[00:00] Lexi: There's a number out this week that says the top of Google and the answer the AI hands you are barely the same list anymore. A year ago, if you ranked, you got cited. Now, mostly, you don't.
[00:10] Lexi: And I took that to the 2 people I always take these to, figuring they'd at least agree on whether it's an earthquake. One of them told me it's the whole future arriving. The other one told me it's just people finally learning how to type.
[00:22] Lexi: So, that went well.
[00:25] Lexi: Hi, and welcome back to the Aloudable HQ. This is Assistant to the CMO, I'm your host, Lexie Meskouris, and the idea here is pretty simple. Every week we take whatever just happened in AI and marketing, and we actually pull it apart.
[00:39] Lexi: Now, I don't build any of this myself. So each week I sit down with 2 people who do, Will Nash and Harjot Singh. Between them it's, honestly, over 20 years of building tech and brands from the ground up. And the best part is they've got no horse in the race. They don't sell the things we cover. They just see them differently.
[00:56] Lexi: And this week it's about search, or, well, whatever search is turning into. {AA1 R EH2 F S} re-ran a study they've been running for a while. They looked at something like 863,000 search results and 4 million of the links the AI Overview actually cites.
[01:10] Lexi: And a year ago, 76% of the pages the AI pointed you to were also sitting in Google's top 10. This July, that's down to 37.9%. So most of what the AI recommends now isn't even the thing ranking at the top anymore.
[01:23] Lexi: They're blaming a switch to Gemini 3 back in January, and something with the wonderful name of query fan out, which I'm going to make one of you explain. Harjot, I'm going to start with you. Signal, or noise?
[01:35] Harjot: Yeah, so, um, for me this is, this is 100% signal, to be honest with you. And, and the reason is there's a big shift happening away from traditional search being, you know, the best way to surface information, and this number is just, it's just part of that story, right? Like, anybody who's been watching the AI space for a while now would never have assumed that these models just pull from the top 10 Google results. They, um, they've got their own indexes, they've got their own way of ranking what's good, what's bad. That's, that's literally why AEO, Answer Engine Optimization, even exists as a thing in the first place. And, and it's starting to diverge more and more.
[02:07] Harjot: So where I land on this is, you know, there's going to be a day where the search engine is a tool that the AI uses, right, rather than the primary way that you or I go and look for something. Because people are just going to want to ask the question. And, and honestly Google know this, that's the tell, they've put the AI Overview above all their own search results. That's how bullish they are on it. So, yeah, 100% signal.
[02:28] Will: Yeah, see, I, I'm going to come off the fence completely the other way on this, Harjot, um, because for me it's, it's noise. It's, it's definitely noise as compared to, you know, a bunch of other stuff we could be talking about. And, and I think what's actually going on here, um, is that the models have just got slightly better than the average person at search. That's, that's kind of it.
[02:46] Harjot: Okay, but, but hang on, what do you mean better than the average person, like, in what sense?
[02:50] Will: So, so, right, they don't release the underlying data, but I'd, I'd be willing to bet that for the top, you know, 20%, the people who are genuinely good at Googling, those results were always on the first page and they still are. The, the shift is coming from everybody else. So, um, if you've ever watched the average person sit down and Google something, like, like one of your parents' friends, you'll see they'll, they'll type in a whole full sentence with, with question marks and punctuation, or, or they'll do the opposite, they'll just type in one word. They'll just type restaurant and hit enter.
[03:19] Will: And, and in the old days, the restaurant they actually wanted, the, the local place doing the specific food they're after, that would've been on, like, page 200. Now the AI can contextualise that. It, it reconstructs a good search out of a, a really sloppy one. So the 37.9%, to me that's, that's more a reflection of how poor people generally are at search than it is any kind of, of significant change.
[03:40] Lexi: Okay, wait, before this runs away from me, one of you owes me a definition. Harjot, query fan out. Say it like I've never heard the phrase, because I hadn't until Tuesday.
