Eric Enge (2:53)

So fundamentally, I'm an entrepreneur at heart. I mean, honestly, that was sort of bred into me as I was growing up, as I watched my father interacting with a lot of different entrepreneurs and saw some very successful people. And, you know, that's entrepreneurs, when they see disruptive events and times of change, rather than being scared of it, they embrace it because that's when new winners are born. And so that energy just comes naturally to me when there's big change on the horizon. And I think we all agree that with what's going on with AI and Gen AI right now, that there is big change on the horizon. So it's a cool time.

Tyson Stockton (3:36)

Well, let's talk a little bit, you know, maybe before, like the application of how we're using this, like from your perspective right now, like, where are you viewing, you know, the level of sophistication in some of the recent tools? Like, not too long ago we had the release of ChatGPT5. So maybe we kind of start with like, what are we working with today in this technology and how are you kind of viewing that?

Eric Enge (4:02)

So, yeah, there's two big things that are happening right now. ChatGPT5, as you mentioned, and also agentic browsers is sort of the next thing coming, and they both have some really interesting implications. So with ChatGPT5, the reason why this is important is one of the big Critiques of where OpenAI was a year ago was that their chances of making the infrastructure required to deliver answers to people via ChatGPT was not fair, fiscally viable, it was too expensive, there was too much processing power. And it's because they had this model of absorbing all the world's information and then, you know, finding a way to be able to spit that back selectively out to people to GROSSLY Simplify it. ChatGPT 5 is different. ChatGPT 5 is smart at knowing how to go find information quickly as targeted to specific topics. So the computational overhead of building the model and running the model should bring them a lot closer to not running so deeply in the red. I don't have enough information now to tell you if it's going to make them profitable per se, but it should bring them. It's headed very much in the right direction, I think is actually the way I want to say it with Agentic browsers. That's also really interesting because the goal there is for both Perplexity and OpenAI that have them, is to provide users with a Surface, which comes their primary interface to the whole web. And you could say, okay, well Chrome is after me now. But the difference is an agentic browser will do things like take high level tasks for you, go research information across a whole bunch of websites, bring some things back to you, ask for some clarifications about what you want, or how to handle some aspects of information it gets and will actually be your coach through the execution of the task. Unlike Chrome, which is a tool you can use to execute the task, but you have to manage all the steps of the process on a gentic browser, it's largely done for you. So they're both pretty significant changes, which.

Tyson Stockton (6:24)

Is like, I feel like there's an intersection and I've always been kind of describing to people as like, there's three main areas that I'm looking at for how these shifts are changing our actions or our approach. And you have the landscape shift of the space that we're competing in, whether it's AI mode, ChatGPTV, perplexity, whatever. And then there's how we're also using this from an execution of work perspective. And then there's kind of like always the stakeholder management, how we're leading the narrative in the organization or responding to executives moving quicker on this, or having someone in their ear to move quicker. And so I think those are three areas. But I like that you're kind of calling out the kind of approach to that agentic area where it's almost like a little bit of an intersection of it, because some of how we use it is that companion coach. And that's been something, you know, Jordan and I have been discussing a lot is like how that then impacts the composition of SEO teams and structures or how you're looking at that kind of like management in the channel. So maybe kind of like going into that areas of how we manage the channel. How are you seeing kind of some of these recent developments impact that?

Eric Enge (7:48)

