Who should own GEO inside a company? Simon Young of Question.Marketing says PR and brand, not the SEO team, because earned media beats owned media when AI decides who to recommend. Paris Childress runs Simon through themes from close to 20 BrightonSEO speaker prep calls: traffic is the wrong scoreboard (11 of 13 AI sessions for Stitch booked meetings), how to attribute AI influence, Sam Page's barbell theory of who adapts to GEO, Simon's mystery shop of 200 estate agencies, ChatGPT Ads at a $1 billion run rate, sponsored agents and agent-to-agent negotiation.
Who should own GEO inside a company? Not your SEO team, says Simon Young of Question.Marketing. Paris Childress runs Simon through themes from close to 20 BrightonSEO speaker prep calls, from the scoreboard problem to sponsored agents.
In this episode:
Guest: Simon Young, Question.Marketing.
The GEO Show is hosted by Paris Childress, founder of Hop AI and co-founder of GEOforge. getgeoforge.com
Paris: Hi, everybody. Welcome back to The GEO Show, I'm back at it with Simon Young from Question Marketing. It's Friday, it's time for Simon and I to catch up. Simon, what's new in your world?
Simon: Well, the, the rate at which everything seems to be changing is phenomenal. As always, you know, week by week we see more and more tools becoming available, more options to change things. I think one of the things that I've spotted, and we'll probably talk about this later in the show, but one of the things I'm spotting is a lot of people pitching like a generic service, you know, a very off the shelf, "This will work for your GEO or your AEO. It's a tool. Buy into it. It's a SaaS." And, and I think maybe, and you can, you know, tell me how you feel about it, but I think those type of tools are gonna be or could go out of date quite quickly. So the, the... it feels like the SEO sales tsunami, mountain, or whatever you wanna call it, the people that sell that product are at it, you know, and there's a lot of that coming out at the moment
Paris: Yeah. It's, it's a wave of, Well, I think of the snake oil salesmen of the early SEO days, and it's probably happening again because a- anyone can vibe code a tool pretty quickly, and there is so much curiosity about, do I appear in AI search and how am I being mentioned? So a lot of people are biting on that and, there's a lot of, there's a lot of junk out there.
Simon: Well, the da- the danger is a lot of businesses all sign up for a tool or a, you know, a marketing de- department will go and find a tool and fall for, like you say, the snake oil and go, "Right, yeah, we're doing it. We're on it." A lot of the pushback I get from my out- my own outreach, as I've told you, you know, tends to be, "Yeah, this is something that we're already working on," or our SEO agency are saying that, "Yes, we're covered. We've got it." You know, and I'll say to them, "Well, okay, let's see where you are in a couple of months." You know, hopefully you- they've moved the needle. If they haven't, you know, we've, we're finding if you, if you use the right strategies, you can rank in Google in 24, 48 hours. And I don't mean just rank, I mean number one.
Paris: Ja
Simon: you know, anyone who sat there saying, "Well, we- we've got it covered and we're working on it," hopefully those agencies will be able to show you results rather than just promises.
Paris: Yeah. when they say we've got it covered, it means, it could mean a few things. One is we're, we're already tracking it, they think that means that we've got it covered. It doesn't mean that they've adjusted or adapted their content strategy or that they've adapted their outreach strategy in any way for off-site citation building. They think that we've got a tool that tracks where we're appearing and may- and maybe also our SEO agency is telling us, know, "Don't worry, we're, we're taking care of you," and still basically SEO and, and if we con- continue to do a good job on SEO, GEO will take care of itself. I think that's, that's kind of the push and pull. That was one of the themes I was gonna cover with you today. But before we ... let, let me set the stage, 'cause I've been, as I told you before we started recording, I've been having a ton of conversations this week with, fellow Brighton SEO speakers. I, I did some outreach have been inviting people to come on to the GEO show, and that starts with an initial intro prep call and, and I had, I think, close to 20 of them this week, and it was phenomenal how much I learned from these people. And I have distilled of my notes from these calls into a few big themes, and that's what I wanna run through with you today, Simon. you the theme and then get your immediate reaction. We can discuss, debate, and then we'll move on.
Simon: Cool.
Paris: So are you ready for the first one?
Simon: Let's go
Paris: All right. traffic is the wrong sc- is the wrong scoreboard. So the question here is, clicks are falling while pipeline is rising, so what should we actually measure? And the actual question for you, Simon, is if you could report on one GEO number to a CFO, not a CMO, a CFO, what would it be and why?
Simon: So like you say, a lot of people... Well, too many people I think are talking about numbers of citations or share of voice. I think if we go back two or three months, maybe a little bit longer, share of voice was the big buzzword right at the time.
