Website traffic decline in 2026: Ryan Law on AI overviews, zero-click search and what to do next

Hello, lovely humans. I hope you're doing great. It's always just an absolute delight, to see you all popping in the chat where you're watching from. I've even seen people making pub plans for tonight, which is a pretty good outcome from one minute of a webinar. So, congratulations, to the three of you for for your pub meeting. I hope you have a lovely time. Hello to Andy in Daventry, my hometown. Hello to Laura, to Lindsay, to Sophie, to Rose, to Sarah, and so many more of you. Thank you all for tuning in. If you haven't already, do pop in the chat feature, where you're watching from. And as, Sarah rightly points out, folks, it's sunny. It's sunny. The sun is shining, so we're having a nice time. Let's get started with introducing today's session. Karine has already put in an address to meet at the pub, which is fantastic. Good for you. So let's get introducing today's session. So in twenty three days, we have the second ever Marketing Meetup Conference. You are very, very much invited if you would like to be. We're actually in a nice position at the moment where the biggest danger, I think, is that we're gonna sell out, which is a a nice problem to have. If you would like to be there, then please do come. It'll be a fantastic day where the team and I have been working so hard to really craft it and make you feel cared for and cared about in an environment where you will, of course, leave with a bunch of marketing knowledge. Second thing second. If you would love to chat with the Marketing Meetup community after today's session, then we have a private community space called Hive. In Hive, you have the opportunity to hang out with the rest of the community and chat with them in between events. So you'd be very much welcome to join Hive too. Thirdly, let's get introducing today's speaker. Today we have a very handsome man. We have Ryan Law and you'll see why he's handsome in a minute because he is the better looking of the bald gentleman with glasses on today's call. Ryan is the director of content marketing over at Ahrefs and for me, and I say this with genuine sincerity, he is one of the top people I follow on LinkedIn right now. Ryan brings thoughtful content, content that is written well, but also content that is insightful. I'm absolutely thrilled to have him involved today. Today we're speaking in the context of what to do when your website traffic declines in twenty twenty six, which is largely caused by stuff like zero click content, GEO, etcetera, etcetera. We'll hopefully investigate a lot of it in the rest of today's session. Before we get started, I just want to say a quick thank you to our sponsors, because without these folks, then we wouldn't be able to make these sessions happen. So the first, sponsor that we want to feature this week is a company called Wistia. Now Wistia is an incredible video marketing platform. They've got webinars hosted in there. You can host your content in there. You can even use that content to then use it for lead generation for the rest of your activity. It's really, really solid as a platform, and they're fantastic sponsors. Right now on screen, we have a webinar repurposing book which Wistia have produced. It's well worth checking out. Also a big thank you to Mailchimp, Cambridge Marketing College, to Frontify and to Planable. All of these folks are incredible. Now with a pub chat going on in the chat, the main thing that I can do is encourage you to keep that going, not just about pub but also about, the content that we share today. If you have any questions do pop it in the Q and A. And with all that said, let me introduce Ryan to screen and say thank you so much for being here mate. It's, it's over to you. No. Absolute pleasure to be here. And, yeah, I think I'm gonna take any chance I can to come on a webinar with you, Joe, because you are the nicest, most flattering human being I've ever met in my life. So I made me feel very good about myself before I get home. Very welcome. No. It's it's a real thrill. There you go, Simon, brothers in the chat already. That's the first comment of the day. We're gonna be having this for the next, hour, I think, mate. Amazing. All right. Well, let me get some fun slides up and we can all commiserate about the loss of traffic to everyone's website. Good. Bringing the joy. Here you go. I took my hat off for one second and and and and we disappeared. Cool. I haven't got your slides for ****, Ryan. I'm very slow. It's all good. Here we go. No worries. Alright, there we go. There we go. Perfection. Right. Over to you, matey. Very cool. So this emoji captures perfectly my sentiment about this topic. I used to have lots of website traffic, and now I have much less, and the world is a slightly bleaker place for it. But we shall all rally around and find some way of dealing with it. Very briefly, most boring slide here. As Joe said, I am the director of content marketing at Ahrefs. I run our blog team primarily. We write lots of hopefully helpful educational SEO content. We do lots of research as well trying to understand how search and online visibility is changing, especially in the wake of all this crazy AI shenanigans. I used to work at a very big content marketing agency, very lucky to work with some very cool companies like Google and Zapier and GoDaddy. Most salient for today is the fact that I carry with me a deep sense of trauma every time I open my Google Search console account at the moment because things are not as rosy as they once were when it came to organic traffic. So put another way, the Ahrefs blog, I am responsible for, is absolutely bleeding clicks at the moment. This is an actual screenshot from the monthly blog report I put together just a way of keeping the team in touch with what's happening on the blog. This is a very, very standard story. And it has been for around about the past two years, a monthly decline in organic clicks. So three point five percent decline in January. And actually something that I noticed as well was that since August, direct traffic, so this is the kind of anonymous bundle of traffic from miscellaneous sources, has actually become the primary traffic source for the HRF's blog instead of organic traffic. That was a huge wake up call for me because, that has primarily been our biggest source of growth and customer acquisition. So now direct traffic is much bigger part of that picture. So basically, under the two years I've been in charge of the HRAS blog, traffic from Google search has been falling almost every single month. It is frankly a miracle that I still have a job. And by all accounts, this should not happen, right? On paper, lots of things are very good about the hrefs blog. We're quite highly regarded as a brand, people recognise us, and they tend to click our content when they encounter it. We've got amazingly smart, wonderful people writing this content as well. This is a bio of Despina from my team. Ten years in SEO. She's an author of a book. She's created courses about SEO. She's been featured in all these amazing brands. She really truly knows her stuff. We have a very intense editorial process. We care a lot about the caliber and the quality of the work we create. We've obviously got access to tons of amazing search data to help inform the content we create. I'm very lucky that we have a d r ninety one website, which basically can rank for anything. We've not been hit or penalised by any Google updates generally speaking, so