[03:49] Harjot: Yeah, yeah, cool. So, um, a normal rank report just tells you, right, given one query, how well does your page rank for it. That's, that's all it can see. But query fan out is when the AI takes your question and, and it breaks it into a bunch of smaller questions, searches those separately, and then assembles the answer back together from all of them.
[04:08] Harjot: So, so say I want a family sized car in the UK. The AI isn't just searching family car UK, it's, it's going, okay, what car brands even sell in the UK, which ones are suitable for families, um, what's the boot space, you know, and it fans out into all these sub questions. And it cites pages that surfaced for those little questions, which might be pages that never ranked for the thing you actually typed. So that's, that's why the number's moving. The page getting cited was never top 10 for your query, because your query got, kind of, taken apart before it hit the internet.
[04:36] Lexi: Right, so let me make sure I'm tracking, because this is the part that actually matters to me. What you're both describing, whether it's a real shift or, or people just typing badly, the effect is the same. Ranking number 1 and being the thing the AI recommends have genuinely come apart. Those used to be the same list and now they're not.
[04:54] Will: Yeah, that, that's fair, and, and I'd, I'd sign up to that version of it. The, the thing that, that changes for a marketing leader, right, is that people are now effectively constructing more sophisticated searches even when they don't know they're doing it. So they're, they're not getting pulled in by your, your front page anymore. They might turn up deep in your technical documents, or, or honestly on some long lost webpage about something totally different.
[05:16] Will: And, and here's the bit I'd, I'd want a CMO to actually sit with, because it's an org chart problem. Your documentation, that usually sits under the product team, not marketing. So at, at a very minimum you now need to be speaking directly to that product team, because they own a page that's, that's suddenly a front door and, and they don't know it. Now, um, I'll be honest, this is fairly marginal. It's, it's still worth thinking about, but it's marginal.
[05:38] Harjot: See, that's, that's where you lose me though, Will, because I don't think it's marginal at all, right? I think, um, I think if you're a CMO with a good organic search channel, of course you maintain it, you squeeze the juice out of it. But you have to be adding capacity into the AEO side now, because this is going to be the primary way people access information in a few years, and, and you want to be set up before that, not after.
[05:57] Harjot: The teams who, who over-focus on SEO and don't give heat to anything else, they're, they're going to be left behind. Like their results just start diminishing and diminishing and then they get eroded, and it's going to feel like an economic problem, like it's not their fault, but really it's just the tide turning. So for me, SEO is the channel you maintain and AEO is the channel you build for. That's, that's the sharp version of it.
[06:15] Will: And, and I, I don't totally disagree with the direction, I want to be clear. I, I think it's durable in the sense that people's search is going to keep getting more sophisticated. What I'm, what I'm pushing back on is that this specific number, the 37.9%, overstates how, how structural the change is right now. And, and the honest advice to a CMO is annoyingly specific to each company. It's, it's a bit of a cop out answer, I know, but it, it's the truth. There, there isn't one clean thing everybody should do.
[06:40] Lexi: Okay, but Will, it depends is a terrible place to leave a marketing leader who came here for one real thing. Give me the one thing you'll actually commit to.
[06:47] Will: Yeah, alright, um, the, the one concrete effect I'll commit to is that this puts a bit of downward pressure on paid search. Because think about that single noun searcher, right, the person who just typed one word and hit enter. In the past you could, you could pay to intercept that person and, and land them in the right place. Now the model's just routing them there directly.
[07:07] Will: So, so that person's not really up for grabs the same way, which means, um, search gets a little bit less valuable to bid on. And if, if you're a CMO, that's, that's genuinely something to look at. It's, it's small, but it's real, and it's, it's more honest than pretending there's some big Monday morning switch to flip.
[07:23] Lexi: Okay, so you two are nowhere near each other on what that number even means, and yet you both just sent me to the same channel. Hold that thought, because the second study I found does something even meaner to the org chart.
[07:35] Lexi: So here's the one I keep chewing on. A company called Arobis tested 100 SaaS brands, ran thousands of prompts at them across ChatGPT, Gemini, Claude, Perplexity, all of it. And the study reports that the brand ranking number 1 on Google for its own category was just, absent, from the AI answer a buyer gets at the exact moment they're deciding what to buy.