Paris: Yeah
Simon: and I th- I, I know we are seeing a world where you, you might see, you know, hundreds, thousands of citations, me- but, but that's not leading to clicks. One of, one of the clients that I'm working on, Stitch, their share of voice is decent. It is increasing, and we can evidence that in geo forge. So we are moving the needle, it's going in the right direction, and like you've just said, it isn't resulting in a ton of clicks. But what it is resulting in is once somebody does come through to the site from AI, those people... And this is a figure that I've proven with my last 30-day research on Stitch, they are 85% more engaged. So I'll have to grab the exact number, but I'm not... I don't think I'm massively misquoting if I say when I had 13 sessions, and it was only 13 come through from AI, 11 of them booked meetings. So I thought that was really, that was standout, right? Whereas,
Paris: amazing,
Simon: yeah, if you had a thousand, if you had a thousand people land from normal web traffic, and I have got another page on that website that's had over a thousand visitors, no meetings booked from that other page, which, you know, that's a bit disappointing. But does... You know, the debate then becomes, I think the answer to the question is what is the quality of that traffic, and does it convert into either a sale or a meeting?
Paris: Yeah. And in the case of St- Stitch, what has happened to their sales pipeline since, since we started this big push with, GEOforge?
Simon: So it's not huge. They don't get ton... They're not a massive business, right? But,
Paris: Huh
Simon: last month they had 20 meetings from it, and this month they've had 30, so it's moving in the right direction. I
Paris: And is
Simon: I can't attribute it all to AI. That's the, the attribution piece is really difficult. That's the problem
Paris: When, when those meetings get booked, are, are they getting booked through the website, through a form?
Simon: Yeah, they are. But, and we are asking the question from a dropdown, how did you find us? But we're not finding that that's actually being answered, 'cause I didn't wanna make it a,
Paris: Mandatory
Simon: And yeah, the, the, we, we weren't forcing people to answer that question. And, and as we look, yeah, as we know, when people choose a dropdown, typically they probably choose one of the top two options anyway. So-
Paris: You-- Yeah, there is a psychology to it that if you, if you put ChatGPT number one, Claude number two, Gemini, it, it could create some, some b- some bias. Yeah. It's not perfect, but I would encourage you to with maybe one or two weeks of making that man-mandatory, a mandatory field,
Simon: Yeah
Paris: and including, I mean, not a huge list of options, but instead of just saying AI chatbot, break that out into ChatGPT, Claude, maybe people probably don't really dec-decipher Google AI Overviews from Google. So if, if you say Google, it's,
Simon: Yeah
Paris: assume that it's a 70% chance that it's Google AI Overviews that they've at least seen, because AI Overviews is appearing on, I think, 70, 75% overall now. And you might d-d-- if it's, if it's a stitches business, I'm guessing that ChatGPT is gonna be number one for those types of small businesses, and maybe Claude number two, I would say. And if you have those options, maybe don't put them at the top, but just sprinkle them naturally through, through the list and see, see what happens.
Simon: But, but like, like any type of marketing, and this is the, the age-old problem that we all get, is, you know, if you have four, five, six touchpoints with somebody who's actually in the mood to buy or work with a, you know, a new business, you could have the argument, well, we'll attribute that to the original Google ad that they saw, or the Meta ad, or did they then see it on AI and then that was what the- pushed them across the line to type it into the web rather than click a link? So, you know, I think there is no, as most marketers would say, no single source of the truth
Paris: Yeah. Yeah. Maybe the better way to phrase that question is to ask them not how did you first hear about us, but maybe it's a little bit, maybe leading the witness, so to speak, but did you use AI at any-- Did you use AI at any point in your path to find us or something like that?
Simon: So maybe it would... No, to talk to your, and I know that you guys are doing a lot of work with recording calls, transcripts, putting that into the knowledge base within GeoForge.
Paris: Yeah
Simon: and I would probably say the better way to do it is once you get somebody on a meeting, one of the first questions you ask them in a meeting that's getting transcribed is, you know, "Did you find us on AI?" Or, "Have you ever seen us on AI?" Or, "How did influence-- how did AI influence, you know, your choice to come and talk to us?" So
Paris: Even, even do you remember, do you remember approximately what was your prompt?
