SEO Jesus: Link Building, AI Agents & Scaling an Agency!

So we've got Paul Truscott here. Welcome here. Hi Julian. Thank you very much for having me.

Thank you. And you're working, I know we were talking before about local SEO, semantic SEO. I'm excited to get into it. What sort of stuff you've been working on recently?

Okay. So recently it's more, um, overviews, LLMs and how they behave. Um, but one of the things I've realized in pursuing that is that all of this to you, to, to grab one of Corey's phrases, all of it's holistic. You can't separate semantic SEO from our overviews from LLMs.

And I put all of that in a local context. It even affects Google. My best, um, a lot. I had a surprise about two weeks ago, some tests that I was doing to manipulate the overview ended up failing in that regard, but put the Google, my business listing of this particular client into the three pack where it hadn't, it wasn't anywhere near for this series of keywords.

So this cluster of keywords that they weren't ranking at all for. Um, so yeah, you can, you can end up getting them a surprising result. So disappointing result on the overview in that particular instance, but a really surprising, nice result for the Google business profile. So all of these things are interlinked, certainly at a local level they are anyway, which kind of isn't surprising because you, you're dealing with the entity.

Now you're not dealing with a website or a Google business profile or a video. They're all part of the brand entity. So it's no surprise that whatever you do to want to affect one is going to affect the others. Yeah.

This is something I've seen with social media as well. It's kind of like the barrier between SEO and social media is like closing in or kind of just links inside one ecosystem for marketing that. Yeah, it does. And it's great for people who've got a background in marketing as well as SEO, because they will get it a lot more quickly.

And I think if you've got a background in marketing, if you've got a background in SEO, and if you've studied semantic SEO, like we were saying earlier, I tip my cap to Coray, like I've studied pretty much everything that Coray has published anywhere. I've read it more than once. And eventually the penny drops with me. I'm a bit slow, but the penny drops eventually.

I'm like, okay, I'll get this now. And you can apply this like, well, let me backtrack for a second. There are that many SEOs out there who actually rubbish Coray. And the reason they do it, I believe is because they're just jealous because he's figured some things out that is above their head.

And so rather than actually learn from the guy, they just want to say, oh, nothing much has changed, which like is ludicrous. It's absolutely crazy. It's crazy. I mean, like, yeah, we had him on the podcast earlier today and he's just, I mean, even the way that he approaches SEO is just so fresh, I think, compared to what everyone else is doing.

And it is so much more complex. And I feel like we have a temptation as SEOs to like simplify, but if you look, it's like all the tiny details. Yeah. Yeah.

I think if you, if the temptation to simplify, it may have worked in the past. I think if you go down that path of least resistance now, you're going to get stuck because the reality is it isn't simple. Google aren't spending tens of billions of dollars every year to produce a simple system. You wouldn't do that, would you?

So it's naturally inherently complex. And so decoding that isn't easy. And like guys like him who are willing to, as I said earlier, off record, guys like Coray who are willing to sift through patents and read them, and then are willing to do a case study based on it and actually, or test it and then do a case study based on the results. Like what a godsend to have somebody like that.

And the SEO industry should be grateful for it. But a lot of them aren't. I've heard some horrible things said about him, but it's all out of, it's just jealousy. They don't understand what he's teaching.

They're not willing to put the time in to learn it. And so the easiest way is to rubbish someone and say, oh no, he doesn't know what he's doing. Yeah, that's it. And I think as well, even sometimes if someone doesn't have that much knowledge and, you know, Coray's like black belt level SEO, obviously, but if you, if whenever I see someone doing SEO or whenever I meet people at events, whatever, it's like, I can learn something from everyone in this room.

Like there's something to learn from everyone. Yeah, I think that's true. I think that's true. It's surprising actually, or maybe it's not surprising that you can learn things from people with very little experience because they're not, because they don't have any, so they don't, they don't come with any assumptions.

So they'll test things that perhaps we would not even think of testing simply because we already know it doesn't work. And then they come back with a surprising result and they make you have to think again. Cause I've had that, there's one guy in particular, who's a carpet cleaning guy in Coventry here in the UK. And I showed him a few things, maybe three, three years ago or so.

And he went away and did some testing and came back with like, he sort of expanded upon what I'd shown him and came back with these test results. And I was like, well, I'll be damned. I never would have even tried that because I didn't think it would possibly work. So yeah, some people can surprise you even, even though they, they haven't necessarily got a huge amount of experience, but they're willing to test stuff.