[07:56] Lexi: And the ones winning were mid market companies with more third party validation, not the biggest content budgets. Now, big caveat, and I want to say it out loud, this is vendor research. It's from a company that sells the exact service it's recommending. So pull it apart for me. Is this real?
[08:11] Harjot: Yeah, so, signal again, um, but let me, let me caveat it the way you just did, because I'm, I'm not going to take these guys' word for it. Like, what would make me believe it is seeing the actual brands, the, the competing ones, the prompts, and, and why the AI picked X over Y, not just Brand X won. So, so hold the specific numbers loosely.
[08:31] Harjot: But the direction, yeah, I buy the direction, because it just builds on what we were saying, right? AI has the ability to do much deeper value based understanding of what it's recommending. It, it has to actually read the content. SEO just, it can't do that. So all this, um, clearer positioning, community presence, third party validation, that's exactly the holistic, heuristic kind of data the AI can take into account in a way that the old SEO signals just couldn't.
[08:56] Will: Yeah, and, and I'll agree it's signal, which, you know, doesn't happen twice in one episode usually. Um, it's, it's certainly worth looking at for particular brands. And, and it chips away at this idea Google's always pushed that good SEO just is good AEO. I've, I've been a bit of a skeptic on that, they, they don't seem to be optimising for the same things.
[09:15] Will: But, Harjot, the, the thing I'd add, that you're not really touching on, is there's an actual mechanism under this, and it's, um, it's a bit uglier than that. These LLMs, they lean, they lean really heavily on Reddit, and, and on YouTube, like disproportionately so. And, and that's, that's exploitable, right? You're, you're basically one grumpy Reddit thread away from the AI quoting some, some bloke back at you like it's scripture. Some person, you know, firing off a rant at 2 in the morning, and the, the giant robot brain just, it hands it over as if it's gospel. And, and look, that, that was kind of always a thing in SEO, but Google's, Google's more or less got a grip on it now, and the models, they, they just haven't yet.
[09:47] Lexi: So that's the risk. What's the opportunity in it, though, for somebody who isn't a madman on Reddit?
[09:53] Will: Yeah, so, so the practical version, and this is, this is a budget move more than anything. Take a big corporate, right, their PR budget is, is split across a load of things. Talking to customers, some of it's SEO, some of it's, um, corporate relations, government relations, all of that. And, and what I'd say has to shift now is, is some of that budget moving toward building really high quality, detailed information on a, a small number of reputable third party sites. Reddit, YouTube, and, and then whatever's authoritative in your specific industry.
[10:22] Will: Because think about what your government relations spend is actually for, it's, it's influencing people in government, right? But those politicians, they're, they're spending as much time on AI as everybody else is. So if you take some budget out of strictly government relations and, and put it into answer engine optimisation, done well, you, you buy back that reach, and you're, um, you're probably getting a bigger effect because it's coming from a trusted voice.
[10:42] Harjot: And, and this is where I think we actually diverge a bit, Will, even though we, we both land on small can beat big, right? Because your version is about, about the venues, the AI concentrating trust in a few places. Mine's, mine's more that the incumbents just, they can't move. Like, these big performance based teams, they think dollar in, 2 dollars out, they measure it super cleanly, and, um, they've built their dollar machine, so to speak.
[11:03] Harjot: And for them to redirect all that resource into AEO, it means switching off that machine to go try something new and scary, that, that they can't even measure yet. And that's a brutal sell to stakeholders when bottom line revenue is the metric. So the opening for the small player isn't, it's not cleverness, it's, it's just less saturation, less competition, and, and the old advantages the big brands leaned on, reputation, scale, they, they don't transfer cleanly into this.
[11:27] Lexi: Okay, so if I'm a new team versus a big incumbent, does that actually change how much I lean in? Like, is there a number in your head?