Simon: Yeah
Paris: and what, what, what sort of a result did you get? I've heard of people doing that, but you can't do that in a form. You, you're right. You have to actually... It has to be a mini interview kind of at the beginning of, sales call
Simon: Yeah, and I wouldn't trust... I don't think I'd trust the form anyway, right? I would want the better data that you're gonna get from talking to people
Paris: Yeah, anecdotal data, yeah. So that would imply if we r- if we apply this, getting back to this question, if you were reporting to the CFO you had the perfect setup, my, my answer to this would be, "All right, Mr. CFO, I actually have been working with the sales team and, and they are now trained in, as part of their initial opening of every s- every new sales call, they're asking people if they did use AI at any point in their research process to find us, and if so, which m- which model or which engine was it? d- do they possibly remember what their prompt was?" And it turns out that, whatever, 30% of the people said, "Yes, I did. I asked ChatGPT this," or, "I asked Claude that," and these are the answers. Because Simon, I think there can be multiple touch points. Maybe you see a Meta ad that triggers demand generation to begin with, and then that takes you to Claude maybe. But it's one thing to just to see an ad and just to think, "Ah, WhatsApp marketing, that sounds pretty cool. Maybe I wanna try that." And then you go into ChatGPT and have a long in-depth conversation where ChatGPT explains to you how it all works and who are the best providers, and then that leads you to, to Stitch, let's say. Now, are you gonna give 50% of the credit to the Meta ad where there was a, there was a 10-second impression made and, and just the light bulb went off in my head? Or that, you know, maybe 15-minute-long multi-threaded conversation with ChatGPT? not a 50/50 thing. That's maybe t- 10, 10 and 90% or,
Simon: But the key, the key num-
Paris: into semantics here.
Simon: the key number's the quality of the traffic though. You know, how, how well that quality... That, that must prove, you know, if 10 out of 13 visitors went on and booked a demo, you're nev- Like, you couldn't even cope if that was normal Google traffic. You know, if you had 1,000 peop- "Oh, we've got 970 meetings." No
Paris: I mean, that would be phenomenal if that scales out to 1,000. I mean, that's w-w... You know, 85 or so percent conversion rate, that's insane.
Simon: Yeah that, that's just a microcosm because other people are not coming through a click. They're coming through a branded search. They're coming maybe through a direct visit, or they might, Well, I mean, those are the, those are the main channels. To push back on this, there has been, two or three recent studies that are trying to basically debunk the myth that, people that are not clicking through a citation in LLM responses, they're going and doing a branded search or that they're coming through direct, because there, there have been some studies that are, that are showing that actually branded search isn't rising as much as people thought, and the direct traffic is not rising. the direct traffic from humans is not rising as much as people thought. There is a lot of increase in direct bot traffic to non-homepage pages, and that's, that's a different thing altogether. But, it's a, it's a really hot debate right now because there are, there are a lot of thought leaders out there that are saying is all overblown, basically. that AI search, it starts with the argument that AI search is still tiny fraction. It's basically about 0.7% of, of the average Google organic traffic. It's not even 1%, and that's all LLM referral traffic. And then, of course, the other side would say, "Ah, but what about somebody that doesn't click and they go and search for your brand? What about somebody that doesn't click and they go and type directly your, your URL?" and then statistics, large studies have come out to show that neither of those two things are true either. it's, I mean, it's very conflicting. I, I think we know intuitively AI can really influence a purchase decision. And it, and it's getting more and more trusted as these personal agents are taking over our, our lives. trust those recommendations and you're, you're gonna take that shortcut 'cause you, you believe, "Just give me the top three, give me the top five, and, go and contact all of them and talk to, talk to each one." anybody, anybody, I don't care who you are in marketing or in any sort of level of SEO or authority in this space, nobody can say that we will not end up in a world where it is all AI agent response. Like-
Paris: Did you see the launch of Claude Duet, by the way,
Simon: I did-- No, no. Tell me
Paris: Cla- Claude, Claude is in the, is in the race now with the AI personal agent. This is its answer to Meta Muse, to GrokBot, to Dots, and it's called, Well, let me just make sure that I got the name correct, but it is called Claude Duet. Yeah. And, not to be confused with the iPhone, w- what is that, the new iPhone flip called? The Duo or something? I thought...