we you know we should be ranking quite consistently. And as Joe said earlier, obviously very talented and handsome content director at the helm, things should be very very good and they are assuredly not at the moment. So what the hell has happened? What gives? I think it is basically AI overviews. By far this is the single biggest culprit that is affecting all of the websites I've come across, all the people I've talked to. And this is a bit of research we actually redid about a month ago. We looked at the impact of an AI overview on the number of clicks the top ranking piece of content gets when an AI overview appears. So basically, if you do the very hard thing of getting to rank number one for your target search query, you've, you know, achieved the gold standard milestone, you've done all you can do, you are gonna see about fifty percent reduction in click through rate if Google decides to put an AI overview on that page. So for every hundred clicks you used to be able to get for doing that, you now only get fifty eight of those. That has a pretty significant impact on total website traffic. And AI overviews themselves, they appear on roughly twenty one percent of all search queries. So we analyzed about one hundred and forty six million search engine results pages to see this. By far, the biggest clearest trend is that almost all AI overview triggering keywords are informational intent. They are designed to help people get answers to questions that they have in the search results without having to click one of those annoying blue links and actually go back to somebody's website. We also found that AI overviews appear for about sixty percent of all question queries and almost half of all very long seven or more word queries. So very clear picture of the role AI overviews are trying to play here. They are basically there to help answer educational informational questions directly in the search results without people having to click away and visit some websites. Unfortunately for me, that means that companies that rely on informational content like the Ahrefs blog are disproportionately very heavily hit by this. It is thankfully not all entirely doom and gloom. Although informational queries are the very much the remit of AI overviews, other types of intent, search intent are not impacted in the same way. Commercial, transactional, navigational. People want to actually buy things, and that's evident in the search query they put in. If they're trying to get to a specific place, generally, these do not trigger AI overviews anywhere near as much as informational queries do. And that is still a very important subset of everyone's website traffic. And even some topics as well are relatively safe. Google does not generally put AI overviews on local searches. So if you're a local business, you get most of your web traffic from local searches. That is fairly safe from AI overviews at the moment. News queries, only six percent. Not safe for work searches if that is your jam, if that's your industry, you're at least fairly inoculated from this problem. And shopping queries as well. In most of these cases, Google has other search features that they think do a much better job at meeting meeting the intent and probably allow them to monetize more directly than AI overviews do, so AI overviews are not a problem in those areas. Now I think a very important point. When I started seeing all this traffic decline happened, my first thought was, you know, is traffic being siphoned away from organic search and going over to chat GPT? Everything I've seen, all the research I've seen suggests that actually ChatGPT search is in some ways additive to traditional search. People are definitely searching more and more in AI search engines like ChatGPT, but that does not seem to massively have been at the expense of search happening in Google search. Google searches are still growing quite significantly at the moment. And actually, from a sheer traffic perspective, AI search is just the tiniest small fry you could imagine. This is from a live website. You can actually check this out. It's chat gpt versus Google dot com with some awkwardly placed hyphens in there. But we track seventy to eighty thousand websites that have our analytics installed anonymously in aggregate. And almost forty one percent of all the traffic those websites get still comes from traditional search. By far, it is the most important source of traffic for these websites, and that has remained surprisingly stable over the past year. By comparison, AI search, so ChatGPT and all of the other AI search engines, they only send zero point two six percent of all traffic. Sending traffic is just not in their nature. It's not something they're designed to do. They're designed to keep people in the the chat experience. So very important thing to internalize, I think, this point. If you are struggling with lots of traffic, you've seen loads of your top of funnel traffic disappear, It is probably not a you problem. It is probably a Google problem. And if you have a boss like me, the lovely Tim Solo, and you occasionally get panic messages at three in the morning saying, where did all of our clicks go? There are lots of things you can say in answer to that. Obviously, our research shows that AI overviews have this impact, but actually so do tons of other studies as well. There's a ton of evidence out there to validate this hypothesis. As I mentioned, lots of different query types are not as impacted in the same way. So unless you have a big top of funnel informational content strategy, you might not be as impacted by this. This is also a trend that did start many, many years ago. If you remember featured snippets, there are lots of these zero click search features as Joe mentioned earlier that have been trying to serve this purpose for a long time, keep people in the Google search results satisfy their search intent there. And most important of all, consumer demand still exists. Even if all the traffic to your website suddenly vanishes, the people who would otherwise be buying your product, the needs they have, the desire to actually go and buy a product to solve their problems, that has not gone away. If anything, what we've seen with ChatGPT is it's now easier to understand which products to buy, easier to have your very strange queries resolved. The important thing is that the role your content plays is going to be different now. Content is less, think increasingly about bringing people to your website and it's more about informing the narrative that these search engines and AI search engines have about your product and your brand and what it's good at. And hopefully, after this webinar, you will have a plan to share with your boss if you need to do something about this. And that plan is to live in the woods. Throw your laptop away, give up content marketing forever, and just become some kind of strange bearded hermit like I plan to do. If for reasons of mortgage or livelihood or children that's not possible, I do have some practical tips that you can do. Probably the most useful and immediate thing you can do is content refreshing. If we can assume that Google is giving us less additional traffic for every article we publish, I think there's a strong incentive to actually go and improve the caliber of the content you already have, get more traffic from your existing content. And that is actually fairly easy to do. We know that readers generally prefer fresher, more accurate content. If you're looking to read a product review