And I think at the end of the day, that's one of the things I love about Dooley, like he just test, test, like, and that's the only way to really find out. That separates the theory from the reality is does it actually move the needle? Yeah, yeah, for sure. Well, tactically, what's your approach to local SEO right now?

Or what's your approach to AI overuse, for example? Right. Well, it's, it's kind of multifaceted now by, by need, you know, because you can't rely any, any more on being a specialist. Like I used to only do Google My Business when it was Google My Business.

The only use I had for a website, like my websites, I wouldn't even show to people because they were horrible, but they were constructed purely to move the needle for the Google My Business system. And I understood how to do that at the time, it was actually quite simple, but there was a specific way to structure them that got results in Google My Business. But the websites themselves didn't have a hope in hell of ranking anywhere, because they weren't, they just weren't structured properly for that. And then I guess it was around three years ago, coincided with, with, with finding Corey.

And that was a friend of mine saw his interview with Matt Diggity, and said, Do you know who this guy is? So now I've never heard of it. This is some Turkish guy that's doing this, you know, he's, he's got this interview with Diggity, and it's like fresh, it's, it's like, I've never heard anything like this. So I watched it.

And it just made me think, maybe this is the answer, because I'd already realised that I had to move into organic, because everything was getting interwoven. You know, Google would like Google wanted to rank brands, we were already seeing the brand, the amount of brand search in the search console was affecting how you rank. And so I couldn't ignore it anymore. Because my results for lead generation were telling me that the old black cat methods of, you know, verifying GMBs and spamming the heck out of everything just wasn't getting the same results anymore, like it used to.

So I had to, by necessity had to start learning. And that took me down the road to where I am today, which is where I'm, I'm now, you know, exploring AI overviews, and how to the cheapest ways to get them ranked, because just like, just like Corey talks about cost of retrieval, and how it's important for Google to keep the cost down, I primarily do lead gen. So same, it's exactly the same mentality, you know, we can't, I can't go spending 1000s of dollars on link for the assets we have, because the payoff isn't there. So you've got to find cheaper ways to rank.

So it naturally forces you to, to be a little bit more innovative in coming up with ways to rank. Yeah. So what I found, though, is that all of these things are interconnected. So I'm learning as much as I can about everything.

But if I'm honest with you, the whole foundation, even Google My Business now, or Google Business Profiles, I can't help calling it Google My Business. But even that side of things, the maps is governed by semantics. Now, too, you know, it's you've got the proximity and everything there, of course, that's something that's unique to the maps platform, you don't have that level of proximal sort of ranking factors that come in with organic, organic, I find basically every city, like if you look on Google, and just Google the city, and it gives you the red dotted red line around the boundary, with the exception of some larger cities, London is different, Manchester is different in the UK and Birmingham is different. And in America, New York is different.

I think New York's the only city in America, because their system's a bit different. But with the exception of those big cities, basically, you draw that line around the city, that's one search for organic. So it doesn't matter where you're searching from within that boundary, you're going to get the same result, you'll only start changing when you go outside of the boundary. London's London's carved into boroughs, Manchester's carved into suburbs.

So it does depend in Manchester where you are. But most cities in the UK, other than Birmingham, Manchester and London's from what I've seen, they've all got one single search, which of course, isn't the case for Google My Business or Google Business Profiles, the maps is, you know, the proximity comes in, it's such an important factor. Interestingly, though, I'm noticing not so much when the AI overview is there, like if there's a maps AI overview, that is pulling in attributes more than proximity. So it's factoring in whether you're a good match for the query in terms of the attributes, your service or company attributes.

So you can, we're kind of going back to the old days a little bit now, I think where you can overcome proximity by just doing things better than other people covering those attributes. This is where again, knowledge of what Corey teaches with the EAV triples, understanding that gives you a massive edge in Google Business Profiles. fast now. What is that for anyone listening?

What does that mean? Or what does that involve? So entity attribute value triples. So for those that don't know, Google's knowledge graph is essentially a massive repository of triples.

So that is things that are joined by the two nodes and an edge, right? The middle part of the triple being the edge that joins those two nodes. So if we think of an entity attribute value, that is an entity which is connected to a value via some attribute. So for example, if you took the entity car, one of the attributes of a car is the tire.

And the tires have different sizes, different widths, different tread. Those would be values for that attribute. So that's how Google stores information in those triples. It does it with what are called SPO triples as well, which is subject, predicate, object.