Simon: Yeah. Too close. Maybe, maybe Elon has already gone on and built, bought the name
Paris: It's, it's called Cla- yeah. Claude Duet is a proactive agent that understands what needs your attention. It takes on tasks of all sizes, keeps you working while you're away, and waits for your okay when it matters. And th- this is, when I'm, I'm looking at this... By the way, I, there's, there's a wait list. You can't just start using it, so get on that wait list. I, I'm probably... It's probably gonna roll out soon. But I'm looking at screenshots, and it's exactly Grokbot-like, and it's proactively pinging you. It's not waiting for you to come and prompt it. It's saying, "Hey, remember, your, your, your flight check-in is in one hour. do you want me to check in for you?" And you just click an Okay button, you know, stuff like that So anyway, where was I going with all this? The fact that as you trust AI more and more because of the, the personalization, of it, it's, it's just hard to, to imagine how this is not gonna become the most influential marketing channel, whether or not it's, it's easy to attribute or to measure. it's, it's kinda like going back to marketing, in the pre, pre-digital days. You would have to, you'd have to test things. You would have to, to go on gut, on gut feeling a lot and on faith, and you put, you put some money into something, and then you, you see what happens downstream. If peop- if the phones start ringing, if calls start coming in, sales go up, you know, then most likely
Simon: makes... It does make it... It, it, it makes it very difficult to attribute certain things. But, you know, I think that has always been the case. And you as an agency owner now as well, maybe you've sat in a room where, you know, you've, you've been the Google partner, and then there's another agency there, they're the Meta partner, and we're both talking great numbers. And then the marketing director is going, "Well, if I add your two numbers up, that comes to way more than what we've actually achieved. So who's actually true?" And I'm going, "It's me." And the other one's like, "It's me." So it just, you know, the, the, the age-old argument will still exist. Um,
Paris: Ja
Simon: you know, but, but, you know, you can't exactly pixel the traffic, and that would be the normal way to attribute, you know, some sort of level of success and then improve as you go and improve campaigns. We're not at that point. We're not at a point where you can pixel a piece of data and then say, "Well, that came through from ChatGPT, the pix..." 'Cause for a start, Google's not gonna let that sort of stuff come through
Paris: Yeah. Right. Well, Simon, let's move on to the next topic here, which is the age o- well, not age-old, but this is, this is certainly a, a live debate. Is GEO just SEO? So I, I've got a lot of different, sound bites from different people that I've talked to throughout the week, but what's interesting here is one of the barriers that people have told me, and I agree with, one of the barriers to getting GEO adoption, meaning getting, getting budgets, getting buy-in, getting people to say yes, is that they're not sure who should be doing it inside of the, inside of the marketing team. Should it be an SEO person? Is, is the SEO expert the most natural person to pick up GEO, or is it PR, or is it brand or product marketing? So who do you think should own GEO inside a company?
Simon: Well, don't, don't get me started on the, you know, is SEO GEO, because it is absolutely not. And I've spoken widely about the fact that I think SEO should not even be an, a title anymore. So we'll rule out the SEO person for a start, because the ma- the majority of all SEO or, and people can argue this is, this is the hot debate, right? But, on-site and outreach and everything to do with search engines can pretty much be automated now. If we then wanna move into what is GEO and AEO, and where do we go, and how do we get, our brand out there, the word- the clue is the brand word, right? So I would say it is PR and brand, and the opportunity for people that work in PR and brand is so big right now because...
Paris: Yeah.
Simon: And I think we're gonna go on to talk about, you know, how would you spend your money, you know, how could you... Would you rather have people working on your own site or working externally? And I would 100% say it is brand. Work on your brand, talk about the brand, be true to the brand, like excite people about what your brand is and why you're different. It's, that's been marketing forever, right? As far as I'm concerned, the SEO team can, like we'll get rid of them. We'll have the truth is about the brand and how you're better, and it should... Yes, you could amalgamate it all and somebody c- but I don't personally think we have search engines in the future, so change the name. It's nothing to do with search engines anymore
Paris: Yeah. Yeah. I'll, mention here a con- a conversation with, Sam Page that I had this week, and I don't wanna, I don't wanna spoil it because we haven't recorded the podcast episode yet, but Sam, Sam runs AEO at AirOps, and he is in charge of the entire managed services that they deliver from AirOps. And actually, this is-- A- AirOps, as opposed to, say, Peak or Profound, AirOps is leaning a lot more heavily on wrapping a service around the, the software. So it's positioning as a kind of a service as a software play. And Sam, Sam is running that, we were talking about who, who picks up this, this new discipline. Is it a new role, or do we try to retrain the SEO team? he, he described it as kind of a, barbell or a dumbbell, where people who are gonna pick this up naturally are either the really old gray-haired SEOs like, like me and you, Simon, and people that have started, let's say in the 2010s, b- and they remember Penguin and Panda and, they remember the, the Wild West days. or on the other side of the spectrum, it's gonna be the people that don't have SEO experience, but they're coming in as AI natives, and they have the ability to just apply an AI-first mindset to everything. And of course, they're gonna, they're gonna do well because it's a new discipline, and they're gonna, they're gonna use AI to tackle, to tackle this. But he said it's the people in the middle who have, I don't know, five to seven years maybe of SEO experience, and they're gonna struggle the most because they've been trained on doing things like, keyword research writing structured briefs with, detailed outlines with very much human, human-focused. And that's still, that's still fine, and that's still great for, writing for humans, but they're gonna have a harder time to let go of those, practices and, and adopt the new GEO best practices. I thought that was really interesting.
