NB 66: Navigating Answer Engine Optimization Approaches
No Brainer · 2025-09-24 · 48 min
Substance score
56 / 100
Five dimensions, 20 points each
What our scoring noted
Our reviewer’s read on each dimension, with quotes from the episode.
Insight Density
There is a genuine cluster of useful tactical points - LLMs relying on mentions not links, the semantic triple concept, the ziptie.dev tracking insight for logged-in users, and SparkToro's audience-AI-usage mapping - but the episode is heavily padded with banter, tangents, and broad validation of obvious points. The ratio of signal to filler is mediocre for a 48-minute episode.
the LLMs rely on mentions, not links. So you need to make sure that you've got your semantic triple
ziptie.dev was the first on the market and they have figured out a way to track logged in Google users...you're going to see seven times more AI reviews popping up
Originality
The core contrarian claim - that AEO/GEO is 'horseshit' rebranding and it's still fundamentally SEO - is refreshing and practically useful, but most of the surrounding advice (allow crawling, HTML over JS, structured data, clean online presence) is recycled SEO 101. The user-embedding framing is mildly novel but not developed rigorously.
I think it's horseshit. I don't know people who created those terms what they think the difference is
Google has 373 times more traffic than ChatGPT right now. So if we're talking about traffic, we're still talking about Google
Guest Caliber
Katherine Ong is a genuine practitioner with a decade running her own boutique consultancy, a Ketchum background, and named client work at cancer.gov and National Academy of Sciences; she references real tools and real client data, which puts her well above the podcast-circuit thought-leader category, though she operates at boutique scale rather than enterprise.
she helped one of my former brands really crush it. She helped us increase traffic by 300%
I have a client that wants their content stolen. They're an open data platform
Specificity & Evidence
The episode earns credit for citing specific numbers (373x Google vs ChatGPT traffic, 67% cookie blocking, 7x AI overview undercount, 3% AI referral share, $12k/year user testing cost, 2% AI mode usage in iPull Rank study) and naming real tools and clients, though some figures are asserted without sourcing and several claims remain hand-wavy.
2% of the people actually used it and most people ditched and didn't stay for very long
It's like 12,000 a year or something
Conversational Craft
The hosts occasionally push productively - asking about traffic drop attribution, measurement under data loss, and the CEO personalization trap - but most questions are soft setups and the episode is dragged down by sustained off-topic banter, political asides, and the hosts frequently answering their own questions rather than pressing the guest.
How much of that is due to the rise of AI versus some other wide range of potential?
to what extent are these AI answers just confirming what an individual already knows or parroting back their behavior?
Conversation analysis
Computed from the transcript - who did the talking, and the verbal tics along the way.
Share of words spoken
- Speaker C61%
- Speaker D25%
- Speaker B12%
- Speaker A3%
Filler words
Episode notes
An exploration of one of marketing AI’s hottest trends with Katherin Watier Ong In this episode of No Brainer, Geoff and Greg dive deep into the evolving landscape of search engine optimization with Katherine Watier Ong, founder of WO Strategies. Katherine brings her 10+ years of experience helping science-based organizations navigate the complexities of modern search - from traditional SEO to the emerging world of AI-powered answers. The conversation explores the hype surrounding Answer Engine Optimization (AEO) and Generative Engine Optimization (GEO), with Katherine providing a refreshingly grounded perspective that it's still fundamentally SEO at its core. She shares practical insights on how to optimize for Google's AI Overviews and discusses why the fundamentals haven't changed as much as the industry buzz might suggest. Katherine reveals compelling data showing that while AI-powered search features are getting attention, they currently represent only about 1-3% of actual traffic for most organizations. She emphasizes that brands should focus on solid SEO foundations before chasing the latest AI optimization trends.
Full transcript
48 minTranscribed and scored by The B2B Podcast Index.
Speaker A: You enjoy podcasts. We know because you're listening to one. Hi, I'm Jason Falls, the executive producer of the Marketing Podcast Network. We love podcasts, too. So much so that in addition to providing you with the great episode you're listening to, we also help businesses, brands and even individuals produce their podcasts. MPN Studios offers podcast consulting and production services to help you get you or your business its very own podcast. We've helped dozens of smart people, advertising agencies and brands develop podcasts.
Speaker B: Podcasts.
Speaker A: All you do is record your content or interview. We take that and deliver a professionally produced audio or video episode, complete with show notes, transcriptions, and even promotional clips to use on social media. And our monthly subscription prices are hard to beat. If you or your company are jonesing to have your own podcast to establish thought leadership, drive business leads, or just have fun talking about your business or topic of choice, we'd love to help learn more@marketingpodcasts.net studio. That's marketingpodcasts.net studio.
Speaker C: I'm not giving up.
Speaker B: I am selling the building.
Speaker C: The final season of FX is the Bear. The restaurant is flooded. Everything's either gonna be okay.
Speaker A: Nope.
Speaker D: Stop.
Speaker C: Or not.
Speaker D: We are outgunned and we are outmanned. But we have each other.
Speaker B: FX's the Bear the final season, all
Speaker D: episodes now streaming on Disney plus. Hey, this is Greg Verdino. Welcome back to another episode of no Brainer. Today. We are recording this on August 27th, and I am here, as always, with my good friend, my bald brother from another mother, Jeffrey Livingston. How are you today?
Speaker B: Great, M. As you know, I went for a late morning run, kind of my lunch break, and it was actually decent out. I'm like, oh, my gosh, we are hitting fall. This is amazing.
Speaker D: It definitely is starting to feel a bit fall. Like, especially after a hot, humid, oppressive summer. As we've discussed many, many weeks in a row.
