GEO and How to Rank in SearchGPT

Summary

The Digital Marketing Podcast: GEO and How to Rank in SearchGPT

Episode Overview Released on November 1, 2024, The Digital Marketing Podcast delves into the emerging concept of Generative Engine Optimization (GEO) and strategies to enhance visibility within generative AI search engines like ChatGPT and SearchGPT. Hosts Ciaran Rogers and Daniel Rowles explore the nuances of optimizing content for these advanced platforms, drawing parallels and distinctions from traditional Search Engine Optimization (SEO).

---

Introduction to Generative Engine Optimization (GEO)

At the outset, Daniel Rowles introduces GEO as the latest three-letter acronym in the digital marketing sphere, standing for Generative Engine Optimization. This new paradigm focuses on optimizing content for generative AI search engines, which operate differently from traditional search engines like Google.

Notable Quote:

"We've got SearchGPT, which is beta testing at the moment. You've got a number of different generative AI Search engines... how do you get visibility in those?"
— Daniel Rowles [01:06]

---

Understanding the Complexity of GEO

The hosts acknowledge the infancy of GEO and the limited understanding of how large language models (LLMs) like ChatGPT and SearchGPT function. Daniel emphasizes the challenge in optimizing for systems where the internal workings are largely opaque.

Key Points:

• GEO is still a developing field with no definitive guidelines.
• Existing SEO strategies offer a foundation, but GEO requires adaptation.
• An academic study by the Allen Institute of AI in Delhi (2023) explores GEO, highlighting the experimental nature of current optimization techniques.

Notable Quote:

"Nobody knows the answer and loads of people are writing articles about it... they just made this up."
— Daniel Rowles [02:30]

---

Practical Strategies for GEO

Daniel outlines actionable strategies to enhance content visibility in generative AI platforms:

1. Optimize for Bing:

   • Since ChatGPT utilizes Bing for real-time data, leveraging Bing Webmaster Tools is essential.
   • Implementing structured markup (e.g., for recipes, articles, authors) aids AI in understanding and indexing content effectively.

2. Create Conversational Content:

   • Adapt content to mirror the conversational queries typical in generative AI environments.
   • Focus on answering questions directly and thoroughly to align with how users interact with AI search tools.

Notable Quote:

"Start thinking about how people search within a generative AI platform... creating content that answers those questions."
— Daniel Rowles [05:25]

---

Enhancing Content Credibility

The discussion highlights the importance of establishing content credibility through:

• Citing Sources:

  • Incorporate bibliographies and link to reputable sources to allow LLMs to verify information, reducing the risk of perpetuating fake news.

• Including Expert Quotes:

  • Featuring quotes from industry experts and linking to their bios enhances both SEO and GEO by showcasing expertise and authority.

• Adding Statistics:

  • Backing claims with verifiable statistics and citing their origins improves content richness and trustworthiness.

Notable Quote:

"Citing sources... adding statistics to back up what you're saying... enhances both credibility and richness of content."
— Daniel Rowles [06:10]

---

Leveraging AI Tools for Content Optimization

Kieran Rogers shares practical experiences with AI tools to streamline content optimization:

• Embedding Relevant Links:

  • Using AI to insert contextually appropriate links to existing content can enhance SEO and GEO without manual effort.
  • Importance of verifying AI-generated links to avoid errors.

• Customizing AI Interactions:

  • Creating custom GPT configurations to prevent unnecessary self-referencing and ensure precise link placements.
  • Utilizing tools like ChatGPT’s Canvas for interactive document enhancements, such as adding quotes and references.

Notable Quote:

"I asked it to insert four or five highly relevant links to content we've already published, and it did it... highlighted the right phrases and embedded the links."
— Kieran Rogers [11:25]

---

Balancing Traditional SEO with GEO

The hosts discuss the synergy between traditional SEO and GEO, emphasizing that many effective SEO practices inherently benefit GEO efforts. By focusing on user-centric content that delivers value, marketers can satisfy both search engines and generative AI platforms.

Key Insights:

• User-Centric Content:

  • Prioritizing the needs and interests of users ensures content remains relevant and authoritative.

• Structured Data:

  • Implementing structured data aids both SEO and GEO by providing clear context to search algorithms and AI models.