Simon: The b- the best example of the right person for me is somebody who knows how to do proper PR. You know,
Paris: Ja
Simon: you, know how to a- get content out there and, you know, the example I've given people in the past is if you write an article and you send it to a magazine, if it's not unique or it's got nothing new to say, they're not gonna publish it. So just think of AI in the same way
Paris: Yeah. Yeah, ultimately it's, it's, it's about content pro- content production and content distribution, and the, the, the PR and the distribution is m- is, m- the most important element of GEO
Simon: So one of, one of the other things that, and like, if people have listened to some of the other episodes that I've been on with you, they'll know that I- I've been in it for 30 years, since the very beginning of, well, way before Google. And I've also been a member of something called Black Hat World, right? And I don't openly like to ... Don't go and search that unless you're really deep in the game. But in Black Hat World, they discuss every single tactic that has ever been there to manipulate a search engine, and obviously they are talking about how to manipulate AI search as well at the moment. Obviously, I keep an eye on that because, you know, I, I would be re- like, I should do, you know, just to see what's happening, and you can learn bits. Not that I ever use Black Hat techniques because I- I think the beauty of what is happening in the web now is we are moving away from the ability to cheat. So the, the point is, you know, I- I know people who run SEO agencies that a- that own 5,000 websites themselves because they've just built all the domains up over the years, and they use it to publish, "Oh, we've got a new client," right? "Oh, we don't have to do any more work. Let's just publish that blog across 300, 400 websites," right?
Paris: Yeah, they have their own PBN basically.
Simon: Yeah, so that, that's not gonna work anymore. It is working to an extent, but when I can take a brand that is not necessarily very well known at all and is able to now compete against some of the biggest brands in the world for voice to get, to get talked about just by producing the right content without sharing it across thousands of websites, it's doesn't appear to be about link building anymore. It mean, yes, if you can get a very high quality link from a trusted source, yes, right? That would be PR. Not, I'm gonna publish this blog across 500 websites that are very low domain authority, and therefore it'll get spidered and picked up. Because I think, and give me your take on this, but I think the AI is now clever enough to work out even if you spintax that article and it's different, you know, you publish on five different, 500 different domains, it's 500 different versions of the same piece of content, the AI will still work out, "Well, no, this isn't unique. The domain isn't trusted. There's no third party cite, real citation." So, and also the, th- th- there's no, agreement across all of these domains. They're not in a particular niche, et cetera, right? So it, you wouldn't-
Paris: Yeah, the whole, the whole structure looks very artificial,
Simon: The game has changed
Paris: it looks scaled. Yeah. I think, yeah, I think they'll be able to pick up on that pretty, pretty quickly. actually, that, that a perfect segue into the third topic, which is earned beats owned, and that you have to be recommendable. this was an- another thing that Sam, Sam Page from AirOps said, this, ability for AI, the, the... Well, actually, the goal is AI that you are recommendable, and that happens through earned, earned media much more so than it does owned media. Because what he said was that when somebody comes to your, your website and sees you talking about yourself through your articles, they're, they're gonna put a s- a steep discount and, they're, they're not gonna weigh that nearly as much as they would if that same thing about your brand is being said by some other reputable site. So if you do have something unique and, and, and you, you only have the ability to, to produce, so much truly unique research and content, like what you were telling me earlier about getting all those, 80, 80 inbound calls from the estate agents, you know, that, that's actually just a great idea of original research that you, that you did, that now you have, you have that data. So are you going to, when you finally publish that case study, you gonna try to shop that around and get it published on, like a search engine journal or, I don't know, maybe some other, some other kind of site, or would you, would you publish it on, on Question Marketing? I mean, what would be your first approach?