Speaker B: I think you're playing on words with the oppression thing. Too soon? Too soon, Jeff?
Speaker D: Too much.
Speaker B: Ah.
Speaker D: But, uh, we do have another good show. Before we dive in and let you introduce our guest for this week's episode, I do want to remind everybody, as always, that your support is very important to us. Please be sure to subscribe to our channel to like, to love, to rate, to link, to share, to do all that good stuff. Jeff's doing a thumbs up. He's hoping that his web camera is going to send balloons and fireworks our way. Uh, but do be sure to express that support if you can. We appreciate it very much. And on that note, Jeffrey, who do we have with us?
Speaker B: Well, we have a really great guest today and she actually is a former vendor of mine. Her name is Katherine Ogg. Hi Katherine. Let me embellish your kudos. Uh, she had a big, I guess it's now the big four, but a big five background with Ketchum. Um, and since then it's gone out. Has her own, uh, SEO, I guess maybe now it's ao, geo, mno, M, mma, whatever. But basically search optimization organization. And she helped one of my former brands really crush it. She helped us increase traffic by 300%. So I ended up having a conversation with Katherine when all this answer engine optimization hubbub started to really, uh, peak, uh, and start hitting the media about four or five months ago. And of course she's up to speed and rocking and rolling for all our clients. So what better person to have on than Katherine? Somebody that actually does this work to understand, help everybody else understand what we're dealing with. Katherine, welcome to the show. Founder of WO Strategies and Search Grid.
Speaker C: Thanks. I know one of it's funny because I posted today because this is my anniversary month for my business. I am at year 10 and I happened to flag my work with cancer.gov and uh, Peter over, who used to be the head of communications over there, actually called me a search Sherpa. I think that's brand new for me. I've been called a search fairy godmother and now a search Sherpa. So there you go.
Speaker B: It's like back to the social media days where everybody had a weird neighbor. Sasha was a social media swami. Nope.
Speaker C: I do remember that actually.
Speaker B: Yeah, yeah, yeah, right. You remember Sasha.
Speaker D: Chris Penn was a ninja something or other. Or something other ninja.
Speaker B: Yeah.
Speaker D: And you were just a cranky old bastard at that time.
Speaker B: Although prematurely old, but still cranky now. Now I fit the beard matches. So. So, Catherine, you know, obviously there's a lot going on and we're in a really weird time. I feel like with information, uh, where people can just say things on LinkedIn and, or even put out studies that have maybe surveyed 40 or 50 people and have finite answers for those that are coming up to speed or trying to understand, um, what are AEO and geo, um, in practical terms. And then let's get into like, kind of how they're disrupting the ecosystem, if you would.
Speaker C: Yeah, the pace is intense. I mean, I got into SEO because I really love learning. And it's always been changing, right. Like Google makes 6 algorithm change of the day. So it's a space we have to love learning to do it as a career. But, um, I got on the Internet in 94 and it's been feeling a lot like 94, where there's just so much new information your brain feels like it's going to explode. Um, and so that is constant across everybody, including folks who've been in the industry for a long time. Um, no, I have not adopted these new acronyms. So Answer Engine Optimization, Generative Engine Optimization, those would be the acronyms because actually underneath it all it's SEO. And SEOs are well positioned to even understand how to optimize with the new stuff we're dealing with. So I actually think SEOs are going to be leading the charge. And it's kind of has been feeling like that tools that have rolled out. Everybody who's building the tools was a former SEO, right? So, um, yeah, so I haven't doubled down on those terms though. It's funny because I will be speaking in front of a client's sort of user group and they are using the terms like we've got our AEO expert or whatever.
Speaker B: Is it, is it geo? A class in high school?
Speaker C: Yeah, exactly. You Google it and it's like, yeah, exactly. Is it geography? You know, I, uh, just don't think we. I mean, ultimately it's going to be the new Google, right? So if we take away all of it, Google's got 373 times more traffic than ChatGPT right now. So if we're talking about traffic, we're still talking about Google. And Google has said you can't block, you know, any crawling for training data. If you're in, you're in. You got to give it its training data and then you could also be maybe in these answers. Um, and of course they're talking about maybe having some agentic future, but it's still, still SEO at the end of the day. It's not like this new thing now, mind you, optimizing for it is radically different and I think it's very interesting. Thank you. Mike king over at iPoolRank. It's very interesting that we had all these changes going on and we all missed it. Like the whole industry. Mike's been writing this book about Google, which is why he discovered it. Between that and the algorithm link leak, he put some stuff together and yeah, like just last week I discovered that Meta, in addition to LLMs and Google Search are using these vector embeddings. Well, they could have been using these vector embeddings for a very long time. We just recently discovered that they were doing it. It did it with the user embedding. So Meta is using user embeddings. And if you have gone to any of my trainings or talks or whatever, I always talk about how creepy it is about how much Google knows about you. And they say they don't personalize using all that data, but this is what Google knows about you. Right? Um, so anyway, I think the user embedding stuff might have been going on for a long time too. It's just. It's more transparent now. People found the patents, um.
Speaker B: Whoops.
Speaker C: Whoops. Exactly. Exactly. But the interesting part for me on Google's end is that, um, I had not connected that Gmail is part of it. So if you want to appear for a person based on all the things Google knows about them, that also includes your email campaigns.
Speaker D: Right?
Speaker C: So that pivots your brain a bit as a marketer about how you're going to make sense.
Speaker B: So, I mean, it's all the same profile. It's all Chrome. It's generally speaking.
Speaker C: Right. And we know they're watching us on Chrome as we move about. That's how they can figure out whether people stay on the websites or.
