Tools Look Backwards: Radina Ivanova on Predictive SEO

The GEO Show
November 14, 2026
Watch this episode on YouTubeListen on Apple Podcasts, Spotify and more

Radina Ivanova, Organic and AI Search Lead at Hallam, explains how predictive SEO uses Reddit, Search Console impressions, and BigQuery to spot demand before keyword tools catch up. She describes prompt-style queries in Search Console, a content piece cited in ChatGPT in under 24 hours, and the KPIs she now takes to clients when traffic falls but rankings and the pipeline hold.

Key takeaways

  • Keyword tools report the past. When ChatGPT was adding millions of users a week, third-party tools still showed only a few hundred monthly searches.
  • An opportunity engine layers Reddit conversations, Search Console impressions, and BigQuery on top of normal keyword research, then checks growing clusters in Ahrefs or Semrush and with the client.
  • Search Console connected to BigQuery surfaces prompt-style queries such as "evaluate this brand versus that brand," which are more useful as category entry points than as prompts to answer one by one.
  • A content piece driven by a category entry point, not by a keyword target, was cited in ChatGPT in less than 24 hours and ranked Radina's client first for that software.
  • F5 Bot works like Google Alerts for Reddit, and SparkToro shows the subreddits an audience already uses. Fresh comments on older posts can matter more than brand-new threads.
  • When clicks fade, she reports brand search, MQLs, SQLs, closed opportunities, and how often AI appears on the conversion path, and she treats AI Overviews as the biggest AI search channel for most people.

Keyword research still matters, but it describes what people already searched for. Radina Ivanova, Organic and AI Search Lead at Hallam, argues that the lag between a trend forming and a keyword tool reporting it is now large enough to miss real demand. Her answer is predictive SEO: layer Reddit conversations, Search Console impressions, and BigQuery on top of the usual tools, then decide whether a topic is worth content before competitors see the same volume spike.

Why do keyword tools lag real demand?

They have to. Third-party tools cannot pick up last-24-hour Search Console data, and they are slow to register a niche topic. Radina's example is the launch of ChatGPT. The product was adding millions of users a week, and the tools still showed a few hundred monthly searches. She has also published a regulatory content piece on a Monday and seen queries and traffic in Search Console that were not yet in Ahrefs, because the brief came from the industry, not from a keyword list.

How does the opportunity engine actually work?

The inputs are Reddit, Search Console impressions, and BigQuery, with Ahrefs or Semrush and a client conversation used as a cross-check. BigQuery matters because it lifts the thousand-row cap in the Search Console interface, even though some queries stay anonymized. She groups queries into topics with AI, then reviews the clusters herself. The examples she showed on the recording were clusters she had not recently optimized, so impression growth was a demand signal, not a reflection of a fresh content push. She also tests topics with paid media first when she wants to know whether a theme drives leads and SQLs before a long content investment.

On Reddit she is looking for insights, not just presence. A cybersecurity client came with a service customers were already asking for. Classic keyword research looked flat. Reddit was already talking about that service. Another example, regulatory intelligence, was a four-day-old conversation with almost no volume in the tools. She uses F5 Bot as a Reddit version of Google Alerts, and SparkToro to find the subreddits an audience actually belongs to. She reads posts with the most comments, not only the newest posts, because a two-year-old thread can still be live.

"AI is really hungry for unique content and unique point of view."

She treats the engine as an addition to keyword research, not a replacement. The goal is to get to those unique topics before the tools, and before competitors, show the same volume spike.

What do prompt-style queries in Search Console look like?

Once Search Console is connected to BigQuery, and sometimes with regex in the interface, she starts to see queries that read like prompts: evaluate this brand versus that brand, how does this compare, what tools predict a given outcome, or a first-person setup such as "I am 30 years old, I am a marketer, give me a list." Volume is often low and clicks are thin. She does not recommend writing a page for every prompt. She looks for the category entry point underneath the set, the job or comparison the person is trying to do, and writes for that.

Can unique content get cited by AI in a day?

Yes, in the case she brought. She still writes for the user first: what the topic is, how it affects that person and their industry, what to do next, and how the company helps. In the week of the recording she tested a piece that was not driven by search volume. It was driven by a category entry point, a problem the brand wanted to be known for solving. In less than 24 hours it was picked up in ChatGPT, cited, and her client was ranked as the first option for that software. Her read is that AI systems are hungry for a unique point of view, which is the same information-gain idea Paris has been arguing with first-party knowledge from sales calls, support logs, and product data.

How should you report GEO when traffic is down and rankings are not?

Education first. AI search is not a like-for-like replacement for organic as a traffic channel. She still tracks referral traffic and prompt presence, and she can see AI referral leads in HubSpot or Salesforce. The numbers she leans on are brand searches, business KPIs (leads, MQLs, SQLs, closed opportunities), and conversion paths: whether organic is a first touch, a mid or late touch, or a follow-up to paid, and how often a path now moves from organic to AI and back, or from paid to AI to email.

She also treats Google AI Overviews as the biggest AI search channel for most people, because that is still where the audience is. For a stretch after passage ranking, Overview clicks carried a parameter and even a passage target, which made reporting easier. That parameter is gone. Search Console now also reports impressions, clicks, and queries for verified YouTube, Instagram, TikTok, and X surfaces, which she uses to build the case for those channels as landing pages. LinkedIn is not in that Search Console set, but LinkedIn articles already appear in the SERP and in B2B prompts, so she treats them as owned surfaces too.

Notable moments

  • 00:02:58 ChatGPT was adding millions of users a week, and keyword tools still showed a few hundred searches.
  • 00:17:34 A category-entry-point article was cited in ChatGPT in under 24 hours and ranked her client first.
  • 00:22:54 F5 Bot as a Google Alerts-style watch on brand, competitors, and subjects in Reddit.
  • 00:29:09 The hard conversation with clients: traffic is down, rankings are not, and the KPIs have to change.
  • 00:34:25 AI Overviews will be your biggest AI search channel, because that is still where most people are.

