into and what they and how easy they are to manipulate or exploit versus uh SEO. And real realistically, Google's algorithm is so much more ahead or so much further ahead than any other algorithm in the digital marketing world or the kind of AI/ internet ecosystem that we have that it's pretty unbelievable. um and you quickly realize that there are just very linear signals in the other algorithms that you can very easily manipulate versus how dynamic it is for SEO and in organic um and even in things like Google discover right so that's been just really interesting kind of looking into that that whole world um and then trying to reverse engineer and piece together uh parts of it and again X has been the test that I've been doing with X as a result of figuring out some of their algorithm has actually been quite successful and they've been very limited so far. So, I'm looking forward to getting some more time to actually commit to that as well. Yeah, really. Access X X has been successful. Yeah. So, I did um the the very first post that I did was to promote that article, right? Um and it was basically an engagement bait piece. So, I reverse engineered a that it was a video that you wanted attached. So, the maximum amount of boost that you could get wasn't an image, wasn't text, whatever. It was video that you posted on there and that would get the most max amount of boost from the algorithm. Um, from there there's actually a clip length measurement. So, there was a measurement of how long they wanted you to interact with the video clip, right? Well, that meant that actually you only wanted a certain length of time of that video clip. So, max like 10 to 20 seconds cuz you wanted a 3 to 5 second interaction with the video, etc., etc. I piece it all together. that one uh that one tweet got I think 25,000 views and it only had about 200 engagements or something within the first few hours which again is really low for how many views it actually ended up getting. There was also other bits and pieces like quote retweet and uh like certain emojis and stuff that you can use and sentiment analysis. So trying to make sure that you have the ultimate sentiment for Grock to process and then output at the end and stuff like that. Um but it's been working phenomenally well. I just want to piece it together and probably again put together like a skills file for Claude or something uh and try and like make it into more of an automated pipeline where we can roll it out for not just my own personal brand but also other company brands as well. Dude, when you posted how important comments are now I make like a habit to comment on all of my friends posts cuz it's going to help it do better. And and it was like what what was it? It was like a 75 times multiplier for comments. It was like the biggest thing by far, whereas like a like was like one or two times. Yeah. And actually bookmarks were more valuable than likes, which was interesting as well in of itself and things like that. So there's a bunch of kind of things that you that I found really interesting. Um and obviously because it's open source, you get a very intricate look at how this algorithm works and you can piece things together whereas we've only got kind of the leaks attributes which don't have the waiting right from the Google algorithm side. This does have the waiting which is actually the most important part really. Dude, I think I ruined uh LinkedIn. I was getting I was so I was I started shouting on LinkedIn about how insane my reach was because I noticed this was like me many months ago. I noticed that I was getting 200,000 impressions per month or sorry per week on LinkedIn. Per week on LinkedIn 200,000 impressions and I I started like writing you guys got to post. Everyone has to post more on LinkedIn. this is the most amazing thing ever. I'm gonna post 10 times a day on LinkedIn. And I did. And I got up to I think it was like 1.4 million impressions per week on LinkedIn. And now everyone has like listened to me and so many people are posting daily on LinkedIn and crushing. But now my reach is down to like 140 or 150,000 uh impressions a week on LinkedIn where and I can't like it's I can't even crack 300,000 anymore cuz like everyone's going so hard on it. Yeah, I I LinkedIn's probably the worst for AI comments as well in terms of replies. um every post is like five to 15 kind of LinkedIn engagement bait comments whatever from from traction and I'm not sure again who how many how much that is like open claw automation and things like that um and how bad those platforms are at filtering it versus how good you know Facebook and X and things are LinkedIn is probably the worst for spam and it always kind of has been um but it's it's kind of interesting to see my my impressions are about the same as yours and again the only way you tend to crack it is by one post just outperforming that specific week, right? And it tends to be newsbait related stuff or uh really good kind of actionable posts where you're saying, you know, use AHF to do this that most people haven't seen before or something along those lines. Um but yeah, it's actually it's quite difficult to kind of consistently grow your LinkedIn engagements. I also found that that algorithm is different to most other social algorithms in that it's spread over a week. So your impressions for per post isn't the first few hours turned into the by the 24 hours, right? It's over the actual seven days of that post, which is really difficult if you're trying to kind of come up with execution plans to grow like viral clips and things. It's actually better to be consistent, like you said, with the kind of posting daily than I think it is to try and just be getting massive kind of engagement based engagement baked post every time. Dude, I've even noticed posts on LinkedIn blowing up after a month or two. Like, it's it's such a it's such a strange algo. Yeah, it's it's by far the weirdest search or sorry, it's by far the weirdest social algorithm out there. Um, and the waiting seems to change quite a lot. Um, and they do seem to be doing a lot of testing, but again, it's it's they don't have any kind of public releases or things like that, which the other systems tend to have. and even Facebook to some extent because they've had so many antitrust trials and TJ testimonial this and stuff. We've actually got a lot of insight into how their algorithms and back end work, but we still haven't got that from LinkedIn. Um, and I don't ever see kind of, you know, the head of LinkedIn being dragged in front of the Senate or anything anytime soon to be honest. Do you see uh many sales coming from LinkedIn? Oh yeah. Yeah. So, yeah, big time. Yeah. Nice. Are you I don't see not that many. No, no. I think uh mostly YouTube and X is like where most and and on the agency side a lot from Facebook ads but LinkedIn not that much. We might get like maybe 5 10k impressions a day something like that a day. Okay. But it's not that's interesting. I wonder why you're not uh you're not popping more from