Semantic SEO for Local Rankings Explained | James Dooley Interviews Panel

James Dooley: Semantic SEO strategies for local SEO.

Today I'm joined with Mike Lovat, Paul Troscott and Luke Baston, who are absolutely incredible when it comes to semantic SEO in local SERPs. So let's get it started. Paul, I'm going to throw straight over to yourself. I'm going to throw you in the deep end.

Paul Troscott: Thanks, James.

James Dooley: For local SEO specifically, with regards to semantics, there is obviously a lot of confusion. There are a lot of people talking about semantic content networks, topical maps and stuff like that, but does it actually work when you're looking to do it in local SERPs?

Paul Troscott: It does, but what I have found is that you have to modify things a little bit.

This will depend on the niche that you're in, but in a lot of the niches I work in, once you start looking at the subtopics, they do not get much search demand. The idea of trying to use an outer section to build historic data does not work. You can end up building a site that is too big and gets no traffic, or at least the outer section of the site gets no traffic. What you have done then is dilute the site focus. You have also increased the site radius, which we know from the API leak is an issue. What I have found with local is that unless it is a niche where there are a lot of subcategories, like plumbing for example, where there is a lot of search volume around subcategories such as boiler faults, it may not work. People search for things like that, so maybe you could do it in that niche. But in a lot of the stuff I do, there is not much search volume outside of the transactional stuff. There might be one or two informational queries that get a bit of volume, but outside of that there is nothing because people are really only interested in the service. That is when I have found it just does not work, because you end up with this giant network that gets no traffic, and you have just increased the site radius for no advantage whatsoever.

James Dooley: Is there anything that you're doing with semantics that does work?

Paul Troscott: Yes, just narrow it down. Keep it to a core of pages.

It is almost like getting rid of the outer section of the site. Keep it really tight. I am more focused now, excuse the pun, on site focus. I want to keep the focus of the site not necessarily narrow, but laser targeted on what the niche is. Then I use cosine similarity. I know Google is using chamfer similarity now a lot because of paragraph indexing and all that stuff, but I'm not sure how much of that is being used in local because they are expensive algorithms to run. So I'm on the fence about whether they are really using this in the local space. I want to look at the cosine similarity between the pages I'm creating and the query vector. So, the distance between those two vectors, and then make sure that all my pages are as close as they can be to the query vector. I do not want to go too far away from it. Then you're going to guarantee that the site focus and site radius are nice and tight. That seems to be working really well, in combination with some old-fashioned stuff like TF-IDF and BM25, which we have discussed before. That still works in local. This idea that people say they do not use keywords at all anymore, I'm not finding that in local. If you do not use your keywords on the page, I'm not seeing anything ranking.

James Dooley: What about you, Luke? Is there anything that you're seeing specifically in local SEO with regards to semantic SEO strategies?

Luke Baston: Yes, I would say I agree with pretty much everything Paul said there. I'm seeing that as well.

I would say I probably have less of a focus on keywords than Paul does. I go from more of an entity perspective when trying to build the site out. Two things I see exacerbating the problems Paul was talking about are the targeting of the site not being based on the right hyponym or the right overall category. You see lots of websites where, if they are in a certain space, say roofing, just to pick an example we all probably know about, they might think roofing services is the main thrust of the site. They then target roofing services, and maybe that is a keyword approach. But in that case, you're overlooking so much of the market because there is repair, maintenance and other services. It is not the right top-level category to go for. Roofing contractor or roofing maintenance, which covers those different subcategories, is probably a better option. So not picking the right hyponym is one thing. The other thing is trying to be too broad on informational pages. It is pretty much what Paul said. Lots of people do want that informational focus as well as transactional, and I do not see too much harm with that because SEO could be part of a more joined-up strategy to build email lists or do things offline. There could be reasons for it, but if you get that too diluted, then it can water down the site overall. In some markets, it is really dangerous as well. Again, sticking to roofing because it is the same example, you can easily stray into things like guttering, window repair and stuff that is tangential to roofing but not within roofing. That can water the whole site down. Those are the two big traps I see over and above what Paul was saying.

James Dooley: What about you, Mike? What are you doing with regards to semantic SEO? Obviously, I know you're doing a lot more than just local. You're doing a lot of iGaming and stuff like that, but what have you seen work in local with semantic SEO strategies?

Mike Lovat: One of the things Paul touched on was site focus.