So you connect a subject to an object via a predicate. So the predicate becomes the edge between the two. And that's how they manage really complex, huge volumes of entities and connections between entities by storing them in triples essentially. And so if you think that way, because that's how the machine processes, you get more out of it.

You get a better understanding of how to put something together, what you should make the subject, because whatever you make the subject is going to be more salient in that content. So just understanding how all of that works just gives you a massive advantage when you're putting together, especially for AI overviews and for LLMs pulling that information. Like how you structure a sentence can be the difference between getting them to cite you or cite someone else. And then when you're creating the content, because you're creating it for like GBP profiles, are you creating the content for on websites as well?

Yeah, yeah. And in fact, the website is where the vast majority of the, what's the word we're going to use? The secret source comes from, if you like, right? That's where most of the magic happens is on the site, because that is the entity home, really.

The website, if it's at least it should be, if it's structured properly, that is going to be the canonical that Google refers to and says, okay, this is everything about this business that we can, or at least be expected to believe is on that website. And so everything else has to be congruent with it. So yeah, so it's, I always think of the, I like to mirror everything on the Google business profile, right? It's very important to mirror it.

So in a way, if you think of the Google business profile as being a synopsized version of your website, because you can't put as much information on there, but you put a synopsis of everything, you're building out services for every service page, your description should, well, I believe it's worked for me, make your description the, pretty much the same as the subordinate content under your H1 on your homepage, then they match. And then Google gets this congruence. So there's no incongruence between all of your entities. If you do this everywhere, wherever Google goes, they said, okay, yeah, they're consistent.

And then after that, it's all about third party. I don't know if you're familiar with Jason Barnard's work, but you know how he says, claim, frame, prove, right? You claim it, you frame it, but then you've got to prove it by having third party validation. Yeah, we had him on the main channel actually a couple of years ago, Jason, I should get him on again.

Yeah, yeah, definitely. I think right now with everything that's going on, Jason is like, should be front and center of everyone's world, really, in terms of, you know, putting what he knows together with what Corey knows, you've got most of it covered in all honesty, between those two guys, there isn't a lot. And there isn't a lot else that you need to know if you really get in depth, like, for me, the key, the thing was understanding when I, when I was pulling Corey's work apart, trying to get my head around why it wasn't working in local because it didn't work for me. And then when that penny dropped, which was actually only fairly recently that I, I managed to get things working, but I didn't understand what, what, what part of what I'd done, it was that changed.

So I figured that out now. And so since figuring it out, well, I figured it out until Corey burst my bubble, when he comes along and says, Oh, no, no, no, you need to consider this then. But until then, I'll, I'll claim and think that I've, I've worked it out. But yeah, so there's a lot of things that you can do in local, they're a bit different, you have to do them a bit different to how you would approach.

I mean, I could go into that a bit if you want to, but it's like, yeah, it's, it's just, it's easy to go wrong with, it's much easier to go wrong with Corey's approach in local than it is in a broader topic, put it that way. So you can like, because you were saying before, I know, before we went on the, on the show, now you were talking about how one idea that you had for a topical map was actually totally different in reality, right? Yes, yes. Yeah, it was to do, it's to do basically, with Corey talks all the time, or all the time, he's spoken up a number of times about most of the topical maps that he sees are not topical maps, their concept maps.

And that is because people build them based on the dictionary definition, right? So they build taxonomy, taxonomies and, and ontologies based on dictionary or encyclopedic approach. So they'll build out all these pages. And they've, they've got fantastic topical coverage from a dictionary perspective.

But the problem is that Google isn't a dictionary. And so Google's world is based on what people searching. So you have to look at query semantics, not dictionary semantics, because if people aren't searching, like if you've got, if you're processing an entity attribute pair on page, but there's no representative query for that entity attribute pair, then it may as well not exist as far as Google's concerned, it's a waste of time processing it, you're actually just bloating the site. And that's what I did, I would increase the site radius, I would increase, like I would dilute my site focus, and I would be bloating the site with pages that are going to get any probably not even impressions, nevermind click.

So just shoot yourself in the foot from day one, which is which I did numerous times. Yeah. And there's also like content velocity to think about as well, right? So I guess you're creating too many pages, then there's also a temptation to create too much at once, which you kind of have to balance already.