Speaker B: Right. We definitely have to discuss that whole crazy perplexity angle at some point. But, yeah. So I feel like I get that it's SEO, but Answer engine optimization and generative engine optimization, what's the difference between an answer engine and a generative engine? Or is it just horseshit?
Speaker C: I think it's horseshit. I, uh, don't know people who created those terms what they think the difference is. Ultimately, when it comes back to Google, it's appearing in Google or it's appearing in bing. And if ChatGPT's search efforts get big enough, it's appearing in ChatGPT. Now, there is a different process for getting there, but underneath it all is top rankings in Google. And there's a bunch of studies that recently came out, like this month. They very clearly show that it's still, for the most part, it's top rankings in Google. So with a lot of the regular old SEO stuff, especially the most recent rollout of ChatGPT, instead of trying to, like, build their own index, they're definitely using RAG and going to Google and pulling from Google. So not surprised, right? Right now, I don't know what Perplexity is doing per se. Um, but they're smaller. And again, I have clients that really want an LLM strategy, but I'm still like, uh, uh-huh, huh. Let's do the SEO stuff first. Right. It's the bigger pie. Let's keep an eye on the LLMs. I know which one to work on next once we get through the SEO stuff, because most of my clients have a lot of SEO stuff to be doing.
Speaker B: Yeah, Perplexity just wants.
Speaker D: Greg, isn't it?
Speaker B: They wanted to buy Chrome.
Speaker D: Yeah, well, yeah, I mean, I don't know how much of that was performative. Right.
Speaker B: Hey, we're Perplexity Media, pay attention.
Speaker D: But I think Claude now has a Chrome, uh, plug in and you have, I mean Perplexity has their own browser now. I don't think anybody uses it. Right. But they've got comment now. So I think it was just they weren't getting enough media coverage or something and Sam Altman was out shouting them or whatever and they said, hey, let's make this like just balls out offer to spend, uh, a ton of money on a Google product. I mean, who knows? Um, you know, it's refreshing, Catherine, to hear you as you come from an SEO perspective and you have no ax to grind, no new product to hype, no venture money to chase, um, and to, you know, kind of sort of just deliver that fundamental, what you know, hopefully is maybe is a truth, right, that this is just SEO. Uh, it's an advancement. Yes, there are different tactics and techniques and technologies, but at the end of the day a lot of the same fundamental principles still apply. What would you say from your perspective, flat out does not change. What are the things that have been good practice all along that remain good practice that people still need to put their energy and effort against?
Speaker C: So you need to make sure that you're allowing crawling. Right. So I have some clients where they're on cloudflare intermittently. Cloudflare will block Googlebot, right. So don't do that, allow crawling, which includes, if you want to be in the LLMs allowed, they're crawling too. Um, and then the interesting part that has changed a little is that though we've been telling people to do this for a while, put the essential stuff in HTML so the LLMs cannot read JavaScript. And as SEOs, if we see a full on JavaScript web page or site, we're going to encourage that the navigation and the text is in HTML anyway. So that's new but not really new, um, writing, well structured. I find it very intriguing that the LLMs rely heavily on things that are also ADA requirements and SEO best practice, which is well nested headers as well as paragraphs underneath the headers that talk about that, you know, like semantically connected, um, structured data still helps from the rag perspective. It helps Google understand who you are. Um, so that's still a thing. Things like cleaning up your online presence universally, which is definitely a requirement for local SEO and then is a requirement for my bigger brands because I want them to have either get a knowledge box or claim the knowledge box and link all that stuff up. Right? So that's a piece of it, which we've been telling people. I've been training folks on what the industry calls fraggles, right? Which are these fragments or paragraphs which is the key to the cosine similarity right now, having a topic that's well matched and closely related to the paragraph underneath. So I've been talking about that whether everyone's been adopting it as something totally different. I know Healthline was definitely adopting it because I've used that as an example. Whether everybody has, who knows. But those, those are still like legit signals to keep an eye on. So yeah, crawling and then making sure that you get in the index and knowing that Google. The big change, I guess might be a year or two ago. I forget when the Google Helpful update rolled out, they changed the requirements for getting in the index. And even with my high domain authority client like National Academy of Sciences, not everybody gets into the index. Indexing has gotten harder over the last six years or so. And it's because your content has to be quality enough and enough search demand to put it in the index. So while they could come and they could crawl and maybe they turn on rendering and they can see everything, they might not actually stick your stuff in the database and you can't rank until it's in the database. So that's sort of a big change for people that aren't following SEO. But it's honestly been a, uh, reality for the last six years. Um, but that's what I usually train people on. I'm like, hey, this is what this update means. And for Google it means having a clear author and having somebody with hands on experience and expertise and authority and making that clear on the website. So for my clients it's somewhat new, maybe if I'm working with a new client, but it's not necessarily new in relation to our SEO checklist kind of stuff we've been doing, I guess the only change is that the LLMs rely on mentions, not links. So you need to make sure that you've got your semantic triple, like, what is your brand about? Like, so if it's Nike, it should be sneakers, right? As one of them and making sure that's mentioned in the trading data and the appropriate places. Um, and before it was like a link thing and the links are still important for the ranking in Google. But the LLMs in particular want to are going to scrape their training data and try to pull people where the brand is next to the topic. That's a little new, but generally I think about, I mean the cosine similarity is new, um, but a lot of it has not changed.
Speaker D: Yeah, because I'm thinking while that's new air quotes for the people who are not watching but listening, um, it also sounds very obvious, right? Why would Nike not associate their brand closely with Sneakers or Cognitive Path, associate closely with AI consulting or you know, your company associate closely with SEO. Right. I mean maybe brands have gotten a little bit too far away and kind of too soft and conceptual and all of that stuff. But it doesn't sound like there's a lot going on that is so earth shaking that people should be banging the drum as many are about the death of search. And now it's all about being in the answer. Yes, you need to be in the answer, but you also need to still do all the fundamental things you've always done or always should have been doing as a search marketer.