LinkedIn. How is X for you? It's very niche specific for sure. So we I I found as an example in I gaming so internet gambling and things like that. Most of the I gaming crowd are on LinkedIn um but then not so much on YouTube or Facebook or a lot of them are on X but a lot of them aren't as well. The biggest crowd is for sure on LinkedIn and that tends to kind of correlate to other spaces as well. So for for example, I've got a lot of uh education related consulting calls from LinkedIn, which is actually kind of bizarre, but you I I would have never thought pre getting these calls about this being a potential client, but you know, universities and ranking lists and uh small kind of degrees and there's all sorts of um kind of weird kind of ed higher levels of education kind of going on. Um and the SERs there are quite competitive and the the consults I get are mostly from LinkedIn. they find me from LinkedIn at least from from all of those kind of referrals that I've got. That's interesting because the conversations that I have on LinkedIn are also typically with brands that are like very serious niches, right? Like if someone runs a university, right? My case is like literally like a local practice for something. Oh, hey, I run uh you know this CPA alligator firm, right? I'm in this like hyper regulated because it's the kind of people that really spend a lot of time on LinkedIn on like proper like corporate level networking and more CMOs than business owners and more than also like agency owners, right? So less people like us basically and more people that are in like corporate spaces for like really serious type of industries. That's most of the the the crowd that I see on LinkedIn that then gravitate to people that putting the kind of content that we're sharing. I wanted to ask you Charles something that you said earlier about uh Twitter if I may Edward excuse me for the invasion of roles. This is good. Um you mentioned on on Twitter comments having a a much larger impact than than likes, right? And so from a from a principal standpoint, right? because you mentioned some of these algos you do have the weights I haven't had the opportunity to look at them so if you have any files that you can send me I'd love to do that right the correlation between signals that are closely tied with human effort versus signals that can be programmatically easier cheated what's what are you seeing there so the the waiting actually is interesting around comments because it's not just if you get replies it's if you reply to your replies. So, you have a time window where you have to reply to the comments that you get on your posts and X gives you a lot bigger boost and it's literally like a 5 to 6x boost versus just getting a reply if you reply to them as well in that window. Um, so I think that's actually really interesting from kind of a social dynamic that it's forcing you to engage more with your audience and vice versa and kind of creating this almost Q&A format I suppose. And it is going harkarking back to Twitter's original uh public town square kind of mentality actually which I think is is interesting in terms of AI effort versus human signals. I think the the best site we can look at so far that has actually done a good job of it at least to some degree is Reddit, right? because they get spammed more than any other platform when it comes to AI comments and posts and that kind of stuff and their automod and to some extent their uh moderation kind of systems and they have bear in mind they've used a lot of AI in that um and a lot of machine learning in the back end to do this um has gotten really really good to the point where you might actually get a comment to stick but two 3 weeks later it will get picked up and and actually automod deleted. So there's a lot of um stuff going on where I think they're trying to see how that account engages over time and it's not just that one initial post that they're trying to analyze. They're trying to look at pattern based behavior. So they're trying to look at if you are really actually a human and the only way that you can do that unfortunately isn't just by looking at one post or one comment. You have to look at how they're engaging with the platform over a specific set of time and then compare that to your data sets of what a human would do versus what a what a bot would do. You know what I find fascinating about that that you have described in Reddit is it's like the you know personal user account equivalent of a quality site. You know what's the frequency of publishing? What's the quality of your average page and your engagements? How much traffic do how much karma are your comments get? Because I had like actual accounts that I use rarely, right, that like I go and I comment in a couple of posts because one week I'm looking at something and then I don't look at anything and ready for another two months. I come back to my account and it's banned. Like what do you mean? I've I've done literally nothing. I just commented like the most inocuous thing two months, two months ago and now I'm banned. Right? So I'm similar to like, oh, I have a website, I publish a few pages and then I ignore it and then now the site goes down. Right? So it's it's quite interesting the the way that they've sort of like replicated one angle into the other Reddit actually if so if you're a moderator as well um you'll know this all day long but whenever you highlight uh or you basically you hover over a user's name in your subreddit they'll give you an AI summary now so it'll actually give you a summary of how that user engages with your subreddit as well as other subreddits on the platform. Um, and it's I'm pretty sure that that summary is how they're effectively uh kind of not only banning based on passive recognition, but also banning based on kind of actual engagement because it can say, "Oh, this user is self-promotional. This user is spamming this website. This user is uh actually an honest, cohesive, comprehensive engagement reply. They know what they're talking about. They're from this city because they constantly talk about, you know, some little city in Florida or something like that." Um, and that that kind of insight I think number one as a moderator you wouldn't have had anyway. But as a platform was basically impossible to generate a few years ago. Um, and now it's becoming way easier that I think again even you as a human can look spammy and promotional and body and stuff right to that sort of an algorithm even from small amounts of engagement. Have you uh ever tried creating a sorry no I was say have you have you guys like created a subreddit at all and tried marketing on the platform for that? Yep. So I own our link building. Sorry. Oh wow. I took Yeah. I took over it a few months ago and now we're ranking. If you do like best link building services we should be number one with the subreddit and things like that. Um, we've been doing a bunch of Reddit ranking using it, but also we're trying to be more subtly. So, there's there's some