Radina is on LinkedIn if you want to continue the conversation on predictive search and category entry points. The GEO Show publishes daily conversations on how brands get visible and recommendable in AI search.

Full transcript

Paris Childress: Hi, everyone, and welcome to the GEO Show. Today my guest is Radina Ivanova. And Radina is Organic and AI Search Lead at Hallam, a UK digital marketing agency. She has spent about 11 years in SEO across B2B, travel, e-commerce, and lead generation. And at BrightonSEO this October, she showed how to wire Reddit Search Console impressions and BigQuery into a predictive SEO opportunity engine that spots demand before the keyword tools do, and I cannot wait to get into that topic. Radina, welcome to the show.

Radina Ivanova: Thank you for having me, Paris. Really excited to be here

Paris Childress: I'm dying to get into this topic, but before we do, can you give us a little bit of an intro about yourself and your background in SEO?

Radina Ivanova: Yeah, absolutely. I've been working in the industry for about 11, 12 years. I started fresh from uni as a digital marketing kind of consultant specialist, part of a bigger team. We did email marketing, SEO paid, and then I quickly wanted to specialize in SEO. So that's, that's where my journey has been in the past 11, 12 years. I mainly worked agency side, so have worked across various different industries and clients and, types of websites from big e-com websites where technical SEO really takes the front to more lead gen sites where content planning, for example, content strategy, is the priority

Paris Childress: And I do need to throw this in 'cause I learned that you are from Plovdiv in Bulgaria, it's a small world, and hopefully we'll even meet each other, at this conference, the Women in Marketing conference coming up next month here in Sofia

Radina Ivanova: Yeah, that was such an incredible thing to find out when you, when we, when we connected and you said you're based in Sofia and I'm from Bulgaria, so yeah. I look forward to seeing you there next month

Paris Childress: All right, so let's get into predictive SEO. This was your, this was your topic in Brighton. Um,

Radina Ivanova: Yeah

Paris Childress: the subplot here is that keyword research always looks backwards. So whatever you're seeing, it's

Radina Ivanova: Yeah

Paris Childress: kind of old news, basically. So can you walk us through this opportunity engine, in terms of what goes in, what comes out, how can you really spot an emerging trend and and, separate that from all the noise that's on Reddit?

Radina Ivanova: Yeah, absolutely. it's a topic I'm very passionate about. I do think, you know, it's absolutely fine for us to do keyword research with, you know, using the mainstream tools. We have to do that. We have to understand where the gaps are, where the highest search volume is. But I've always felt it's looking in the past because there is always a lag between a trend kind of appearing and starting to brew, and the tools actually picking it up and, and showing you the true value of the potential traffic you can get. And so, example of this is, I think I talked to you about it, when ChatGPT launched, and everybody was talking about it and they were... ChatGPT was gaining millions and millions of users, of users every week. I would, I would go to tools and search volume would still be, you know, a couple of hundred per month, which was just not, it was just not realistic to what was actually happening in the market. And in the past I've worked with, kind of teams where we've looked at more, we've taken more a forward approach. So looking at an industry, looking at what's coming down from a regulations point of view, looking at what, the market is doing, and really think about what people would be searching for in that environment if something changes, and we've seen really good success with that. and most most recently I've been doing a lot of work with Reddit, just mainly monitoring. There are a lot of people doing amazing work with actually presence on Reddit. What I focus more on is actually getting insights from Reddit and what the community's talking about. And to be honest, every time I've struck gold, it's, it's just been amazing because these niche topics that you find in Reddit are actually great content opportunity pieces

Paris Childress: Yeah, that's fascinating. so you've built a system where you can pull in, you can pull in Reddit data, Google Search Console impressions data, and then the...

Radina Ivanova: Yeah

Paris Childress: bring that into Google BigQuery, if I'm not mistaken. Is that right?

Radina Ivanova: Yeah, yeah. It, it, the... Yeah, yeah. So at the moment, the process is kind of still separate, but yeah, the system is on the way to bring this all together. But the idea is that, well, we should all be looking at BigQuery data anyway because it's incredibly important, and it removes some of the limitations we have with Search Console. Naturally, we'll still have queries that are anonymized. Google is not gonna show us everything, but it's gonna show you a lot more than what's in the thousand rows of data in Search Console that you see by default. And so, actually, I'll show you a few slides on what I mean when I, when I talk about kind of wiring Search Console with BigQuery. So What happens with BigQuery is that once you start, once you start, kind of understanding the data on a monthly basis, you can start building your reporting. You can start understanding, okay, this this keyword here, shown very little impressions in March, but actually I can see the topic growing. And I wanna caveat here that, of course, we'll see topics grow if you recreate new content on these topics or if we optimize existing pages, but the examples I've given here are actually specific to a cluster of content and pages that we have not optimized. We, we have just left for them to be because they have not been a focus. But we can now see, I can now see when I look at three to four months worth of data, how impressions, kind of change for a certain topic. So this is one for impressions for business market intelligence.