Speaker B: Right.
Speaker C: Yeah. I think the big, um, the big change which we haven't topped on, which has nothing to do with the LLMs. There's two, three really big changes that have happened in the last three years that are really impacting all of this. One is that social has been zero click for a while. Like you put a link in, you get depressed. Now searches and zero clicks, it's worse
Speaker D: than zero clicks I think.
Speaker C: Right? Exactly.
Speaker A: Yeah.
Speaker C: Uh, the other thing is that people actually aren't going to Google to solve their problems. They're solving their problems on channels we can't track very well, like chatting with their friends or a Facebook group or whatever. And when they come to Google search, most of the time they're searching brand queries. So that's a big change. And then our marketing data has radically changed. GA4 really changed the playing field. And then 67% of US adults stitch cookies. That's a huge number. And so we're just losing marketing data. And these LLMs go into direct the AI overviews and AI mode also direct the measurement is getting really mushy. Right. So that's a huge challenge for marketers and they should be adjusting their KPIs this year if they have not. Those aren't necessarily driven by this LLM answer thing, aio, whatever. Just more larger things that are unfortunately happening at exactly the same time which make this year super challenging.
Speaker B: But AI is an easy foil for that, right?
Speaker C: Yeah, yeah, as usual.
Speaker B: Um, so, but with that in mind and the inability to um, execute with precision kind of like we used to, what are some of the answers and can AI help with that?
Speaker C: Yeah, I mean I have integrated AI into a lot of my workflow and I'm a training oriented consultancy. So I train all of my clients a technical SEO and how to write for SEO because they're their expertise, subject matter experts. I'm not writing their content. Um, so I'm helping them rework their process and speeding it up by using AI. I think that AI does a great job at writing meta descriptions. For instance, they can help with creating content brief, they can help you with the brainstorming. I mean I use them in a variety of places and have ever since AI came, uh, on the playing field. What's the best part of the question? You were talking about integrating AI.
Speaker B: Yeah, I mean is AI, uh, part of the solution also analyzing how to get more precise and to hit the right result?
Speaker C: Yeah. So the, the latter part of that is the maybe rank tracking. Right. So there's, I have like a spreadsheet. I feel like every week there's a new LLM rank tracker I've not had a chance to play with. Profound, which is the one everyone loves because they're enterprise pricing. I just want to make sure everybody realizes that for keyword research in an SEO tool, we have a database from Google we're pulling on. We actually know people are searching for these terms for prompting. We have no idea whether there's people actually using these prompts.
Speaker B: Right. The prompt tracking thing is a big thing on the PR side too, where.
Speaker C: Yeah, and I mean figure it out. I know, I think with Ahrefs or Semrush, they are pulling from people also ask. So clearly if people query on Google a question and we see it in the people also ask, perhaps they're prompting that way and that's sort of like how they're jumping over that gap. But I do worry a little bit about uh, putting in a prompt and you might be using a question nobody's actually using. I mean my big recommendation though is as you start tracking AI mode and AI overviews, I really recommend using zip tie. So ziptie.dev was the first on the market and they have figured out a way to track logged in Google users, which I think is most of the people on the planet. And for information based clients like mine that are getting really hit by these AI overviews, you're going to see seven times more AI reviews popping up than the tools that don't track the logged in Google users. But you do need a separate tool. Some of the SEO tools, ah, are like I said, are rolling out AI tracking inside their tool set. So you know, you can get it now in ahrefs. But again I worry about like setting up the prompt, like is this the prompt someone actually using or not.
Speaker B: Right.
Speaker C: It's a big question mark.
Speaker B: Plus also it may be performative based off of what's happening in the news cycle or what's the trend of the week. Right. I mean it seems to be something that moves very quickly.
Speaker C: Yeah, yeah.
Speaker B: But even with all this and you were saying that it's just like 1% of the traffic. So if it's that, I mean are we just basically distracted by this right now?
Speaker C: Yes. I have a client that um, wants their content stolen. They're an open data platform.
Speaker B: Sure.
Speaker C: For them, I want to say that it's a developer audience and so they've got a little bit more AI activity than what would be the norm. But we're talking about like maybe 3% or something. It's still not the big traffic driver.
Speaker B: It's uh, like Yahoo. Yeah, yeah, Remember that? What was that? Yahoo. What was it?
Speaker C: What I loved about Yahoo is what you can actually get in SparkToro right now. Yahoo used to be able to had a free tool where you could put a keyword in and it would tell you the demographic of the person using the keyword, which was like super helpful there for a while until they spun it down. But yeah, I Recommend People use SparkToro because SparkToro will actually put a keyword in and it can tell you whether or not that audience is over indexing on using AI platforms and which one that's handy.
Speaker B: Yeah, for sure. Then it becomes relevant. So really, unless you have a very AI centric audience, it's really kind of an overhyped thing right now.
Speaker C: It's an overhyped thing. Also the people getting obsessed about it haven't done the blocking and tackling on the SEO side in my experience that I have talked to anyway, I can't
Speaker B: believe that they're the same ones out on LinkedIn every day. Right.
Speaker C: Let's just do the Google stuff first.