or something like that, you are gonna prefer the thing that was published last month and not three years ago. In the same way, Google has an obvious preference for fresh content in many areas. If readers are clicking your fresh content, that can often translate into better rankings because Google uses user signals, which I'll talk a little bit about later. Google also has this idea of query deserves freshness. There are some topics that Google knows the top ranking content should be updated very regularly because the information changes so often. In those cases, updating your content regularly will have a huge impact on your visibility. And just a quick illustrative screenshot, I think this was something Louise from my team did. One article on very simple update refresh to the content, three hundred percent uplift in the months that followed to organic traffic, which was not too shabby. And very importantly, AI. If anything, AI actually loves fresh content even more than traditional search. We looked at, it was like seventeen million citations in our massive index of AI conversations. And we looked at the average age of the URLs they cited and compared them to the average age of those URLs for the same search query in Google. And what we found was that AI assistants, they generally cite things that were about twenty six percent fresher than what traditional search would do. And this makes a lot of sense given what we understand about how AI search functions. Most LLMs will try and answer a question from their base knowledge, their their model knowledge. If they can't do that, they have to go out into the web and get fresher, more up to date information that is not in their training data. So most of the times you will actually see a click, you will get visibility, you will be cited is because they're going out to find explicitly fresh content. And some very smart SEOs, Chris Long, he found that Chattypity actually appends the current year to a very significant percentage of its fan out queries, which are the search queries it generates to go and retrieve information to bring back to the user. And Mehtahan Yasiliert as well, he found that ChatGPT actually has a documented URL freshness score that plays some role in the selection process when it's choosing URLs. So AI really, really does care for fresh content. It's quite easy to find refresh opportunities as well. You don't have to use href to do this, but it's very easy if you can. You are basically looking for pages that have seen a sustained decline in organic traffic over some period of time, say six months or a year where you have not made any changes to the content. In Ahrefs, we have literally this top pages report that lets you see an ordered list of all the pages you have on your website, and you can sort them by how much traffic they've lost over a period of time. And we have this content changes column, which is quite handy because you can see whether any changes were made to the page content recently or not to allow you to know whether they're good candidates for refreshing. And in terms of actually doing that, actually managing this refreshing process, there are tons of very simple things you can do that will have a big benefit. You can obviously correct any obviously outdated information. Something I do regularly is I run a very simple site search in Google like this, hrefs dot com forward slash blog and then twenty twenty four. And that will show you any articles that have in the title that outdated year. These are obviously contenders for me to go out and refresh them, update them, and bring them up to scratch in the year twenty twenty six. You can use Screaming Frog to do more advanced versions of this, HF Site Audit, but this is generally pretty useful for everyone to do. You can also test new titles. I have seen extremely small changes to titles have a dramatic impact on search rankings and performance. And actually, I don't know if anyone was following the Department of Justice trial where we learned a ton about Google through all the leaked data. Google actually uses user signals in its re ranking process. So for years, Google said, we don't use clicks in ranking. And that was strictly true because it turns out they use clicks in a re ranking process, which is slightly different than what they were talking about. But effectively, if you have a better title and that encourages few more people to click in a given period, that sends much stronger user signals to Google, and that can result in stronger rankings for you as well. A few things you can do to improve those titles, try creating a curiosity gap. Promise some kind of benefit that has to be clicked into the article to unlock. You can try differentiating from the competing articles as well. Your URL is not being viewed in isolation. It's being viewed alongside other competing websites. So respond to those. Offer things that the other titles don't do. Choose a different format. We've even experimented with responding to some of the titles. I think SEMrush had an article that was like, it was posing a question about like homepage SEO and our response article ranking beneath it was actually yes. The answer is this. We gave the answer straight away in the title. And this is just a screenshot from one of my blog posts where I changed literally a single word in the title and saw it jump from position, what is that, eight up to position one because it was better optimized for the target query. Speaking of which, you can sometimes just target a slightly different keyword. Sometimes you will publish something and to your surprise you will find it will rank better for a slightly different keyword to the one that you intended. Very good example of this, we have a page which is about the HRF certification and its top ranking keyword is actually hrefs course. This is one of those situations where minor tweaks to optimize that for this actual keyword hrefs course would probably help it perform a little bit better. Google is very good at understanding similarities between the word course and certification, but I have still seen big benefits to exact matching keywords with this kind of process. And probably most important and consequential of all, filling topic gaps. Google generally expects there to be some element of consensus among the top ranking content. It wants a similar story to be played out across all of the top ranking results in most cases. So one of ways you can improve your ranking is to go and look at those top ranking articles and see if they cover any important topics that you have not covered in your article. You can do that manually. You can read that content. You can use, you know, human gray matter, heaven forbid, to do that process. We also have a tool in HRFs called AI Content Helper that will do that for you. Scrapes the SERP for those keywords, it analyzes them and it presents you a list of topics and keywords that you might benefit from including. And a very quick nerdier side for you, but I actually vibe coded this content freshness dashboard for myself with considering making this into a tool if people would find it useful, but you can get very nerdy with this stuff. This lets you calculate the average and median article age for all of your content. You can analyze the distribution of articles. So I can see that we've got four hundred and twenty articles that are less than a year old compared to a hundred that are three to four years. You can look at the difference between the potential traffic you could have given the keywords