One thing I have definitely seen over the last few years is that years ago, everyone was talking about content marketing. If you were a local dentist or roofer, then you would have a blog or guide section and add a lot of content. The problem is, with how many dentists there are in the UK or the USA, they are all going to end up with quite similar blog posts over time. Maybe not specific to the local area, but about things like the cost of treatment, how something works or what it involves. There are going to be a lot of similarities. So how are you meant to stand out? I had a national e-commerce client where we added loads of content. The category pages themselves were okay, but rather than spending money on conversion rate optimisation, link building or some sort of external signals, which we know are important with third-party cooperation, they spent loads of money on getting more blog posts. I guess that is a KPI for some people. They want to publish 15 articles a month, so they covered every kind of entity and attribute within their niche. It was ideas-based content, and they could get traffic from Pinterest, which you would normally think would work fine, especially if it was an informational site. They were selling flooring and things like that, so they added loads of content relating to flooring ideas that then spanned into other interior designs. It would be something like red kitchen with white floor tile ideas and inspiration. There are so many different permutations of that, and they ended up with around 250 blog posts. They were all pretty good. But I do think that skews the site focus away from being an e-commerce site that sells products to being more of a guide with a shop on the side. There are probably a lot of local sites where they might have one core service and three pages for the different variations, whether that is new roof installation, roof repair or something like that. Then, if you have 100 blog posts on roofing in general, is that going to skew your site focus away from being an actual local business and towards being a reference point? I do think there is something in that. Years ago, we used backlinks as a way of validating how real something was. Then, as SEOs, we ruined that by spamming the internet. Writing quality content became more of a way of validating things, but now there are enough freelance writers and AI tools around to produce a huge amount of content. I think traffic now becomes more of a validation. If you have a local site and 95% of your traffic is going to blog posts, then that is validating you as more of an informational website rather than a local business. I have seen that play out. I have seen it play out on my affiliate site as well. Years ago, people thought Google was going after affiliate sites, so they needed to add non-commercial pages too. So you add blog posts and guides. I did that with my affiliate site. I had around 250 pages of guides, which then dwarfed the main core pages of the topical map that I wanted to rank. Instead of looking like an affiliate site, a review site or a comparison site, it looked more like an educational guide that wanted to make money through ads. Then you see that the main keywords you want to rank for suffer, despite there being internal links and good signals going through. If 90% of the traffic is validating your site as an educational resource, then I do think that skews it in Google's eyes. I would say the same in local. If all the traffic and focus is on outer pages of the topical map rather than the core pages, is that going to skew site focus into being more of a reference point rather than the go-to person to get your roof repaired in the local area?

James Dooley: It is crazy that you all talk about site focus and site radius, and trying to keep it as on point as it can be.

It does go against a little bit of Koray's methodology of going wide. I know you can go wide or deep, and I presume you're all saying stay narrow and go deep on that one topic. Do not really go for all the additional terms and stuff. If there is one key takeaway for someone watching this, they are in local SEO and they think they need to understand semantic SEO more, what would it be? What one key takeaway would you give to someone listening or watching this? If I were you and I were in local SEO, this is what I would be doing with regards to semantics. Paul, you mentioned BM25 and TF-IDF. Is there anything on that you would be doing? People have previously said to use tools like Surfer SEO, MarketMuse and OnPage.ai. Are there any tools that anyone is using that can help them semantically get in the entities like you mentioned, Luke? What one key takeaway would you give to someone who has a local site and wants to improve semantic SEO?

Paul Troscott: For me, it would be geo-topical and geo-contextualising the content, which we can do so easily now with AI in a way we could not easily do before.

Think in terms of if you were really working in that place, if you really went out and did jobs in that place, what would you know about that area that somebody who does not do that would not know? Try to link your service with how it is affected by where you perform it. That could be local by-laws, local constraints, local regulations, that kind of thing. That is working extraordinarily well for pushing the Google Business Profile. The content on the site, particularly the landing page, pushes back to the Google Business Profile and works like gangbusters. That used to be hard to do.

Luke Baston: I would say the biggest thing I have found, which has made the biggest difference, is getting really well researched and deep on what the commercial attributes are for that particular market.