actually like platform level rules about promotion. Um, where you're totally fine if you, for example, like put your logo in the uh cover photo, but if you try and actively put like anything that is anformational subreddit and is then actively pushing a brand, that can get deemed spammy and have your uh subreddit banned. So, there's quite a tight rope you need to walk with it. But, if you own the subreddit and you're commenting and engaging, um, you can just approve your comments and they'll never get deleted or banned. So, it's it's a fantastic way to basically uh to be to market. And I do think that there's a lot of um subreddits out there that are either inactive but still have a fairly decent weekly business account that you can try and take over rather than kind of building your own because it's really difficult to get traction on new subreddits these days. Um, and the algorithm for Reddit is actually making it more difficult to push fresher subreddits. Basically, there is some kind there's there's almost a Reddit sandbox going on for new subreddits, right? Just like Google guys, they're all learning. Yeah, we um we started a new one in January last year. You're right. There's like this sandbox period and then the trajectory afterwards is insane. It's like it's growing. I've never seen anything grow month after month after month relentlessly like that. I I'm a big believer that it's based on positive user accounts engaging. So if you have a bunch of high karma accounts that have got positive engagement signals based like high quality score signals with Reddit's algorithm and they start engaging the quicker you get past that that kind of threshold basically. Um and but yeah and the amount of traffic you can get is pretty unbelievable. Like I think our subreddit the first month we started fully promoting it got 55,000 visits or something just on Reddit which was crazy, right? Yeah. Julian, how are you growing your subreddit? We just basically all our content we'll post on there and then we'll turn our content into articles as well and post it on there too. Um and like all our editors will post as well. So, it's like it's pretty easy to to create a lot of content and then we just found the more content you push, the the more it grows and the quicker it grows. Yeah. Is it ranking well? Yeah, I think it ranked okay. I think we might have taken a look at it like last podcast or something like that. It's ranking okay, but it's more like on the actual platform that it gets most of the views. But, I mean, to put it in perspective, like we went from zero like January last year to 300,000 views a month now. And it's like it's growing more and more each month. Yeah. There's a uh a friend of the podcast, someone who's been on several times, not going to say this person's name, but people will probably know who I'm talking about, and uh has his own subreddit, and he um he put up like a a post ranking like the best geo agencies or something, and this is like getting him recommended number one, and it's this post on his own brand subreddit, ranking the best geo agencies is kind of like a self-promotionalistical thing on his subreddit. It's doing so well. And I mean it's and it's crazy. And also like people are creating subreddits for competitor brands before the competitors have had a chance to create their own subreddits. The stuff that that is happening with Reddit and then it's so hard to get these things taken down is honestly insane. It's scary. the most of the foundational kind of link building and SEO SOPs will tell you to create a Reddit profile, but they don't tell you to create the subreddit, right? And I've seen that loads and loads of times where they've gone and create this the actual profile, but they haven't created a subreddit. And a lot of the times, if you create the profile and you don't actually put any posts or content or anything on there, then it's not actually that worthwhile of an asset versus if you build your own subreddit and it's active and there's posts and links to your site. And bear in mind, if you have a subreddit, it's a do follow from the sidebar. So profile links are no follow whereas subreddit actual sidebar links are do follow. So you can actually get a lot stronger signal from building your own subreddit than you can from uh from doing your own profile on there as well. I wanted to ask you Charles earlier uh when you were mentioning about the the quality of the accounts that actually engage with your subreddit right and since you've been tracking ready for a while you see any changes on the maturity of the algorithm since Google and Reddit did the OG partnership a good while back have you seen them basically refine things any basically any transfer of knowledge because it it's basically like nav boost in a way what we Yeah. So, I I I wouldn't say necessarily uh the algorithm. I would say anti-PAM protections for sure. There seems to be some there's there's a massive kind of uptick in ability and cap and capability um since then. And it it used to be the fact or the the occasion whatever that they used to use recapture. So, it used to be Google's own kind of capturing service to actually bypass things anyway. And I remember having to do Reddit comments and getting that capture and that's no longer a thing anymore. um it's mostly their own kind of anti-PAM systems and and pattern based things and stuff like that. When it comes to the algorithm, it's not actually changed from what we can see that much. Um and a lot of the waiting and stuff that they had previously still seems to be similar. Um and actually we found that from that partnership, it's more the other way round. So Google have actually got direct ranking signals from like Reddit up votes. And Jackie Chow put this on Twitter yesterday. I've talked about it before as well. Um, you literally if you can get a Reddit thread to have more up votes than the existing Reddit thread, you can often outrank that Reddit thread just by getting more up votes, right? Or more engagements, more comments, that kind of thing. So, there's definitely some correlation from Reddit side going into Google. And Google have, from what we can tell, switched off some anti-PAM systems or done something along those lines for Reddit itself specifically. Um, because you can get away with a lot of things where they're trusting Reddit's moderation to kind of do some of the heavy lifting for the Google algorithm. Basically, Julian, what what are you excited about in uh your SEO and marketing right now? Marketing in general, like YouTube is just is super fun and and Twitter as well. I mean, Twitter is like, I'll give you an example. So, one thing that we've done with Twitter is like we we look at all of our best performing content and we take the hooks and then we plug it into a custom uh clawed project, train it on what's working and also what flopped on Twitter. And then anytime we create a video, for example, or any piece of content or an idea, we plug it through the claw project with the transcript to the