So wow, that's a multi layered question. So I'm going to start with the first part which is the scope of impact. I mean, let's face it, Google is still the big dog. OpenAI and ChatGPT are certainly getting enormous amount of press and getting mind share. And I think it's fair to say that the impact on the younger generations is stronger than it is on the people who have deeply ingrained habits of using Google for decades as Google and Chrome together, you know, as the surface they use to approach the web. And so those habits are going to die hard. And I think the thing that we can't overlook, in spite of the jump that OpenAI and ChatGPT got on things, is that Google caught up with that quicker than people thought they would. And you see this, by the way, reflected in their stock price because the analysts are still seeing Google growing and they're not forecasting any doom and gloom around Google. So Google's still the big dog. And so there's a lot that needs to happen to change that. And the way we need to think about this is for a change to be truly disruptive needs to be a big enough change that it's worth people changing deeply ingrained habits. That's not trivial. And there are all kinds of things that have happened through history where people launched something that everybody thought was going to be the next big thing and it just wasn't a big enough thing for people to adopt it. This one's a little different. ChatGPT, you know, launched November 2022, as we all know, and then Google responding. So this is a situation where Google has embraced the change and you know, AI overviews, which are a lighter weight version of Gemini, which you can read is costless to run and provide answers. They are public, after all, they have that burden of delivering profits, but you know, nonetheless that will give that experience to a lot of people. And the other thing I expect is I'd be shocked if sometime in the next six to nine months Google doesn't take some portion of their queries, call it 20%, 25%, 30%, and fling them into full AI mode. And if it were me, I would seriously be thinking about doing that, right? Because I mean, Gemini is pretty good. So when we talk about the shifts in the market, and this is where you started with this question, and I want to lose that part of the question, certainly it's happening to a degree, but even when you think about what might happen with agentic browsers, a wild success would be that they have 15% market share in three years. You know, it's not going to be, you know, 70% next year. So it's going to happen incrementally. And the same with, you know, ChatGPT, which is nowhere near 15% of search, it's going to take time for, for them to build their share. So while we think about what we do in our day to day work, and this is my philosophy, I'm sure other people would espouse different things because there are people out there who like to emphasize how big the world changes are, but that's not me. I try to be very Pragmatic about things. The primary thing you still have to focus on for getting traffic and visibility is Google and you know, AI overviews and traditional search and you know, that's the primary platform still. And it's important to remember that there are some other changes which I think we're going to end up talking about, which is how much more we know about how all these algorithms work. And that's a whole nother conversation which is good for optimizing across all the platforms.

Tyson Stockton (11:49)

I definitely want to get into that aspect of it from the previous piece though. To me it sounds like there's kind of like two pieces kind of at play where you have on one hand the strength and our advantage that Google has with dominance in the market size. With that also comes volume of data that they're working with. And then they also have a challenge in the sense of you do have a publicly chained company aspect. You made kind of the comments of, hey, that's going to influence how they might approach some of these releases and next steps. So I mean, if I'm kind of reading between the lines there, it feels like Google still has the advantage and they would be like the favorite if you're boiling this down into a thing. But there is enough of a threat. And like, at least from my experience, this feels like the most significant actual threat that they've had in a sense of how they're reacting and responding to it. And I do think it's smart that they're leaning into this change rather than just trying to continue to it. And so it's not like it feels like they're being naive to some of the shifts in the market, but at the same time there is, you know, some public traded company elements about it.

Eric Enge (13:09)

Yeah, no, they definitely felt the pressure. And as I understand the early part of the story is that, you know, ChatGPT came out, but they, you know, weren't yet prepared to go and try to disrupt the search experience. And as I understand it, Samsung forced their hand because Samsung was prepared to dump Google to go with Bing in order to get access to what's now called Copilot because they wanted to have, you know, the generative AI experience on their phones. And suddenly the willingness to change became. Look, it's entirely possible that, you know, how I ended up hearing that story may have been overstated, but actually market pressure is absolutely in play here and Samsung may well have played a part, maybe even a big part of. So it's important enough, right? You know, I talked about how disruptive a disruptive change needed to be. It isn't so disruptive that everybody changed overnight, right? Well, changed overnight is relative. But you know, you look back when there were railroad companies and all the transportation was done by railroad. When you're going, you know more than you could get to on your horse, so called horseless carriages came out and the railroad companies didn't adapt, right? But cars were personal, right? That was a huge change. You had your own thing sitting right up outside your door. You walk out, hop in and go. That was a massive disruptive change. And it was so disruptive that the, the victory of that was inevitable over a very short period of time. I think that ultimately the generative AI stuff will rise to that level of disruption when we're said and done. It's just going to take a while yet. And there's many reasons for that. It's still, they still make mistakes. People talk about it less, but I got a couple images created the other day and then I was like tearing my hair over trying to get it to fix the typos without adding new ones. It's still a little bit stupid in its own way because it's basically providing the next best word based on the sources it uses. And you know, that's not knowledge, that's homogenization.

