What is GEO (Generative Engine Optimization)?

Hey, what's up, and welcome to our channel. In this video, we're going to be looking at generative engine optimization. We're going to look at what it is, how is it different from SEO, and what are some things that we can do today to make sure that our content is being surfaced in these new areas of opportunity, where more and more people are going to get their questions answered, and how we can best serve them with the content that we've already either produced or the content that we're working on now. And we want to make sure that it's not only surfacing in traditional searches, but also within ChatGPT Search, Perplexity, Google AIO, and all the other opportunities that we have today. All right, so let's talk about generative engine optimization. So generative AI, it is changing how people find answers online. There's a lot less friction. It's a little bit more of a natural interaction as we do this question and answering. So these AI search engines, they're growing, right? We all know that. We're seeing them pop up. There's new releases. I mean, we've seen some massive improvements just over the last few months with all the new different models coming out. So a lot of people are asking the question, what do we do about it? How do we approach this as businesses? How do we make sure that the content, the investments that we're making into our website, into our marketing, is allowing us to be surfaced on these types of engines and these types of large language models? So before we get into the nitty gritty, let's talk a little bit about the evolution. So in November 2022, ChatGPT, they released their training-based model. And this was really the revolution of conversational AI. This took conversational AI to a whole new level. Now in 2022 of December, we saw Perplexity launching. And they focused more on providing cited answers. So remember when ChatGPT started coming out, one of the big issues that people were having was citations, right? And as Google was even surfacing these generative AI results, people were having issues with no citations. So we weren't actually able to say, hey, this is the person who created the content. It just was like the chatbot was saying, hey, here's the answer I know about. So people had issues with attribution, and Perplexity was working on solving that issue. Now in 2023, we saw Claude and Lama. They started to come up and be released with their training-based models. And we saw an expansion, right? And then, of course, in December of 2024, we saw this hybrid of AI and Search with SearchGPT. So now we were combining this large language model of chat with a web search AI. And again, kind of revolutionizing how search is being used by an end user. So GEO is not just SEO. It doesn't work the same way. As you can see, we've got some AI-generated images on the side here. It's where they're having issues spelling. But search engines, right, like Google, they do leverage things like AI overviews. Now a lot of those AI overviews are highly influenced by traditional SEO with some additional considerations, right? It's not just about having good SEO on a page that's going to allow you to be there. But foundational SEO is at play. Structured data, as we found out so far, has a pretty large impact in that as well, as well as content structure, credibility of the site, credibility of the domain, all of those things that you would expect to see from a search engine. Now things like Claude or Lama, they rely on training data, which requires a long time to build that brand because these aren't being updated every single day. There's batches of training data that are built into these systems and that these systems don't have access to that real-time data like a search engine would. So if you're trying to optimize in one of those platforms, it might take a little bit longer. Now search AI, like Perplexity and ChatSearch or SearchGPT, they do integrate search directly. And so you do have some similarities of traditional SEO, but it's not exactly the same because they've got their own models as well. And they also rely on past data within their databases, that training data that they use as well. And then you've got kind of these hybrid systems where you've got ChatGPT and Gemini, which are kind of moving in between their training data and the web index. Now Gemini has direct access to Google's web index, which is a ton and ton and ton of information, of course. But Chat is working with large search engines as well. So think about Bing, right? Think about their relationship with Apple. It's giving them more real-time data that they can use as well. So there's a little bit of a mixture when it comes to these generative AI platforms and how we might want to optimize for them and the different levers that we can pool in order to surface our content. It's not going to be just one set of plans that's going to work across the board. So we've got to really mix in some traditional SEO elements. But then we also have to think differently when it comes to things like training data, legacy data, data that's maybe not being refreshed in real time like you would on a search engine, which is not real-time either. Of course, a page needs to be indexed. There's a lot of things that go into it. But it's a little bit slower when you're looking at these models that are completely reliant on training data or the ones that are hybrid. All right, so let's talk about the difference between influence and optimization. So the goal of SEO is to drive people from a search engine to your website. That's the main goal. That's what we're trying to achieve. When it comes to generative optimization, the biggest thing we're trying to do is make our brand surface in AI responses. We want to make sure that our brand is part of the AI conversation. So that means our content should be approached a little bit differently. In an SEO standpoint, our content is being optimized to surface on search engines. When it comes to optimizing within these large language models, we're educating those models with our content so that we'll include our entity in relevant results. We want to make sure that we are known for certain topics, niches, problems, answers to questions, right? So that's a slightly different view of content. And the metrics we use are going to be different. Traditionally, SEO, we're going to be looking at things like organic traffic, keyword rankings, click-through rates, things of that nature, where we're going to have a little bit more brand visibility-type metrics when it comes to GEO and the whole goal of influencing a language model. So what