you're targeting and the actual traffic you've got. You can compare the average age of your content to your competitors and benchmark against their refreshing process and end up with basically a prioritized list of content to go out and update. And this is not too hard to build. I spent maybe an hour in lovable, building this. You could definitely do the same thing yourself. Cool. So content refreshing, very important. Another really important thing to do if you do not have as much website traffic as you used to have, we obviously want to get more sales from the content that we do have. When I worked at a content marketing agency, all the time people would come to me and they would say, I want more traffic from more sales from the content I already have. How can I do that? The answer is you can't. You can't unfortunately change your button color and double your sign ups as much as I would love that to be the case. But what you can do is begin this gradual process of moving towards content that does a better job naturally promoting and segueing to your product. This is something we do at Ahrefs. They this is one of the reasons I joined the team was because they were so smart and pragmatic about this. They use a framework called business potential to do it. So every topic we publish is vet accorded vetted according to this framework, and we score it from a topic of zero where there's absolutely no way we can mention our product whatsoever and probably shouldn't be writing about it, all the way up to three where we're basically saying you cannot cover this topic without talking about HRFs. Obviously if anyone reads that there's a very good chance they may need to and want to become a HRFs customer afterwards. And we literally apply this every single week to every single topic we do. This is another screenshot from the monthly report I put together in Slack and you can see that I have scored all of the articles we published according to the estimated business potential they have just so I can very quickly grok whether we're doing a good job of writing about our product enough. Generally, we try and make sure that most of the topics we cover are a two or a three on the business potential scale. So things that would be good would be, you know, tracking your chat GBT visibility with hrefs. Obviously, of chances to talk about our product there. How to rank in AI overviews, lots of chances to talk about our products even if it's not strictly necessary to the topic versus AI marketing examples. It's actually quite hard for us to segue and introduce hrefs there. It'd be a bit tenuous if we do that. So we try not to publish too many of those articles. Now this is a very boring thing to do, I'll be honest. And I'm quite inconsistent with it. And we don't always agree about whether we've categorized these things right. This is something that AI is very good for and you probably should use it for. So at the moment I have a skill file I wrote for Claude very simply. It contains this business potential rubric. It contains a summary of all the products we offer at HREFs and it concludes an example of me previously doing this exercise, how I scored some products. And every time I want to generate this now, I just dump a bunch of URLs into this Claw Chat and it scores them for me. Very, very helpful. Now business potential is important, but there are other things that matter in marketing. We do use this as part of a broader marketing framework that, Tim, my boss, came up with called Brew. So business potential, we've covered. Things that also matter include the reach. Do we have the potential to reach a huge audience with this, piece of content? The effort, how much time and resource does it actually require to create? Is it easy or is it very hard? And who? Is there a person who can actually run this and take ownership of it? This is important because sometimes we do want to publish something that has low business potential but has a ton of reach or is very easy to create or, you know, maybe somebody just really wants to make this thing. Having a kind of broader rubric for this allows us to balance the pros and cons of each of these and make sure we're not missing out on opportunities to grow that we should be doing. So we've been quite doom and gloom. I've been talking a lot about how the kind of content playbook is dying a little bit. But as one playbook does die, I think others do emerge if you are open to and attuned to them developing. So my colleague Louise, she did this lovely reflection post where she looked at the content that performed best for our blog last year. This was a list of our top five blogs. And this was a great exercise because it made me realize there were three archetypes of content are actually performing very very well for us at the moment. These are things we want to spend more energy in. Research. So I think three of these five articles are research articles. Us using our data to try and answer questions that people have about how marketing is changing. Thought leadership as well. Dasbina, my colleague, shared her own personal process for AI content creation. This content is dime a dozen out there, but this was her unique interpretation and stance on that, and that performed incredibly. And trending topics. There are some topics that are just so hot. People are so interested in almost anything you publish will perform well if you are able to identify them. So research content, this is probably our biggest area of focus on the blog team at the moment. We do a lot of this. This one article about AI overviews, I think maybe less than eight hundred words took me a day to write had ninety six thousand views last year, over a thousand referring domains, and it got cited by the BBC, by CNN, by The Atlantic. This type of content performs extremely well on social media, it's the kind of thing that not only do journalists love to cite and link to, and that can be great for distribution, but AI loves to cite primary sources like this as well. So this is a great strategy for AI search if you can do it. And you may be thinking that's all very well and good. You are HRFs obviously. I don't have a data science team at my disposal. How the hell can I do this kind of thing? You can actually start very very simply. One of the best things to do, the best motivation is to try and answer the questions you have about how marketing is changing. What data do you have available to you through your company, through your product, through your customers that you may be able to look to to answer those questions? There are also quite a few public data sources out there that you can mine and pour through, This is made a lot easier with Claude or ChatGPT actually analyzing these and coming up with interesting insights from it. There's a website called Kaggle that a lot of data scientists use to get example datasets. I've created a few articles in the past just going through these public data sets to see what was interesting. Or just run a very simple survey. If you can get fifty people from your industry to answer some questions that other people are interested in, you have the makings for very compelling research right from the get go and you just need a Google form and an email account to get started. I also recommend my friend Cassie. She runs a content marketing agency. This was a big post she said the other day of all the simple easy to get started types of research she's helped customers with. I recommend checking this post out and looking through some of those ideas because