Measure the impact AI has on your Traffic (US/SA)

before writing the actual content. Instead, you can scroll down to the SER overview table for a given keyword and click on identify intents. This will tell you in simple terms what people are actually looking for when searching for a keyword. Use them as hints as to what content angles to cover. You can go all in on the top search intent or target multiple ones in the same article. The choice is yours. Fast forward, you already wrote an article based on the previous keyword research. It's ranking. Hooray. But you can do better. What kind of optimizations should you start with, though? Just plug in the article's URL into our AI content grader along with its primary keyword and target country and voila, our tool will automatically identify the main subtopics covered by the top three ranking articles, compare them side by side and score them on a scale to see how well or bad you've covered them. You'll also get suggestions on how to best strengthen your article for each subtopic with real life examples of the words and phrases used. All right, you're crushing it with your English blog. Your next big step is to localize your content. Normally, you'd have to hire experienced and expensive translators, make sure the translated keywords are localized properly, and create a localized content plan. What a mess. With HF's AI translation feature, you only need to select a preferred language or dialect and hit translate. We'll not only translate your saved keyword list, but provide translated alternatives for each keyword, so you can choose the version with the highest search volume. Are you more of a technical SEO rather than a content marketer? Say no more. With patches, you'll soon be able to create fixes with the help of AI and push them by yourself, right? within site audit with no developers needed. And there are also a handful of other AI powered tools made by hrefs. Word counter goes beyond counting characters and words and helps you extract keywords and entities out of a paragraph for SEO purposes. And we also got over 40 standalone AI writing tools which are primarily designed for various copywriting tasks, not just SEO related ones. So go ahead, try these HFCI features and ace your SEO and marketing tasks today. Ever done a site audit only to find out there's a 100 issues to fix? So now you have to export issues, add fixing instructions, pass everything to your developers, and then have to hunt them down over the next two months to make sure everything need some help with your daily SEO tasks? Ace dome with HF's AI features. Let's start with keyword research. Quite challenging if you don't know where to start, right? Don't worry. Get unstuck with the Ask AI feature in Keywords Explorer. Just choose from a list of presets. Type in a seat keyword and you'll get a bunch of uncommon keywords that both you and your competitors might have never thought of. Did I mention this took us less than 5 seconds? The best part, you can also generate keyword ideas in other languages as long as you add in your desired language at the end of the initial keyword. Great. So, you now have a list of potential keywords, but you still have to manually check the top results for each keyword to determine their search intent before writing the actual content. Instead, you can scroll down to the SER overview table for a given keyword and click on identify intents. This will tell you in simple terms what people are actually looking for when searching for a keyword. Use them as hints as to what content angles to cover. You can go all in on the top search intent or target multiple ones in the same article. The choice is yours. Fast forward. You already wrote an article based on the previous keyword research. It's ranking. Hooray. But you can do better. What kind of optimizations should you start with though? Just plug in the articles URL into our Hello everyone. How are we all doing? Uh my name is Gordon. I'm a product marketer here at Hrefs. I am joined in the chat by my lovely colleague Michelle. Uh she'll be here helping uh fielding some questions you've got during the chat. As always, we will pick some maybe interesting questions throughout the presentation that we want to answer. Uh but generally we'll leave most to the end. Um so yeah, to kick things off, uh love to hear in the chat, uh where you're from, um what kind of role you're working in and uh what kind of experience you have. Oh my goodness, there is uh there is something over my face. Right, everyone give us a wave. Everyone feeling shy today? Not seeing a lot of the messages. Thanks, Matt. Marco from California. Ohio. Okay. Vitali, will this webinar be available for a rewatch later? No, this is a one-time thing. Then it's gone. Of course. Yeah, I'm joking. We will send you the recording as we always do and it will actually be on our YouTube channel for posterity for you to watch whenever you like. Philippines, Washington. Okay. Minnesota. We got a good um a good US contingent with us. Um and if you want to dump in there your job roles um or if you've got experience with HFS, that will be interesting as well. I think there's always a bit of lag on this uh on this chat, which is always fun for me. Okay. Well, while we're waiting on those conversations