Go back

How to get your brand mentioned in AI search results with Charlie Marchant, CEO of Exposure Ninja

60m 16s

How to get your brand mentioned in AI search results with Charlie Marchant, CEO of Exposure Ninja

The conversation between James and Charlie, CEO of Exposition Ninja, explores SEO in an AI-driven world, focusing on how brands earn mentions and citations in platforms like ChatGPT, Gemini, and Claude. James introduces the session, noting Charlie’s ability to simplify complex topics like SEO and GEO. Charlie defines key terms: mentions (brand recommendations in AI answers), citations (links to sources), and sources (reference articles). She clarifies that while SEO and GEO share tactics like technical optimization and content, their strategies differ because each AI platform has unique algorithms, similar to social media channels like TikTok vs. LinkedIn. For beginners, Charlie recommends tracking AI referral traffic via Google Analytics to identify which platform drives the most conversions (e.g., ChatGPT for most businesses, Gemini for Google Workspace users). She advises focusing on middle- and bottom-funnel topics, as AI queries are more verbose (averaging 60 words) than Google searches. Despite low click-through rates on citations (around 2% of organic traffic), Charlie emphasizes their value for brand awareness and trust, especially as AI usage grows. She highlights three core pillars: technical SEO, on-site content optimization, and off-site digital PR (reviews, forums, news). Nuances exist between platforms—Google Gemini favors Google reviews—but good SEO work generally benefits all AI systems.