there's some good stuff in there. Thought leadership content as well. This performs very well for us. So we're talking sharing the team's opinions, our analyses of industry events, even the experiments we run, whether they succeed or fail. This does not get recurring traffic in the way that traditional search traffic does. You quite often see a very quick plateau if you're lucky enough to get any traffic to it, but it does have extremely high credibility. People, they see you as an expert when you publish this type of content, and it does have very high social engagement as well, which is very useful. It's also very useful if you're trying to build a personal brand. This can basically become a recurring distribution channel for your own ideas, something that we use a huge amount at HRFs. I think lots of people have a bit of an aversion to thought leadership very understandably. The idea of sharing opinions and personal things on the Internet makes many of us wanna shrivel up and die. My best advice is basically just to find some version of thought leadership that doesn't feel bad to you, that doesn't feel gross. There are lots of ways to achieve this. You can share your opinions if you're a very opinionated person. Maybe there are counter narratives. What do you see happening in the world that from your experience does not stand to be true? What experiments have you been running? What processes have you built to solve your own problems? That is fantastically compelling content. News aggregation and commentary. I follow people like Lily Ray and the leader Solis in SEO because they basically consolidate all of the news and they provide context and opinion on top of it that makes my life much easier. Or you can do what I think of as yes and thought leadership. Find an idea you agree with and then add some value on top of that, an additional opinion, some extra research, some additional insight to further substantiate their ideas. And most important of all, just keep trying. The only reason I've ever published any vaguely good articles was because I published a hundred very very bad articles before that and iteratively learned from that process. I actually wrote an article called how to become a thought leader in just six incredibly difficult steps because it's not complicated, but it is hard to do because it requires just a ton of content creation basically. And trending topics. If we can assume that Google is sending less traffic to the average article, there is more competition, people are using AI to publish content and compete with you in the search results. There is a benefit to covering topics before they get big, before there is, data to convince people to cover them. This is high risk, but it's high reward in the sense that they may come to nothing, but if they do pan out, they can generate a ton of traffic like this article did about LLMs dot TXT. When we wrote this, there was, basically no search volume for this keyword at all, but I saw that enough people were interested in it. I had a hunch that there would be search volume quite quickly and that certainly ended up being the case. That became one of our most visited articles. Plenty of ways to find these topics as well. You can use the trending filter in HREFs if you're a HREFs customer. We literally two clicks and you can filter any type of topic to find keywords that are forecasted to grow over the next three months or six months or twelve months. You can lurk on social media, you know, somewhere like, Hive as well, see what your peers are talking about, user generated content sites like Reddit, industry Slack groups, all these kinds of things. And sometimes you just have to trust your intuition if you think there is a good reason to do a topic. If you suspect it will become more important in the future, it's sometimes just worth taking a pun and creating something on that. And last of all, I think social distribution. This is another really important thing that we can use to compensate for this loss of clicks. The great thing about SEO historically was that it was predictable distribution for content. You could publish something and generally a few months down the line, you would get more traffic. And then a few further you'd get even more traffic. Assuming that's not the case, we have to look to other distribution channels. We are investing a ton in social distribution for our content. This is a screenshot from my LinkedIn account last year, almost two million impressions from what I think was a very sporadic and not very good LinkedIn strategy. Myself, Tim, and Patrick from my team, we racked up about eleven million organic impressions on LinkedIn between the three of us last year just from sharing things in a fairly ad hoc, disorganized way. So there's a ton of benefit to be gained on LinkedIn at the moment. Some of the things we've learned that really help with this process, don't withhold value. Nobody wants to be dragged away from the social media site that they have taken time to go to to engage with people in order to read some random blog posts on a SaaS company's website. People generally want the ideas, the takeaways directly in their social feeds. So be willing to give away all the value in the things that you create directly in social media. We did this big research report last year. We surveyed like nine hundred people, and we decided to just publish the PDF directly in LinkedIn. No gating, nothing like that. And I think we had far better distribution on the back of that than we would have done had we gated it on our own website. That was very worthwhile. You can build your personal brand as well. I kind of briefly alluded to this, but people generally like to follow other people. If you can muster up the courage to share content on your personal account as yourself, you will probably, I think, get better results than doing it behind your brand account. These are not quite archetypal posts, but like fairly representative, you know, personal post from me with a thousand likes versus forty four likes for a Ahrefs branded account despite Ahrefs having an order of magnitude more followers than I do. I will caveat this with the fact that we do want to improve our branded social game. We're actually hiring a social media manager specifically for this at the moment. So if you think you can do a better job of this, you probably can do a better job than me. Yeah. Please check out that job listing. We'd love to love to hear from you. And sometimes you do just have to play to the algorithm. There are little idiosyncrasies and quirks of all social media platforms that it can sometimes be worthwhile, subscribing to and engaging with even if it may feel slightly icky and gross to do at times. LinkedIn, things that we've seen work very well. PDF carousels still unfortunately work disproportionately well. Something I make a point of doing is replying to all the post comments on my posts. That has a dramatic impact onto whether an article will be continually resurfaced in other people's feeds because you are continuing the discussion. LinkedIn wants to encourage that. In a similar vein, this whole comment to get x format you might have seen, my colleague Dasbina wanted to test this out so she put together a checklist and said, hey, comment e e a t in order to get this. And she had six hundred and sixty seven comments on her post. So that is still very much a successful useful thing to do. And native video content as well. Actually recording and publishing videos directly into your LinkedIn feed, I've