Paris Childress: here? And what is

Radina Ivanova: Of course-

Paris Childress: for clustering the data in Search Console into these topics? how do you do that? you're just getting a whole lot of, query data

Radina Ivanova: Yeah, we-- I use AI, definitely. This is not something... And I do not recommend anyone to go and manually audit every piece of keyword they see in BigQuery, especially if you're dealing with a big site. You would want AI to help you with analyzing, summarizing this data. Of course, review it, of course, categorize it maybe yourself in the different topics, whether that's, you know, based on product, if you're running an e-com site, or based on industry, if you're working with lead gen. but definitely use AI. And actually, what I wanted to also show you is th-these-- essentially, these slides are just showing you that for a particular keyword and topic, we are seeing an increase in impressions without us really doing anything. What I do with that data is then I cross-reference it with Ahrefs or Semrush. I speak to the client. I understand any analytics data to see whether any of these pages are receiving more traffic. and then we make a decision from then on. What you can see here is what this data represents. So the examples I gave you are very specific, but actually, when you plot them on a graph for a longer period of time for, you know, thousands of keywords, you can understand easily, okay, which are the keywords that actually are growing in the past three, four months, and which are the keywords that are just starting to form a trend. And so actually here what you see is the bigger circles. The, um-- It means that kind of-- it represents the total number of impressions, and the color changes means how much they've changed from the beginning of the period I'm looking at to the end. And you can kind of understand what people are searching for

Paris Childress: Yeah, this is fascinating. so many questions come out of this. First is, so you're, you're, you're cross-referencing the Search Console data, the queries with, let's say Ahrefs keyword research tool data. Do you see a consistency there, or do you... How does, how does Ahrefs actually perform you put it up against the actual Search Console data?

Radina Ivanova: there is a lag. There, there is definitely a lag. I, am a big fan of seasonality, which is another topic we'll cover later today, but there is- there will also be a lag in seasonality. it's natural for there to be a lag. I don't, I, you know, I don't think that, especially with Ahrefs and BigQuery, we now have data from the past 24 hours. It's impossible for these tools to pick it up, to pick up demand so quickly, but you get the data. I was creating a... I was working with a client on a content piece actually this week about regulatory changes coming into the industry. We published a content piece, I think on Monday, and I can already see traffic forming, in Search Console and queries coming in that were not in Ahrefs because actually we were driven by what the industry is doing and what demand might be as opposed to what's currently there

Paris Childress: Have you ever also pulled in, keyword research data from Google Ads Keyword Planner?

Radina Ivanova: yes. Yes. So I do use Keyword Planner quite a lot. but I find, actually it groups a lot of keywords, and so the granularity that you get with Search Console, Reddit or, Ahrefs or, you know, the key- or the keywords that have less search volume but actually are very high intent, I find them not to be there. But what I do, I'm a big proponent of, and I work with the paid media teams, is understanding what terms they're driving, performance on. And so if, for example, for organic, we want to push certain topic, we can test it with paid, the paid media first, and we can understand actually is this topic driving leads and also SQLs before we invest a lot of time in creating content. But unified search. Yeah, unified search is something we work with, at Hallam almost on a daily basis. So we communicate to understand where can we pull paid media for organic to win, and actually what's paid media doing really well that organic can also catch up on.

Paris Childress: absolutely. And I think that's gonna be an interesting relationship as we move into the AI ads era as well, because we're not going to really have a keyword research tool for prompt tracking. I don't think that's gonna, exist. but there is gonna still be a very close relationship, between organic, let's say organic or GEO and paid. I don't know if it's gonna be called paid GEO or just paid... it's m- mainly, ChatGPT ads now Yeah

Radina Ivanova: Yeah, I mean, I'm starting to see ChatGPT ads when I, when I do testing and wh- when I do search. I find them still not very native, if I have to be honest. I find them a bit out of place. They kind of appear sometimes out of context. So I think we're still very early on in the journey, but we're for sure unifying a lot of the channels. They're just becoming more branding channels, in my opinion. We're moving a little bit away from organic beco- being in the performance channel into more of a brand channel

Paris Childress: And I-- Yeah, I've been saying this a lot too, that it's, it's a little bit unfortunate that GEO has such a close association with SEO the natural, the natural instinct is that, all right, this is, this is, an evolution of SEO and it's performance marketing. another algorithm that we're trying to optimize for. But it is much bigger than, th-than an optimization, algorithmic endeavor. it is, it is a lot about branding and positioning your brand in the-- in AI's mental model

Radina Ivanova: 100%, I agree. I just agree wholeheartedly. W- I am-- I actually, I'm very excited about it because it means that I think SEOs are moving away from purely optimization to more marketing in general. I mean, I'm encouraging people to read more general marketing books, for example. like How Brands Grow is a big It's-- I think it's having a resurgence now by Byron Sharp. But a lot of people are reading it, and it's amazing, and it talks about category entry points. and, you know, I'm working with our head of brand at Hallam to actually develop that a little bit further to understand what are the category entry points for companies and how are they communicated across their, I suppose, content and communication and website and social media presence

Paris Childress: one other quick question I was gonna ask you about the inputs to this data and this real-time approach. Another tool comes to mind, which is Google Trends, and I think that is based

Radina Ivanova: Yeah

Paris Childress: It is connected with a keyword planner in some way, but have you used that? Is that helpful?

Radina Ivanova: Absolutely, yeah. Yeah, I use it. I cross-reference it. I think it's really helpful for bigger topics. I think anything direct to consumer and e-com is very valuable. I do get the updates in my email, which again, is very useful. I think it-- th-this, I guess when I say predictive SEO, we can't, of course, predict the future. We can just pick up the signals that we are seeing from different platforms to understand whether it's worth creating that piece of content. but it's not about using one tool, it's about using everything at your disposal to actually say, "Okay, we want to invest in this, in this content." Google BigQuery is one of those, places because you can connect your search console data, you can get a lot more insights. I'm very also eager and excited to talk to you about that. In Search Console, you can... If you use Regex or you connect it to BigQuery, you can start seeing more prompt-type keywords. I don't-- Have you-- Is that something you've worked with?

