Latest / Semantic SEO Podcast / Common Mistakes With Semantic SEO: Building True Topical Authority (with Germans Frolovs)
Transcript
- James Dooley: The most common errors with semantic SEO. Today I'm joined with Germans Frolovs, and the amount of people that seem to get semantic SEO wrong. It could be topical maps. It could be semantic content networks. It could be not providing the source context. Could be to do with how you write the content briefs. It could be the micro-semantics. There's so many different things within semantic SEO that you need to do right and connect things together. But in your opinion, Germans, what are the most common errors that you see others do wrong with semantic SEO? Germans Frolovs: Uh, I think related to most common errors, I think most kind of, like, maybe fundamentally understand it... Well, before teaching, like, understanding a little bit more that Koray is teaching and so on, is that the semantic SEO, by treating it maybe more like a slightly more advanced version of traditional keyword research, when it actually is quite a big shift from the strings to things, kind of, like, motto and transition from maybe matching the character sequences to actually presenting the knowledge and the meaning of your domain. Uh, I think we already mentioned that there's quite a difference between a concept map and a real topical map. Or the concept map would be more like a dictionary of a knowledge domain. In a real topical map, we're going to go into quite a lot more detail on different processes and how, uh, and have a deeper, uh, let's say, consideration of our business. In semantic SEO, we also try to cover the information gaps rather than just keyword gaps that you might maybe find in a tool like Ahrefs. Uh, of course, keyword gaps are still important to consider, but information gaps are more important. Also, things like relevance and responsiveness, that also probably are maybe misunderstood and not necessarily taken into account, uh, when maybe just starting semantic. I was also kind of, like, guilty of that. That, of course, we are all trained on how to be relevant, uh, with our text copy. But is it actually responsive from the textual point of view, but also from, let's say, visual semantic point of view? Is it first, uh, answering the question or giving the answer immediately without any fluff? Uh, maybe even matching the, uh, uh, question format with the answer format. But then also, does it provide the function or the action that the user is looking for to satisfy that and, and have that satisfied search engine user session and, uh, that, that, that positive signal? Uh, we also mentioned the whole query deserves a page concept. That is also probably largely misunderstood. But, uh, I think if people doing traditional SEO have been coming from keyword clustering, uh, background, then, uh, it is helpful. Uh, it's also not a one-time strategy, like we briefly discussed. It is something that is always changing. Uh, and, uh, essentially something that is also quite apparent in the last couple of years, that we are transforming into ranking web entities rather than websites. So actually how the web entities are presented across the web is becoming increasingly more important rather than just, rather just than just ranking that one specific website. James Dooley: Yeah, for sure. Oh, I mean, obviously there's a... There seems to be quite a big divide, um, between, let's say, local SEO, where people are using, like, BM25-type optimisation strategies, which could be using, like, PageOptimizer Pro. It could be using Surfer SEO. It could be using things like MarketMuse and stuff like that, where they are just literally adding in specific queries and entities onto a page. And because they've got more entities on the page, they feel it's better optimised. Now, the issue is, because of a lot of other reasons, like freshness of content and stuff like that, you can see jumps. And with certain algorithms, you can see those jumps and say, "That's it. I know how to write content. There you are. See you in a bit. I know what I'm doing." The issue is when it becomes into more like BERT and MUM, and the different algorithms come along, and the more, like, qualitative algorithms. This is where the information semantic SEO comes. Well, what can you tell people about... Semantic SEO isn't just about adding missing words or cosine similarity words or synonyms or entities on a page. And it is more about feeding information to the algorithm that improves NLP. Germans Frolovs: Yeah. So, uh, it is related to adding entities and so on. But I think the biggest difference is adding those entities and those synonyms and co-occurrences within the right context. Many of the tools are providing these details based on all the competitor, let's say, uh, summarised information, TF and so on. Basically get an aggregated report on what others are doing within that, let's say, uh, within that SERP. And this is probably what you should do to outcompete, to outrank them, right? I also myself used many of these tools before. But once you, let's say, uh, yeah, do, do semantic SEO, let's say, how Koray teaches it, I realised that, like, by doing that manually, you kind of already optimise and hit all the check marks from those point... From whatever that tool, let's say, shows from Surfer, MarketMuse or PageOptimizer and so on. Uh, and then after a while, I just stopped using them because the results are not... Or the tool is not that much useful anymore because already, uh, it is optimised if you follow those semantic principles. And I think, yeah, the biggest difference is that adding those, uh, those, those elements, those co-occurrences, synonyms and entities, uh, and those NLP, let's say, optimised terms, how they refer to it, within the context. And, uh, covering those information gaps rather than just word gaps or keyword gaps. James Dooley: Yeah, for sure. I mean, it definitely taught me with regards to the content briefs, with regards to the