Notable Quote:

"Focus on the user. If you do that, you're doing good stuff for the user and helping out those models and the search engines as well."
— Daniel Rowles [14:20]

---

Challenges and Considerations

Daniel addresses potential challenges in GEO, such as:

• Competition with Large Organizations:

  • Small businesses may struggle to compete against larger entities that dominate key generative search queries.

• Niche Focus:

  • Identifying and targeting niche areas where smaller players can excel without competing directly with industry giants.

Notable Quote:

"The risk is that when you get these tools to look things up online, what they bring back is the top results. So any small organization competing with big organizations is going to struggle."
— Daniel Rowles [09:45]

---

Community Engagement and Future Directions

The episode concludes with an invitation for listeners to engage and share their experiences with GEO. Daniel promotes upcoming masterclasses and encourages feedback to shape future discussions.

Call to Action:

• Join Target Internet's membership for exclusive masterclasses.
• Share success stories, best practices, and challenges related to GEO.
• Utilize resources available at targetinternet.com/podcast for further learning.

Notable Quote:

"We would absolutely love to hear what you've been up to and how you've been playing around with it or whether you found any flaws in it or any workarounds."
— Kieran Rogers [16:12]

---

Conclusion

The Digital Marketing Podcast episode on GEO provides a comprehensive introduction to optimizing content for generative AI search engines. By blending traditional SEO techniques with emerging GEO strategies, marketers can navigate the evolving digital landscape effectively. The hosts emphasize the importance of user-focused content, credible sourcing, and leveraging AI tools to enhance visibility across both traditional and generative platforms.

---

For more insights and resources, visit Target Internet Podcast and stay tuned for future episodes covering the latest in digital marketing trends and strategies.

Transcript

[Kieran Rogers]

Welcome to the Digital Marketing Podcast brought to you by targetinternet.com hello and welcome back to the Digital Marketing Podcast. My name is Kieran Rogers and I'm Daniel Rolls. And today we are talking about Geo. Okay, right, confession time. I haven't got a clue what that is.

[Daniel Rolls]

Yeah, I found this on purpose.

[Kieran Rogers]

I tried Googling it because you said that much. He wouldn't let me look at the coggle. I'm like, what? What? He's like, no, I don't want you to know. So I googled it and I like, don't. I don't think it's any of the things that Google thought it might be nice.

[Daniel Rolls]

And it's a, a new thing which might not be a thing. So first of all, we should say Louise isn't here because she is at the Semrush conference in Amsterdam. So she will be back in a couple of days and we'll be doing a summary of that. So you just lumbered with the old, the old school today. And I want to go Kieran to go into this blind because I'm not sure if he's gonna love this or hate it, but I'm pretty sure he's gonna hate it. So first of all, pretty sure.

[Kieran Rogers]

Basically we're gonna be Kieran baiting today. Yeah.

[Daniel Rolls]

There's kind of the aim, it's kind of the end. Yeah. So what we've done is as an industry, we love a three letter acronym. Right. And we've, I've gone on about this for years. We're just obsessed with three letter acronyms. Well, now we've got a new one and this is Generative Engine Optimization. So you've got Search engine optimization.

[Kieran Rogers]

It's like a boy racer thing.

[Daniel Rolls]

Generative Engine optimization.

[Kieran Rogers]

Dig out my fluffy dice and my Tracy and Daniel.

[Daniel Rolls]

Tracy and Daniel.

[Kieran Rogers]

Oh, I don't know. That was rude, isn't it? Because that's your name.

[Daniel Rolls]