Paris Childress: we have, yeah, we've done a little bit of that with one client where we tried to spot some patterns. We were trying to identify the query fan out queries and to

Radina Ivanova: Yeah

Paris Childress: or to segment them apart from all the other queries because Console doesn't yet... I hope that they will do this in the future, but they don't yet break out the specific query fan out queries the way that Bing Webmaster Tools does, which is really nice. so we used Regex in that example to try to parse out... It really was, trying to parse out the longer... I mean, these query fan outs are very long, so they don't, they don't look at all like a real human search query. And so we were look- we used it for that basically. But can you tell us more about how you

Radina Ivanova: Yeah

Paris Childress: you, how you use it?

Radina Ivanova: Yeah, yeah, of course. What, what you said is absolutely right. Bing have kind of outdone Google on that front, but, I think Google will follow suit soon. it's more the fact that I'm noticing in the past few months when you really look into the data, and as you say, we're not gonna have a keyword research tools for prompts. They're just so unique, the experience is so unique, it's... We're just not going to have the same level of data we ever had with SEO and keywords and search volume and competitiveness. and so what I'm looking for in search console is actually topics with fewer impressions, probably not a lot of clicks. And so when you filter the data that way, you start seeing queries in the form of evaluate this brand versus this brand, or evaluate my brand versus this brand, how does this compare? Or, what tools are you using to predict X, Y, Z? Or actually very self-explanatory prompts in the sense of, I am, you know, 30 years old, I'm a marketer, this is the-- these are the tools I'm looking for, limit, you know, give me, give me a list, for example, of, of the top tools in that industry. So you're starting to see that across a set of keywords. And it's about actually combining them and understanding is there an underlying category entry point you can create content around as opposed to actually answering every prompt, which is not what I recommend.

Paris Childress: And then once you identify a trending topic, decide that we're gonna either... Let's say there's gonna be a new, a new piece because this is a new, a new topic, so we're not gonna do an update or a refresh the existing. How do you think about content formats? Have you changed, the way that you approach, formatting if you really wanna try to get this content picked up and cited by AI?

Radina Ivanova: yes. Yes and no. I haven't changed in the fact that I always put the user in the front, so the customer and the audience. So I just try and envision if I were in this person's shoes reading this piece of content, what would I want to know? What's the order of the information I wanna receive? It's probably something like very, you know, explaining what the topic is, how would that impact me and the industry I'm in? What do I do next, and how does the company I'm actually reading the blog from is going to help me resolve that issue? So it's very much user-centric and driven, and of course, if there are any keywords I can include, I will include those. but putting the, the center, the, the user at the center and the front. And actually, I have an example from this week where I've tested content. Less than 24 hours it got picked up in AI search in ChatGPT and actually ranked my client as the first option for that particular software.

Paris Childress: Oh, great

Radina Ivanova: I think AI is... And that was a content piece that was not driven by search. It was driven by a category entry point. It was driven by a problem that we just wanted to solve, and we wanted to be the, the brand that talks about the solution. Within less than 24 hours it was picked up, it was cited, and it was in the answer, which then shows to me that AI is really hungry for unique content and unique point of view

Paris Childress: absolutely. This was, one of the main themes that I talked about at Brighton was the concept of information gain, used the ana- the

Radina Ivanova: Yeah

Paris Childress: truffles, that this is, this is the stuff that is often buried a little bit deeper, and it's brand proprietary knowledge and, data, it takes a little bit of, a little bit more effort. It's, it's typically not stuff that's being published on the company's website or anywhere on the web. So AI hasn't seen it, ingested and crawled it, and it just represents net new learning. And the,

Radina Ivanova: Absolutely.

Paris Childress: make the effort to dig out these truffles, so to speak, that could be, sending sales call recordings, customer support logs, digging out the the product data and getting some original research ideas out of that. They're going to be rewarded with visibility because AI is really hungry. AI is the truffle hunter, so to speak, looking for information gain and looking to expand its knowledge of the world. Um-

Radina Ivanova: I c- yeah, I couldn't agree more with you, and that's exactly what the category entry point research is. You know, just actually speaking to your customers, looking at those transcripts, understanding their language and what they might not be searching for. They might just be telling you or the sales representative, or reviews. I'm a big proponent of, channeling your inner Rory Sutherland where you can and actually focus on why don't you do research on where your competitors are bad at and focus on resolving that issue? 'Cause that's what AI is actually also really good at. You can summarize, and you can understand the the weaknesses of your competitors and communicate that in your either content or social media messaging or LinkedIn posts.

Paris Childress: it's as easy as just asking AI, "What do you see

Radina Ivanova: Yeah

Paris Childress: strengths and where can I

Radina Ivanova: Yeah

Paris Childress: these competitors? We- well, what are the, what are the

Radina Ivanova: Yeah

Paris Childress: weaknesses of these competitors and help me, communicate how to, how to exploit that?" Mm-hmm.

Radina Ivanova: Exactly. Yeah. Yeah

Paris Childress: and then so getting back to Reddit, so you've got...

Radina Ivanova: Yeah. I

Paris Childress: you're using Reddit also to spot the trends, and do you incorporate any of that conversation into the content itself when you're writing it?

Radina Ivanova: yes, it's, it's more informing the topics as opposed to a little paragraph in one of the topics, for example. But yeah, I absolutely love Reddit for this. It's, it's proven useful on so many occasions. I have an example. I have so many examples from one recently from a cybersecurity client, is that they were, They came to us and they said, "Oh, this is what our customers are talking about. This is what they're coming to us for. Can we do some keyword research on it?" And we did, and actually, you know, it was kind of plateaued, like nothing really, we didn't really see a spike. But actually going to Reddit, we can see very recently that people are exactly talking about this service. That's what they're going and saying, "Oh, I'm looking for this service, and I'm looking..." So it's your industry talking about either what they want to get out of a product or their frustrations that you can resolve with your product or service. and so the data just aligns. I also have, examples of, a lot of, Actually, I'll, I have a few ex- I have a few examples on some keywords where, if you, if you go to Reddit and actually look at them, and then compare them to, ... You go to Ahrefs and you understand, okay, is there actually search volume for these topics? And there isn't, but you can see that on Reddit people are talking about them. I mean, exploring regulatory intelligence, that, for example, was four days old when I, when I recently took this screenshot. It's not enough, it's not enough time for, you know, something so niche to make their way into Ahrefs or Semrush. And actually when you look at those topics in those tools, they show very little, very little search volume. But this is where potential content opportunities hide that you can, you can go and explore before your competitors do.