video. And it's optimized for what works and what doesn't. And so like we just we're growing it more and more to the point where I mean a typical day on Twitter will reach like minimum 500,000 people like in one day. Yeah. Yeah. So it's it's grown a lot. And um so that's pretty fun. And then also the same on YouTube. I'll give you another example. So on YouTube, we we did the same with the scripts for YouTube. So we have like a a process of like figuring out, okay, here's what worked previously, here's what didn't. Here's an idea, like literally an idea. Like for example, let's say Claude Mythos was was trending a couple of days ago. Plug that through the script writer on Claude and then I can easily like create a script in like one click that's literally like I wouldn't change anything on it. It's it's perfectly optimized. Um, and and then for example, like I did that on my main channel this morning, just recorded it through the teleprompter, the actual script from Claude. It's already had like 8,000 views today, something like that. Um, and then I So, this is a very scalable system, right? Cuz it's just like you sit you sit down for 7 minutes, you record the script through a teleprompter on your camera, and then you upload it or you can get your editors to upload it. And so I started a a new channel for the podcast. In fact, you were on it, Edward, right? Yeah. Yeah. And uh and so like now every day I record a video using that script on a trending topic on the other channel and that other channel is getting more views than a lot of my videos on my main channel. The other channel has 5K subscribers. My main channel has like 360. Um and so like the point is like it's just kind of like a cheat code for getting a lot of views, reaching a lot of people and just generating as much traffic as you want. Um, yeah. And it's, oh man, I mean, like, for example, Charles, you you knew me before, like my YouTube channel grew a lot and and so did you, Alejandro, and it was like it took me two years to get to 2K subscribers. Now, it's taken me literally a month to got to 5K. Just grow so fast. Yeah. So, that's fun cuz I can do that for me and I can do that for the avatar, too. For my AI avatar, cannot underestimate the value of the many reps that you put into it. Yes. Yes, for sure. It's like you you can already I already have a understanding of what's a good thumbnail, what's a good title, what's a good hook, etc. Yeah, I know the game. So, it's it's far easier, but to achieve what took me like two or three years in a month, I was pretty happy with that. Can you break that down again? like the step by step of how you got to you said it was 5k subs or yeah it was and and more reach than your than your channel that has almost 400,000 subscribers. Yes. So for example a lot of the videos that I create on the podcast channel they get more views than the videos on my main channel um a lot of them. And so the process is basically I find a trending keyword in AI because that's what I want to generate traffic to. I want to generate traffic to my community. So I'll plug that keyword for a Claude project like a custom GBT but on Claude and it has my whole process for writing a good script. So like for example writing a third grade language or for example um creating an amazing hook based on what's worked before or like adding open loops during the script to keep people watching and then halfway through how to have a CTA that's customized to the actual topic. So like it automates the script in one click from the keyword that will check it quickly. Then put it onto my teleprompter just read it out and it's literally takes me like 7 minutes to read out and I can read it perfectly from start to finish. Then I send that recording which is just a link to my editor. She gets a thumbnail and the video edited and uploaded and that's the whole process. And then we'll post X as well. Yeah. That's crazy. Also, the open loops, we've been working on that a lot with the podcast. It's like putting in We've noticed that the drop there's like a drop off around the 4 minute and 30 second mark. And so, like, we're working on putting open loops at that point. Like, oh, you're never going to believe what's going to happen next. And I got that. I saw that Mr. Beast does that like all of his videos. I'm like, why aren't I doing that? Ah, yeah. Yeah, for sure. I mean like he would be a great one to like you look at all his best videos, you you plug those scripts and the transcripts through Claude and you say based on the techniques that have worked here, apply it to my industry and write me a script. I should do that. I'll probably do that tonight now. Oh man. Yeah. That's sick, dude. That's so sick. Also, the reach that you're getting on X. Oh my gosh. 500K a day with And it's with AI content. Yeah. Exactly. Yeah. Yeah. Yeah. So, it's my team that that posts out the content and usually they automate the captions and everything else using the content from the video. The the weird thing is though, it's like I don't know about whe you've seen this as well, but like on Twitter, if we start a new account and do exactly that same process, does not work. Doesn't work like at all. But for some reason, there's so much momentum on the main account, the it can grow as much as we want it. It's really insane. It's just like the more we post, the more views and the more reach we get. Yeah, I have seen that. Starting a new account on X is a huge pain. That might be one of the hardest. Like if you compare YouTube, YouTube, you can get thousands of views from day one. X is super hard. Yeah. Sorry, Charles. You were going to say, no. Yeah, there there's a there's a hack that I wasn't going to cover, but I might as well for the thousandth episode. Edward, you get the you get the secret two source. I'm getting the goods. Um, yeah, it's it's not that secret source, but uh if a tweet doesn't perform, you just get AI to quote retweet it until it performs basically. Like that's the whole hack. So quote retweet has a 3x ampl amplifier of the original post. So if you can get AI to rewrite variations of that original post, quote retweets it, and if you have again multiple accounts, you can get it to quote retweet. One of those is hopefully going to amplify at 10x the original post uh volume. So you can basically kind of continually do that and if one post or if one quote retweet doesn't perform, you delete it and do it again through another account and it actually will tend to form on that account which is kind of what Julian was saying. Some accounts perform, some accounts don't. It's really kind of niche down with this new croc algorithm as well. So is that is that like you have to quote retweet it with a different account, not the So you can use your own account, but you'll have less amplification on the same account than you will on another account, right? It's it's reduced by like a third. So, it's still a large amount of amplification, but and obviously if you have more followers on the main account, you should still theoretically get