The Great Organic Traffic Decoupling & What to do about it - #SEOFOMO TL;DR, June 23, 2025

The latest in SEO and AI search with today's SEO FOMO to London read, going through the top news from the last edition of the SEO FOMO newsletter. Joining me today, I have Jono Alderson, very well-known and experienced technical SEO consultant to go with me through those top stories.

Hello, I'm great. Thank you so much for having me. It is a treat to be here.

Likewise, thank you very much for coming over. And before starting, I want to thank SimilarWeb, who's our today's sponsor. And I want to dedicate today's edition also to the memory of Jim Wallen, a pioneer in the SEO industry who sadly passed in the last few days. Thank you for everything that you did, Jim. This goes for you. And to start with the first story, the coverage from Barry Schwartz about the Great Decoupling explained by a Googler, Martin Splitt, who was at the Google Search Central last week. Where he explained the trend going on due to the AI overviews that I guess has been even further accelerated in many scenarios with the AI mode, and we can expect is going to move further in the future. Decoupling the traffic, the clicks that we see in the Google Search Console from the impressions of the site. Many more impressions, fewer clicks. Due to the type of conversational engagement and interaction that users have. The term, by the way, the Great Decoupling was coined by Darwin Santos. And I think that it's very descriptive of what we can see. It's interesting to say though, Jonah, I don't know if you have seen the same with your clients depending on the industry where they are. I work a lot with e-commerce and marketplaces and SaaS. So very transactional-oriented websites. And I haven't yet seen a trend as extreme as the one Martin is showing in his deck. I am seeing those much more on informational sections of the websites that I work on. Thankfully, because they are very commercially oriented and most of the activity doesn't happen in there, the impact is not quite as high, but I understand that it's very, very different. with publishers, publications, and those players that rely a lot on informational content, especially if their business model is completely reliant on that.

Yeah, very much the same. And I think this decoupling is such a great— it's definitely the new normal. And I'm seeing across a range of different sites. I thought what was really interesting is obviously this is very scary and it's very new. And if it is the new normal, that raises a lot of really big questions around the whole economic model of the web, right? And one of the things that Martin said, and has been echoed in various other places, this idea that even if you're getting fewer clicks, they're going to be higher value, they're going to be more likely to convert, etc. That bit makes me super nervous. I think that feels like quite a reductive ecommerce-centric lens. And I think it assumes that every site is trying to sell something or should have like an explicit user outcome. And it's almost like they're saying, don't worry, we've, we've stripped out all of those pesky curious audience members who weren't quite ready to hit convert or buy now. And we're only sending you people who are right at the bottom of the funnel, which may be great if you're treating this as a performance space. But we all know that brand and general marketing and audience cultivation, et cetera, is so much bigger than just get them to click my thing and convert and buy now. So we're kind of ceding all of that trust building and audience stuff to Google. That's scary.

And even if you are not a publication, if you are not relying on content as a business model, if you are a merchant, if you are an online store, what are the incentives then now to create information that is very well optimized, that targets and connects with the way that the user search to clarify the doubts, to help them through the whole customer journey? It's zero because now if everything will go just to the bottom of the funnel directly and If most of the traffic will eventually be referred to the PDP and you don't invest in informational content. And we know that it's like chicken and egg type of situation because it's not only traditional search, but also AI search, LLMs. They rely on informational content, citations. Hello, even a bigger role there in order to gauge authority and relevance at the end of the day. So take the balance of where we were and the incentives are not there. There's a shift certainly happening, but indeed, we're at this point still at the early stages, but let's see what are those lights. And this ties very well with the second story. A way to be able to assess and gauge that light or lack of will be to have a really good access to measure the impact of that.

The challenge, you've got to build business cases if you want to adapt to this new world, and you've got to have a very difficult conversation with your senior stakeholders and budget holders and say, look, We have to take a gamble on helping people at the top of the funnel when we can't measure it, on winning hearts and minds in ways that might convert in ways we can't actually. Those are hard conversations. And to do that without any meaningful data and without a way to breathe those stories. Yeah, that's bizarre.