seen great results from that as well. LinkedIn wants more video content. Finally, if you are lucky enough to work on a team, make it a team effort. Support each other, borrow each other's ideas if one of you is lucky enough to have a good idea, share and engage with each other's content. Something Tim, my boss did was he actually vibe coded this LinkedIn leaderboard just to try and gamify the process a little bit and play on the competitive nature that many of us have. Obviously, took this screenshot in particular because it shows me having more engagement and likes than Tim does, which I was very proud of because he's got like four times the followers I do. But this is a nice way of rewarding people and encouraging them to actually turn it into a game. So covered a lot. I would just briefly recap it to help your poor battered brains. If you are in the same boat as me and as HRFs, if you are seeing some significant decline in your organic traffic, first of all, it's not you. It is very much a Google thing. This is documented across many different websites in many industries. AIOs hit informational content the hardest. If informational content has been a big part of your marketing strategy, you are probably gonna feel the pain of this sooner rather than later. One of the best things to do in response is to spend more time updating content that you really care about both for traditional search and for AI search. It's probably a good idea to start prioritizing topics that have very high business potential that relate very strongly to your product and the thing that your company actually sells. You can experiment and embrace new content formats. I think better to do it now before it becomes too late and you have to do this. Better to invest a bit of speculative energy while you can. And embrace platform native social content. We are seeing great results from social media at the moment. I think it's worth your effort. And most important of all, despite these scary weird times we find ourselves in, people are still buying. Consumer demand is still there. Content is still the medium through which AI and traditional search engines learn about the world. Content is gonna be incredibly important. It's just gonna be in a slightly different format, and I think we need to be a bit more flexible than we used to be. And that was all he wrote. Thank you so much Ryan. That was blooming wonderful. Absolutely wonderful mate. Was eight minutes in, I was just reflecting that I'd already learned a bunch. So the fact that we had eight minutes plus the rest was so incredible. Like I can see already in the chat feature, you got people like Kelsey saying furiously in capital letters taking notes this whole time. Thank you so much. We've got Paul who's saying that was an amazing talk. Thank you Ryan. Informative yet interested. Interesting. We've got Alexandra exceptional and trying to keep up with notes. Mate we've got Catherine saying that was so good. I really appreciate the actual insights and actions and like let me mirror that by sort of saying the way that you started that by just being so open was really, really appreciated actually like just absolutely fantastic. I'll do one last one from Jenny because there are many more and thank you everyone in the community as well for being so supportive and so wonderful. But we've got Jenny here who says great talk, the best I've listened to in a long time. Can carry that through. We've got eighteen minutes left mate and I can already see eighteen open questions. We haven't opened the Q and A section just yet. So let me sort of encourage folks if you do have questions do drop them in the Q and A. I've tried to start higher level with this first question and then we'll nail into some more specific tactics if that's okay. So there's no right answer to all these things so please don't feel like you're on the spot but we definitely value how it goes through your brain. So we have Fiona who's the first question. He says I know we want website hits but if we have a tracking link that monitors where our website has been crueled for error responses wouldn't this still be a good thing? And I guess really part of the the crux of this question for me is like, how are you adapting in your own head to this new world of like, oh, our website traffic is going down and that feels bad. But we're getting like these new sources from elsewhere like social AI etc etc. I mean there's an emotional answer to this as much as there is a marketing answer to this I think. Yeah, I mean so certainly on the emotional side of things you know, I spent my entire career with a very concrete set of like signs that I'm doing well in my head and by far the most important of those was traffic go up. Like if can't do traffic go up, I'm not good at my job. And actually joining a company as esteemed as HRFs, I had a lot to live up to. Now I think, have I messed something up? Something I try and periodically do is I ask other people in the company like, what have I messed up? Is there something I can fix? Because I want to make sure that for all my bravado about AI overviews, I still have a lot of self doubt about that. But very pragmatically, you know, we add to the actual question about attribution. Yes, if you can get some kind of little iceberg of attribution that is no matter how imperfect it is, I think it is very directionally useful to know that If you can provide a little bit of confidence to yourself, to your boss, to your managers that actually, yes, you know, we're seeing the amount of times our content gets referenced in AI conversations increasing, or we've looked at the log files on our website and we know that these bots are accessing our website, that is the start of a very early process of attribution and building a picture to say that this is working in these new and weird channels. So yeah, I think there's no definitive way to prove this is working yet. There's no playbook. But if you've any kind of analytics and you can see that improving, I think you're in a very good spot. You're in a better situation than most companies probably are. Nice. I love that, mate. And could could I follow-up on that? Because like we don't need to go too deep into it because it's not the remit of this this conversation necessarily. But do you have a few sentences or any tips for anyone wanting to start that journey of tracking more widely, you know, in this sort of AI space? Yeah, I mean, there's a whole kind of there's a whole industry spinning up that wants you to buy AI tracking tools at the moment, Right? Yeah. I think by far the simplest thing you can do is to start any chat model you have access to, ask it what it knows about your company and your products and your brands. It's not a perfect, starting point by any means, but you begin this process of trying to mine the model's understanding of your business and your products and just begin to see where you might have blind spots. And then eventually you can progress beyond that. Maybe you come up with a set of questions and topics that you care a lot about that you will track over time and see how the answers change and whether you get mentioned all the way up to stuff that we do at Ahrefs where we have, you know, millions and millions of index of search prompts and you can and AI outputs and you can search through that and try and understand the underlying characteristics of these models and the brands that get mentioned in it. Yeah, just talk to an LLM to get started, think, very good starting point. That's perfect