Transcription

10470 Words, 57696 Characters

English
The marchin is the CEO of Exposition Ninja. And I first found out about Charlie on a social media clip, I was just telling her this a minute ago backstage. And I was blown away with how articulate she is when it comes to explaining topics that frankly I don't know enough about. And she distilled them into really simple concepts that made the idea of SEO and Geo and all things AI a lot less daunting as a result. So I think we're going to have a great session today. And another thing to mention is Hive. This is a space that we run for free where you can stay connected after the event. You can get some port from other marketers, learn between the events. And we also have RCA expert every fortnight. So it's a place that you can hopefully learn more and connect and build your confidence. And that's all headed up by Sarah, who will know that I'll be in the chat somewhere. And then finally, because the sessions are brought to you free, we have sponsors who help the sessions come to life. And today we're going to talk about one sponsor in particular, which is planable. And I've put a link up in that top right corner there. Planable have for years work with us and they're always coming up with really, really useful resources. And the resource that I put there in the corner is about the visibility that your social media has in LLMs and AI and how social media can make a real impact on those things. Now, because planable and our part of SE ranking, they've crossed their knowledge of social media with SE rankings, knowledge of search, and they've given you some definitive answers. And so I think that will be a really useful resource, especially if you're here for the topic of search and citations today. Also a huge thank you to all our other partners that will talk about the future weeks. We've got Canva, MailChimp, Wistia, Frontify, and Prismic. That is my introduction done. I'm going to remove that from stage and I'm going to bring forward the very brilliant Charlie. Hello, Charlie. Hey, James. Thank you so much for having me on. I knew I'd do something wrong at the beginning, but hey, we are where we are. Today, I didn't actually say today is going to run as a Q&A. But if folks, if you've got questions for Charlie, drop them if you see in the top right-hand corner, you've got the chat function and then you've got the Q&A function. So if you pop those in there, after Charlie and I have had the chat, we will then go and pick some questions from there and hopefully answer those for you. So Charlie, I'm not particularly verse with SEO and Geo. Like, I'm fairly low down on the knowledge scale of this topic. It's really important in a conversation like this to set the table stakes. So we all know what we're sort of speaking about. The title of this session is SEO in an AI summary world, how to earn mentions and citations. So I wanted to ask you the best way to have this conversation when we speak about the master at hand. Is there a nuance that wants to be understood in stuff such as Google overviews versus LLMs, such as chat GPT, Claude? And really at a really basic level, what do we mean by mentions and citations? Like can we just get the groundwork in first? Let's get the definitions done so that we all know what we're speaking about. Absolutely, James. So the important thing to know about mentions, citations, and sources, these are probably the three that come up the most. A mention is in an AI answer when the response names you, so it names a brand, a product, a service, as a recommendation in an answer to the user's query. So that's going to be something like I'm searching for the best of edgeboxes in London. And then chat GPT might recommend me Riverford, Abel and Cole, London edgebox. Those are mentions. Then we also get citations. And this comes up in chat GPT and other LLMs answers and AI overviews as well. So citations are a little bit different. They are the references or the links that sometimes appear in those AI generated responses. They usually point to the original source of the information. So that might be an article, but sometimes it might be your own website. So that in that example I just gave, it might be Riverford's own website, but it also might be an article like BBC Good Food, or a blogger who's doing a roundup of the best edgeboxes, or someone who's reviewed a specific kind of edgebox. So sources, I like to think of a little bit like when we're at university right, and we had to write essays, dissertations, and then we had to take a whole chunk of time to write at the end of that essay, every article that we'd referenced, every quote that we'd call them, or that came from, yeah, those are the days. Well, very fortunately though, AI also does this. So chat GPT answers, Gemini, complexity, chords, all of those. If they have had to use their search index, they've had to do a live search in order to pull sources. So they're not just answering from their pre-trained data, which they'll do for very simple questions. Then they will provide those sources. If you haven't seen them before in a chat GPT answer, you just scroll down to the bottom. There's a little button in the bottom footer that says sources. You click it, and that opens up all of the articles that were used as sources, used as reference points in that answer. They won't always appear as in-text citation links, because sometimes it may have just lifted specific parts of content rather than directly quoting them and putting them in. So that list, depending what you're asking, that list can be quite long. Yeah, yeah. And presumably, when you are worried that maybe you're reading a hallucination, like a false answer, this gives you a bit of validation to say that this is actually a logic, because you can click through and go to the original source. Yes, absolutely. And if you ever read something, you think, that seems a bit. I'm not sure about that. Then you can actually check the sources to see what were the reference articles. And you can then decide yourself, do I trust that reference article? It might be something from the Guardian or the New York Times. Do you think that's really reliable? Absolutely. Or you read it and you think, "Hmm, I have no idea what that website is." Or who that blogger is. Or if that data is really legitimate, now that I'm actually going into the source link and having a look at it. Yeah, that makes sense. Okay, that's cool. So we've had a few seasons on AI. And we've had talks before that kind of have talked about search. And I tend to simplify things. So in my head, I was having a conversation with Joe. And I said, "I hear people talking about Geo and the principles of how to rank and appear." But they sound very similar to SEO to me. Yeah. And where do you see the meaningful differences between the two, if there are any? Because in my head, I was like, "Did people talk about Geo?" "This is just SEO. Is it as simple as that?" Or I'm guessing not. Yes and no. I think this is probably one of the biggest misunderstandings about AI search. So there are clear overlaps between SEO and Geo. And this is because the tactics are the same. The things that we actually do in order to influence showing up in answers in chat, GPT or Gemini, they are things we know from SEO. The things like technical optimizations, the things like our schema and our structure data on our website, optimizing content, doing digital PR online, which is hugely influential in AI search. But the strategies are different because we're optimizing for different platforms. I think we've become so keen on Google that we think of SEO as meaning just for Google. Yeah. But really, we should be saying SEO for Google. SEO for chat GPT. Because now we have so many different platforms. We're trying to optimize for Google. We might be trying to optimize for search on YouTube, AI mode, Gemini, chat GPT for Plexity, Claude, all of these different LLMs. And if we compare it to social media, for example, no one ever says, "Oh, well, you just posted something on TikTok and you posted exactly the same thing on LinkedIn. It's all just social media." But 100% of us would know that the strategies we're actually doing for LinkedIn and TikTok are going to be incredibly different. If you posted exactly the same piece of content on LinkedIn and TikTok, it's going to perform better on one than the other, depending what you're posting. So I kind of like to think of it that way. We're actually optimizing for different search channels. They don't use the same algorithms. So even though we're employing the same tactics, we are actually following different strategies to get to a different result. Yeah, that's a great way of explaining it. Because I think people have been on social media long enough to know that, probably back in the day, we would just post the same thing on every social channel and now it's completely different. So I love that. That's really useful. Absolutely. And like the tactics of social media, they're all very similar. Essentially, we are just writing texts, doing graphics, videos, posting content, schedule and content, interacting with other people's content. But our strategies are very, very different between those different social platforms. And I think we're seeing exactly the same now as search. It's just that many of us have for decades been optimizing for Google. And now we're coming to terms with the fact that actually there's a lot more search platforms coming into the mainstream now. - Yeah, got you. That's really interesting. I, like