Speaker B: Getting serial clicks too, I mean, but
Speaker C: the challenge is they're still getting decreased clicks. I can't do anything about that. I mean so the most recent update from Google, literally two days ago, is that they are. And it's interesting because this was on the heels of Ipull Rank doing a user study about AI mode where people basically did not click. Let me see if I remember. So 2% of the people actually used it and most people ditched and didn't stay for very long. So Google is now putting, uh, more links in and hopefully not links to themselves because they've been putting a lot of links themselves in there and we've been trying to tell them nobody likes that. And that study actually showed that people were annoyed to go, like back to Google when you were hoping to get something else. Right. So if they put more, uh, link references, maybe we'll get more clicks. Generally with the AI overviews we haven't seen. Depends on the topic. If it's. We have seen that financial and healthcare. Sometimes people research the tiniest bit more, but oftentimes they just unfortunately take that answer, which could be completely bloody made up and wrong, and run with it.
Speaker B: And run with it. Yeah, it's really comforting. I have this oil on my leg. ChatGPT, what should I do?
Speaker C: It's horrible. I mean, I, I support science folks and it really concerns me that we've got everyone relying on an LLM that could fully make up crap.
Speaker D: Yeah, absolutely.
Speaker B: It does. Regularly.
Speaker C: It does regularly.
Speaker D: It's not making up crap. It's something that came off of like Rihanna's Wikipedia page or something.
Speaker B: Maybe ChatGPT should run for president. It would probably do a better job.
Speaker C: Oh, nice.
Speaker B: We might have to delete that one. I don't know. That might get edited out.
Speaker D: So I'm trying to thread the needle here because there's clearly a change in behavior that happens when people start getting served up AI high answers, as you've just described. Um, but to your point earlier, the amount of usage is relatively low relative to traditional search, let's say. But a lot of marketers are claiming massive drops in website traffic.
Speaker C: Oh yeah.
Speaker D: How much of that is due to the rise of AI versus some other wide range of potential?
Speaker C: Again, like this week, SparkToro released a study using click data showing that Google search volume in general has not decreased. People are still using it just as much as they have before. If you're a heavy ChatGPT user, you've actually searched on Google a little bit more than you did before. So these things are coexisting, if that makes sense. So you could be an active ChatGPT user, but it has not decreased People using Google Search. Now what has happened is the stupid AI overviews for every informational based query. And then a couple of those your money or your life queries, like the health stuff, where just fully disagree, they should be doing an AI overview. When people see those overviews, they take the answer because it's a full, a full, complete, nicely stolen answer, I would like to say, for instance. And then they just move on and they don't click on anything. They start a new query and everyone's seeing decreases across the board.
Speaker B: Now I know what happened to my phone.
Speaker D: I know I'm a focus group of one and I mean probably everyone. You know, all three of us on this, on this, uh, podcast are atypical in our behavior because of what we do for a living and what we know.
Speaker B: Right?
Speaker D: I mean, we're all atypical in a lot of ways. But when I think about my own personal behavior, I do the kind of thing you're describing, which is I might do a quick query in ChatGPT, let's say, to get, uh, an instant answer on something, but I'm smart enough to know that's not a definitive answer on anything. So then I will hop over to Google, I will do the search and because I know what's likely to end up in that AI answer, I just, I'm blind to it. Unless it's something super, super simple, I just jump right past it and do the same thing that I've always done, which is go through the blue links and you know, look at multiple sources and come to my own conclusion. Find the best information that I think I've found and you know, take an action, make a decision, do my research, make a purchase, whatever.
Speaker C: But you're literate and 60% of the US reads it as 6th gr. Great level.
Speaker D: Exactly. And you know, that's probably the other thing that's in my head is. And I know we're not here to talk necessarily about these sort of like wide ranging philosophical, you know, thought experiments, but, um, what scares me about a lot of this is the fact that we have a generally media illiterate population, um, that is likely to take that answer at face value and whether it's for a brand and you know, you think about brand misinformation, brand misrepresentation, all that stuff, or something far more important than what are the best running sneakers or whatever. It seems that there is a tremendous risk to the entire information ecosystem.
Speaker C: Yes.
Speaker D: When you have a population that's ready, willing and able to just accept this garbage response at the top because it looks like it's prepackaged, ready to go.
Speaker C: And they don't know that it's a autocomplete system. Right. It's just a mad libs. That's all LLMs are. They think it's like somehow more intelligent than that, which it's not. And I mean I'm in a couple like chronic health condition type Facebook groups and dear God, the number of people sticking their health information into ChatGPT and then either one, sticking their health information into ChatGPT, which is the problem, or two, like taking the answer as like gold without triple checking anywhere. And I'm, I am the one that's like, do you know how LLMs work? Like, um, and, and people don't. And, and my mother, who's like 77, well educated, but of the generation that just believes everything from Google right now, in theory, these, the LLM that creates the answer is trained on Google's quality rater system, including the human raters that are involved in the search system. So in theory, some of these go through a process of making sure they're legit. That said, you hang out on the SEO community on X at any point and people are always sharing these answers that are just cray cray.
Speaker B: I mean, that's the other thing too. If there's a fair fraction of content that gets indexed, it's intentionally misinformation and disinformation. And that in its own, um. Right. Is not discerned by these same algorithms.
Speaker C: Well, and on the LLM side, if you are. Because I consider myself very white hat SEO, because the clients I work with, I've always been above board. But if you're going to do nefarious black hat stuff and trying to get a quick win, a lot of that works right now on the LLMs. So like the white text, a white background, kind of old school. It's easy at the moment. Yeah.
Speaker B: And it's because there's no rules, regulations or anything.
Speaker C: No to this whole spamming thing. School's been doing the spamming thing for a very long time and they just rolled out a spam update this week. They're trying to stay ahead of the spammers and I don't know that these LLMs realized how much of a problem that would be. Right, uh, manpower they might need. And maybe they don't care.
Speaker B: They don't care until they're called out on it.