more visibility if you have obviously like half because of the third less whatever maths being mass. Um, we want to try and ideally get it on other accounts, but you can use the same account. And there's also the timing gap as well. So if the first post was posted 2 hours ago and didn't succeed, then you still want to wait another 4 hours to give it the 6 hour time window. So there's a there's this whole kind of like time time thing that you have to do with it. But quote reaching on other accounts, you can do it at any point and it it really does uh amplify it massively. Oh, I might try that. How many X accounts do you do you have? Me or Julian? No, you. Me? Oh, okay. I I only have my main one, Press Whiz account, SEO careers, par SEO, a couple of other ones. I have like a couple I only have my main one like this one plus like seven more. And then I have some like farm accounts that we've set up as well. And actually, I have got account like I've got a accounts that we own but haven't been posted on yet. Right. So, yeah. Alejandra, you were saying something? Yeah, I was going to ask you Charles. So the quote retweet are you always retweeting the same OG post or chaining their retweets? No, always doing the same post bec so the if you do the same post it will amplify that post visibility as well. So so the quote retweet vis impressions count as the impression for the main tweet as well which helps get the main tweet additional visibility boosted essentially. Um which is why we always want to do the the main one first. you you can do chains to for user engagement reasons. So So if you're wanting to try and get somebody to click through onto the other quote retweet, then you can quote repeat that one and be like, "Look at this crazy [ __ ] going on, like stuff like that." Um, and you're trying to get users to click on it with other accounts. But again, this is where it starts getting into like really big scaled manipulation kind of stuff, you know? Yeah, world government's trying to manipulate the outcome of opinions on X or something, you know. um you know repost with your own thoughts kind of in LinkedIn you see the same no so not at all yeah so LinkedIn does and also LinkedIn repost with your own thoughts doesn't allow media attachments and um one of the number one things on LinkedIn for LinkedIn's algorithm is media attachments so having images carousels videos that kind of stuff is a massive amplifier for LinkedIn but the fact that you can't get that with the quote repost for some reason that just reduces the visibility and it doesn't have the same amplifier mechanics that a quote retweet would do. Do you uh you speak to so many marketers? What have you seen that's blown your mind recently? Oh, me or Charles? Yeah. Yeah. Oh, yeah. Oh, yeah. Oh my gosh. taking positive reviews and putting them with different language that people use to search for reviews of your product on the top social platforms. Instagram actually like recently has been doing a lot better uh within search engines um on or Facebook groups. Oh my god. Of a Facebook page. You take a review, you say like is your product a scam? this is what this customer said and you or or oh my in LinkedIn posts doing super super well and then what you're seeing you're seeing from first of all like just anyone who's doing due diligence on your products a lot of people don't even want to read reviews on your site and will be a lot happier reading the same reviews in a different place the same reviews except it's on your LinkedIn page or it's on your it's on your LinkedIn profile or it's on your Facebook page they're a lot happier reading it there finding it on Google and reading it there than they are on your site. That's number one. Number two is we are seeing LLMs do much bigger query fanouts. So for people who don't know what a query, all of you guys know what a query fan out is, but for listeners who don't know what a query fan out is, you put in a prompt, the LLM is going to a search engine. It's doing several different searches in order to get its information. That's a query fan out. But now, this doesn't happen with every prompt, but there's a growing trend that LLMs are doing more grounding queries. So, they're learning about brands from the initial query fan out. And then the LLM is doing more searches to look into those brands that it found. So, it's actually going and it's searching, it's doing due diligence on the brand. What is it going to find? It's hopefully you have a ton of positive reviews delusing the search results for your brand name and it's going to find all of this stuff that that you are putting that you are putting out. Whereas if you just have it in one place on your site, it's not going to do as well. So I'm for and just like in general, I am constantly shocked at how many SEO keywords this podcast ranks for. It's ins like I don't want to say any of them cuz I don't I found I don't know if you guys have found this but when you when you actually give like rankings and keywords they kind of just go poof kind of like so I don't I don't I don't like doing that publicly. Um but like there will be lots of times where I'm looking for a a certain image to use in content in SEO and I'll search that on Google images and it's like oh there's my face. I'll search it on Google. Ah there's my face. I'm just like that's crazy. is is how well social is doing or using reviews. That's what I'm really hyped about right now. And targeting just targeting query fanos in general, that whole process, but like yeah, repurposing reviews and how well social is performing. That has me so excited. Did you see Instagram's new real option? What's the new real option? So when you can upload reels now, there's an option that you don't show it to any of your followers. So it will only show Oh, the trials. The trials. Yeah. Trials. Oh, yeah. So people were people were abusing that for a long time. So what happened was there was like tri when you put out a trial reel, it would actually do a lot better than like normal reels. It would get pushed really hard. And so people would take their best performing reels, and I was doing this too, and just spam them on trial reels, put them up for a few days, take them up, put them up again, take them down, put them up again, and it was working really really well. But I I stopped doing that cuz for me, I just like and I used to repost a lot of my content, a lot more on Instagram, and now for me, I just try to put up one new piece a day. And like that's I think if you optimize for like one new piece of content today, like you'll be shocked what will happen for the average I mean all of you all of you are doing it but for like the average listener it's it's crazy the difference that it will make in your life. Nikita buyer has talked about bringing the option to X and then there's some talk of it going to Tik Tok as well. So, it could be interesting if kind of all the platforms start uploading it and you can actually kind of test content again. If it does just get abused where you're