when you talk about strategies, I kind of wanted to ask like if you were a marketer starting at, you know, has zero geo strategy at the moment. You're in a, you know, traditional business or a business that hasn't looked at geo before. I was going to ask what the, what the first thing is they should, they should do. But I guess I'm kind of asking what are the questions they should be asking themselves as well, because, you know, I was talking to somebody the other day who was looking to help their partner launch a wedding business and they wanted to do everything. All it went, looks like slow down. Like you're not going to be able to do that. You're in the startup mode. You need to be strategic with what you do and what have you. How would you approach that when it comes to geo? - I love that question. And I think this can feel like the same thing. It's like a really overwhelming topic and you think I'll just try and do everything at once and optimize for every platform. The very first thing I would do is if you're a business with a website that's already operating, it's already up and running. And you've got your Google Analytics up and running. One of the things a lot of people don't know is that you can actually see your referral, trafficking conversions that have come from AI, so from ChatGbT, Gemini, PlexT, other LLMs, by creating a channel group in your Google Analytics. That's an amazing first starting point because even though we don't get 100% complete data from Google Analytics, we'll only see the people that actually clicked through from AI searches. A lot of people, of course, have their conversation AI and then go to Google and type something in. So we don't see those people. But we can at least get an initial starting idea of which of those AI platforms is driving traffic to your website and which is driving conversions and what percentages that of your overall organic traffic and has it been growing over the past couple of months, over the past 12 months, ideally we'd be looking, I think. From there, you can then actually decide what platforms, what AI search platforms are my target customers actually using, which then means you don't need to think about optimizing for every single AI platform. You can focus in on the one that is right for you. And the majority of businesses, the majority of websites are going to see that ChatGbT tends to be the one driving the most referral traffic and conversions. It's sometimes Gemini. And if you're B2B, it's sometimes co-pilot because if you have customers who have to work on Microsoft's suite at work, they're using co-pilot instead. The same if you have customers who have to work on Google Workspace, they're using Gemini instead. But that gives you a great starting point to decide, okay, what do I even want to optimize for? And then the second question is, what phrases do I care about showing up for, what topics do I care about showing up for in those AI searches? And it tends to be the middle and bottom of funnel topics of the ones we really care about. We know this from SEO keyword research as well, right? Top of funnel is great, particularly for brand awareness. It means people understand and learn who you are. But middle and bottom of the funnel are the ones that actually drive conversions. And if we're a for-profit business, which I'm sure most of us are, they're the ones that actually drive the leads or the sales that we want. And so you can use that keyword research to then decide the prompts. And I'm saying prompts because people don't just type keywords into a chat TVT. The stat I read was that the average number of words someone writes into chat TVT when they're asking questions about 60 words, which is huge compared to the average on Google is about four words. So we're much more verbose when we're asking chat TVT because we feel like we can give so much more context and it's going to understand. Which means-- - That brings you interesting. - We need to think more broadly about topics that we want to show up for rather than just this specific keyword that we want to show up for. And then from there you start building out your strategy around what kind of answers does chat TVT or Gemini give me back. And am I showing up in them or am I not? And does it describe my business how I want to be known? Is it positive about me or not? Does it say I'm expensive? But actually I'm really good value. Whatever it is that you're positioning us. - That is really interesting. I hadn't thought actually that when you're on Google, it's like best cafe. - Yeah. - You just want it to just give you an answer. That is opposed to learning a lot deeper. In my head when you mention chat TVT, I think it's the one that my dad talks about it. Like he's mid 70s and he's that, although he calls it chat PT or something. Everything I got the word right. It's terrifying that he's started to use it. But as a business we've switched from chat TVT to Claude. And if you'd asked us 12 months ago, we were learning rapidly about using chat TVT, how it can help the business, what we can do. So all of a sudden that's no longer a thing. And we're on to Claude. With the work that somebody would put in to gain citations and mentions, still be just as valid for something as similar as chat TVT and Claude and Gemini. Like are there big differences between those platforms? - This is a great question. And I would say there's nuances between those platforms rather than big differences. So the core principles of how you're optimizing a very similar, I like to think of it as they're being three core pillars. The first is your technical, what we know from technical SEO, your technical side of your website being excellent. The second is your on-site content, how well optimized it is. There are nuances and differences between the types of content that each AI system likes to surface and what they believe to be factual or most relevant to the query. And then the third pillar is the off-site side, the digital PR, the story that is told about our brand online. So all of these AI platforms are pulling in information sources from across the web. So if I ask, you know, is the marketing meetup a great place for me to learn from as a marketer, as someone who's just come into the SEO sector, for example? It's not just pulling that information from the marketing meetup's website. It's also looking for reviews. It's looking for people's first-person experiences across forums like Reddit and Quora, across social media platforms, across review sites. It might look at events, it might look at the types of speakers that have been on. It might look at, you know, the marketing meetups mentions in news publishers and any awards that they've won. All of that type of thing is being brought in. - It's different. - But different. - This webinar as well. - Yes, exactly. - Yes, it is a great place. - Yes, it's wonderful. All those different sources are being pulled in. And there's nuances between which sources platforms like to reference. And so you can make some educated guesses as well, right? Google, Gemini and AI mode are part of Google's product stack. And so Google loves to keep people in the Google ecosystem. So they're always going to reference Google reviews. A lot more than they would be referencing trust pilot reviews, for example. Those nuances are there for each of them. But pretty much I would say all good work that you do, the same as all good SEO, it's going to benefit you overall. - We're good. Okay, that's really nice and clear. And to keep us on the citations and mentions topic, one of the things that whenever I've used NLLM to get some information, I, it says, what I've written down is, some of the data seems to suggest that click-through rates on citations are actually quite low. And therefore, should we be optimizing that content, for that content, if people are getting the information from us, but not knowing that it's coming from us, and then just disappearing. Like is there value in gaining citations? If they're not ever kind of gave you with a brand, if that makes sense? And people often tell me that click-through to citations are low, even though we don't know the full data on this. Again, you can see how many clicks through you've had from the referral traffic that comes from those AI platforms. And it tends to be, we work with clients across all different sectors in B2B and B2C. And across the 60 or so clients that we've worked with in the past year, I would say there's around 2% of their organic traffic. It's coming through as referral traffic in their analytics. So that's only a small percentage. The interesting thing about it is it's growing, right? ChatGbT, Gemini, FlexT, these are all new channels. Google's been with us for decades. So of course they're bringing through the most traffic, what we're interested in is how much of it's growing. Whether the citations are a value to get though is kind of a two-part story. The first part story is if it's something very top of funnel, you're unlikely to get click-throughs anyway. The same as the Google search results now, because AI overviews take up such a significant portion of informational searches that the likelihood of someone clicking through to your website is so much lower. But also that person was probably never really converting on your website anyway. It was a bit of a punt, right, that they might-- convert on your website. If they learned enough information about the brand and you happen to be exactly what they were looking for. But if they were just searching, you