Paris Childress: there'll be almost

Radina Ivanova: It's about ... Exactly.

Paris Childress: They're just looking

Radina Ivanova: Exactly. So yeah.

Paris Childress: And I think Ahrefs

Radina Ivanova: Yeah.

Paris Childress: evergreen, but I mean, it's just not, it's not the right tool if you wanna, you wanna tap into real-time trending topics

Radina Ivanova: Yeah, absolutely. And what I would recommend is treat Reddit as a Google Alert, kind of bot. So I use F5 Bot, which I absolutely love, and it's something like Google Alerts but for Reddit. So after you've gone and s- researched your subreddit for your own industry, you don't have to go and scrape it again, for example, do the same research, but you can stay on top of the topics or your brand or your competition. So what I do is I track, my brands, I track the competition, I track the main subjects that I care about, and I can see it live, in real time when people are picking up conversations. And then I can decide, do I participate in the conversation as part of the brand, or do I actually see is there a trend coming up, and can I form a content piece around that?

Paris Childress: This is fascinating. I'm, I'm thinking that, I wanna check out this tool 'cause that, it was gonna be my next question, which is that are you going straight into Reddit and just searching in Reddit? 'Cause that seems like it wouldn't be good enough. But this is the tool. It's called F5 Bot. Interesting

Radina Ivanova: Yeah. Yeah. I w- I w- I do sometimes search in Reddit if it's something super quick, but not, of course, that's not scalable. we work with a co- a tool called, SparkToro, which is amazing. you've probably heard of it, yeah. It, it gives you a lot of audience data. It also gives you what your audience is interested on s- on Reddit, so you can see the subreddits that they are participating in or part of, or part of the community. once you know that or, you know, you would know which subreddits are important for your industry, you can scrape the data. I would suggest not just looking at the top, the most recent posts. I would suggest looking at the posts with the most comments because a topic could be two years old, but if the comments are fresh, it means that it's still live, in people's hearts. And then understand, okay, actually d- d- cross-reference that with search volume in Ahrefs or Semrush or other tools, r- run it by your marketing team. Okay, this is, this is what people are talking about. I think part of, in my session at Brighton, the gentleman before me was talking about how we as SEOs undersell our, ourselves, and I really love that because we have so much data about the market. It's market intelligence data of what people are talking about online for a product or a service, and we can, we need to use that to communicate to marketing and sales teams. This is how people are buying your product and are talking about the product and the industry, and we need to be part of that conversation

Paris Childress: Yeah. And once you... So you're leveraging these insights to jump on topics well ahead of competitors that are, that are just doing classic keyword research. publishing. Do you then... Is there any way to connect the content that you've just created back into those Reddit threads, or do you have any f- actual forum participation in those threads after you have the content? I know there's a fine line to walk, and you shouldn't, you shouldn't be going in and salesy at all in a Reddit thread. But now you've actually just rounded up, know, you've put your finger on the pulse of a conversation in Reddit, maybe across several threads, and you've, you've even supported that with keyword research and with outside data and research. So now you have a really, a phenomenal asset to bring into those conversations. Do you, do you go in then and do that kind of UGC participation the content?

Radina Ivanova: I don't at the moment, I have to be honest. I am... I ca- I see other brands doing it and it, and it's so, and it's not natural. It, it looks shoehorned. It looks, "Look at my content that I've just created." I think people go to Reddit sometimes because they can't find the answer online, so this is what I try and focus on. What I try and do with Reddit on brands is actually invite them to be part of the conversation in a very informational way. So when we wanna contribute to a topic, we do it from a very pure learning and development informational, supportive role where we just talk about, "This is our, this is our perspective and this is what we think you should do," as opposed to, "We've written this content piece based on what you were searching for." Because I just don't think it works well with the, with the Reddit audience. I think they sniff it out immediately, if you're trying to sell something to them. So I think if brands are on there, they have to be helpful and they have to be part of the community as opposed to trying to sell you something immediately

Paris Childress: Definitely. let's, let's pivot over to another topic about just measurement now, the measurement challenges that we're facing in GEO that have made things harder compared to SEO. It, it comes down to the click I guess, sunsetted in a way. Clicks, clicks are fading, we're, we're moving into the zero-click world. traffic rankings may be still fine, and Search Console can show rankings, but the other stuff is becoming less trustworthy. Even your search impression data, I often see big spikes search impressions, and that is, that's all AI. And, and I do wonder, yeah, how many, how many of these search impressions now are actually coming from, just a, retrieval query fan-out activity versus real, humans pulling up SERPs. given all of this now, how are, how are you navigating this new conversation with clients that have been conditioned to be looking at SEO reports for years and looking at traffic and how is my traffic trending? and they may see that now their traffic is going down despite the fact that rankings are solid, maybe rankings are even up, and even their sales pipeline is doing well. I mean, the business is still getting good leads. how do you parse out and tell that story now? How do you shift their focus to other KPIs that matter?