Speaker C: But at that point, somebody like the kid who just killed himself because he got counseled by ChatGPT to do so, like there's real impacts. Mhm. For like, this kind of behavior. It just fully bothers me. I mean, we had a scientifically illiterate base of folks here in the US Anyway, and now you're making it that much harder for me to get accurate science information to the end user. And sometimes you're gobbling it. You're totally making stuff up.
Speaker B: It just, it's painful.
Speaker C: It's painful. And then politically, it's painful. Everybody has a user embedding related to the content that they get off the Internet, like their social profiles, they follow, and all of them. And so the filter bubble is significantly worse than we thought it was.
Speaker B: Yeah, right.
Speaker A: Ah.
Speaker C: Oh.
Speaker B: You could see it all the time on any social network. It's just unbelievable. I mean, it's the same with Google, obviously.
Speaker D: And to what extent does that bubble get bubblier when you're looking at AI answers? I think about, like, when I use ChatGPT, the number of times ChatGPT has served up my stealth in the answer. Like, if I say no, who are the top keynote speakers on such and such a topic and my name is in there, or who are the top AI consultancies? And Cognitive Path will be in there. Which is true, by the way, for anybody listening, uh, you should have called
Speaker B: us last week, not this week.
Speaker D: However, if I then challenge it and say, hey, by any, you know, because it'll be like, the Top consultancies for AI are McKinsey, BCG and, you know, Cognitive Path. And I'm like. And I'll be like, gee, Cognitive Path seems like an outlier there. Why did you mention that? And it's like, well, because of our previous transactions, I get the sense that you were involved with Cognitive Path, so I wanted to make like. So to what extent are these AI answers just confirming, uh, what an individual already knows or parroting back their behavior?
Speaker C: They're all personalized to you and your use of the tool. Um, so yes, they will. And then they're also not so much the Google answers, but the LLM answers. Uh, the actual LLM has been trained to flatter you.
Speaker D: Right?
Speaker B: Of course. Yeah. The sycophant complex.
Speaker C: Right, exactly right.
Speaker D: And that's where I start to wonder when we even talk about optimizing for something like ChatGPT, if ChatGPT is going to customize those responses to the individual based on its prior interactions with that individual, how well can somebody, even an organization, a brand, a, um, nonprofit or whatever, break through that bubble when the LLM is just trying to be a
Speaker C: people pleaser Anyway, so my suggestion if you have budget is you run a user test with something like user testing, starting from search with your target demographic and an example of something you want them to get done without leading. Right. And watch what they do and have them talk out loud about how they're interacting with the interface. Then you're going to actually figure out what websites they visit, what the social is, how they think about how you structured on the page. It's so insightful. Okay, not everybody has that budget, but that's like the, I think the gold standard, if you could do it.
Speaker B: And what's the cost of that roughly, for somebody that's listening?
Speaker C: Oh, it's like 12,000 a year or something. There's a couple cheaper ones out there, but the concept of like getting a
Speaker B: drop in the bucket for most marketers.
Speaker C: Yeah. So like start from search with a scenario and just let them talk out loud and see what you get. Um, so that is absolutely the best plan. Now if you can afford that, Sparktorio is getting you halfway there. So if you, you can either put in your whole domain or if you haven't optimized for anything yet, you could put in the keyword. Right. So if I put in a particular keyword, you're then going to see all the places these people go online and other topics they're interested in. I just did this for a client and I knew that they had like a developer audience and like a government policy audience. And as I was looking at it for one particular keyword, I saw that that person goes to every single other military website. And I was like, oh, there is a military person who cares about the website. I did not know that. Um, but then you know what podcasts they go to, what YouTube channels they're on, and you can start a surround sound strategy which would influence their user embedding. But you need to take your old Persona plan and put it on the shelf. And instead I'm, I'm calling them user groupings. But think about grouping users together based on where they hang out online. It's not going to be perfect because everybody's like fully personalized to the like, inst degree, but it's closer than what we've been doing. M. Right. So now you know, you have to hang out on Reddit, you try to get on this guy's YouTube channel, you got to get on this podcast, you make sure that person gets on your email, so you're getting into their Gmail account, etc. Rinse and repeat for every bigger user grouping that you have. So Marktoro also only works in English, unfortunately. Canada, the US and the uk. I don't have a solution outside of that. And if anybody's listening and has a solution, let me know because I need that. But at the moment, I think those would be your two plans for attempting to tackle that user embedding part.
Speaker B: Mhm.
Speaker C: But you can never measure it, right? You can never say I definitively am getting in front of this person unless you somehow rework the user testing test again or something. Ah, again, it wouldn't be definitive. You're like, hey, now we're showing up more for this demographic, a new demographic that's similar that we recruited. Maybe. Um,
Speaker B: how are you measuring success now, given everything?
Speaker C: I mean. So for my clients, all we're doing is tracking in GA4 as a new channel, all the LLMs coming in to start to see how much traffic that actually looks like. Um, and then, uh, for some of them we're looking at rank this prompt, rank tracking option. And I think for AI mode and AI overview, it's pretty straightforward to kind of figure that out. It's the ChatGPT and not knowing if anybody is using the prompt that get the little wishy washy. But that's all we're doing. I don't have anybody that's doing whatever the next step would be because again, they still have basic blocking and tackling on the search side to tackle. Yeah,
Speaker D: something I'm thinking about, um, you know, brand. It's 2025, end of August, 2025. Um, God, we feel it, don't we? But, um, is it 2028? The um, you know, brands are still very much focused on kind of controlling a very carefully constructed narrative about the brand. Now they should have learned 20 years ago that maybe more than that, but certainly 20 years ago when social came onto the scene, they should, it should have dawned on them because people like Jeff and I have been telling them for that long, um, that you don't control your narrative. Right.