mate, thank you very much. That's a really direct and practical useful answer, so I really appreciate that. Thank you. Let's take us to way back the beginning when we were speaking about refreshing content because Sarah has a great question, which is not shown entirely on screen, but it says for smaller teams, how to balance publishing new content to build authority versus refreshing existing content. I'm curious about where you sort of sit in that world. Honestly, I'm probably a very bad person to answer this because I should be much better at refreshing our content. I know it's important. I really should be doing it. I just find the shiny new interesting content much more alluring to me. I think I almost there's a bit of a thought experiment I was talking to Tim about. We were talking about maybe we should stop publishing any new content for three months and we should just focus on content updating and see what happens. Because I do think for many companies, if you have, you know, you've been doing content marketing for a little while, that will probably have a bigger impact than publishing, you know, five, ten net new blog posts. So that's one way to do it. Maybe treat it as like a discrete sprint that you do in like a a big block and then move on to the fun stuff. Yeah. Otherwise, just, you know, maybe once a month you pick your the articles that have lost the most traffic and you go in there and you try and improve them a bit. I am trying to build an automation with Claude code that can actually automate content updating for us. It's obviously this is quite speculative and slightly out there, but, you know having something that can at least update old statistics and that kind of thing you could probably automate that now I think. Nice I mean that's that's dead interesting and I guess we're always speaking about you know when it comes to AI rarely using it for the base content you know like always you know sort of doing what you can to produce as much as you can and sure refine it with AI etc but there's something really strong there. A very quick question from Simon. And I know that Simon's sort of involved in the world of AI, because you've mentioned a few little projects. You've mentioned chat GPT and cord a few times. What's your go to LLM of choice and why? Great question. I was using ChatGPT for the longest time. I think since I learned more about Claude code in particular that has switched me over to, I don't know if it's the dark side or the good side depending on your perspective but I've been using Claude Code a lot the past couple of weeks and I'm kind of slowly edging more towards the Anthropic ecosystem now. Yeah. Yeah, they both have their strengths and their weaknesses I think by the looks of it. Yeah, No. Totally. So, on the refreshing front, one of the things you did, showing partway through, your presentation was showing the lovable dashboard that you built. And there were quite a few people who were like, oh, I need access to that. So I think you got a SaaS business in itself. Is that something folks can access or is that something you've purely built for yourself? So what the plan is, we we basically spun out a kind of independent like arm of the company called hrefs labs, and it's basically the marketing team. We are vibe coding tools that we want to use. Then our actual developers legitimize them and improve them and fix all the hideous security issues we've caused through our vibe coding and the intent is then to release them. Probably is free tools to get started with just to gauge some interest there. So I don't have a timeline on that. Glen and our team is kind of leading the charge with that but yeah we're trying to solve our problems with these tools and then share them more widely and help other people. That's how great companies are built so I love that and that's fabulous and it is you know for folks who haven't we're in no way affiliated with a lovable or a rep lit or whatever but like they're they are really interesting tools you know to sort of play around with and so if you haven't played around with them then it's definitely worth checking them out for exactly as you're doing for solving your own need because it's really, really strong and just really interesting. It was my Christmas project and I definitely got lost in a few of them. So let's take the next one from Matthew. So Matthew says, what is the play on blog creation dates if you update an existing blog? And there's a couple of questions in the Q and A about blog creation dates. So just to acknowledge the other one as well. Do you leave the original creation date or update it to the current date when updating? I've seen conflicting recommendations and I appreciate it's always hard to give a definitive this is what it is but I'll be interested. You know it's funny if you look at the HTS blog we have two dates on every article. We have a published date and we have like a way of we have a last updated date as well. So I think even you know we don't even have a definitive answer to that. I generally think if you have made substantive changes to an article, republish it, change the date to whatever date it is when you publish it. That's probably the clearest signal to Google that this has been updated. This is fresh content. Yeah. I think that's probably the simplest thing to do. Nice. I love that. And and it's a really good signal for the user as well as much as anything else. It's like a nice signal, which is which is fabulous. I'm sorry, I clicked on this one a little bit early, but it came up on screen. So Inez is asking, and I think this speaks to some of the comments that we had earlier on about sort of the emotional sort of conversations part of all this as well. But our website's seeing conversion rates decreased or increased with lower traffic volumes. I suspect our CFO might say, so what about declining traffic now as long as we're the downstream KPIs are fine, I. E. Sales, etcetera? And I guess this is an interesting one to put to you specifically because your world is probably higher in the funnel to a certain extent, but maybe not. So so how are you handling those conversations either personally, internally, whichever way you want to take it? Yeah. Love this. This is such a great question. I have a lot of thoughts on this. I think it's very pragmatic to say we've lost all this traffic, but the business is still growing. Therefore, what was the value of that traffic? Let's move on. Let's not worry about it. My suspicion is that, you know, top of funnel traffic, it does not bring in loads of people in the short term who will buy your product. So you're not gonna feel the impact of this going away in the near term. Everyone that is still searching for a product like yours now, they will still find you because it's not impacted in the same way. However, I do I'm a big believer in this kind of halo effect of top of funnel content. There is a massive benefit to having things in the world that have massive audiences and expose your brand and your brand values and, you know, your company vision at a very high level to people. That is not gonna manifest in losing sales now when that goes away. But six months, a year from now, I think a lot of companies will probably see a decline in sales and customer acquisition on the back of that. There's just a lagging effect behind it. That's one of the reasons that HRFs we do loads of top of funnel stuff that is not just content. Like we have a conference now and an event you know obviously you know value of Joe. So yeah I think