just re reposting stuff, it's not great, but I can see some really interesting kind of proof of concept style stuff that you can do to actually see if things perform in the greater uh ecosystem. Yeah, I knew uh I I do know some people who are like they're using um this software that takes the original video cuz Instagram knows, right? if it's a duplicate or not. Like it it will tell you, okay, you've posted a duplicate. We're not going to show it to as many people on a trial reel. So, what you can do is you can get this software that basically splices it up and just adds new filters and slight percentage differences so that it looks like a new video even though it's the original same video to Instagram. And then like once they find something that pops off on trial reels, they'll just create like 200 versions of that and post 20. Yeah. And that that works like really really well. Yeah, I I've seen people using AI as well to resize and reformat the videos so that it's for the specific platform, right? Which I think is really clever and I don't see why more brands don't do that. I see so many brands that are just posting Tik Tok videos to YouTube or to X or to LinkedIn and it's this same Tik Tok format into the Tik Tok bubble, the whole thing, right? But ideally, if you could just use an AI to rewrite the whole thing so that it looks like it's customuilt for that specific platform and then it's uploaded with the custom dimensions, it looks like a unique piece of content for the algorithm, the whole thing, you should theoretically get a lot better performance out of it than kind of uploading the same content to each platform or the same file at least. Yeah, I've seen like there's um there's platforms like Vizard, right, that take like an original video and it could be like a landscape video and then it will repurpose it for Instagram for Tik Tok and you can even get an API for it. So you could get one of your AI agents to handle it for you. And then you you have 10 different accounts for each platform and then you test variations and then we start HB testing get social media algorithms. I see. And then you quit SEO, shut down the agency. Yeah. One thing I wanted to ask on the on the query find out to you guys what you have seen there right because I started going uh deep into this actually when I saw your original video about it Edward um is I'm seeing a lot more of like site colon and then the keyword on the very first fan out and interestingly is the sites that we know like either seed list or like a jump away from seed list and and how just like it's just getting rigged like do you have the authority because if you have the authority you've already been checked like you're going to appear just just going to appear. The key is getting into that list. You guys seen that happening more in in more are you seeing that you're seeing that on the first query fan out or after like the after the initial like best of search for example I've seen it more in Y and well so depending on how precise how not precise the opposite how broadish the query is any anything else health health related right if you put it on the thinking on the research mode it's usually going to go like healthline site colon webmd and then look for your path which you know in in some niches I completely disqualify as anything that doesn't meet that bar in other niches I see as as a refinement but the one that you were talking about about the reviews I almost always see that as the second step whatever it finds whatever brands it is mentions it be like cool let me look at the reviews of these people and then I've seen cases where the reviews are literally not there right what like you google the keyword right and it says okay this these guys are appearing especially in like hyper local stuff these guys are appearing but This guy here has like really low reviews. It gets cut out from from the answer that the AI ultimately chooses. It doesn't have enough confidence. Right. Well, that's why I love So, Prompt recently shared the types of content that is doing best within Chacht. Listicles are number one by I don't remember the percentage. I maybe it was like 19% or something 20%. Then was SEO landing pages at 15.2%. 2% and then was SEO product pages at 14%. And I always share like I love creating SEO landing pages and SEO product pages because when you create one of these things, you are also saying why you are the perfect fit for the search term that you are targeting. These are sales pages that are made for SEO. And so when the query fanout is doing the due diligence with site colon and looking if you have these things, you've already created a page that has said why you are the best for that. And so yeah, I'm that's why like uh you're you meant to to also mention the site colon and Chris Long shared that. Um, but yeah, if you're putting up if you're or or if you have like let's say you're just doing listicles, but you have a ton of documentation on your site, like every company should have a ridiculous amount of documentation on their site about what they offer, why they're great at it, and all that stuff, and then it's going to be like a no-brainer. feel like these search engines, I mean, sorry, these LLMs are burning through so much cash and there's such like a fight right now to upgrade and update all the systems. It's almost like search is not really something they're that focused on. Do you know what I mean? Like for example, um you know, you got like Sora shutting down the video platform from OpenAI and then you got Claude who are trying to IPO this year. It's like I feel like, you know, for example, Charles, if you if you're saying there's a lot of loopholes or things that are working that shouldn't be working but have come back from 10 years ago, that's going to be wide open for quite a while. I I I think literally Google's mentality is AI will fix this, right? But until it becomes the default, so so it's it's kind of twofold. Number one, does AI fix this if you still need that algorithm as the sourcing algorithm, right? is is is if that algorithm is still sourcing the sources that that the AI is compiling and then summarizing you still are left open to manipulation from the AI outputs by the algorithm itself right um so it doesn't really solve it secondarily it's only going to solve it when you default the experience so when the entire experience becomes AI mode becomes AI overviews becomes whatever interface that might look like in the next year 2 years 3 years um that's the only reason and again if it's still using the underlying blue links algorithm to source it it's it's still got a quality issue um I think I've seen some site colon stuff in patents which was around like factchecking and validation for the LLMs so it's just making sure that for specific queries that that product pricing for example is the actual pricing still then it will go and double check by doing a site query and checking the pricing page um but I haven't actually seen it in initial queries that much yet. Um I would