That's what I thought. I mean, and now you're gonna have people asking, what, did you have an ex girlfriend called Tracy and I never did. And then you're gonna get Susanna my wife suspicious and it's not gonna go well for me. So thanks for that anyway, we'll move on. It is basically, how do you get to the top of these generative search engines? So, you know, we've got, we've got Chat GPT and we've got Search GPT, which is beta testing at the moment. You've got a number of different generative AI Search. You've got Search Generative experience built into Google and it's basically people asking how you get visibility in those. Now I'm going to start with a bit of a caveat. Nobody knows the answer and loads of people are writing articles about it, so. And it's clearly that thing of people are just getting stuck into it and going, yeah, well, I imagine it will be like this. And I saw this brilliant table in the article that comes number one in Google when you dig into it a bit and it said these are factors that could be having an impact. And it's like you've just made this up. You just literally made it up. Now the more I dug into it, the more I started to go, well, there are some things we can untangle from this. Now what I came across is that where this, this term was kind of coined in 2023, there was an academic study written about this and it was in the Allen Institute of AI in Delhi and it's titled By a goal and all 2023. And we'll put the link into the show [e-mail] forward slash podcast. But they basically gone and said, okay, look, there's loads of these new large language models out there. How our business is going to get listed in those because with search engine optimization we know roughly what we're doing and we can make sure we optimize and it gives us an opportunity to, to kind of get our content and our businesses featured. But actually these large language models are somewhat of a black box and therefore what do we know? And they've gone through and they've done a very scientific study into what works and what doesn't work and they, they looked at around 10,000 different search queries as well. The problem is, and I've, I was reading the paper with my kind of academic head on, trying to just go through and understand the study and it's a good study except for they based it on a large language model in particular, but they're all different. So yeah, this one, but it might not work for Chat GPT, it might not work for Claude, it might not work for Gemini. So it's like, oh, okay, so even this is a little bit not sure at the moment. Then I've spoken to someone that knows quite a lot about large language models that has some good connections and OpenAI and they're saying, well, if we don't know how the large language model works, truly the connections it's making, its neural networks, how could we tell you this is the way to get to the top of it? They can Some key principles. But what I thought was interesting is if you just think about this from practical terms, how does Google do SEO? They've been building an algorithm for decades basically now of trying to tease out what's really good and what isn't. And the thing they've really lent into recently is twofold. One is user signals. Does the user scroll? How long they stay on the page, do they bounce back to Google? The large language model will not have any of that stuff. It's just not something to have and then eat, which is the experience, expertise, authoritativeness and trust. Well, the large language model can untangle some of that stuff to some extent. But even, you know, we don't even have absolute clarity on how Google's doing it, but we can kind of read between the lines in, in some cases. So what I then thought is actually one of the things about Search GPT is that it is very much looking at real time data as well, that real time data is going to come from Bing because that's what ChatGPT does. It looks stuff up in Bing so suddenly, and you've been saying this for, I don't know, 10 years, we really should be looking at Bing more Bing Webmaster tools, which is the equivalent of Google Search Console is something you should really be looking at and going, absolutely. What are the opportunities in there to make sure that I'm, I'm doing everything I can to help Bing find my stuff? Because that's going to definitely have an impact. But then within Bing you can do structured markup like you can in Google. So you can say this is a recipe, these are the ingredients, this is an article, this is the author. This structured markup is going to help to some extent as well because it will help the AIs understand the content and it's going to help you show up in Bing. So absolute no brainer Bing. Look at it. Get your webmaster tool set up as well. Then think about how people search within generative AI platforms like ChatGPT and Gemini and if they ask questions. So actually thinking about that more conversational type of content can have a big impact. We spoke about this phases people started going on about, I think it was 2022, maybe it was the year it started happening. Oh, because of AI voice search, it's really important that you answer with. And everyone's like yeah, that makes sense. And then you go, didn't make any sense because Siri and Alexa were absolutely terrible. You say do you know the answer to this? And they go, I'll look something up for you. I found this on the web and then it would read out this thing that was completely irrelevant to what I just asked it to do.

[Kieran Rogers]

Yeah.

[Daniel Rolls]