If you have a business in the solar panel space, just thinking of examples where I can give real examples of what I have found, a lot of those services touch on what Paul was talking about, but more on the service side than the location side. One of the biggest things is credit. Do you offer financing as a business? It is amazing how many businesses pay lip service to it or do not even mention that finance is available, or that tax credits are available for certain types of solar solution. It is those commercial attributes and where they are on the page that can really categorise a website in the right way. So I would say understand and prioritise the commercial attributes.

James Dooley: What about you, Mike?

Mike Lovat: Similar to what Paul says, by being hyper geo-specific.

Anyone can spin up a local SEO site about roofing or plumbing, but some web browsers are going to have specific questions. For example, what month of the year can they get their roof done? In certain parts of the world, it is going to be impossible in certain months due to monsoon rain, wind, snow and so on. There is also accessibility. Some people live on the edge of a mountain, so if you have really generic advice that says you will visit any day of the week, it does not apply to everyone. Some people might be on a small island that is only available by car ferry twice a week. By covering that ground, you will get those highly specific attributes covered. The best way to think about it is something someone once said to me. The best way to fill out your site with decent content that is specific to your niche is to imagine every server in the world went down except yours, and ChatGPT could only index your website. If someone had a question about roofing, and asked how much it would cost to get their roof repaired or what time of the year they could get their roof repaired, would the LLM be able to answer the user's query just based on your site alone?

James Dooley: That is a really good way of looking at it.

For me, I have had quite a lot of interesting conversations with Koray, who obviously all four of us know very well. I was trying to play devil's advocate, which I'm sure all three of you know I love to do, to try and balance things off. The biggest one for me was that I was trying to tell Koray that his methodology does not work in local. I was debating with him left, right and centre, saying it does not work in local because I feel like it is too informational. Then he actually sat down with me and went through a load of different things that I did not know and I do not hear a lot of people talk about. It was about wasteful domains, as he called them. These are other domains that he owns. You could call it a PBN. I would call it a PBN. Mike, where you spoke about informational-based terms, he would still go and do a topical map and map out a semantic content network. But if he felt there was not much outer section, as you spoke about Paul, or if he felt that it should not be on his main transactional money website, he would create the whole topical map and define what goes on the site and what goes off the site. He has an off-page topical map, which is what I have been saying for ages. Why not do an off-page topical map? You need to do an off-page topical map for the third-party corroboration that comes back to the site. He does do one, and he calls it wasteful domains. I had never heard him talk about wasteful domains before, and that is what he does. He still does the topical map. He defines what goes on the site and what goes off the site, exactly like what you're talking about there with site radius and site focus. He keeps that focus. He keeps that radius. He makes certain he does not fan too far away. But does he want to cover that topic? Yes, he does. That goes on a different domain and points back to the main site. It says something like, by the way, you might need to do this, this and this. If you're looking for a contractor that can do it, here is the best company, and it points back to it. I thought that was genius when he told me because I had been trying to debate with him that his methodology of these informational-based articles was not good enough. He still wants to cover it and still wants to cover it with the brand being connected to those topics. He calls it wasteful domains. I call it off-page topical maps. Whatever you want to call it, Jason Barnard calls it third-party corroboration, which we all talk about. For me, that was the bit where I thought, okay, now I get it. It does fall in line with everything you're saying there. It covers all the entity attributes. Your site focus score is brilliant. The radius is all bang on. I'm exactly the same as what you're saying there, Paul and Luke. Stay on topic and be transactional. Mike, you were saying if you go too far away with a blog, it becomes more informational. Does that dilute your main transactional pages? Yes, I have always found that it does, like you have been saying. I feel that this kind of method works. Have you all done the course? Have you ever heard Koray talking about that?

Paul Troscott: Not using those words, but conceptually, yes. It is pretty much a summary of what we have all been saying in different parts today.

Luke Baston: I have seen it in the wild. I have seen people use subdomains instead of wasteful domains as a completely different domain, and it probably has a similar technical effect.

HubSpot is an example from a few years ago. If you did not have a website that was a HubSpot website, but you wanted to hook up HubSpot CRM to your website, you used to have to have the blog as a subdomain. I do not know if that is still the case, as I have not looked at it for a couple of years, but that system used to work fairly well from a blogging perspective. I did not know this at the time, and this would have been about six years ago the last time I looked at it, but that would make sense to me because it is a subdomain and you have this off-page and on-page dynamic.

James Dooley: For sure. Anyone watching this, I hope you liked the topic about semantic SEO specifically for local SERPs and local SEO.

Paul, Mike and Luke, it has been an absolute pleasure.