Yeah, especially now with, with, with AI, like, you've got to be particularly careful with velocity, because that's like, you know, I think Google understands if you're a new publisher, especially in your small, you know, you start with a small site, and then all of a sudden, you're publishing all these pages, that's almost the invitation to come and smack you, you know, it's, it's so obvious that you're using AI to do that. So that's a dead giveaway. And I think Corey's talked about this, that it used to work, you used to be able to publish really, really fast, and get traction quickly as a result. And now you're likely to just get a penalty quickly as a result.

But this is where his visual semantics idea, I think Google are probably turning more to that now, because it's easier to focus on that than it is to try and detect that content. So maybe Google is saying, well, to hell with it, we don't care whether people use AI. But if we look at the amount of effort that has gone into being not clever with the site, but specific with it, so that as I said to you off camera, that I've been dealing, I've been working with something for the last year and a half that I called visual task nodes, which is basically, to me, like, a user is happy when task fulfillment is achieved, right? So they've got whatever they're trying to do with your page, or whatever they're trying to do when they search.

If your page gets that done for them, that's task fulfillment. So I'd separate the page out into a series of task nodes, and a lot of those task nodes would be visual. So I didn't understand anything about visual semantics. I didn't even know the term existed until recently when Corey's put it out there.

But it's kind of the same thing, like Google are looking at the structure of the site, and does the structure of the site meet the search intent? And that's probably easier for them to tell. Because the temptation now would be just throw up an article, don't put much time and effort into the structure of the page, but it's got with bang content on there. I mean, everyone's an expert in everything now, because you can just use AI.

You can write content so easily. Yeah. So I don't think it has any value anymore. So then they have to turn their attention to what does have value.

So maybe the structure of the site, and it's harder to fake that. Yeah. And then like, I mean, for example, one case study that Corey was showing was one website was like a tool, like audio to text, for example. Yeah.

And I grew from like zero to 8,000 clicks a day or something like that. Yeah. That's crazy for what is essentially... Well, it's not...

I mean, he said single page site. It's not really, is it? Because he's got a page for multiple... I think there's 11 languages.

So he's gathering clicks from all those languages, but still that's impressive. And yeah, he's basically signaling to Google immediately that the search intent is going to be satisfied, because the exact thing that you're searching for, you can do immediate on that page. The one thing I did... Sorry, carry on, Jimmy.

Yeah. And I was going to say, like, so you use visual semantics to do that, right? Like in terms of the structure of the site. Yeah.

And what was particularly interesting when he said he had, I think, the area where you could upload your file to was the full width of the page. And that was on purpose so that you're occupying more pixels, so that you're signaling to Google that this is important. Yeah. So all of those little nuances, they're like, it's easy to miss that.

And I would have missed that. Had I been reading that case study two years ago, I'd have completely missed that point. Yeah. Another interesting point he mentioned was, for example, when you're building that website with AI, if you're using AI to build it, then it's kind of...

the same code that all these other AI website builders are using. So you kind of have to clean out as well in terms of the CSS, the HTML. You've got to clean it and also put in some features that other people just aren't doing, which even if you're using AI is still human effort. One second, one second.

No worries. My TV is threatening to switch off. I've got to just press a button on it. It's all good now.

Yeah, that's it. Are you testing anything with that? Are you building out or experimenting with visual? Yeah, we have.

I have done. I'm going to get a lot more aggressive with it now that I'm learning a bit more. And once Kari's course comes out, probably going to be a game changer in terms of my testing because he'll be teaching me things that I don't currently or haven't currently thought of. But one of the things, for example, that has worked well is taking the SERP filter chip.

If you look at a transactional search, you almost always get SERP filter chips come up. So I'll take whichever of those are relevant to my particular service, and I'll use those as jump links in a sticky header so that I'm signaling to Google that I'm covering these attributes that Google has determined are important. So I think they, I'm not a hundred percent certain, but I think that they determine those from Google business, like from maps, because people aren't searching these things. You see top rated, you see reviews, you see prices open now within some, whatever range, within two miles, within 0.2 miles, within half a mile, whatever it is, you see things like that.

And online estimates is a common one as well. So what I would do is signal with the jump links that I'm going to be covering all those things on the page. So I'll have jump links that say reviews, opening hours, service area. So they kind of reflect what Google puts in those SERP filter chips, if that makes sense.

So by doing that, you're telling Google, right, hey, nobody else bothers to do things like that. And you're also signaling to Google that you're covering these important attributes on the page. Yeah, that's really interesting. Actually, that's one thing I was going to ask was with all of this, because it's so in depth and it's all about like the thousand tiny little details when it comes to semantics, how do you systemize that?