know, what is a mortgage? And they've got the answer to the question because they haven't yet taken out their first mortgage and they're sort of scratching their head about what to do and what can they afford a house for the first time or how much money can they take out based on what their salary is at the moment and they're trying to multiply it by five. They're just trying to get the information. They're not actually ready to apply for a mortgage at that stage. And this will be exactly the same in Chatchy BT. People have very top of funnel conversations when they're in research stage. The core difference, though, is that these AI platforms, because they're conversational, people actually move through the funnel much quicker. Right. Because they're learning asking the question back, rather than when we're searching on Google, we used to have to create all these different searches. We've suddenly got 10 browser tabs open and you think, you know what, I'm just going to get a coffee because I'm feeling overwhelmed. And we actually came back to it maybe a week later, rather than just finishing the chat. And then the second part of this story is the middle and bottom of funnel terms. If I am searching something like what I mentioned earlier, the best veg box in London or the best veg box near me, I need a veg box. I'm trying to sort it out. I need a veg delivery. I don't have high, am I just want my organic carrots coming in every Friday. I'm trying to avoid going to Sainsbury's again because it's driving me crazy and I can't go to the supermarket after work for a day this week. That the value is really high, right? It's the same as if we think about the CPCs in the search engine results. You bid for your ads on on much higher value commercial terms because you're more likely to actually get a sale from it and even though we're not really at the stage of sophisticated ads or and certainly not mainstream ads with with chat, CBT and other AI platforms, for organic, it's an amazing opportunity to actually get those types of conversions that you want to get. And this is why we see studies from time to time coming up showing that conversion rates from chat, CBT and other LLMs can actually be higher percentages than Google. But it's because someone's just had that whole research conversation all in one go and then eventually they've got to the recommendation for your brand. Yeah, yeah. I like sort of 10 years ago plus God, it's moving very quickly. I was in an agency that that had an SEO department and I just definitely remember them talking about things like Trust Schools on the on the website. I don't even know whether that's still a thing in SEO but like the website had like a school with a Google as well, it was well trusted. Is there a similar thing within LLMs where it's picking up citations to answer questions, but actually it's doing that because it trusts the source more or is that is that not there? Yeah, we have no concrete measurement of this, but I would say something similar to that is definitely happening in the background. So the algorithms or the way that each AI platform works is slightly different and they're all different from Google. We could argue that Google is in some sense is much more sophisticated because they've been optimizing their algorithm rolling out updates that have sent the community into a spiral for years for decades for a really long time. Google has been doing that and trying to make the search results as objective and as best that they can and there are ways that people I think try and manipulate AI search in the short term sense with listicle articles and top 10 articles and putting out that type of content because those algorithms are a bit newer and they're still learning what sources to trust but there is definitely some sense of AI is trying to use sources that it feels trustworthy but it's also just trying to answer your question really really well and if you're asking something very specific it's going to pull from sources that have been able to answer that specific question and intent as well. Yeah, yeah, yeah, it's I guess it's looking at the yeah just finding the finding the best answer for you as the main priority, yeah, of course. Yeah, I guess this is so this the question I was going to ask is around the types of content that are getting cited right now and whether there are certain types of content that teams are wasting times on. So whether it's like listicles are actually a thing of the past and like are there things that that marketers are doing that you think are a waste of time or are there opportunities that the people are missing that not enough people are doing. I can't be interested to know either under the scale whether you've got. Yeah, I love this question because there's there's a lot to get into I think what a team's wasting time on when it comes to AI search platforms and I'm talking about LLM's the Chatchee BT Gemini per Plasty Claude. Yeah, there tends to be an over focus on our own websites which is what we know from optimizing for Google because when we're optimizing for Google we're actually trying to get our own web pages from our website to rank in the 10 results in the 10 blue links and in the AI overview as well. Now when we're optimizing to appear in Chatchee BT answers, we want our website to be mentioned but remember those mentions are influenced by sources and a lot of the time the sources aren't content from our own websites. So it's that the stat is around 85% of the time even when your brand is showing up it's actually third party websites being about your brand that's being used as the sources in AI. So a huge over focus on our own website and actually the best way to manage that is to think about what pages from our specific website are actually showing up. So for example in B2B SaaS their pricing comparison pages are going to show up because everyone is comparing prices of SaaS products before they buy. So something like zero if you're looking for an accounting software versus competitors we know that their pricing plans are going to show up alternatives and comparison articles are going to show up from their website. So it's well worth optimizing those specific pages on the website. This is actually a big problem as well because for branded search when we're thinking okay I'm putting exposure ninja into AI checking if they're a good company. It can be really difficult to put in lots of different searches you're like it's a nightmare. You're manually putting in all these different searches into Chatchee BT Gemini Complexity to try and get an idea of your brand and where it's pulling citations from. We actually just launched a tool yesterday I hope you don't mind me saying James. No no good. Called mindmybrands.com. Mining is in you know Snow White in the 7th of all of a sudden mining with a pickaxe mindmybrand.com. And what it does is it pulls the citations and AI's view across four major AI platforms of your brand. So how you're seeing in AI and what specific pages are influencing those branded searches about your business which then allows you to actually make changes to the specific pages without going on a wild goose hunt of thinking I'm just going to put hundreds of pages on my blog and hope that they show up in AI which is highly unlikely to actually happen. You want to be targeted. Going back to those third party sources though and the common types of content that's actually coming up. This hugely depends on the niche and the kind of target money phrases. The kind of middle and bottom of funnel terms that you really want to actually show up in. So I mentioned Best of List. They're showing up a lot. But this is because it mirrors how we tend to search for something. So if I'm looking for the Best iPad Aircase, I am just writing the Best iPad Aircase when I'm searching for that. I might be putting things like most durable or it's not going to break when I drop it because I drop it every day or under 30 pounds or whatever it is or I want it to be yellow and have sunshine and rainbows on it. Those kinds of modifiers as well but a lot of the time we are just looking for the best one and we want to get there quickly. But actually you can optimize for those in third party websites but the third party websites that tend to get referenced are niche. So for something like Best iPad Aircase is, it's going to be Best of List on on wired, on ZDNet, on tech review sites. If I was looking for the best hyaluronic acid or moisturizer for my skin type, those websites are not going to be showing up in the sources. It's going to be things like glamour and vogue and cult beauty, probably red it as well. Yeah, so paying a lot of attention to those specific sources is really important and paying a lot of attention to same with your keyword research in SEO. What are people actually searching for and what kind of content is going to ask that query and remembering only about 15 percent of that content might be from your website. 