probably estimate that it we will move to a larger and larger and larger kind of summarization source as context becomes cheaper, right? So as kind of the input tokens become a lot cheaper for the AI overviews, the AI models, whatever it is, claude, chat, perplexity, whichever model you're using to to get that output from the search from grounding, um it will become you'll just have more and more availability because the context windows will be bigger and they'll be cheaper to to use. So, I only see there being more sourcing and more summarization from those sources, which is only a good thing for link builders and publish publishers and off- page SEOs because you need more and more consensus. Um, and the only way you do that is by building from from third-party publishers because at some point, even if you're doing a multi-dommain strategy, if they are actually using all 100 results from the original sourcing algorithm, it doesn't matter how big your multi-dommain strategy is, you need a 100 pages to try and tune uh to what you want the output to say. Earl, what do you think SEO looks like in three years from now? It's uh it's really difficult, right? Because you have to predict what Google will do. Um, and the only kind of way you predict their movements is if you're like tuning into their earnings calls and you're seeing what the company decision-m is going to be, right? Like what their actually end goals are because they don't lie on calls. They will they'll actively tell you this quarter we're going to try and raise our KPIs for our search ads by 25%. Well, where are they going to get that 25% guys? You know, like it's coming from organic, right? you you kind of quite quite realize where the company's trajectory is based on what their corporate decision-m is. Um and unfortunately their corporate decision-m the last year, the last two years has been very reactionary. So it's been based on okay, we're losing to Open AI, we're losing to Chat GPT. Let's try and do anything we can. This is code red. We're going to try and actively take over that market. And to some extent they have had success but I really think Anthropic and Claude have outperformed them to to a large extent and they've had a lot bigger market share taken from Claude and Anthropic than Gemini and Google have been able to do. So has that investment shown an ROI? I probably don't think it has and I think that they will at least in some way change and again be more reactionary going forward and kind of change tact. So, I'm not 100% sure what the search interface looks like in 3 years, but I think it will look like whatever Google's profiteering eyes are showing them, it will look like basically wherever they see dollar signs, that is where the trajectory of that company is headed 100%. And they've shown that routinely for the last the last decade or so. You're very Oh, go on. I was going to say they actually own a percentage of anthropic as well, right? It's like 10 or 14% something like that. Yeah. Yeah. Yep. So again, they they've actually hedged their bets as well and wherever open air have lost space. It does seem like at some way or some corner Google has taken some of it away from them. But again, have they had an ROI from that? Have they been able to make, you know, the profit from that investment at least yet? No. Nowhere near. They've invested hundreds of billions of dollars and they have seen maybe $10 billion in revenue back from their AI product. So we'll have to see in the next few years, One thing that I've been thinking about right with with Google and what you were pointing at like the island like and you know when HCU helpful we all know how helpful it was right and we see what was happening there like you know the DOJ declarations and the internal emails that were left of like no we have to make more money from ads we don't care if the quality of search goes down right and now what's happening with AI. I think one of the challenges that they have is that they're trying to make the AI for everybody. I be like Chant, but you know, Chibd is being bankrolled by basically everything. And you can see that was like on their crawl graphite. Chantity will hit a website and eat 40,000 pages of garbage. Google reach three and says, "Nah, all right. I I have what I what I need." Right? And then you you look at anthropic models like I'm sorry, I don't think it's even a fair comparison. and you Gemini 3.1 against Claude Opus 4.6 is like makes Gemini almost look like you're working with GPT40. It's just ridiculous that the jump, right? But I think it's also like because anthropy had chosen like no, we're going to go this direction and then narrow it down and then obviously the models is here. But I think Google is trying to like have AI for absolutely everyone and I'm I'm not sure how that's going to play out. again especially if they try to take the the rag pipeline of the actual organic results which if they move all the attention into AI the organic results will go down and then there's a disconnect between the quality of the source so like yeah I don't know what happens there what do you guys see the guys that I see winning are uh Nvidia out of all of this it's like whatever happens they make money it's like they own every single layer of the infrastructure and it's like you know if If if Gemini or or Claude win, no problem. They're they're going to have more hardware to sell. If if Nemoclaw wins and all their local models win, it's great because people will need more compute and power. Um but yeah, I I agree with you. I mean, like Chat GBT as well, they're burning through cash. Like they're they're forecast to like run out of cash within two years, right? Um so it's it's looking pretty bad for them. I think both chat GBT and Anthropic are planning to IPO this year. So that'll change things a lot if that happens. Yeah, I think OpenAI have changed their tune so much which is the problem. Like originally no NSFW videos then you know you can make as much as you want. No ads going to roll out ads to be the first AI company to do so. there's all this kind of flip-flopping from Sam Alman. And I think again he's he he was the set the trail setter, the the trailblazer, whatever you want to call it. Um and then quickly he's now turned into this reactionary mode which again didn't lead Google down a good path. I don't know why you're trying to go down this path either. Um Anthropic have just kind of sat back, focused on enterprise and cleaned up the cake because of it. Um and I and you do really like you were saying Andrew, you really see the difference in outputs. Even when you ask very basic prompts, you know, like make a documentation SOP PDF about this simple thing for SEO, it will come out as just a little prompt in Gemini. You get an actual doc in claw, which is what you ask for, right? You don't get the same experience in certain AIs that you do in other ones. And I don't see Meta really competing to be honest. Manis is just way too expensive now. and they don't really have uh frontier models