Sponsor Voice (15:34)

Time for a one minute break to hear from our sponsor, Pre Visible. So you're looking for SEO help and you got a couple of options. You could start replying to spam from agencies that claim they can get you to rank number one on Google. You can pay an hourly rate for a consultant who will inevitably nickel and dime you with hourly charges. Or you can work with a cookie cutter agency to quickly launch a strategy less project with low success rate. None of those sound very good now do they? Well, that's where Pre Visible's integrated consulting model comes in. Pre Visible draws From a collective 40 years of SEO and digital marketing experience to unlock your organic growth opportunities. They build custom solutions that combine strategy, technical expertise, content and reporting to effectively operationalize SEO for your business. Pre Physical's four stage approach ensures that your SEO programs thrive by starting off with a strategy first approach. Then they support you in your efforts to create quality content, help you identify technical issues, and most importantly, they'll work with your cross functional teams to integrate your SEO strategies to make sure that your SEO budget actually drives results, not just your agency's bottom line. So join brands like Yelp, eBay, Canva, Atlassian Square, all who rely on the SEO consultants at Pre Visible. For more information, go to Pre visible IO that's pre visible P R E V I S I b l e.IO.

Tyson Stockton (17:02)

True. Which is always a nice reminder. But I mean to that point, like let's, let's dive in a little bit to, you know, what are some of those factors? It's a lot of questions and I feel like conversations that I'm hearing of like are we completely changing our strategy, our prioritization based on how these systems are working or is it more a reordering of it? Like there's clearly a lot of overlap kind of between the two. So I've always been advocating to people it's not an all or none kind of like we're only going to do this, we're only going to do that.

Eric Enge (17:39)

Correct.

Tyson Stockton (17:39)

But maybe like how are you viewing kind of the factors that can lead to success in these systems over maybe like traditional approaches?

Eric Enge (17:51)

So I'm going to say something that I know a lot of people out there hate. I want to emphasize how different everything is. But there is actually a lot of commonality in the algorithms used like ton of commonality in the algorithms used to determine relevance of a piece of content. And relevance is a fundamental ranking factor no matter what platform you're using. So that's a big part of the challenge. And even the way that authority is determined between LLMs and traditional Google search isn't like remarkably different. You know, it kind of comes down to, to there's a set of algorithms called embeddings which allow us to create vectors, usually very large dimensional vectors. I think we believe that Google uses their text embedding 5 model, which is 768 dimensional vector, to determine relevance of a piece of content, which could be a word, a phrase, a passage, which is, you could roughly think of that as being a paragraph, a whole page or a whole website. And all of those can be reduced to the same 768 dimensional vector and then they can be compared to each other. And you can then take a search phrase or a prompt and you can find content that is relevant to that based on using this, you know, this vector math. And for those who want more of the gory details and you can research this uses an algorithm called cosine similarity to measure how similar one vector is to another. And so conceptually, just to make it simple for people who aren't familiar with how this works, just to focus on two words for a moment. The words hat and cap sound like they're extremely similar to each other, especially if you're talking about this kind of cap Right here. With that in mind, the vectors for those, if we just reduce it to two dimensional space for a bit, will be extremely close to each other. So what that means is the cosine similarity score will be very close to one, which means nearly identical. Right? Whereas if I take a different word, like refrigerator, pretty different than either of those things, you don't wear a refrigerator on your head, for example. In fact, you don't wear one at all. And so this underlying math is what's being used in every platform that tries to respond to a question, be it a search query or a prompt, and what that means because we can do this for ourselves. Screaming Frog gives us a way to get at, you know, the how similar a page is to a specific, specific phrase and give you similarity scores. And other people have developed their own. The algorithms aren't remarkably difficult to implement, as it turns out. So we can do this for ourselves and we can optimize our content in ways that we didn't have available to us before. And this works for both Google search and LLMs at the same time. So clearly there's some differences. Right? But the differences I think are mainly driven not by the similarity comparison process or the relevance assessment process, as it is by the objective of the platform on the output side, which is where the bigger differences come into play. But understanding relevance is a huge part of the game and I can make my content more relevant to LLMs and Google search at the same time, and that's really, really powerful.