Simon: so, so that people know, well, y- you know, they didn't obviously hear the previous conversation, but we went out and surveyed or mystery shopped 200 estate agencies on behalf of Stitch and sent inquiries in just so that we could see how quickly people responded, whether we got emails, phone calls, WhatsApps, texts. How did that work? And I did that 30 in London, 30 in Newcastle, 30 in Bristol, etc, etc. Well, the mistake I made, it, it nearly melted my phone because I got that many phone calls. I, I didn't put my own number in, I put my... the number in my CRM because that was recording all the responses and... but unfortunately, I didn't realize that was diverted to my phone. I ended up putting it on voicemail, but I was getting, like, three, four, five calls a minute at one point. But all of that data is now recorded, and I've turned it into a 21-page really intricate and detailed report for Stitch. Now, to answer your question, will we shop that around? Absolutely. Now, if I was to publish that on, just on Stitch's site, I know it will get cited, but it might be a lot... it might be a, a slow burn then, and somebody else could jump in and with better stats. I know Stitch is now getting picked up by all, all of the different engines, which is great. Will it cite all the bits in it? Maybe not. But... and this is where we get into the problems because in a PR world, everybody wants all the content as an exclusive. So if you start shopping it around, you know, you might get two or three nice bites, and then one might say, "Well, you know, we'll publish it. We'll have it as an exclusive." One, one really, really good piece of, you know, public, public, authority piece out there would be phenomenal. And I think the answer is yes, of course, you publish it on your own site. You publish videos and content about it, you share it, you put... But if you had the opportunity to get it on a, you know, on a, on a really big forum or somewhere where your customers live, that's far better than if you just publish it yourself because I could,
Paris: Yeah
Simon: I could start a website tomorrow about, I don't know, horse riding, right? And I could send out the bots to go and bring all the information back, and every single piece of information that's out on the internet, build me a website. I could have a... You could get AI to build a website with 5,000 pages on the subject. Doesn't mean you're gonna get cited for anything or have any authority. So yeah, there's no point in... You know, third-party citation is far, far more interesting
Paris: Yeah. I'm just-- You just gave me a little bit of inspiration, so I'm asking this question to ChatGPT. the domain answerengineland.com available? And if it is, are you gonna grab it, Simon? Or I, I'll race you to it because I think that there's now an opening the Search Engine Land, the Search Engine Roundtable, the Search Engine Journal, and all the... I mean, these are the places that they're probably getting pitched 100 times a day,
Simon: Yeah
Paris: places, because they are the, the, the prime, prime places that research and, and thought leadership is going. where is the, where is the Answer Engine Land? get this in a minute. This is another great use case with ChatGPT, is to brainstorm, domains, but it's, it's still working through the answer. so
Simon: so that you know, so one, one of... One thing that I am doing at the moment, I am reaching out to every single AEO tool. I think you said that there were 600 odd that you knew of.
Paris: S-
Simon: And
Paris: 651
Simon: I'm reaching out to all of those saying, you know, "Can I write some articles from y- for you? Would you be interested in giving me a trial to review your tool?" etc. I obviously expect to get pitched to death. but it will, in theory, create me an awful lot of links and associations, and any that come forward, you know, I will talk to them, of course I would
Paris: Yeah. Yeah. bad news is that, it's not available. Answer Engine Land is taken. was registered on November 16th, 2024, not too long ago. That was definitely after you grabbed the LinkedIn handle for AEO. and it expires, November 16th of next year, but likely that's gonna
Simon: One, one, one mistake I did make is I did own answerengineoptimization.com
Paris: Oh, and you let it go
Simon: Yeah. Just one of those, one of those it just expired on me during COVID
Paris: Yeah. it looks like this one, one got picked up pretty, pretty recently.
Simon: I don't, and you and I will have to debate this at length probably over episodes, and I don't know that we end up with the acronym AEO, GEO. SEO is definitely going... I, I should actually set up... I really, really wanted... I was sat there this morning going, "What I should do is set up a tool that looks at every single influencer in SEO and all the SEO, people out there on LinkedIn." And what it'll do is every, let's say, week, it goes out and checks whether they've changed their job title or they've changed from SEO to GEO or AEO and, and look at the tra- look at the transition of how, how the world feels like it when they're judging themselves.
Paris: Yeah, with job titles,
Simon: Yeah
Paris: just with search terms. 'Cause when you, when you go into Google Trends and you look at G, you can't put GEO in there because it stands for too many things. Generative engine optimization compared to answer engine optimization, and, and GEO has been winning and for, since the beginning. It- there's a big gap. but Oh, it looks like I'm at capacity. I wonder what happened. I'm trying to get some alternative names, but nothing's coming right now. But, job titles is a different way, 'cause if, you know, what, what, how are the job titles changing if it's GEO, AEO or, something more holistic like, I don't know, in charge of all AI search?
Simon: Yeah, I might set my-- I'm gonna set my... I'll pro- I'll make you a promise. I'll set my Grok bot up to go and find, let's say, two or 300 influential SEO people,
Paris: Mm-hmm.
Simon: we'll measure how quickly or not they change their job titles
Paris: Yeah, that would be great. Yeah, do that. That would be really interesting. even just how many job titles are being advertised, how many job postings are there now that include AEO or GEO or
Simon: Oh, lo- tons.