are these new metrics that we need to be paying attention when it comes to generative engine optimization success? Well, the first thing we have to look at is brand visibility. We need to track our mentions across multiple queries and multiple platforms instead of just ranking. So instead of just looking at Google and seeing where we are in positions 1 through 10, we need to be thinking about where are we being surfaced on certain questions across Gemini, Chat GPT search, Perplexity, Clawed, AI overviews. All of those things have to be in play to give us a better idea of how visible our brand is within the ecosystem. We also need to think more about opportunity size. What markets or what language models or what tools is our audience using? How big is that opportunity there? And where can we surface for more queries? This is very different than looking at click-through rates or search volume calculations. These things won't work with generative search engines. We have to think differently about the opportunities that we have. Traffic patterns are changing drastically. So searchers may click on a link in one of these tools, or they may research right within AI and then directly search either for your brand or use your URL, right? So we need to pay attention to brand search because it could have a large correlation to what people are doing on these large language platforms. And Google today is becoming much more of a navigational tool as complex queries, questions, move into these AI spaces. Now, again, this is cutting edge, right? We are not here at the large, mass front end of the market. If we look at a technology adoption lifecycle, we're in that early adopters phase where people are starting to kind of leverage these tools a lot unlike they ever have before, which is one of the reasons that they're growing so fast. So what are some of the SEO pillars that also apply to GEO? Well, content matters, right? Website, they're not just data feeds that go to indexing. They also help train large language models. So they're not just a destination for people, they're a destination that trains the language model. So we need to have good content that accurately describes not just the services, not just our products, but how we solve complex questions. Technical SEO still matters. Structured data is important. Large language models can leverage structured data much more easily than unstructured data, right? We're talking about computers again. It's still a computer crawling and looking at your information. So allowing these bots to crawl your content. Google recently just updated some of their own best practices for websites to include not blocking these types of bots if you want to surface in their AI overviews. Authority still matters. Now, large language models calculate this differently. So you'd have to customize your approach. It's not just a one size fits all. If you've got a lot of mentions or you've got a lot of backlinks, that it's gonna have a direct result within the large language models. And also just because you're ranking for keywords doesn't necessarily mean that that's true either. While there is some correlation and some tests have shown that, there's also other data and specifically some data points that we've found recently that it doesn't match up one for one or as high of a correlation in all markets. Some markets, yes, but in other markets, not so much. So we need to think about how that language model is calculating authority and what we need to do to help influence them. And lastly, user experience. This information, we don't really understand it fully, but engagement does impact language models, especially as they're going to get more data and there's gonna be more data in the feedback loop. I know this is a big deal when it comes to Google and their AI overviews. Are people getting the answers to the questions? And if they click on the link, is that information solving the problem? So they're getting that user engagement feedback and you can definitely be sure that tools like ChatGPT, Perplexity that offer this for free are using your information to better educate them on user experience. So don't forget about the people that are involved in this process. It's very easy to think about Google's computers and the large language models and training data get very technical, but we can't forget about the people because they're gonna have major influence into how these tools are shaped and used because that's gonna allow these large language models to improve themselves. So should we take a unique approach for each model? Yes, based on our own in-house research, it appears that the Google AIO rankings are heavily influenced by more traditional SEO metrics, but not only by traditional SEO metrics. So you still need to make sure that your content is optimized and you need to be paying attention to how you're answering questions in that conversational way. ChatGPT definitely relies more on contextual relevance or that training data bias. So it may take a little bit more time for you to influence. Now, they do have connections with real-time search data, especially depending on the model you're using, but some of the models, it could take some time. You need to create deep content that will help these models understand the knowledge and expertise that you provide. Now, Perplexity has some similarities to ChatGPT when it comes to contextual relevance or that training data, but it also has live search data access. So once again, deep context with a mix of SEO best practice, they are needed here. But we've also found in some of our studies that Perplexity tends to favor more niche content than broad content. And just a little chart below here, some of the platforms and how often those core updates are coming into play. Now, ChatGPT has been doing a lot more updates lately. Perplexity is not updated daily, but it is integrated with live search. Google AIO, it does have a underlying LLM that is updated occasionally, but not daily, right? So it talks a little bit about the daily needs, you know, how you're leveraging these different tools and then how fresh the content is. And Perplexity and Google, as well as ChatSearch are gonna be more real-time, but some of the other language models like Claude, Lama, and just maybe just a regular GPT model, it's not gonna be as real-time. Hey, thanks for checking out this video. Hope that you found it helpful. If you wanna add to the conversation, please comment below. We'd love to continue that with you here on the channel. Don't forget to like and subscribe and share it if you found it helpful and until next time, happy marketing. ♪♪♪