Tyson Stockton (21:34)

I feel like to a lot of SEOs listening to this, that should also be like refreshing, a little bit of.

Eric Enge (21:40)

A relief, a little wave of relief, right?

Tyson Stockton (21:43)

Because I mean, some people, it's like, yeah, if you're new to the industry and you're not familiar with vectors, embeddings and some of those pieces that you were just mentioning sounds like a lot. There's plenty of material out there and there's plenty of ways to kind of get up to speed that SEOs that have been in the industry for a while is like, great, that's all familiar ground for me. And so I think one, it's like you can take a little, again, relief, some confidence that the fundamental approaches aren't, you know, night and day different.

Eric Enge (22:18)

That's right.

Tyson Stockton (22:19)

Where we're still looking at relevance. And I think even from like an approach of this, like hypothesis test and like that, I mean, we've been talking about it for years, is like that test mentality to search, like, to me, all of that is still very much intact and it's just, hey, we have some nuance here. We have some new spaces that we're competing in. And if anything, I mean, we start in the conversation. It's an exciting time in SEO rather than more of a fearful time.

Eric Enge (22:49)

Well, yeah, I mean, I can't emphasize that enough. I mean, if you've been in the search industry, then you have to be in a role where you are continuously embracing change. So if that has been causing you anxiety, then you've had a long life of anxiety because the change is constant. And I mean, that's just the reality of it. But really try to feel the joy of it because it's an opportunity to take advantage of 100%.

Tyson Stockton (23:22)

You were also kind of like touching in on Screaming Frog, how the barrier to entry of looking at some of these pieces is now a little bit lower, more approachable. How does that or elements like that, how does that change how you'd be thinking about building an SEO program team, like, operationalizing the function, like, has that shifted and changed significantly from you if, like, you were leading SEO in a, you know, Fortune 500 business?

Eric Enge (23:58)

So, yeah, for me, every organization who is publishing content or updating content on a regular basis should make sure they have tools in hand to understand the relevance of their pages, their web pages to specific target terms, which hopefully in 98 or more percent of the people listening to this call have title tags that are well tuned to the target search term. Right? And assuming that that's the case, you can just look at the title tag. You can run an algorithm to tell you what your similarity is. You can run that same algorithm on the competitor pages that are ranking, you know, hopefully behind you, but maybe in front of you and see who gets higher similarity scores. And then you can learn from their pages, or you can just apply some basic principles at reviewing your content. So I can't emphasize this enough. This is simply money lying on the table waiting for you to pick it up. That's all it is. If you want more money for the business, then reach over to the table and pick it up. And what that means is you've got to just set it up. And by the way, if you want to not rely on Screaming Frog, you can build a tool for yourself that does these kinds of measurements in a few days of developer time. It's not actually remarkably difficult, as it turns out, especially since the text embedding models like Google's text embedding5 is publicly available, you can use it. So you should be setting up a process for all your most Important pages, good place to start. The ones that aren't ranking as well as you'd like, you probably have something wrong that's holding you back. And then, you know, run the model, see how you score, see how the competitors score. Review the pages based on how these algorithms work. We can talk a little bit more about that as to things you might do and optimize them to make the score go up. And it doesn't necessarily mean you're going to jump from position 9 to position 1 overnight. There are other aspects of the algorithms and here's where it's bigger difference between LLMs and search. Search is very tied to user engagement as measured by the chrome browser. And LLMs don't have that available to it, but they can see back and forth interaction with the LLM surface that the user is interacting through as well as authority signals, a whole nother conversation. But this is just something you can decide to do it now and you can have it in place in two weeks and be, you know, rolling on down the highway and doing things that improve your chances of ranking in all these platforms now.

Tyson Stockton (26:50)

And I'm going to go a little off script and it'll be a tangent here, but I feel like I can't pass on the opportunity.

Eric Enge (26:57)

There you go. We gotta get up.