85 percent of it is going to be about you on third party websites. It was actually going to be my next question which was around, is it only realistic for big brands to, when we talk about owning original data and having interactive tools on blog posts and all this sort of thing which seems to be a popular thing. If you're a one person marketing team, you don't have the time or the resource to do that but it it sounds like if you can. and if you can be smart and strategic with getting your brand out into other places. So, you know, exposure and injure appearing on the marketing meetup, for instance, like could be a way of, you know, not having to kind of open the evidence. - And it's like a differentiator for you, right? If you can do, this is the same with all marketing. Like it becomes saturated and you wanna find something that makes you a little bit different or the same with sources in AI search. If you, just because a certain 10 sources have been referenced, if you're thinking, well, I'm not in any of those articles. I mean, step one is you could try and get in them. Step two is you could try and gizamp them. You could do something more interesting, different, better, more targeted to a specific customer. I have a great example of this from my own searches. Someone asked me, a customer asked me, do you know what the overlap is between Google page one rankings and chat GPT visibility? If I'm in the top three on Google, how likely am I gonna be to be in the top three in chat GPT? And we generally know from research that Google has a completely different algorithm. So page one rankings don't guarantee you a top position in AI search. That's the whole reason GEO has even come to be a thing is, just because your position one in Google doesn't mean you're even mentioned in chat GPT. But it was really difficult to find actual data of a good study that has done that. And actually, there was only one person, one website that published this, they're called chatoptic, who I'd never heard of before. They're a small publisher. They only have a relatively small scale data study given sort of the size of this topic. And they did estimate a 62% overlap, which is reasonable, but not that high actually, because then there's a whole 40% that is totally unrelated. That's not even on Google page one that's actually GPT references. But that, because I was asking these specific questions in chat GPT, chatoptic were coming up for it because there aren't other sources with that interesting first party data. So from my perspective, this is, if you're a smaller brand team, you can absolutely compete here. Because yeah, okay, big brands could have thought to do all of this and put that data on their website. But a lot of the time it's just about who has the idea. And it tends to be, I find particularly if you work in lead generation or anything where you have customer calls. If you can get customer call transcripts from like 10 customers that your sales team spoke to you next week, put them into AI and find the commonalities, you will find interesting questions that probably haven't been answered in people's content. Because you're actually, because you're looking for, okay, what are the gems, what are the diamonds here that would make us have something a little bit different from somebody else, from one of our competitors, or even through informational sites in your sector? - That's great. So it's favoring creativity in a sense. Like if you're a marketer sitting there going, how do I compete with all these bigger competitors with bigger spends, it's out thinking them rather than trying to outspend them. And you can get rewarded by the sounds of things. Yeah, it's just knowing what your target customers are interested in searching for. And there's loads of interesting tools, original data and research, yes, particularly in B2B as well, where that's a really big thing. But there's also things like interactive tools that work really well for big brands and small brands. Like, why is this currency converter calculated? It's probably one of the most famous ones that people go to wise to use a currency converter calculator, even if they're not doing transfers on wise. They could just do that in the Google Serps, but the UX is so much nicer on the wise one. I mean, people have come to trust it. And we have a fantastic client who is a mortgage provider in the US. So they're not a big bank, they're not a huge national player. But some of their biggest drivers from organic trafficking leads were lone calculators that we built. We did variations of articles that are, how much house can I afford with a 60K salary, an 80K salary, a 100K salary? Variations that people actually are asking ARA, they might be asking Google some of those questions as well, but they're definitely having lengthy conversations with ChatGPT about how much mortgage they can get or what price of house they can get. And even things like downloadable eBooks about mortgages, you would think who's downloading eBooks about mortgages? Thousands of people are apparently doing that. So it kind of makes sense if you're in that world and that's a problem you need to solve, those tools are great. And they support the topical area you care about. So if you want to be known as a fantastic customer-focused mortgage provider and you're trying to undercut the big banks with personal service, you're supporting those topical areas and ChatGPT other AI platforms come to know you because of those topical areas that you've gone after. - Yeah, I love that. And you're a, because whenever I think about the prompts that, I mean, I'm no expert at prompting AI's, but when I think about a really common thing that I'll type in is like, I want you to act as, let's say, a mortgage broker, as like the world's best mortgage broker. (laughs) Give me some of what, ask me the questions that are relevant that I need to be asking and that it is, so I guess if you're acting like that as a brand, then you'll provide in the answers for the LLM as a bridge to just getting what word can we rank for? - Yes, exactly. And I think that's part of it is thinking beyond just what keywords can we get in there for? And thinking more about who are the customers that we want? How do we want to sound as a brand? And what are the USPs? How can we do something a little bit different from the biggest big dogs who are dominating the marketplace so that we can actually reach the customers we want? And AI does that great because it's so specific to the user's intent. Remember, we get personalized results if we're a long-term chat LGBT user, a long-term Claude user, it's personalizing to us. So it's trying to actually give us the answer that we want and that's not going to be the same as Joe down the road or Harry next door. - Yeah, yeah, of course. Sarah's put in the chat and she says, I was one of those ebook mortgage people when we purchased that. - There you go. (laughing) - It's true. It's, that's fascinating. I've got a question around the sort of, so a lot of people will be having to explain their time and investment into search engines and LLMs to their management and leadership. How should marketers explain something like declining traffic to leadership when impressions are growing? Like what's the reframe that like, how would you go about doing it if you were kind of mentoring someone through that? - Yeah, this is a fantastic question and I think it can be a real challenge for marketers who are on the search side right now because everything has started to change since AI became a mainstay of the search results and that's the most common thing that we're seeing with the zero click search situation. Impressions are up, clicks are down because AI overviews, I think they're in more than one in 10 of the search results pages now. They sometimes eat over half the page, the entire above the fold if you're on mobile. - Okay, yeah. Yes, the thing I like to focus on rather than thinking about how many clicks you've been losing or what your traffic figures look like compared to the golden days of 2022 and we had mainly 10 blue links and a few featured snippets is thinking about the conversions that you're actually getting. A lot of brands that we work with, their traffic might maybe down, their clicks may be down but their conversions have stayed the same and are growing. And that's the most important thing to be tying it back to for a business because your conversions, your leads or your sales if your e-commerce is ultimately what matters to the business and ultimately what matters to senior stakeholders is being able to tie the marketing to the bottom line, to the money that's coming through the door for them. And I think sometimes as marketers we get really fixated on what I like to think of more as progress markers, things like traffic which doesn't dictate the revenue or the sales of the business. It's just something that we use to understand the likely direction of our work and the impact it's going to have. And when we speak to our senior stakeholders we should always be thinking about tying the impact to the bottom line. So when it comes to things like chat, LGBT, Gemini, Claude, Poplexity and all of these platforms that's exactly why we want to know how many conversions in our Google Analytics are coming. More interestingly though, because these are new channels and senior stakeholders understand things that are new to market. That's part of the work they do when they're running a business is bringing new things to market. If that's growing, if you can see that that's growing without even having invested significant time in it it's showing you a trend to optimize for which is exactly what we want to see, right? So for me, this is all about reframing the work that we do, how conversions are impacted, how we can actually be increasing those conversions and deciding as a marketing manager, as a CMO marketing director, which channels are going to be the most fruitful for the business, which are going to get the results that we need to get. And I guess highlighting whether the challenges are decreased in