So but actually now suddenly, okay, yeah, those conversational searches, asking and answering, basically answering questions in your content directly, that's going to become more important. So start thinking about run, get your Bing in line. Make sure you've done that. If you can do any structured markup that's going to help as well. But then thinking about how do people search differently within a kind of chatgpt environment to how they would in the search engine and then start creating content that kind of answers those things. I think that that certainly can't, can't hurt in the slightest. The other thing that came out from this study that I thought was, was pretty interesting as well. So they came up with a whole load of evaluation tactics and it came out the ones that were most important from this study. Specifically our top performing methods. Citing sources. So when you're writing about something it's quite easy to say with great improvement in Instagram, you know, the growth of Instagram but actually citing yourself, showing how much it's growing so that, that can be good. Adding more of a bibliography or you know, linking through to your sources so that the large language model can validate where this input insight is coming from. Because one of the things we've realized with these large language models is that they're trying not to perpetuate fake news. So in order to do that they're looking at sources of the insights and trying to marry those together. So that can help adding quotes from experts. So actually having a little breakout within any particular article with a quote and the name of the person and linking person through to that person's bio. Now actually that's a really good tactic for eat for Google anyway because it shows you've got the expertise and you've got the right, the right people involved. So that that makes sense as well. I don't think can't do any harm. And then it's an adding statistics and again so adding stats to back up what you're saying, getting citations to those stats and putting those in in that case as well. And they said with that achieved a relative improvement of 30 to 40% these methods require minimal changes but significantly improve the visibility in GE generative engine responses enhancing both credibility and richness of content. So to me those are really good examples of what Google's been looking for as well. So. So I just think the overlap between search engine optimization and generative engine optimization is going to be huge. And that the large language models would love all that insight that Google has got from Google Chrome and Google Analytics. But they don't have it, but they've got all the Bing stuff or they're probably going to have access to that Bing stuff. Now, bearing in mind that Microsoft is the biggest shareholder in ChatGPT, there's quite tight connections between the organizations. They maybe could share some of that data potentially as well. So I think immediately in there there's a bit of a step twist. Think about the conversational kind of questions and content. Make sure that you're going through demonstrating eat by adding citations, adding statistics, getting quotations, people. That's going to help Google as well. And then making sure you're looking at Bing at this stage and making sure you're appearing in those places, then you're in a kind of stronger position. The risk is that when you get these tools to look things up online, what they bring back is the top results. So it means any small organization competing with big organizations is going to struggle with this. So trying to work out what your niche is, is becomes even more important. And that was always the case in search engine optimization. But at least I knew then you might go to page two of the search results in some cases. Not many people, but actually I can't do that with a large language model while using ChatGPT. So making sure that, okay, if I can't dominate those phrases, having a strategy that says, but this stuff I can dominate because no one else is doing this appropriately as well, what can I say?

[Kieran Rogers]

I mean, I'm just a bit disappointed, Daniel.

[Daniel Rolls]

Why is that?

[Kieran Rogers]

Well, because this hasn't pressed any of my buttons at all.

[Daniel Rolls]

I thought another three letter acronym would upset you.

[Kieran Rogers]

No, not really. It makes sense, I think. Look, it's always been like this, like the sweet spot is somewhere between what a machine likes and what actual people like. Yeah, it's a win. That's a win. Win. And that this ticks boxes on both accounts really, isn't it? And actually I think makes for better content. You know, this is one of the big problems with kind of AI slop really, that it's just pumping it out. Is that the phrase that is the phrase now, yeah. Do you know I picked it up on a marketing call that I was on and I've been obsessed with it since. It's like, yeah, AI slop.

[Daniel Rolls]

As opposed to Jay Swederson's tsunami of word that I used before.

[Kieran Rogers]

Yes, AI slop. Yes, we can say that. Can't we and still retain.

[Daniel Rolls]

And it does add value to it. It does go through and say, okay, let's, let's go through and improve that stuff.

[Kieran Rogers]

Something I tried recently and I was amazed how well it worked. I was using an AI large language model. I won't specify which one, but it's quite well known. And I asked it who.

[Daniel Rolls]

I mentioned ChatGPT about 30 times already. So I don't think I asked Chat.

[Kieran Rogers]

GPT to go through the article that I just created for the blog and I gave it my company web address and I said, could you insert four or five highly relevant links to content we've already published on this website?

[Daniel Rolls]

And it's.

[Kieran Rogers]

And it did it.

[Daniel Rolls]

Yeah. Right.

[Kieran Rogers]

And it didn't just pop the links in, it highlighted the right phrases and embedded the links in into it.

[Daniel Rolls]

Yeah.

[Kieran Rogers]

And that's it.

[Daniel Rolls]

Yeah.

[Kieran Rogers]

And if you want to like, if you have a go at doing this, don't do what I did and cut and paste what was on the screen because that doesn't work. I just embit the link. They just, you just get empty links and anything. Then you have to actually copy it from. From Chat GPT using the copy button. And then you have to, don't get caught out with this. You have to remove the little intro bit and the bit at the bottom that you basically inside. So but you know, but this could be done with other things as well. The golden rule though, like with anything before you start inserting links automatically, you have to go through and check them before you hit publish. Otherwise, you know, you could just be furthering all sorts of shenanigans.