entity, covering all the different attributes of the entity with all the different subheadings, and then answering it in a specific way with semantic triples. I literally used to just put words on the page and go, "This needs to be on 16 times." And it was almost like a TF, kind of, keyword veloc... Like, the velocity and stuff like that of what needed to be done. Um, and obviously we've moved away from there. But with regards to another question, with regards to a common, kind of, misconception with semantic SEO, can you explain the difference between topical coverage and topical authority? Germans Frolovs: Yeah. So topical authority is the ultimate thing, uh, what we try to achieve, and we actually get it. Uh, but there's also, like, this famous formula that Koray mentioned a few times. And then I think also Lewis expanded it with, uh, one other element. So topical coverage is essentially how well you define the topic with their element, entities, subtopics, attributes and so on, with proper depth and vastness within that knowledge domain. Uh, times historical data, divided by cost of retrieval. Historical data is that earned search engine trust thanks to all the positive, let's say, user interactions, clicks, impressions, positive search engine sessions and so on. And cost of retrieval is essentially that, uh, uh, computational power, uh, that the machine, Google, needs to spend, let's say, to understand, uh, uh, your website, to crawl, index, understand your website and then rank it. It is based on, uh, time and money, kind of, like, computational effort, you can say. So topical authority is, uh, equal to topical coverage times historical data divided by cost of retrieval. That is the biggest difference. James Dooley: Yeah. So obviously going into that, so topical coverage, covering as many different topics as possible, multiplied by historical data, meaning the traffic that that's getting, the engagement that's getting, the signals that's getting, and then divided by the cost of information retrieval. Because it's easy for people to go, "Okay, I'm going to cover the whole topic," and you go and do 10,000 pages that get no traffic. And the cost of information retrieval, it's very expensive for them to put the compute power in. Therefore, they're not gaining the topical authority. How important do you believe topical authority is, that formula? I love, I love the idea that Lewis added that divided by cost of information retrieval in. It made complete sense. Otherwise, you could just spam opening up new pages, opening up new pages, opening up new pages all the time. It would cost a lot more. How important is topical authority, um, within semantic SEO? Germans Frolovs: I think that formula presents the whole concept in quite a beautiful way. Uh, and, uh, I also tested it a few times. When with more historical data and positive historical data, you will get, uh, more topical authority. But the start is topical coverage. You need to get that first with the least amount, with the least computational effort possible. Uh, which makes sense. Google is all about money. They, they spend every time you do a search and so on. Uh, and to crawl, understand those results, yeah, requires some, yeah, massive amounts of money and computational efforts. So topical coverage comes first. After that, basically, you get the historical data, and as a result, you're getting rewarded. Typically, I have seen also across different websites that it takes about three to four months from the launch of the content network in the beginning. Once some historical data is accumulated, three to four months, typically, on average, what I see after that, you see a big jump ranking and you become the topical authority in that segment that you have covered. James Dooley: Yeah, for sure. And then if we're looking at here, the most common errors, um, with semantic SEO, then that formula is going to be what's answering this last question. But the last question that I've got on here is, why does publishing more content not automatically build more topical authority? Germans Frolovs: Yeah. It's very important to understand that, yeah, publishing more does not mean automatically you're going to get more authority. Because search engines, uh, do not evaluate a website simply just based on how many documents they have, but rather than, uh, yeah, that intersection with historical data and the information that you provide. Uh, if it is actually, uh, let's say, uh, getting a higher quality threshold than any other source. Does it, does it provide that more unique information and so on? So while that publication frequency momentum is important, it's, uh, necessary to capture that attention. It's not everything. If you have too many pages, uh, with lower quality of information, that's not going to work out. Uh, so that would be, uh, yeah, why it's important. Also, ranking signal dilution might happen if you have too many pages, you publish too fast. If, uh, you are being perceived as just a, I don't know, automated content machine without that much human effort, that's also probably not, not a good thing. Uh, and, uh, the source context and brand identity. I think we already talked about that, but with massive amounts of publishing, that could be neglected. Uh, and, uh, also I have seen that, like I mentioned in an example, you need to have historical data, uh, and some kind of, like, let's say, traction before you can expand a lot more. Uh, if you don't get that with the publication of your first content batch, I think it's going to be hard to, uh, let's say, uh, get ranking if you just continue publishing hundreds of pages, uh, every week. James Dooley: Yeah, for sure. Anyone who's listening to this, I strongly recommend also going and checking out the episode about the common misconceptions about topical maps. But we hope you like this episode on the most common errors that Germans comes across with regards to semantic SEO. Germans, it's been an absolute pleasure. Germans Frolovs: Likewise. Thanks.