Radina Ivanova: It's a difficult pill to swallow. I find that education is key, so we need to be educating our clients and audience that things are just changing, and they're not changing for like for like. They're just changing for a completely different channel with a completely different, kind of a performance indicator. AI search, I don't think will ever be the traffic-driven channel that organic was. It's just not how people use it. It's not... Before, people would go through ten articles and ten pages to find out which one they want to interact with and which company to contact. It's just not how AI searches works. I use it, I use it myself. It's not how I work. when it comes to B2B and attribution, of course, we have the soft metrics for referral traffic and, you know, we can track a few prompts and understand whether we come up there, and we can see, referral leads in, in HubSpot or Salesforce because they'll pick up the referral AI traffic. I think what we need to do, at least what I'm doing, is relying a lot more on, brand awareness. So I'm monitoring brand searches. Google Keyword Plan is really good for that. impressions, yes, it's, I... Very sadly, Google Search Console is also becoming not the most reliable data source because of the impression spikes we're seeing because Google is also fighting AI scraping, so that data is no longer the purest that we used to think it was. But if I had to focus would be, just business KPIs, going back to leads, MQLs, SQLs, closed opportunities, and how organic and AI search are influencing this, of course, directly and of course then indirectly. So conversion paths is something I'm a big fan of. I always look at them, how is organic, you know, what type of role organic plays when it comes to the conversions. Is it a follow-up channel from paid awareness? Is, is it just kind of mid and late touch points? Or actually, is it the channel that brings up leads on its own, within its own right? Is it a first touch, channel? I am starting to see actually AI play into this, so I can see that people are moving from organic to AI back to organic or paid to AI back to email, for example. So that's, that's become also my... I just look at it almost on a daily basis now.

Paris Childress: I love that, analysis on the conversion paths. One of the things that also confuses this is that even organic search now reported in GA4 is very often AI overviews. So maybe people have an AI overview and then clicked into a citation in the citation panel of the AI overview, or maybe they've read an, the AI overview and gotten influenced in a way or gotten motivated in a way to even scroll down and click on a, the first

Radina Ivanova: Nie

Paris Childress: If it's a branded search now, AI overviews now comes up for branded searches, and that means that your, you know, third party-- what third parties are saying about you now are even higher up on the screen than your own web-- than your own homepage, which is probably ranked number one organically, but that's now further down. I guess it's another reason why I think every brand should be running Google Ads on their, on their brand name just to make sure that they're just claiming, above the fold screen real estate on these SERPs. but I think it's hard to parse out even organic search now because of the influence of AI overviews, which is now statistically, I think around overall like something like eighty-eight percent of all searches now. It, it started with mostly top of funnel, and that's when it really just started to take away clicks on informational queries, educational type of, pages,

Radina Ivanova: Yes

Paris Childress: it started to move down and down and now it's it's definitely appearing for, product searches, brand searches, high intent, B2B searches. So it's, it's there. So I think that it-- that really muddies the attribution picture. But your approach is interesting that when you look at a conversion path, you can see some start with organic and then convert right away, so it's both first touch and last touch. may start with organic and then go to AI and then come back to organic.

Radina Ivanova: Yeah

Paris Childress: but I think what's important report to a client here is just to look at the, percentage of times that, let's say, AI was a part of the conversion path overall. We, we don't have to necessarily into more detail than that. But if we had, I mean, if we had a hundred conversion paths, and we saw AI touch points in thirty percent of those, that already is meaningful show and to try to defend the ROI behind it. then compounding with the point that I made about organic search, that a lot of that organic search is now the first thing that they're all seeing is an AI overview. So that is that AI actually? I mean, is that, you know, how do we

Radina Ivanova: Yeah. I would say, I would say, yeah, when it comes to AI search, Google is still the biggest player, so I actually- AI Overviews will be your biggest AI search channel because that's where 90% of people still are. I, yeah, absolutely agree with you. Looking at, looking at how many times AI was part of the, was part of the journey. back when AI Overviews started, actually, we could track them, which it was so amazing to see traffic from them because they always ended up on the parameter. So Google would send, the traffic from AI Overviews. Is, is- back when they introduced passage ranking, so they could they would send that traffic to a passage so we could track that, and for about six, eight months, we could track AI traffic from AI Overviews. It was amazing. We could see

Paris Childress: when they called it Search Generative Experience?

Radina Ivanova: also afterwards. They actually only dropped it maybe earlier this year, they dropped the parameter, but up until then we could see the que- we could see also the everything after the hashtag, so we can see which paragraph they send them to. It was really good data. No longer exists, but at the time it was really good to report on that.

Paris Childress: and it would include even

Radina Ivanova: Yeah.

Paris Childress: parameter as well? The

Radina Ivanova: No, it included the passage ranking parameter, so we could kind of then gauge what they might have searched for based on that because they would be sent to a particular passage ranking on that page. Yeah

Paris Childress: another thing that I think is a, is a huge opportunity when it comes to AI Overviews is YouTube videos as well. So ChatGPT we talked about, and I think ChatGPT, even though it was recently reported that it's been dialed back a bit on ChatGPT, I think it still is pretty prominent. I have generally been associating ChatGPT, a lot with... I'm sorry. I've been associating, Reddit citations a lot with ChatGPT, and then Google AI Overviews top citation is YouTube. and I'm thinking, or, maybe a question for you is do you, do you also think about repurposing that content for YouTube videos? only because YouTube itself is a great, it's a great channel by itself. YouTube search is the number two search engine, now you have the added bonus that YouTube's, YouTube videos are getting about twenty-five percent of the citation share in AI Overviews. So do you have any tips or any workflows into, repurposing like the real-time, that real-time trending content that you're, you're grabbing and you're putting your finger on the pulse and then getting that as a YouTube video out as well?