[Daniel Rolls]

Yeah, a couple of tips then with that, in order to do that, to get it writing and not putting in those introductions. And I'll write this for you now. If you create a custom GPT and you tell it, do not self reference, do not explain what you're doing, it won't put that stuff in. It will just write for you as well. The other thing is, if you create a custom GPT, you can give it a file that goes through and maybe like all the sections of your website you'd like it to look at, you can put it into that document and it can use that to look up all the different things on your website. You can categorize them. So you can say, you know, go to my SEO category and find me five references from this and you can list it that way as well.

[Kieran Rogers]

So it's also really good for finding original sources that you can reference. So one of my content writers had referenced A study by a particular body. So I just asked it, can you find this reference?

[Daniel Rolls]

Yeah.

[Kieran Rogers]

And yeah, it'll go through the finding because it's the new one's pretty good. It's like, it's pretty up to date. Yeah, this is just content just published back in July. So I was pleased it did that.

[Daniel Rolls]

Yeah.

[Kieran Rogers]

So like there's lots of help now for doing this. Whereas previously this could be a very time consuming thing. Actually it's just a mechanical process you can go through to, to aid your content creation.

[Daniel Rolls]

The other thing I would do as well is if you use the latest, the paid version of ChatGPT, you've got 4.0 with Canvas and Canvas allows you to interact with the documents. You can select a section and then say add quotes to this section and extend this section and add references and so on. So it makes creating those, those kind of documents that are going to really help with search engine optimization, but also going to help with generative engine optimization as well. I'm in two minds. I don't know if I like this as a separate term, but I think the reality is we're going to have to because there's going to be more and more of these things out there. But the overlap between what's the search engine and what's the generative AI engine is going to. Because you've got things like Perplexity AI and Perplexity can use multiple large language models, but it brings things in from different sources. So all of them are going to have slightly different ways of identifying. But Google has been at this for decades and therefore they're kind of going to be ahead of the curve in a lot of cases in doing this, when Google Gemini is hooking up to Google more as well, suddenly maybe they're going to take a leap forward. So I think it's like we've had an SEO for years. You never really know what the rules are. You've got to keep an eye on all the rumors that are out there, but just focus on the user. If you do that, you find yourself in a situation where actually you're doing good stuff, the user and you're helping out those, those, those models and the search engines as well.

[Kieran Rogers]

And focus on Bing because it's cool and it just might give you the edge. Just might, just might.

[Daniel Rolls]

You never know.

[Kieran Rogers]

I can see some argument in that.

[Daniel Rolls]

So we have got lots of masterclasses. We've got another SEO masterclass coming up. We're going to get into this a lot more as well. So just a Reminder that if you're a member of target Internet.com you get these monthly half day masterclasses. They're live taught by me and you can attend and ask lots of questions. So where else are you going to get that for £20amonth? The half day training course. So go to Target Internet. If you're a 6 or 12 month member, you get that. And if you've been a member for six months on a monthly membership, you also get it as well. I'm going to put this academic paper into and a couple of articles into the show notes as well. So target Internet.com forward/podcast, you can take a look at that and we'd love to hear your opinions as ever. Do you think this is a thing that's here to stay? Is it going to merge into SEO? What's the, what's the future of this? Let us know your opinions and we do share them within the podcast as well.

[Kieran Rogers]

Request for me. I'd love to know for anybody of you out there if you had success doing this. It's really handy to get you guys to share, you know, your best practice and we would absolutely love to hear what you've been up to and how you've been playing around with it or, you know, whether you found any flaws in it or any workarounds. You know, share that with us. It'd be great.

[Daniel Rolls]

Tell you what else can help. The AI HubSpot grader, which is them trying to show your visibility in the large language models with ChatGPT in this particular case. It's quite a good analysis and actually it'll end up being quite a good GEO tool as well from that point of view. So we'll put, put that into the show notes as well.

[Kieran Rogers]

We just created a nightmare for any Wikipedia entry editors. Yeah, that's where you need to get it, people.

[Daniel Rolls]

That's it.

[Kieran Rogers]

That's it. Good luck. Good luck with that. They are ferocious in defending their content.

[Daniel Rolls]

Yeah, they are, yeah. Wikipedia is not an easy thing to get yourself into, but that's a, that's another episode in its own right. So thank you for listening to Digital Marketing Podcast. Good luck with your SEO Geo and any other EO efforts that you're making and we will speak to you again very soon. For more episodes resources to leave a review or to get in contact, go to targetinternet.com forward/podcast.