To clarify, this second story is about that Google has now added AI mode traffic to the Google Search Console reports. By Matt Giessodern from Search Engine Journal. However, the big thing here is that it's all blended. It's all blended with the organics, the traditional organic search metrics. And what I have seen that a few have started to do is to create reports or to filter reports with regex only showcasing queries more than 5 words or so. But we know that Google optimizes a lot of the more specific, longer tail type of terms. On one hand, I will have much more prefer, of course, that Google provided filters like they do for manual read thresholds. So we need to work around this. There are other solutions now, AI search trackers, for example, that are now like integrating further with analytics and other tools, third parties. I believe that if Google and not only Google, but other platforms like OpenAI, I haven't seen so far that they have taken any actions to give clarity and visibility. on the user search activity or engagement and traffic they refer. Hopefully we are able in the future. And there's also another piece. It seems that everybody is very well aligned here on what is this uncertainty going on right now in our sector, in our community that raises a lot of concerns. The guide from Ahrefs that has gotten a lot of engagement showcases that. Isabel, welcome to the zero-click search. Please leave your traffic at the door. Indeed. Yes. Louise Lanhan goes through the reality, what is causing zero-click searches, and we have AI overviews, AI mode. So she goes through all of the culprits for this and what should we do to try to navigate this in a way that is actually not negative for the final outcome or goal marketing-wise. She goes through different ways that We can tackle this particular challenge, developing products or user-focused content, creating deliberately human content, going method with customer research, diversifying to new channels, going all in on brand and PR.

I really love this. It's such a great overview. And I think all of these are the kind of difficult conversations we need to be having because they are all quite a step change in how we've perceived and done SEO. I think annoyingly as a technical SEO, the bit that resonated with me most was this idea that we should be doubling down on brand and PR, and that some of the most influential factors for fighting to survive in this change are citations, links, references, et cetera. And that makes a ton of sense, right? Because either you're purely what these AI systems and LLMs are understanding and believing about your product proposition. Or instead of trying to say, how do we rank better for things that people are searching for? You change the narrative and you say, actually, what, what if we did a whole bunch of marketing that changes how people search and changes what they search for? Or we change, we double the number of people who search for our brand instead of trying to make this keyword rank 5th instead of 6th, etc. It's such a more grown-up way of thinking about search as an ecosystem and less as a kind of awkward performance channel. I think I think there's a lot in this that feels like this is how we should always have done it, but didn't need to because perhaps we were a little bit lazy. But this is a great view on what does the future look like for sure.

The thing is that this is far less, let's say, checklist oriented and tactical that we know that SEOs love, unfortunately. And, and, but it's much more strategic and needs to align with a real brand, real positioning, understanding of what you are about, what is your advantages, disadvantages, who are your target market, who are not. And there's another point here, right, that I think is going to become even more important in the future with the level of personalizations that AI search slash LLMs will get in the future. It's not anymore a matter of, oh, you're not competing versus other 9 ranked pages anymore that are below or above you. You have here all the citations that are all blended together pretty much. Or even in, in product packs, they are all, and you have like 5 different products in the same row showcased as, as big as you do. So who do you think that users would click on? They would click on the brand that they recognize, the ones that they have more affinity with, the one that they know what they are about and they trust. And to build that trust, you should have had to pretty much been showcased before when they were like right at the beginning of their journey. Looking for alternative for their problems and engaging there with results. So you will have much more chances also for the LLMs to suggest you later on in the customer journey and the search journey. So it's the chicken and egg type of situation indeed, and it has more implications in the future with the level of personalization that we will see with the engagement in LLMs. Kevin Indig also wrote this piece that was particularly popular too in the latest edition, providing another, let's say, angle of this same particular challenge and going through also many screenshots here of websites with the great decoupling trend happening for them. He recommends on taking certain steps in order to still get visibility, still get influence in those queries that actually matter for you, those that have actually the right intent that you want going through paid media, social and community, product marketing, content, and GTM.