traffic or whether actually you've got a load of friction points on your website once they're found you and they've got to check out and the button doesn't work or you know, that's the reason you're not converting is because you know, you're complete button is just off your page or you know, yeah, and I have to say if you're e-commerce, even if you're sass, one of the best ways to test that is to have AI try and do it, right? You can just use agent mode in chat, if you're T, I can see in the chat here there's loads of Claude cowork lovers, just try and get them to check out on your website, like set them the mission of buying, you know, that coffee subscription and trying to get it into their cart to make sure there isn't that friction because you're exactly right, James, a lot of the time it's something like our conversion rates or silly things we don't even know that are broken on our website that causes friction or even things like the intent when someone has clicked through from the search results, what they read in the title and the meta description and what they're reading on the page or what they read on an ad if we were running a Google ad and then what they're reading on the page doesn't line up, like that user flow is so so important to the conversions. Love that. I don't know how we've got to 44 minutes into the webinar and I'm still jibring on our questions. There are loads of submitted questions. I'm going to bring one to the stage from Hannah here who says everyone always gets panicky about changes in SEO but I always come back to the golden rules of writing good quality content and relevant content. Do we actually need to make deeper changes this time? Love this because I totally agree with you Hannah that getting panicky doesn't help anyone with anything. What helps people with something is having a plan. Writing good quality and relevant content, incredibly important. The strategic shift I would be making in my mind about AI search platforms is that going from just writing it on your own website to understanding where you need to publish on third party websites and that's going to be your digital PR, your off-site work, your awards, accreditations, review sites, user generated content, what's being said about you in random Reddit forums which does happen even if you weren't controlling it. All of those areas when we're thinking about the content we create, we need to focus on those so much more than we did before. That's cool, great answer. Let's keep whipping through these because there's absolute loads of them. Samantha's put, "I've heard different explanations for Geo, one being geographic related and the other being generative related. Are both still commonly used or generative taken over?" I think generative engine optimization is generally how people think this acronym means and what it relates to. When we think about geographic, I always kind of include that in the definition anyway because just like Google, we've had, you know, I'm looking for the best plumber near me. We've had near me for 10 years that that's been working on Google and it works in AI as well. If anyone hasn't realized it, I hate to break it to you but AI knows where you are. Your location is already known and so does Google. Yeah, exactly, no, that's cool. But it's interesting right because the, I guess we've always just instinctively known that Google will, I mean, you even tell you that you've like at the top of Google that it's you're in such and such post-code but I guess I haven't even seen that in chat GPT. I just presumed I was a novice location wise. So you saying that's neat? It just learns from the conversations you're having, right? It understands so much more about you than I think we even realize when we're having those conversations with AI. Yeah, yeah. Yeah, it's memory is taking all in. It's going to start picking up Bromley Accent soon and then it'll be obvious. Talking about Bromley Accents, we've got Simon here who is one chat from the Birmingham community. He's put, are there any AI reports that are like exploding topics that suggest trending topics and insights? Actually, Simon doesn't have a Bromley Accent, but he is basically Birmingham. This is a great question and I actually haven't seen any specific AI reports for this, which is because chat GPT and other AI platforms don't release this specific data. So we don't know search volumes in those platforms. Any platform and there are tools out there so things like Profound, Squanch AI, PKI, these platforms, they're actually using Google search volume data to estimate what people are searching in those platforms and they're overlaying other data that they have as well and then they're running the prompts from their own platforms in the backgrounds through those platforms. So there's no specific way to know actually how many people search something in chat GPT for sure, they're all guesstimous. Same, I would use exploding topics and also Google trends because those are going to be the closest that we can match up. Yeah, and I guess this is such a fast moving thing that we're, it's, you know, ask that question in six months time and you can completely, you have a completely different answer. You just, well, I hope so James because we all know, well, we may not know that Microsoft now have analytics in your in your Bing webmaster tools, you can, if you've got Bing webmaster tools, not everyone has got that set up, but if you haven't set it up, it only takes five minutes. You can go in there and in your Bing webmaster tools, you can actually see just like in search console for Google, what queries are being searched and bringing people through in co-pilot, which is the most people co-pilot's quite small unless you're B2B, but actually it's real data about what people have been searching in those platforms that has then brought people to your website. So my hope is that chat GPT and Gemini in particular as well as Cored will actually then start giving this data as well and we'll go in the same direction that Microsoft has gone here with actually giving us something useful. Yeah, some more insight into where the heck it's getting its answers. Yeah. I don't fully understand this question. I'm hoping you do, but as we move towards the genetic commerce, how should brand/businesses balance traditional SEO with MCP, I don't know why MCP is implementation to ensure shopping agents like operator and psychic accurately interpret product logic alongside content. That sounds like the most complicated wordy question that I'm sure Divina fully understands but is way beyond me. Could you break that down for the audience and then first take it down? Yeah, I'll give it my best shot and Divina perhaps can add to it if she would like to. So agente commerce is what's happening for e-commerce brands at the moment is that we're now moving towards agente commerce. So Shopify for example, have just released an app with chat GPT that will allow people to check out straight in chat GPT. They're still checking out with Shopify as the merchant but it's becoming integrated so we don't have to have a chat GPT conversation then go to Google and type in a website and then go to the website to check out which is what a lot of people are still doing and we also. Yeah, feel like TikTok shop, exactly. We're moving in that direction and so this is about to become more and more normal for e-commerce brands. Unfortunately, there's actually a huge 4% fee on customers that have come from from chat GPT via Shopify agente storefronts at the moment. So for other AI platforms there isn't that huge fee. So in some senses this is great because we're actually moving towards a future where some of this is automated where there's less manual checking out we're doing less clicks to actually get what we want to get to. So that's the situation there. How should brands balance traditional SEO with this implementation? So MCP model context protocol is essentially AI models connecting with external data and tools and software so that you can check out. And Davina, if you want to add to this absolutely go through it because it sounds like you know significantly more than me about the implementation which is amazing. Balancing those, I mean the thing that I always like to think about when we're talking about agents coming to our website. Similar to what we were just speaking about earlier with you know how good is your UX is someone going to convert is that agents have less patience than humans even. They are going to a website trying to check out for you. They're browsing through if there's a broken link they don't care who you are what your brand is. They're going to the next website and they're going to check out there instead. As human users we ultimately also want our experiences and journeys through websites to be really easy right. We also want check out to be easy. We also don't want friction just like you said James we don't want to be trying to press a button and it's not working. that's super annoying. We might have a little more patience if it's a brand that we love and we specifically want to buy that pair of shoes from this brand, more so than an agent. But when I think about how we optimize for both, I always think you're balancing human and agents and actually quite often the goals are the same. It's just there's some more nuances in, you know, how long a person might stick with a specific brand and how they feel about brand affinity that AI doesn't really have that brand affinity, but it knows if people trust other brands more