Speaker B: United Breaks Guitars.
Speaker D: Exactly. Um, and I know if, if you're familiar with like Pete Blackshaw, um, and you know, he talks a lot about how in the eyes of the LLM, your narrative is no more or less important than everything else everybody else is saying about your brand, which of course is a social media concept brought into the age of LLMs, right. You might have carefully crafted your entire brand narrative. You might have spent a ton of money on a perfectly phrased tagline or whatever. But at the end of the day, if people on Reddit are saying that claim's not true or this product didn't work or whatever, that's going to have far more influence on, let's say, a chatgpt answer and that, uh, you need to be out there doing that stuff and you need to be engaging in Reddit communities. And you said this too, on the right podcasts, on the right YouTube channels, et cetera, et cetera. Um, which sounds almost obvious because for anybody that was doing social media marketing or media relations before that, or content marketing, that's always been part of the ethos. Um, but for some reason that still hasn't sunk in with a lot of organizations. Um, so I'm not sure what the question is, but I'd love to kind of hear your reaction to that. Is it doubling down on those things you should have been doing all along, just like SEO basics? Um, is it doing something new and different? And why doesn't this stuff ever sink in with marketers?
Speaker C: But I mean, it sort of is doing something new and different. So when I was at Ketchum, I ran social campaigns. I was sitting in a PR firm, right. And before that I was an in house marketer. So I've done all the channels. Um, but I think what is different is that social people don't particularly think about keywords like we do. Right. We're very obsessed about the exact term being in the exact right place. PR people definitely do not. They're going to create flowery language and they're not thinking about the right term in the right place. And also these people are not coordinated generally. Um, and so I think what needs to happen, also, I don't think most brands, um, get a social media monitoring tool. Frankly, they all should. And I know it's because it's expensive. So if I ruled the world and I had all the budget in the world, I would definitely have something like Brand Watch running. I particularly Brand Watch. There's probably other tools on the market that do the same thing. But the reason I like them is they had, back in the day they had Moz Metrics and you could, you could track by topic. So instead of just tracking my brand, I could be like, yeah, okay, but everybody talking about soda is Coca Cola mentioned. Right. I could actually use a keyword and track this keyword conversation. So I would get a social media monitoring tool that lets you track brand and topic that you want to appear for to make sure that if around topic if you're missing. Right. Which you will not get if you just track your brand term. Uh, this is New to PR people, though. So I feel like I have to really emphasize this. Then integrate the search and social, because right now, like, Google's got a feed from Reddit. It's showing a ton of social in search. And so sometimes in order to appear in search, you actually need to answer a question on Quora. Right. SEO, uh, people can handle this just fine, but if it's a different team. And then just the other day I spotted Google Trends set up for my clients and I spotted a press release that went out, but it didn't have the right terms on it. Do you know what I mean? It was like created by the PR firm and I was like, no, no, no, no. This is going to become a mention on the Internet. It's going to matter that we make your semantic triple. Like on that place, it's getting published that matters. Which is a level of detail that PR people have not been focused on. It had been hard getting them to do the backlink in the first place. Like, great, you've got coverage on weather.com now I want the backlink. Right. Which was. That was our ask before. And now it's like, no, no, no. I just really need to make sure that the brand and the topic are close together in that paragraph, not separated by paragraphs. For instance, that's going to matter. So that's new and probably going to be more challenging to get all these folks integrated.
Speaker D: But that's a human challenge. It's a human.
Speaker B: Yeah, it's a huge business threat behavior
Speaker D: and skill set and.
Speaker C: Well, and finding the budget for the tool, the social media monitor, because I really have not run into a lot of people that do good monitoring or they've picked a tool that's just brand focused and it doesn't let you do this topical monitoring.
Speaker B: Or they have a tool and they
Speaker C: don't use it, or they have no idea how to use it. Yes, that has also happened too. Or they don't have the manpower to respond. You were just talking about brand misinformation. So part of this new environment is having somebody on your handle who could respond, having some sort of alert system set up, having a protocol, which, by the way, I have on my website because I had to create this for New York State of Health. So if this is new to you, come get the free resource. But yeah, like have people trained and they know how to respond so they don't fan the flames. And there's a protocol if somebody goes off the handle on social and you know how to handle it. All of that. Most Brands don't have in place.
Speaker B: Well, they've automated them with agents for
Speaker C: that, which might not be a good plan either. Yeah.
Speaker B: And it lies half the time, right?
Speaker C: Exactly.
Speaker B: Oh boy. Um, yeah, it's problematic on a lot of levels.
Speaker C: And they don't have the budget for it. They never thought they needed it.
Speaker B: Right.
Speaker C: But you're right. If you get somebody. I've even had some of my top level brands, um, have gotten like foreign state attacks, put it that way. I've had some negative SEO going on with some of my brands where it's very clear that foreign governments are trying to change their rankings. For instance, those types of brands actually possibly need some social media monitoring and response protocol. Right. And they possibly have not. It's a whole new budget line they haven't thought about. Um, but yeah, if they want to keep their, uh, mentions clean so that the training data doesn't get polluted with misinformation, that would be the new requirement.
Speaker B: Bringing me back to my postal service days of that topic. Foreign actors attacking.
Speaker C: Yeah, yeah, yeah.
Speaker B: Before we, uh, because we did promise we'd get you out of here on time for another meeting. Um, before we do that, is there any kind of a counsel that you would offer people that are really concerned about the AI impacts on search, generally speaking, that we have not brought up yet, no.
Speaker C: But I have a guide that walks you through everything. So it's free and you should go grab it.