Radina Ivanova: Yeah. So, this is why I think AI s- this is why AI search is more than SEO and becomes a brand marketing channel because you're talking about YouTube and Instagram and Reddit and and TikTok and all of the other channels actually appear in ChatGPT or Perplexity or AI Overviews. I'm a big fan of repurposing content, whether that's for LinkedIn or YouTube. The, the wall I'm coming up against is that I think some brands are yet to realize that they have to have that presence and a spokesperson and someone who wants to be in front of the camera and actually, and actually drive that message. so the answer is not yet. That is on the cards, but it's not fully... You know, it's something that we're trying to communicate to clients of how important it is and how much actually of a marketing spend and budget needs to then be repurposed to YouTube because of these reasons. we're now able to see, clicks, impressions, keywords in Search Console from YouTube, from Instagram, from TikTok, which then is becoming more of a... I suppose it's easier to build a business case to a, to a, to a client to say, "Well, look how much YouTube is driving actually from Google, how much Instagram is driving from Google search to your brand.

Paris Childress: Yeah,

Radina Ivanova: This is why we need to invest in AI."

Paris Childress: that Search Console recently introduced these other social media

Radina Ivanova: Yeah.

Paris Childress: and,

Radina Ivanova: Yeah, very interesting

Paris Childress: is interesting. I have a question on that though, 'cause when YouTube impressions are reported, that could be, that could be in the main SERP, that could be in the video, in the video tab of the SERP. But could that also be the, the YouTube videos that are appearing in citations and AI overviews? Is that included?

Radina Ivanova: I think so. Yes. Yes, we see that. I see that with Instagram. I think at the moment the data in Search Console is inclusive of AI Overviews and AI Mode because it's from Google. I think that's where I can see those prompts forming, I think is people either going to the search bar typing a prompt but actually ending up in AI Mode. and so I think it's the same with Instagram. We're seeing it across some seasonal, posts around Christmas, and when you test it actually it's the AI Overview and AI Mode that comes up. Um,

Paris Childress: That,

Radina Ivanova: so yeah

Paris Childress: for, I think, for reporting because if that-- I think that's a great way... That, that's a, that's another great layer for your reporting if you're trying to urge a client

Radina Ivanova: Absolutely.

Paris Childress: more videos and to invest more in video. You can obviously show them the performance of their YouTube channel and the videos in YouTube, but now you have all the search impressions. So they're showing search impressions and clicks too into those videos, right?

Radina Ivanova: And to- and keywords, yeah. So for whatever term that video came up, or post on Instagram reel or YouTube or TikTok reel. Yeah, it's, it's very interesting. Or, or X as well. So these are the four platforms Google has verified, and this is why we need to move to actually not just reporting on Google, not just reporting on AI, reporting on everything on every platform you have to be on. So if you're B2B, you'd be very key for you. If you're B2C direct to consumer, maybe, you know, Instagram, YouTube as well, but maybe Instagram Meta will be more important. And then you have to report on the whole, on everything you have because the journeys are just becoming so, mixed

Paris Childress: Unfortunately, the LinkedIn is not there yet in Search Console. It

Radina Ivanova: No

Paris Childress: will be because that's Microsoft, and they're not gonna share the data with Google, unfortunately. that would, that would be great. And I completely agree about the the more holistic nature of reporting now, 'cause yeah, the users are moving across so many touch points. And

Radina Ivanova: Yeah. Yeah, LinkedIn is actually something I work on for across B2B in terms of LinkedIn articles and thought leadership pieces. and we, and we're testing it, and we're seeing it come up in the SERP as well. So again, the journey is, it's not just straightforward from, you know, from Google to the website. It's from Google to LinkedIn, or from Google to YouTube, or from Google to Instagram. It's just... And then that traffic never shows actually in your website, but actually it's gone to another social media platform that you're on. And so that needs to be as optimized as your website is because it's a landing- it could be, your LinkedIn post or your LinkedIn article, it now becomes a landing page in Google's eyes

Paris Childress: should treat all these channels as owned media, in fact. The, most of the,

Radina Ivanova: Yeah

Paris Childress: the owned media term usually references your website, but it's, it really should be your your website and all of your social media properties that... Because it is owned. You, you fully control everything there. And yeah, a LinkedIn

Radina Ivanova: Yeah

Paris Childress: be just as effective or more effective as a landing page your own website.

Radina Ivanova: Yeah

Paris Childress: And well, LinkedIn articles also, LinkedIn articles and posts are very popular for

Radina Ivanova: Yeah

Paris Childress: citations in B2B, for B2B prompts. You know, we're seeing that. And I repurpose all of these episodes of this podcast, they get repurposed into a LinkedIn article, and then they go out through a newsletter. I mean, the newsletter is also great too, because you get access to people's inbox that way, and they get,

Radina Ivanova: Ja, absolut.

Paris Childress: they get an email update that there's a new edition of, for the newsletter.

Radina Ivanova: Yeah

Paris Childress: Yeah

Radina Ivanova: Yeah, there you go. The, the audience is not just on one channel. They could be on multiple channels at the same time

Paris Childress: Yeah. then back to the video thing, it is difficult to get people to get in front of the camera, especially the leaders in a, in a company. But you also have opportunities to do faceless video, and the tools around that are getting better. Claude's new motion artifact type is really interesting, we've been playing around with, Higgsfield, and some other video tools. Now, the faceless videos are not as good as having, an expert, not an avatar, but the real person getting in front of the camera, but that's, that's harder to do. But think that I mean, is it still worth it if you can just use, let's say, Claude Motion to generate a nice video with some nice graphics that covers the topic, but you don't have any, human actors in the video?