anymore. Like their models are multiple generations behind basically everybody else and the open- source models aren't competing with the Chinese models anymore. So I see consolidation happening industrywide and I think that was preAI anyway. HCU for me was more just consolidation. It was just more okay, who's the biggest three benefactors? Reddit, YouTube, Kora, right? Well, who's the biggest losers? independent publishers and experts and actual people who are creating quality internet content and arguably the same content will rank on Reddit and YouTube but it won't rank on the experts blog. Well then there's there's HCU or the name doesn't actually translate into anything in reality. Um so I just see it continuing like that and I do think anthropic will probably win the enterprise race. I think Google will probably win the image video race and I think chat GPT will just be the default experience for most people because they've won the name. They have like my little brother and sister will say, "Oh, go and ask chat, right? They're not they don't mean live chat. They mean go and ask chat TPC." It's the same as let's go Google it, right? It's the first one to kind of have that uh societal dominance. They become the the default experience. And I do see that happening. But again, can OpenAI turn that into an ROI? Right now, they it's it's a negative on that so many broad people are using it and burning through all their tokens every day. They have they're having to shut down store for a reason. Um, and I think that that broad user base will actually come to bite them in the future because anytime they launch anything, it's used by millions and millions of people. Anytime Claude launches anything, it's used by the people who power users and experts who are going to give you the feedback and the beta and the suggestions and stuff to improve it. whereas it's just going to be a massively broader user experience for for chat. But they could also that that does mean that they can sell data. So you know the the ultimate experience data is the ultimate commodity and I do think that chat will have more data than than anybody going forward probably just because of that experience. Like nobody goes and puts their like therapy questions into Google but people do it into chat GPT all day long you know all day long. Charles what what aspects of your SEO are you automating? Uh also so we try to be more of a manual approach for highle campaigns um until it's kind of necessary right so most of the content for like guest post and things like that that's going to be all automated most of the topic creation alignment making sure that as an example we did a um I gaming campaign the other day but they wanted to go on a bunch of the sports domains so we have to overlap that something to do with this these sports players and gambling, right? And we get AI to kind of create that topic and they could the AI can do back research. It can research the players. Do those players have any links to other things and stuff like that. Are they from specific countries that maybe Backerat was invented in or something along those lines, right? There's all sorts of things that you can go really nitty-gritty granular on to get that content really uh nicely tightly um linked over and and overlapped. But um for most other things, it's I would say 90% human, 10% AI. So I'm trying to get my team to go and do the work and then ask the AI what did I miss? What could I improve? What could I update? What could I add to this? How could I do anything else? Have I could I use a specific tool to make this better, faster, more efficient, get more data, all that kind of stuff, right? Rather than having the AI do stuff for them, I just want them to basically augment and improve their work. Um, and from the people who listen to me and actually do that in my team, they fly, right? Whereas the people who try to get AI to do stuff, it's very obvious, which is the other problem. And we just fire them within, you know, very, very quickly. I'm I'm one of those mentalities where it's higher fast, fire faster. Um, just to make sure that you have people that are in the correct positions and that are going to be able to scale with you because otherwise they tend to be detriments to the company unfortunately in the long run. Julian, I saw you shake your head when uh Charles was like, "Yeah, 90% human, 10% AI." But you're like the AI person. Yeah. Well, it depends. Like, for example, if it's anything for me personally, like content, um, SEO, blog post, anything like that, I'm totally fine automating it. When it comes to our clients, though, it's like you have to be so careful. And I think that's probably where Charles is coming from, too. It's like we a a big cost that we have is like quality approval, quality checking, quality control. Um and this is super manual. Like we can automate outreach all day long, that's no problem. You know, sending out emails and stuff like that. Um writing the the outreach email templates and everything, that's totally fine. But if it's like content that's going to go on a client's website or if it's for example even like the anchor text, you know, and the page and and the content that we'll insert into that one sentence where we insert the link, we we have to be careful on that stuff because if we're not clients will see it and it's like I'd rather we check it than our clients check it. You know what I mean? So, um, it's one of those things where it's it's a little bit frustrating cuz it's like I know how to automate every single part of that process, but I don't want to because at the same time it it's about quality, right? It's like I don't want to be producing those AI slop whether it's link building or content for our clients. So, you do have to be careful on that side. However, from my side, for my website, totally fine. Yeah. Exactly. Exactly. That's what they're paying for, right? It's like um but on my side like I said like if it's articles on X, if it's articles on my website, if it's ranking content on Reddit, no problem. Like I don't mind. Just build really good prompts. Are you automating anything? Yeah, fair bits. Um I got a a a new obsession around rack pipelines and just um it's it's breaking my machine but I'm trying to basically collect as many data points as possible for any given thing that I do right so if I'm looking at a website right normally you would go okay set console let me get some exports ahs let me get some exports let me run a link gap up analysis let me look at this page through maybe a custom GPD for uh on page quality. Let me write through pop. And then you you still have to assemble all that together and come up with a strategy. Okay, well this page is going to need this. Let's prioritize these pages. Right? Basic, right? Like basically like macro level thinking of SEO. And then you also have to put into context what the clients are saying. Hey, we want to focus on this cluster of pages because…

Transcript truncated. Watch the full video for the complete content.