There's a phrase towards the bottom, which might be the best thing I've ever read in an SEO context about needing higher insight density per paragraph. And it's just such a beautiful encapsulation of all of this, right? If we're going to be assessed by these systems, if we're going to be fighting for fewer clicks, all of these things boil down to We have to make our stuff so much better with so much less surface area. I love that concept so much. More insight in every paragraph is such a great way of thinking about it. I think also, I really like this point on conductor mode, the idea that we're orchestrating multiple teams across an ecosystem, et cetera. And I think maybe one of the things that people aren't thinking about enough here is that that extends beyond just your website. So one of the One of the key things about all of this AI mode search and ChatGPT and all these others is they are quietly in the background evaluating multiple websites and sources and reading articles and looking at comparison tables. So it's not enough to say, where do I rank and how many impressions did I get? Because if some other site in that ecosystem says your product is terrible or disagrees with statements you're making or has a cheaper product that's also open on Tuesdays and closer, You might not show up at all. You might not— like, these systems are going to disregard you, and they're going to pick something else in ways that Google Search Console will never tell you. Because you're like, who cares how many yours— you don't see the impressions you didn't get. That makes sense. So you get silently miss out on all of that. So I think the idea that you have to orchestrate this whole ecosystem and really think about, to your point, when a user starts searching, when they first start to say, what are these questions and problems I have? What do they see? What do they experience? What do they feel? Even if your site isn't the best fit for them, then the journey they go on and an AI will go on doing the same research determines whether you show up right at that last minute. It's a very different type of content strategy. It's a very different type of brand and PR strategy that you have to be there influencing. And I don't think anyone's really doing that yet. We're all focused very much on where does my site rank and how do I convince the systems that I'm good? It's bigger than that.

And if they do for the bigger brands, it's all the areas. That unfortunately many SEOs, even at the enterprise level, don't align much with, as you mentioned, digital PR, social community management, et cetera. Like from a sentiment or reputation or citation standpoint, I want to understand what they are doing in Reddit if Reddit is supposedly being taken so much into account by LLMs in answers, right? So 100%, totally agree with this. The next piece is the AI search. content optimization checklist that I created to facilitate the understanding, grounded level of actionability to start taking steps. Because I think that there are like a lot of debates going on in social that we can see very clearly. On one hand, it was like, oh, but we are already doing all of this in SEO, so nothing should change. We already were thinking on better embeddings rather than purely keyword matching, isn't it? Yes, but it's not about that. It's not purely about the concepts or the areas of optimization, but how we optimize them because of the retrieval methodology changes and the behavior of the users changes. And because the interface is different, we need to measure it also a little bit different. And the outcome that we will expect, as we have discussed so far, is different too. So what I did was to summarize. The top actions connected with every area that I saw was relevant to try to understand, to optimize your content for AI search. Chunk-level retrieval, trying to optimize your passages, your snippets of content to make them semantically tied to really answer, to really address specific topics, specific facets of queries rather than trying to rank at page-level optimization, not necessarily anymore. Optimize for answer synthesis. Citation worthiness, topical breadth and depth. So I go through all of this explaining the why, telling the how, and then with examples of doing this right or wrong to hopefully make it much more actionable and accessible for everybody. So I believe that that is why it is important to showcase examples to allow people to understand how it should fit with content created for users. And for your brand, for the sake of a real business behind, not trying to rank for everything and anything just because it's a facet of a query and because of the fan-out technique, it might get triggered at some point. No, no, no, it's not that, right? And João, what are the biggest questions that you have gotten with clients or that you have seen around populating and potential misunderstandings that you see about Optimizing for AI?

Very much the same kind of mistakes and challenges they have with conventional SEO, but harder, right? So we already, you mentioned the kind of the journey from page-level thinking to paragraph-level thinking. I think people already struggle with page-level thinking. They assume that everybody neatly enters a website on the homepage, reads everything about the business, and then linearly navigates through a site and makes a purchase decision when the reality is chaotic. People. Arrive on a random deep page and they don't know anything about the company or the product. And the content on that page almost always assumes context they don't have and isn't written to educate and help people go on those kinds of journeys. So businesses are already struggling to get that right. But now there's this extra level of you've got to take it down to paragraph level. And we have all the same kinds of challenges. Should this be one page or 2 for a topic that becomes, should this be one paragraph or 2? It's far harder. Um, but I think this is what we should always have been aspiring to anyway. And it's a far better version of that. Like if my, my product page should facilitate people who arrive blind to it directly, have never been to my organization, and it should still tell that story and take them on that journey. We've just never really done that well. I mean, that's how humans have interacted with books and libraries and resources forever, right? It's been a limitation of the web that it's been page-centric, a limitation of Google that that the web has been very page-centric till now. So exciting to see what comes next.