Speaker B: We'll put a link to that.
Speaker C: And then, so if you do, uh, bit ly. So bit ly writingforai mode, you should find it, um, and it goes through everything I've talked about, like how to write for it, how to measure it, even how to like get a starter plan going for LLM search. So yeah, feel free to grab that sounds awesome.
Speaker D: Fantastic. Um, I mean we'll check it out for sure, but we'll, as Jeff said, we're going to put a link to that in the show notes as well. It sounds like that's a really good resource for people because I think a lot of people are asking the same kinds of dumb questions we're asking you. It's so new. Right. And nobody knows what to do. Uh, they just have this sense they're being, uh, severely impacted by it, even if they're not yet. And um, you know, they're, they're asking the question.
Speaker C: So also for marketers listening, you might have missed the piece, but it's all personalized to the end user. So when your CEO starts flipping out because something is appearing for him in ChatGPT. Be aware that it's possibly only appearing for him. Right. Like you might need to demonstrate that. But just, I've had to do a lot of this in my SEO career where, because even search is personalized to you, where people are like, oh, my God, why is this thing appearing for me? I'm like, you keep.
Speaker B: Why is my brand associated with pork rinds?
Speaker C: I mean, keep clicking on it, stop clicking on it.
Speaker A: Right.
Speaker D: And the thing is, this is, this is a perennial CEO issue. I mean, I think back to my very first job back in 1990. I was a traditional media planner. And I distinctly remember I was working with Jaguar dealer co op and we would have to buy for this one particular dealer in one market, billboards along the route that he drove between his house and his dealership. So that when he got to work, he could be like, you know, Jaguar's really supporting us with lots of out of home advertising. This is great. Even though we were advertising nowhere else in the market. And, you know, and so I think the other thing, the opposite of what you just described is the false positive, where the CEO gets super excited that we're doing something. Right.
Speaker C: Yes.
Speaker D: When in fact you're not doing anything. Anything. And the, you know, the net effect of that becomes a negative, which is they deprioritize it. They don't fund the tools you need, the people you need to do the job well, because they think for their ChatGPT, they've seen the right answer, you're good. Um, and they don't.
Speaker C: Or they think it's just SEO software and you don't have to talk to your users. We are in an environment where you definitely need to talk to your users. You cannot do this blind, otherwise you're going to miss them completely. Right.
Speaker D: Yeah.
Speaker B: Which I think is.
Speaker D: That alone is great counsel. Worth the price of the podcast 100%.
Speaker B: Right. So with that, I think, Katherine, thank you so much. Where can everybody find you?
Speaker C: Oh, so I'm wstrategies.com and then otherwise you can find me. I think It's Katherine W Ong on LinkedIn. I'm also hanging out on X still. I'm, um, on Blue Sky. You're on X. I am, I am. The SEO folks are still there.
Speaker D: I haven't left yet.
Speaker C: Uh, I tried Blue sky and you just cannot get the same cutting edge news.
Speaker B: No, it's not quite the same.
Speaker C: Some of that's still on X, unfortunately.
Speaker B: I don't know. I left in 2017, so I'm just perennially lost.
Speaker C: Yeah.
Speaker B: Yeah. All right. Very good.
Speaker C: Well, thanks for having me on the show. It was great.
Speaker B: You were amazing. This has been really insightful and now I know to take the AEO Geo trend a little less seriously as result of it. Although not to ignore it. So thank you.
Speaker C: Great.
Speaker B: Yeah. Mr. Greg, you want to take us on out?
Speaker D: Absolutely. Same, uh, way we started. We are going to end, uh, and, uh, I will remind everyone again to please be sure to subscribe or follow the channel, depending on where you listen or watch, like, uh, rate review. Uh, we do like to take your suggestions, your questions, your comments and whatnot. So please hit no Brainer podcast, which you're going to do anyway because you're going to want to, uh, get the show notes for this one. Lots and lots of good links that you're going to really want to follow, especially Catherine Scott. Right, for Catherine. Uh, but while you're on that site, do hit the email icon and be sure to give us any kind of subject, uh, content, topic, ideas, questions, comments and whatnot. Uh, we do appreciate that. We appreciate you spending this, uh, hour or so with us, Catherine. Uh, thank you. Thank you very much. And we also appreciate everyone who's been listening spending their time with us as well. And with that, we are, thankfully for you, not for us, but for you. We are out of here until next time. Take care. Cheers.
Speaker C: Cheers.
Speaker A: Uh, you may know you're listening to this show along the Marketing Podcast Network, but did you know there are other great shows on MPN to help your business? Heather Ek hosts an amazing show called your Radiant Spirit. Heather, tell listeners about the show.
Speaker C: What if the colors you're drawn to, the creative urges you ignore, and the quiet, intuitive hits you brush off are actually trying to tell you something? Your Radiant Spirit is the podcast that helps you listen and live with greater clarity and purpose.
Speaker A: And where can people subscribe?
Speaker C: You can find and subscribe at heatherech. Com YourRadiantsPirit on marketingpodcast. Net or search for it wherever you get your podcasts.
Speaker A: You heard her, Go subscribe. This podcast is heard along the Marketing Podcast Network. For more great marketing Podcasts, visit marketingpodcasts. Net.
More from No Brainer
All episodes →- NB67 - The Last Rant (Final Episode)28 / 100
- NB65 - AI & the Future of MarTech with chiefmartec Scott Brinker47 / 100
- NB64 - AI As Culture, Not Just Code with the Artificiality Institute's Helen and Dave Edwards
- NB63 - Smart Resilience for the AI Age
- NB 62 - Navigating AI in Business: Integration, Leadership Challenges, and Future Perspectives