that's very true but also maybe slightly shortsighted and I wouldn't want people to get caught out in a year from now. Yeah, yeah, totally, totally. It's really, really tricky because you know we all feel like those demands are on us short term you know and often it is that that sort of thing isn't it you know is it going to work tomorrow? Yes, no, etcetera etcetera. There's there's a it feeds into the wider conversation here. There's a lovely question that comes through from from Zoe who's speaking about using Google Ads, for example, to counteract falling web traffic, and whether that's part of the strategy. And I guess it feeds into this wider conversation that we're having about sort of like top of funnel driving people further down. So in your mind, are you looking to compensate for this traffic or is it a different headspace? And I appreciate I'm asking quite a lot about principles and headspace more than I am specific tactics, but I think it's quite a big shift for us all to sort of go through in way. Yeah, I mean as a company we have definitely started investing more in paid advertising of all forms. Think kind of what we were talking about earlier, there was a playbook that worked very well. That playbook is now a question. We do have to then flail around a little bit to work out what the next playbook is gonna be. So a lot of these things we've dabbled in, are testing and dabbling a bit further to work out whether it is worth scaling it. I actually had this very interesting conversation with a very big company about AI visibility and their perspective blew my mind. They said we're not worried about whether we show up or not. We're not going to create any content. We don't care if we're in chat GBT because they will release ads and we will pay for those ads and we will get customers through that mechanism. And that was their entire approach to AI visibility. In some ways I thought that was the cleverest thing I'd heard, know, cut the Gordian knot, none of this uncertainty, they will just wait for ads and pay for ads and they will, if it works it works, if it doesn't it doesn't. But yeah I think yeah paid advertising is is worth testing for many companies I think. Nice yeah champagne problems right there. Let's take a step into the tactical a little bit more with this fabulous question that's come through. Question is, for new business websites and blogs, what would you suggest? So I guess it's quite lovely as a thought exercise really just to sort of think about blank sheet of paper, where should we take it? Does creating a lot of articles in a short time frame help or not, frequency of content? But I'd also say there's other questions here around types of content that you would prioritise. So maybe we just put it as a general scenario to you Ryan, if you're starting today, what would you be focusing on specifically if you're a new or smaller business? Yeah I loved thinking about this. I think I would probably sit down and I would work out what are the twenty topics I would want to cover even if they got no traffic from search, none whatsoever. What are the topics that are so important to me and my business? What are the questions I get asked all the time? What are the kind of principles and guiding ethos that I want to share with the world? What is like the one informational topic that I know better than anybody else that this entire business predicated on, and I would write that content. And almost as a kind of post hoc process, I would look to see what is the search volume for these queries, is there a way I can frame them around search volume. I think you will always need that content and I think it's a very good idea to think from first principles and not build it because of organic traffic because you cannot guarantee that now. And inversely, something I have seen get loads of websites in trouble, brand new domain, AI content, trying to cover this entire universe of related topics, publishing five hundred articles. That's something that Google can detect programmatically very easily, and I have seen many businesses get dinged for doing that and get penalized. So I would be a bit wary of scaling content before you actually have a product and some brand searches and all this kind of thing. Nice. I love that so much mate. It's endlessly helpful like even just how you frame her answers so I really really appreciate it. We've got a few more sort of tactical questions and I'll take this one from Lucy because you were very humble towards the end there where you spoke about personal brand. But as mentioned, I think you're a phenomenal follow. And whether you apply this to yourself or to content in general, do you see a difference between video content on LinkedIn versus text based posts? And I guess probably the wider question that I'm interested in right now is like format styles. Like are you posting a lot more videos on the Ahrefs blog, for example, you know, and sort of doubling down on video or text or infographics or whatever? Where are you seeing the biggest wins at the moment? I'm definitely trying to do more video and the only reason I haven't done more is that I'm lazy and busy and making excuses for myself. But there is a very interesting point where you know video is a medium that LLMs like ChatGPT cannot yet deal in that you know there are video generation models but you can't access video content on the fly in the way that you can text content. So all of our research we've seen YouTube videos, YouTube shorts, other video platforms, Wistia, Vimeo, they get cited disproportionately heavily in AI outputs because they deliver something that LLMs can't give to people otherwise. I think the same thing on social media. They want, they're all trying to become TikTok to some extent. They want people to engage with short form video and they boost it and they elevate it in the feeds as a result. So yeah, if anything I should be doing more video content than I am and I will try and do more every week. So we'll check-in mate, we'll check-in on you. Yeah, yeah. We'll start putting in your comments section on LinkedIn if it all goes wrong. Without meaning to feel like I'm coming across plugging too hard but Wistia has come out with some lovely features which are LLM optimized for video, which is great. Think that you point out. Like it's, yeah, it's an exciting space to be in and the platforms are reacting as well. Ryan, that's our hour mate. So it's flown by but like thank you so much for being so open and vulnerable as you were at the beginning, practical and tactical as you were throughout the middle and then so useful, with your Q and A answers as well. Just a really properly brilliant hour. Thank you very, very much. We've, we've got people like Matthew just saying top man, absolute top men. So thank you very, very much, Ryan. And thank you everyone in the community once again for being endlessly phenomenal. This stuff is a joy to do because of all of you. So it's a real pleasure to be able to show up and sort of feel like we're helping. We'll be back next week with something which is in a very different spirit. We're looking at how we can use AI but have an environmental sort of angle in mind and integrate it into our workflows. So that should be a really interesting chat. If you're interested, you are of course very welcome. So that's all done for this week. A big thank you to Wistia and every one of our sponsors. Big thank you to Ryan. Big thank you to everyone in the community. We'll see you, next week everyone. Take care. Cheers Ryan. Oh thanks.