Tyson Stockton (26:59)

Well, this is something that I've just been kind of like, I don't know, curious about, so I'd love to hear your perspective on it. Knowing that piece with Google and Chrome and so it was like we had the Google API league, the DOJ trials, great, we have confirmation nav boot like we have that information if that has been such a piece. And I mean, obviously it has been. I'm curious about how does that shift and change when more and more of the experience is being captured within the serp, whether it's in AI mode or in even Gemini aios. Do you think Google or even other competitors in this sense will have to change how they're using engagement metrics if more and more experience is not happening within the page but within their own.

Eric Enge (27:48)

Environment, such as in a Gentic browser? So yeah, well, it will change. It'll be. And this is true for ChatGPT and, you know, Claude and Perplexity and Copilot and all of them is that they currently have this experience which is kind of parallel to before Google had Chrome available and they would measure interactions between the user and a search result. So people speculated for years, if somebody went to a website and bounced back, was that a negative signal? And that particular interaction probably is asked, probably wasn't directly used as a signal, but things related to it were because there are certainly queries where user bouncing back fast is actually positive because they got what they wanted so quickly. And so there's a lot of nuance in all of that. But Google has been using the interaction with the search results like going back and doing a clarifying search, which means that the question wasn't answered by whatever sites the person used. And so that is available right now to OpenAI and anthropic and you know, and all the rest. Right. And so I'm sure they're using it.

Tyson Stockton (29:07)

I think that query like refinement is a great example too because it's, and I mean this is just observational so I don't know how much we can tie to it. But I feel like I have noticed more and more, especially in ChatGPT the like prompting for that next question that really feels in this like assistant vein of like great, you asked me this, what's the next action? What's the next thing that you're trying to do? And so I've been kind of curious in that same piece of like, is the number of times that you're taking that follow up prompt or kind of like initiation, how is that being used to measure even kind of like effectiveness, quality of response, those kind of pieces?

Eric Enge (29:52)

Yeah, it's a great question because the reality is that the big, big difference in the LLMs in terms of the output which actually does relate to your question, don't worry, I'm not going in and entirely different direction is that they try to anticipate the next question and they try to in the responses they provide, not just simply directly answer the question asked, but the next one and the next one, the next one. And that's fundamental to how the LLM algorithms work. And so this prompt, this follow on prompt, and I've been seeing it in ChatGPT too. I was trying to actually create an image yesterday for a presentation where I was talking about agentic browsers and I asked it to create an agentic workflow for a specific question that I had a client that was in a particular market and I asked, you know, for a summation of the workflow. And then, and then I asked him, you know, and I got a nice text list and I said, you know, put that in an image for me. And it did that. And then it came back and asked me, would you like to customize this image to a specific question? And then I did, because that's actually where I wanted to go. And so they're collecting data every which way they can. They're collecting data because they understand how important it is. I mean, the Google SERPs, which, you know, criticized at many levels over the years, it would have been a horror show if they weren't using user information because the algorithms to actually determine relevance well enough to determine which was the best answer, and links and citations to the degree that they used them wasn't enough. They needed more data.

Tyson Stockton (31:42)

Yeah, and it's almost like a scorecard, especially when you're doing this. It's like you need some sort of measurement or KPI of are we getting it right or wrong? To me, it just makes perfect sense that it's like, yeah, you need a scorecard.

Eric Enge (31:56)

Well, and they're measuring two things. Right. What's the actual intent? They're trying to understand the actual intent of a question or the range of intents. Obviously it's a range of intent for most queries or prompts, and so they're trying to understand that, and then they're trying to understand what satisfies those intents. So it's a furball of a math problem a little more complicated to implement than when I talked about building a cosine similarity measurement tool for yourself.

Tyson Stockton (32:27)

Absolutely. Well, with that, that's going to wrap up this episode of the Voice of Search podcast. Thanks again, Eric, for joining us. It's always a pleasure appreciative of the time that you spend with us on here. And if you'd like to contact Eric, you can find a link to his LinkedIn profile in the show notes, or you can be sure to check out his company's website@pilotholding.com if you haven't subscribed yet and you want a daily stream of SEO content marketing knowledge in your podcast feed, hit the subscribe button in your podcast app or on YouTube and we'll be back in your feed soon. So with that, that's all for today. Thanks for stopping by and we'll see you in the next episode.

Eric Enge (33:11)

Sa.