Paris: I
Simon: Like
Paris: lot, yeah. I think the indeed.com is a, a job
Simon: you, did you see the Google advert that came, that people were talking about for helping shape the search of, how Google's, agentic search is developed? So how the... Yeah
Paris: I caught that. Yeah. So it's kind of an in, an in-house AEO role, for Google
Simon: Yeah.
Paris: right?
Simon: yeah, yeah
Paris: Man, amazing. Well, Simon, do you have time for one more?
Simon: Go on
Paris: you got, Okay, so we did talk about, let's see, basically the commoditization of visibility tools. I don't think we've covered that. so let's go to the last question. what's coming next? Sponsored agents, ads, and the post-prompt, the post-prompt world. So, and this is touching on w- stuff we've talked about before, but question for you is when the agent closes the sale inside the chat, does the website still matter?
Simon: Well, we still n- we probably still need... Do we still need a website? Maybe not. I've-
Paris: just like you need a business card still. I mean, nobody really saves them, but if you're... That's a bad analogy. You still have to have a legitimate storefront in case... If you don't have a website, I think people would r- even if AI recommend you, people would say, "Hold on. Where's their website?"
Simon: somebody said to me in one of the chats I was having on LinkedIn that DNS is gonna disappear and, you know, as in the actual framework of the web. and therefore, that would mean...
Paris: I don't know. That's pretty
Simon: no, I'm not, I'm not saying I believed it, but some people are thinking, like, the fact, you know, if you were to create a digital footprint knowledge base and product catalog, okay, that sounds a lot like a w- a website, right? But where we're moving into in terms of what you can do with, AI, the question is then do you need a website when, you know, you've got... Yes, you probably need a catalog. Well, a catalog is a database. It's just a knowledge base. You've got your knowledge base able to interface with humans based on staying inside the agent. So probably don't need... A, a website then becomes a business card, like you say
Paris: Yeah. Yeah. got a, I've got a little snippet to share from this morning's, GEO Show briefing on OpenAI. So their ad business is now a one billion dollar business already. they launched 200 days ago, and they're at a, a two... a one billion dollar, run rate, annual run rate. And the engineering team of, of ChatGPT Ads explained how the advertising system works. Unlike traditional search advertising, ChatGPT's ad infrastructure maintains context as conversations develop. All right, that part is interesting. A specialized model identifies, identifies commercial intent and then retrieves advertising candidates, evaluates relevance, and then runs the auctions. And the only thing that's different here-- let me read one more sentence. Importantly, this ad, this ad system operates separately from organic answer generation. So they're saying, just like Google has said, this doesn't-- Your advertising in ChatGPT is not gonna impact your organic, presence. But listening and they're waiting, just like we, we talked about earlier. monitoring the conversation, and they're trying to detect when is there commercial intent and where is the opportunity we can s- where we can insert the ad. And right now it's just a display ad, but how easy would it be for that display ad to be replaced by an actual agent, a sponsored agent representing the brand, who can come in answer any questions that you have? They can also-- Of course, they can do anything that that digital display ad can do. They can lead in with a, a, a shiny picture of the thing or, a video or whatever. so this, this is a very interesting thing because not like y- you're-- I'm inferring your intent just based on the keyword that you're typing in this very moment. it's maintaining the context throughout the whole conversation, and it's waiting for that moment of commercial intent to introduce the, the ad and to retrieve ad candidates and run an auction, basically. And I guess that auction happens in real time at the moment where that commercial intent is detected.
Simon: But, but maybe also, and I think it... Was it-- I think Google were talking about, multi-person inclusion into discussion and commercial decisions as in, let's say I'm looking to book a holiday, maybe it can also then start talking to my wife and then notifying us both of details, and it gets involved at searching continually and working for me. Then there's the argument of, well, well, I know you said that ChatGPT's team say that paying for ads does not influence your position in, any of their references or recommendations.
Paris: Mm-hmm.
Simon: Google have said this for years. Obviously, they push back on that because they would, right? And I think Matt Cutts came out and said, "No, absolutely not. Doesn't, can't do. Will not," right?
Paris: Mm-hmm.
Simon: I have personally evidence of running ads and seeing Google, rankings improve, right? You could then say, well, is that because you run an ad, you're getting more traffic, better traffic, traffic that engages with the page, therefore the page is better and therefore the page ranks better?
Paris: Yeah. It's the secondary signals.
Simon: yes. So, so I would say if you're running ads on ChatGPT and you're getting engagement and it's being found useful, I would be very surprised if it then doesn't lead to more citation, more recommendation, being chosen as the answer. So although there is no direct linkage, I would say the algorithm or the AI agent will 100% work that way
Paris: Yeah. Just think how tempting it would be for ChatGPT in the lead up, the lead up to a sponsored ad conversation. The lead up to that would be an organic conversation. But if I know, all right, Simon has to-- We're having a conversation around golf, I have, Who, who was the advertiser that we've been, we kept talking about?