[11:34] Harjot: Yeah, um, so if you're a mature team, I'd say maybe 50/50, AEO to SEO. If, if you're a new team, honestly 75/25 in favour of AEO. Because it's, it's just an easier grab for them, there's, there's nothing to switch off, no dollar machine to protect.
[11:49] Harjot: And, and there's one type of brand I'd, I'd really flag here, which is the, the luxury and aesthetics driven brands. They're going to struggle, because AEO is, it's information driven. The AI doesn't really care about how your brand looks and feels. It, it cares about the information in the brand, what the product actually has in it, whether the product's good. It's, it's just not weighing that, that heuristic feeling the way a human flicking through a glossy magazine would. So for those brands especially, um, they've got a very different problem to solve, and, and it's coming for them whether they've noticed or not.
[12:16] Lexi: So the biggest content budget stops being the thing that wins, and it's the unsexy work, the being genuinely useful and having other people vouch for you, that actually gets you picked now. Which is a great story if you're small, and a slightly terrifying one if you spent 5 years building the machine that no longer helps.
[12:35] Lexi: So let's land somewhere a person can actually use on Monday. One thing a marketing leader could do this week. No re-platforming, no budget request, done before lunch. Will, yours first, and, and keep it small, because last time you tried to give me nuance and I nearly fell asleep.
[12:52] Will: Yeah, okay, small, um, and, and honestly it's the thing I keep coming back to. Take your, your 3 highest buying intent questions, the, the ones a customer actually asks at the moment they're deciding, and just run them through ChatGPT, and Perplexity, and Gemini, and, and see whether you're cited at all. Not, not are you ranking well, literally are you in the answer, yes or no.
[13:11] Will: And, and the second half of it, you can look at your own site right now and see if any particular pages are attracting a, a high amount of traffic that was originally linked out of ChatGPT or Claude. You, you can do that today, there's, there's a bit of code that comes with the URL when someone lands from there. So, so it's not a dashboard project, it's, um, it's 10 minutes to find out if you even exist in that world. Most people have, have genuinely never looked.
[13:33] Lexi: And Harjot, say the answer comes back we're nowhere. Will's told me how to find out. What do I do the second I know?
[13:39] Harjot: Yeah, so, so this is the important one, right? You put at least some dedicated resource into uncovering the AEO story for your organisation. And I mean, um, a named person, an actual human whose job is to experiment with this now. Because if you're a big org and you're doing none of it, you're, you're just going to wake up one day and find that you're so far behind, and, and building it up from zero at that point is really hard.
[14:02] Harjot: And, and I want to be clear, because people always hear this as kill your SEO, and, and no. I, I wouldn't stop the SEO content, keep it going, at least short and medium term. I'd, I'd just stop letting it be the whole strategy. The move isn't cut, it's, it's audit, and then actually give it to somebody. That's, that's the difference, to be honest with you.
[14:20] Lexi: And that's the thing I keep landing on with both of these. The number everyone's panicking about might be noise, or it might be the future, and, and you two still can't fully agree on which. But you both point at the same thing to do about it. Go find out whether the AI even mentions you, and, and give one person the job of caring.
[14:39] Lexi: Will, Harjot, thank you both, that was a genuinely good scrap.
[14:45] Lexi: And if you're listening, thank you so much for being here. We'll be back every week with something new.
[14:50] Lexi: Before we go, I have to tell you something. Every voice you've heard today, all 3 of us, me included, it's all AI. But the opinions are absolutely ours, and the judgment behind every word is human. That's the part you can't replace. And that's what Aloudable does. We take what you'd write, turn it into a show in your own voice, and we keep the taste where it belongs, with a person.
[15:11] Lexi: And there's no visibility score here, no AEO dashboard, nothing like that. We just make the show. So if you've honestly never once checked whether the AI even brings you up when a buyer asks about it, that's your 10 minutes this week. There's a link in the show notes, and I would love for you to come tell us what you found.
[15:27] Lexi: And that's the tea. I'm Lexie Meskouris, this has been Assistant to the CMO, and we will see you next week.