than others. It can, the same way Google can, there's an idea of how popular something is from what is sourcing across the web. Yeah, that's super interesting. So super technical question. Yeah, I hope I've gone some way to answer that. Yeah, well, let's move on to the next one. Jacqueline says, as people are moving through the funnel quicker due to LLM's which she talked about earlier, is there potentially an opportunity just now to target what might have been the lower value search terms, but will now gain more traffic as people are missing the first part of the funnel by using LLM's? I think there has been a focus traditionally of people going, well, it's just go for the bottom of the funnel stuff because that's where all the money is. Is it speeding that process up? Yes, it's definitely speeding that process up because people can have those conversations move from the top and the bottom, the top to the bottom of the funnel much quicker. The thing I would say about opportunities with the top of funnel is it depends what top of funnel term it is. So from my perspective, there's going to be top of funnel terms that are top of funnel, like what is a mortgage, where someone is very unlikely actually to even be navigating through to a brand. When you're thinking about those top of funnel terms, if you're coming up there, great, there might be some brand awareness happening for you. Some people might be like, oh, it was, you know, Barclays who answered that question or it was this brand that answered that question. And so you remember them, but it's for me, that's very transient amount of memory that's going on. It's very limited. So ideally, you do want to focus still towards middle and bottom because that's when people then start seeing you coming up a lot in lots of related questions, but they are already on the buying journey by the time they're in the middle of the funnel. They are actually investigating, whereas top of funnel, you might just be trying to find like a recipe to cook for dinner. You're not in the market for a hedge box. Yeah. And I guess there is, so I'm sitting here thinking that, you know, when I think about the stuff that I bought off TikTok, which I've deleted TikTok, by the way, because I rubbish from yeah, anyway, there is there's very little sort of brand that goes on for me. It's like, oh, this is cool widget thing. There's a laser level, for instance. I'm doing a house renovation. I was like, oh, cool. There's a five pound laser level here, which will be useless, which I ended up trying. I didn't buy a DeWaltz, a Stanley, a, you know, any other kind of well-known DIY brand. It was just rubbish from China. Like I appreciate that that's a low-cost purchase. But how? Yeah. Like there's still presumably a lot of value in building brand, even if you are turning up amongst a load of other suggestions, bottom of the funnel for best waterproof jacket, and someone's going, oh, you want something for running, and it's got to be tape seams. Actually, there's these three. If you're down to three, you still have to be, have invested in brand. Right? Yeah, absolutely. And I think this comes down to the priorities in your marketing budget, and the resource that you have, right? If you've already dominated in your definitely coming up in those bottom funnel terms, you're going to be in the top three, the top five, even the top 10. Then you're at a stage where you're able to invest more and more in brand. But if you try and do it the other way around and you're investing in brand first, but actually no one has any idea who you are to begin with, and you're not even in the right searches at that point, you're almost building the wrong way around. One of my favourite examples of this, so James is a client that we previously worked with, the ordinary, who are really well-known skincare brand, who are part of the S.D. Lorde Group. And way before, Chuck T.P.T. Gemini Complexity became mainstream, or were even on the scene, they had very, very clear brand positioning statements. And skincare is an extremely saturated market. If anyone's been into boots and looked at the skincare shelves recently, like, there's, you're going to be in there for a long time if you don't know what brand you're looking for. And they went very, very strongly after two brand positioning statements. And that is best value skincare and science-backed skincare. And that was on their website. It's in their PR, it's in their investor reports, it's blooming everywhere. And that is how most people now associate the ordinary. Any time you go into Chuck T.P.T. or another type of AI platform, if you're doing any search query related to best value, higher or I guess, best value, retinal, best value, anti-aging moisturizer, best value SPFs, the ordinary is always in the top results because that's where they put their stake in their brand. So this again comes down to, I think, deciding on the specific, it's not even just a niche, that's a USP that you're going to be focusing on. And building your brand around those topical areas, rather than like hoping for the best and going super broad and trying to get everything in your market at the top of funnel, at the top of the middle and the bottom, which is going to be like, we'll all be exhausted by the time. We'll be praying for the Christmas holidays. If we started that way, whereas actually if you can own a whole topic very related to strong positioning, that's probably the best way from my perspective to be building brand, but also dominating the middle and the bottom of the funnel in a specific area. Yeah. And that goes right back to the business strategy. Like, why are you here? What do you stand for and like who are you there to serve? I love that. That's that's kind of a real shoulder lift shoulder drop moment for me because, you know, when I think about our own company, we're trying to create like kind, friendly, welcoming spaces for for marketers, where we're trying to help them build confidence. We talk about that all the time. And so I guess if somebody nervous about networking or wanting to find free resources, then that means that we start to appear. And that's exactly how you're known, right? And a lot of that brand has just spun off James from who you and Joe are. Like you are genuinely friendly, welcoming, lovely people. And so that's if you're starting a new brand, if you're a startup, even if you're in a in a bigger company and you're a team, the best brands are the ones that actually reflect the people that work there. Like if you think about Monzo Bank, we know some of that Monzo team are a little bit silly. They are not the same guys and girls that are working over at Barclays and that Western Santander. They are different. They are younger. They are those kinds of people and that comes across in the brand. And if you build your positioning around the people you've got or the values you have, if you're a startup, that is how you can create a unique space, right? Like I don't think there's many other people competing on you and just being the friendliest place for marketers. I don't think so. Like we're trying to, and we're talking about us now, but we're trying to prove that kindness is a business model. Like it has a special benefit. Yeah. So yeah, that's that's really reassuring. We're at time already. I don't know how we did that. That was amazing. It's fun. Where can people find you, Charlie? Where's the best the best place to find you? Yeah, the best place to find me is on LinkedIn. And if you have more questions, please, please ask me. I don't think there are any stupid questions when it comes to AI Search. Everything is so new. Yeah, you can find me just by searching Charlie Martin on LinkedIn. And I also, yeah, thank you. I also speak at conferences really regularly. So you can find me at digital marketing conference Europe this month. I'll be at Norwich XL in a couple of months time as well. And I'll be at internet and focus if you happen to be Europe and Scandinavia based in May. And the other place you can find me is the Exposure Ninja podcast. I have the absolute pleasure of podcasting every week. You can find that on our YouTube channel or you can find it on, you know, all good podcast providers was fire apple podcasts and so on. That's a week. Charlie, you've just been like everything that I expected you when I first saw you on social, I was like, we need this person. We need her on on our on our webinar. You've absolutely delivered. I could talk to you for hours because I think this is you've made a topic that I'm not that interested in. Super, super interesting. And I think everybody will agree on here. It's been it's been a great session and we really appreciate you. So yeah, look forward to catching up with you again soon. James, thank you so much for having me and thank you everyone else for being involved in this great discussion. I absolutely loved all of the questions. Just what a community of switched on marketers we're a part of. It's so nice. Yeah, that's wicked. I'll send you some of the the ones we didn't get through as well afterwards. Maybe some link takes off. Please do. Yeah, good stuff. That's all right. Have a wonderful Easter, everyone, and we'll be around next week talking everything about linked in. We have got Phil Osay and his This topic is how to use LinkedIn without turning into a personal brand. So Joe will be back, you'll hear the music for the intro next week and I will be watching along with everybody else. So have a great one everyone. Talk to you soon. Cheers.