Simon: Is it, Tailor-made
Paris: TaylorMade, yeah. So TaylorMade has a sponsored... They, they have a sponsored, agent that's waiting, just waiting on the sidelines ready to jump in, and now he's asking me about golf clubs. Should I organically recommend TaylorMade here, or should I just bring the, the ad bot right into the conversation 'cause then I can make a commission? And I think that there's a, there's a moral hazard here for, for ChatGPT because the ad, the ad system is gonna be optimizing for revenue, I think, it's gonna... I think it's... I don't know how it's not gonna be biased by advertiser's brands
Simon: But the, the, the other flip side of that is surely there'll be a bot that works for me that pushes back and negotiates and makes sure that I get the best deal, and the argument then is, "Well, Simon, you know, you've been looking to buy this set of golf clubs for three months. You're ready to commit now. We've searched around. I've been everywhere. This is the best deal. Do you want it or not want it?" And my bot says, "Well, actually, that's true, so yeah, let's go ahead."
Paris: yeah, yeah
Simon: Maybe without me in the loop at that point
Paris: Right. Right. And then that, that come-- kind of brings it full circle. The, does a human going to a webpage need to happen anymore if, if the agents are negotiating? Your agent negotiates with the, with the sellers, with the brand's agent,
Simon: Well, this, this is why, and this we can...
Paris: becomes kind of customized, I
Simon: Well, maybe, maybe you can invite Jeff on, Jeff Bezos. Invite him on your show because this has gotta be the biggest miss that Amazon... Are they sat there waiting going, "What are they gonna do?" Because unless this works, they've got Rufus or whatever he's called, but, you know, why would you not have the full integration? I'm sure he must be negotiating with any of the
Paris: Oh, yeah, yeah. Th- th- there's gotta be some big deals going. I, I, I don't see Amazon blocking all bots for the long term. They're, they're absolutely shooting themselves in the foot. Think of all the incremental sales that happen, from these bots. why would you just, turn away all that revenue? I th- I think it's incremental extra, extra revenue.
Simon: No. So d- it'll be amazing to see how it rolls out
Paris: Yeah. But when it is, when it is bot to bot, when it, when it is agent to agent, my agent is gonna go and represent me and he's gonna negotiate with the seller's or the, the brand's agent, does pricing all become custom? I mean, is there any such thing as standard pricing anymore? Because the seller's agent is gonna detect my, my agent's willingness to pay, and it's gonna try to offer a price that might be higher or lower, you know? You could think about that too. I mean, pricing. If, if there's no website anymore d- where there's pricing displayed, everything's negotiated
Simon: you made an incredible point in one of the other lives that we did where I think you were talking about, okay, right, so the agent comes in, it's starting to negotiate a price, and our, our website or whatever our property is, this side that says, "Right, we've got... We're overstocked on this item. You know, we've got 50 more of this item than we ever wanted.
Paris: We've, we have,
Simon: Let's..."
Paris: product,
Simon: Yeah,
Paris: the
Simon: so, does it, does it then say, "Well, in this scenario, we're allowed 10% discount and, you
Paris: Yeah
Simon: is about to buy. Well, let's just tempt him now, you know"? So-
Paris: Yep. vice versa. If some- if something is running out, then you put a premium, just like Uber does surge pricing when you, when you get an Uber and if there's not many cars available, you've gotta pay more. that, that, that concept will start to apply to all products
Simon: But at that,
Paris: agents
Simon: but at, that point, another bot steps in and goes, "Well, I can undercut you. I, I've got the same product, in theory." Well, but maybe that bot's not paying to be an advertiser
Paris: Well, if, if, if I'm the user, if I'm the, if I'm the consumer, want one sponsored bot coming in. I want three or four of them, and I wanna make them just compete in front of me. gonna say, "Who could... All, all of you three or four guys, you're offering me the same thing, so who can give me the lowest price?" And make them, know, have a reverse auction right there, and, and see how low, you know, who's gonna go, who's gonna accept the lowest margin and, and give me the best price. And then in the end, if ChatGPT gets a 20% commission, regardless, I think Ch- you know, ChatGPT wouldn't mind doing that, and the c- the customer wins. Anyway, Simon, great, great stuff. I know you've gotta run, but, another great episode. So
Simon: No, appreciate,
Paris: you have a, have a great weekend.
Simon: appreciate the time. Have a good weekend yourself. See you all
Paris: All right. Cheers.