I love how you connect this and compare this with libraries, which at the end of the day, this is, yeah, this is the online version of this. Then last but not least, I wanted to get here your piece. I love your thought-provoking pieces, how you take a different angle or thinking bigger into many of the issues or the topics and challenges. That we're talking. We also now went through here in this other piece talking about making more human content, and you highlighted how funny it was with a little bit of more of your take. This is the thing, we have AI, we can leverage AI in a way that is actually useful, strategic, that is relevant in our businesses, and there are things that we should realistically expect or not. And if the technology capacities are there, someone will leverage it in certain ways. So here you talk about the rise of synthetic authority and what does this mean?

This is super exciting. I completely agree with the, um, Ahrefs article, by the way, about human authoritative content, et cetera. It makes sense. What I'm thinking about is tomorrow. So we use the word human when we talk about human content, et cetera, some kind of barrier and a threshold of quality. As if we assume that non-human can't be valuable. And that makes sense because as humans, we like and recognize human faces and opinion is a big differentiator. And a whole bunch of Google's EEAT stuff rightly focuses on evidence that qualified, capable humans have seen the thing, understood the thing, touched the thing, et cetera. All of that makes sense until you start to realize 2 things. I think one is that Most of the human content on the web either isn't very human or isn't very good. Like a lot of it is just listicles and articles and SEO filler. And even worse, many, many websites and their blogs and their content lie a little bit or a lot about the capabilities of their authors, or they generate them and fake them entirely. And actually, in aggregate, a lot of human stuff actually isn't great. And then if you fast forward ever so slightly to tomorrow-ish, We get very quickly to a place where AI systems, whether they're LLMs or other things, can actually fill a lot of those gaps just as well and maybe better than humans. And when you deconstruct the EEAT stuff about experience, expertise, authority, and trust, all of those things you can get close to and maybe surpass humans with a machine. So for example, I don't know, I want to buy a new air fryer. And today I definitely want to read a bunch of reviews by an expert who has touched 100 of them and done that work. And has formed opinions. Hard in theory to say, I'm going to write a piece of software that tests and evaluates the airflow of 1,000 air fryers a million times and uses AI to evaluate enough spot patterns and say, actually, this one's really good for salmon. That one's really good for whatever. It's not hard to imagine a scenario where it does a far better job of the expert than they ever could because of a limited human with finite hours. These systems suddenly can do those things better. So what then? That content is more trustworthy. It is more expert. It is more authoritative. At what point do we start to accept that and maybe normalize it? And maybe that stuff is inherently better than the human stuff. The swap still exists, but human becomes average and this stuff is the really great stuff. So I think it's worth us starting to think about and asking those questions.

I love that, like, especially for websites that have already a lot of UGC. Adding an AI level there that is able to eliminate the bias, for example, and identify the patterns for a summary and balance pros and cons, for example. I have seen in a much more, let's say, primitive way, or the potentially like the first version of that taking place already across online source with a lot of thousands of reviews in PDPs and AI summaries at the top. Uh, going through the positive and the negatives, uh, mentioned in the reviews below. And, and more and more tools like AirOps, for example, or Wordlift that integrate with your own content, with your own database, and based on what they learn from them in a very personalized ways and based on your own workflows and criteria, they end up generating content. That can, well, reflect the data, the insights that, that you have, right? We're still, I believe, in early stages of that, but we, we already have platforms. We already have tools that are capable to do that. What is important at this point still is to have a layer, let's say, or a phase of validation for the sake of accuracy and to provide feedback and for the machine learning to work well. However, I can definitely see it and only becoming more and more important in the future. 100%. Amazing. Thank you very much for opening up to possibilities and making us think bigger, for sure. And I want to thank again Similarweb, who sponsored today's episode. They have the new reports to track AI chatbot traffic. So if you haven't yet taken a look at it, go and check it out. They have also the top prompts per pages, which I have found to be particularly useful in order to better understand and to ground a little bit and compare the difference in behavior of users with LLMs and AI search versus traditional search engines. So go and take a look at it. And if you want to keep with the latest news without having to wait for the next newsletter or for our next episode, go and take a look at the SEO FOMO News aggregator, where you can not only check the latest happenings, but also you can submit your own news, your own resources. And we will check it out and take into account for the next edition of SEO FOMO. Thanks again, Jono. It was a pleasure to have you. I really appreciate your insights. And thank you for watching too. Remember to like the video if you like it, to subscribe to the channel to avoid missing out the next episodes. And until the next one, bye-bye.

Bye-bye. Thanks for having me.