Podcast Summary

Key Points:

  1. SEO and GEO (Generative Engine Optimization) are distinct strategies for different AI platforms (e.g., Google, ChatGPT, Gemini), though they share similar tactics like technical optimization and content.
  2. Mentions are AI responses naming a brand, product, or service; citations are links to original sources; sources are referenced articles in AI answers.
  3. Marketers should start by tracking AI referral traffic via Google Analytics, then focus on optimizing for the AI platform driving the most conversions (e.g., ChatGPT for most businesses).
  4. AI queries are more verbose (average 60 words) than Google searches (4 words), requiring broader topic strategies rather than specific keywords.
  5. Citations have low click-through rates (~2% of organic traffic), but they build brand awareness and trust, especially as AI usage grows.

Summary:

The conversation between James and Charlie, CEO of Exposition Ninja, explores SEO in an AI-driven world, focusing on how brands earn mentions and citations in platforms like ChatGPT, Gemini, and Claude. James introduces the session, noting Charlie’s ability to simplify complex topics like SEO and GEO. Charlie defines key terms: mentions (brand recommendations in AI answers), citations (links to sources), and sources (reference articles).

She clarifies that while SEO and GEO share tactics like technical optimization and content, their strategies differ because each AI platform has unique algorithms, similar to social media channels like TikTok vs. LinkedIn. , ChatGPT for most businesses, Gemini for Google Workspace users).

She advises focusing on middle- and bottom-funnel topics, as AI queries are more verbose (averaging 60 words) than Google searches. Despite low click-through rates on citations (around 2% of organic traffic), Charlie emphasizes their value for brand awareness and trust, especially as AI usage grows. She highlights three core pillars: technical SEO, on-site content optimization, and off-site digital PR (reviews, forums, news).

Nuances exist between platforms—Google Gemini favors Google reviews—but good SEO work generally benefits all AI systems.

FAQs

A mention is when an AI response names a brand, product, or service as a recommendation. Citations are references or links in AI answers pointing to original sources. Sources are the list of articles used as reference points, accessible via a button in AI answers like ChatGPT.

SEO and GEO share similar tactics like technical optimization and content creation, but their strategies differ because they optimize for different platforms—Google for SEO and AI platforms like ChatGPT or Gemini for GEO. This is similar to how social media strategies vary between LinkedIn and TikTok.

First, set up a channel group in Google Analytics to see referral traffic and conversions from AI platforms like ChatGPT or Gemini. This helps identify which platforms your target customers use, so you can focus on optimizing for those instead of all AI search platforms.

AI search queries are more verbose, with an average of 60 words per prompt, compared to about four words on Google. This means you need to think about broader topics rather than specific keywords to show up in AI answers.

There are nuances rather than big differences. Core principles of optimization—technical, on-site content, and off-site digital PR—are similar across platforms. However, each platform has preferences for sources, like Gemini favoring Google reviews, so some fine-tuning is needed.

Citations still provide value as they build brand awareness and credibility, especially for top-of-funnel queries where clicks are unlikely anyway. Also, referral traffic from AI platforms is growing, so it's a new channel to invest in for future conversions.

Chat with AI

Loading...

Pro features

Go deeper with this episode

Unlock creator-grade tools that turn any transcript into show notes and subtitle files.