How do you get a team to implement that for you? It's yeah, that's not easy. And like we've used, we're transitioning now to doing a lot of this with automation, as much of it as we can. Topical maps, I don't think you can automate for the reasons that I said earlier.

I think once you start trying to automate a topical map, it will by default become a concept map. Because unless you will, you could do it, you could do it if you can pull in, you've got to be able to pull in SEO data so that you can look at search volumes. Because without knowing the search volume or having some indication that there is search demand for the canonical query, then you don't want to be building pages for it. So you'll just go, if you don't have that information, you're just going to build a concept map and you're going to get the same really, really terrible results I got for local.

As your topic becomes broader, the difference between a concept map and a topical map will diminish if that makes sense. Because even if you just build, even if you just go by AI and let it build out a concept map for you, the chances are that a lot of those topics it comes up with, those entity attribute pairs are going to have some search demand. But in local, the chances are they absolutely won't. Yeah.

So you just end up, you've got to be a lot more, you've got to be a lot more precise with local to make sure that everything you're doing does have search demand behind it. And then if there's something that you need to cover on a site to bridge a topical gap, just find a page that you could put on and bridge over to the other page. You don't have to have a page for it. If there's no search demand, I just don't open a page for it.

And that seems to work pretty well. I'm sure it can be improved. But yeah, to go back to your original point though, yes, actually systemizing this is a nightmare. And I'm not even going to pretend that we've got it down because we haven't.

I wish, I probably would, if I was you, I probably would have done by now because you're a bit always with the automation, which I'm not. Oh, you know, I was thinking like, there's probably so much detail in there that even where we've called and, and, you know, these fable five and frontier models like this, I still don't think that they'd be able to, to implement it to the point that someone like you or Corey would be satisfied. Well, I look, I read just from the content perspective when I get, you know, I put together systems in Claude that do a pretty good job of writing the content when we guide it properly. But even then, when I read it back, I'm like, Oh, that's got to change.

So scaling, it becomes more problematic. Although I mean, I know you said this to me before, you got to stop being a perfectionist because it's, it kills you, you know, but there's a part of me that always wants to tweak. And, you know, so I've got, I've got to, I'm getting better with that. You know, I'm getting better.

I think that's what attracts us to SEO in the first place, isn't it? It's, it's being an optimizer. It's being someone who likes to tweak with algorithms and That can work against you, you know, in some instances, like you can go too far with it. So knowing, it's a bit like an artist, I suppose, knowing when the painting's finished, you don't need any more on that one because you get to, it's like Pareto's principle, isn't it?

You know, you get to a point where you've got diminishing returns. Yes, you could improve it, but are you just better off letting that one go now and moving on? Because you'll get more out of the 80% of the next one than you will out of the 20% that you're going to fiddle with, if that makes sense. So, yeah, but I'm not good at doing that.

I will confess. I'm, yeah, I get stuck on, especially if I watch another Corey video, I'm like, Oh no, no, now he's giving me, now I can go back and tweak this. Like, yeah, it's just, yeah. Yeah, I can imagine.

I can imagine. Yeah. It can be a bit of a mental nemesis. Awesome.

Well, thanks for being on Paul and sharing. No worries. It's been, yeah, it's been a pleasure to chat with you and, yeah, anytime I'll be more than happy to talk to you again. Yeah.

Well, it'd be great to, you know, once you've experimented more with visual semantics as well, see some of the case studies or some of the data you get back. Yeah. Yeah. I'll come, I'll come with, with some actual data next time after we've done some testing and then we can give some, some actual evidential proof that things are working and that'd be better for people.

Awesome. Is there anything we can link to in the description for you? No, not really. I don't have anything to sell.

I'm just happy to help people. There's, I don't have a product or anything that I'm selling. Most of my work is lead generation. So there's nothing that I'm trying to, I'm not looking for clients.

I'm not trying to, to sell anything. So, but people can get hold of me through LinkedIn if they want to, or I'm on LinkedIn so they can, they can get me that way. Awesome. We'll, we'll link to that.

And yeah, just thanks so much for coming on and sharing everything. And it's very rare that I don't think I've ever had someone come on and say they haven't got anything to sell. So I really respect that. No, listen, I'm just, I, I've been given a lot of help, you know, over the years by people.

And I like to just pay that forward wherever I can. So if anyone gets anything out of this, that's fantastic. Awesome. All right.

Thanks so much. Thank you, Julian.