Radina Ivanova: I, I think you can test to see whether it will go down well with your audience. different audience will prefer different things. I find the videos that have... You know, some are very obviously AI generated. The voice is just glitchy or too fast and the graphics are not great. I think if you have an in-house, you know, person who is really good at video, maybe they can make it a bit better. but I always go back to the audience. Would they, would they, would this be something they connect with? Because if, because if it's not, then can you... how can you make it better so they connect with it? I think there is so much AI slop. There is so much AI content. There's so many faceless brands now because it's so easy to just create a tool and start selling, that I think the human touch and the element and actually the face of the brand is becoming more important, and that will be more of a differentiator factor than anything else. Because people would wanna connect with someone who is real and someone who brings something else to the table, as opposed to something that's very obviously AI generated. I think potentially for maybe, a tool, if you're walking through a tool, if, for example, you just want to show how the tool works, then that type of video would work well. If you wanna connect with your audience on a more,

Paris Childress: Yeah. Or

Radina Ivanova: level, probably not

Paris Childress: really needs really needs a person and a face.

Radina Ivanova: A leader.

Paris Childress: Yeah, product videos, demos, walkthroughs, stuff like that, that is fine to go with faceless. But how

Radina Ivanova: Yeah. Yeah

Paris Childress: really exist? At some point, you, what you really wanna scale out is the thought leadership. Yeah.

Radina Ivanova: Yeah

Paris Childress: actually I think what we're doing right now, I mean, podcasting is such a natural way to do it because it's, at some point you're recording long form video, and then from that you have opportunities to repurpose it into a lot of different short form. And then you do have the thought leadership

Radina Ivanova: Yeah

Paris Childress: comes through it. And, for me it's, that, that's been the easiest way. It's so much easier for me as a CEO to just commit to doing a podcast on a regular basis, getting guests on like you and going on other shows rather than thinking about, all right, when I get in front of the camera with a script and it feels a little bit more, it just feels a

Radina Ivanova: Yeah

Paris Childress: way and it's more

Radina Ivanova: Great

Paris Childress: I really don't do it. And I think a lot of, um,

Radina Ivanova: Yeah

Paris Childress: same probably with a lot of other, C- busy CEOs and C levels that really do need to get their thought leadership out there. Um,

Radina Ivanova: It's a lot more natural, yes, I agree. Less pressure

Paris Childress: Yeah. Another thing actually, I mean, in terms of just tapping into thought leadership and to connect this back to, information gain and a brand's proprietary knowledge, a lot of that comes out of meetings and, that these could be sales calls

Radina Ivanova: Yeah.

Paris Childress: where the prospects are

Radina Ivanova: Yeah

Paris Childress: questions, and these questions aren't in your FAQ pages or your FAQ section, but you wanna capture them and the, and the salespeople are giving great answers or even internal meetings and strategy meetings. I know companies don't wanna give that up, and it's considered way too sensitive. but I think that the companies and brands that are willing to leverage their meetings 'cause they mostly now all are bringing their note takers into these meetings, so they are getting recorded, but they're not getting, they're not really getting leverage for marketing. And I think there's a big opportunity there. And it may not be for

Radina Ivanova: I agree

Paris Childress: at least you would still get by capturing the audio and getting the transcripts, you still can get the essence of the thought leadership

Radina Ivanova: Absolutely. It's where the real value is. And yeah, they need to look closer to home, then combine this with search analytics data, and then, and then produce content pieces that are repurposed for different, for the audience, for the different platforms the audience is on really. Yeah. I find, there is a, there is a trend that I feel it's forming for the past few months. I find that, every time I look into SparkToro or website, I see that less... I see fewer people being interested in mainstream social media, and I see a lot of people spending time on Medium and Substack, and more where thought leaders are, maybe less AI slop. but it's more where the conversation is and where long form potentially may be coming back as well, where people just want to, to sit and read and follow people with kind of a unique point of view.

Paris Childress: Yeah. Yeah, just to try to

Radina Ivanova: Yeah.

Paris Childress: out and block all the noise. Yeah, the short form is

Radina Ivanova: Yeah

Paris Childress: And yeah, I do think there's a lot of value in honing in on a f- a smaller number of influencers who are committing to long form content and going deeper,

Radina Ivanova: Yeah

Paris Childress: with it. Yeah, Substack is doing really well and Medium. Yeah, these are al- also great opportunities. I think overall it's

Radina Ivanova: They also picked up an AI search, so they should be part of a total search strategy if your audience is actually spending time on these platforms.

Paris Childress: Yeah. It, it, I think it's, it's really about maximizing the entire surface area that's out there. And

Radina Ivanova: Yeah

Paris Childress: yeah, SparkToro is great at showing you some of these niche, more niche, non-mainstream social media where your audience is spending

Radina Ivanova: Yeah.

Paris Childress: and it doesn't, doesn't need to be so scientific. If, if you even have a hunch that, you know, some of your ICP audience is gonna be on Substack, then get on Substack and at least start to repurpose your content there and have a presence. Yeah, just put your, put your flag in the ground in as many places as possible and then try to keep feeding that and nurturing it

Radina Ivanova: Absolutely. Yeah

Paris Childress: Wonderful. Thanks for the, for the great conversation. I can't wait to get my hands on F5 Bot, by the way, but that'll be,

Radina Ivanova: Yes

Paris Childress: So where can our audience find you and connect with you online?

Radina Ivanova: Thank you very much. It's been so great chatting to you. I'm on LinkedIn, so if anyone wants to reach out, talk about predictive search, industry news coming up, category entry points, I think that's one of my new favorite things to discuss and talk about. yeah, please free- feel free to reach out.

Paris Childress: Thanks, Radina, and ha- have a great weekend ahead. And, looking forward to meeting you in person

Radina Ivanova: Thank you.

Paris Childress: in Sofia maybe

Radina Ivanova: Yeah, absolutely. Next month. Yes. Thank you very much for having me

See how AI answers describe your brand

GEOforge measures how often ChatGPT, Google AI Overviews and AI Mode cite you, then builds the content and citations that change it.