coming in, um I will cover what we're going to talk about today. Measuring the impact of AI with hrefs. So AI and search as we all know is a it's a very hot topic. I think we're all being asked to from different perspective what's our plan uh what's the impact going to be? How are we going to build strategies around it? And there's a lot of unknowns. But I think one of the core things sometimes we overlook is first of all addressing what is the impact? Is it having an impact and how can we measure that? And that's really what we want to address today. So the agenda for today is we're going to just talk about the great decoupling between impressions and traffic where people are actually discovering you because that those searches are not going away. People are still searching. um AI uh visibility tools and hrefs, how we can help you measure the impact and discover um different ways you're being uh talked about in AI tools. And then finally, we are going to touch on this, but it's not going to be a full webinar. We're going to talk about some of the ways you can get mentioned by AI. Now, this could be a whole series of webinars, so we do not plan to go into this into any great detail. We're going to cover um a couple of tips on this and then give you some supporting information. And there's a our our team have done some really good work on this and we will probably do a dedicated webinar on this uh at some point in the in the near future. Um, okay. So, I'm just going to jump back into the chat. We have a mix of content marketers, growth marketers, SEO specialists, account managers using HFS, and quite a lot of HF's experience. Okay, that's great. So, yeah, you guys are in for a treat. We're going to go through some of the tools um you can use to measure your impact on AI. So let's start with the elephant in the room. Um the great decoupling really it's the problem is that most of our Google search consoles have started to look like this. And this is what I think a lot in the SEO are calling the great decoupling where uh impressions that line in purple is going up uh but traffic that line in blue is going down um showing people are still seeing your brand but they're not coming to your website. And this is a big problem for a lot of us that have built reporting metrics based on the amount of traffic growth over time. Um and in some cases it's a crisis for some businesses that rely on that traffic for ad spend. So we're going to talk about how do we measure that? Who's responsible? Um I think we know who responsible is and what we can do about it. So the culprit, the big culprit uh amongst many things is AI overviews. They have become more common in across all sorts of search and basically people get what they want and they leave without ever entering your website. I think when they first rolled out, they tended to be for longtail keywords or more informational keywords, you know, um things about things that were really pulled from Wikipedia or history or how-tos, things like that. Whereas now you can see they're even appearing for relatively commercial intents like best baby carrier. And the problem is is that um you have no way of tracking this. So if you are Erggo baby, it's great for you that you appear on that AI answer and that's locked in my brain and maybe I then go directly to your site or maybe I visit the store, but I have no way of knowing uh whether that's providing any value. The or the um the the article that was used to site that is is performing. So it's a it's a it's a problem and a solution. Well, a good thing at the same time. So this isn't just based on vibes. This is based on some research we've done. We looked at over 300,000 keywords and we found that where AI overviews have appeared in the last year, you're seeing a 34% reduction reduction in click-through rates compared to last year. Um, AI when AI appears, we can predict the CTR drops from 0.04 to 0.02 is the actual numbers there in case you're wondering. And this is a massive impact particularly for big websites um across the web. I would say again that this is just a take. I think that um older generations um myself included people are maybe uh above above 34 are more inclined to click and look at the actual link. Okay, what's the source? Particularly if you work in SEO. I think um as we be digital natives grow up with AI overviews that will become less and less common and they'll be happy to accept the information at face value. So even just looking at the IO views, not anything else that Google's rolling out, this is a problem that's only going to continue and probably the gravity is going to continue to um get worse on this. So we know people are discovering you and we've mentioned AI overviews, but um where where else are they looking if all this traffic's gone? And the answer, of course, is LLMs. Um it's interesting because this is our LLM traffic and you can see it's spiking. Um, and actually this is now sending us more traffic than um, Reddit. And this is obviously a massive uh, boon for us, but we're missing a lot of the information that comes with Google over the years. We know we're being discovered, but we're not sure how and why or when. Uh, but we knew but we know we are getting LLM traffic. We will go into later um, what actually shows up as direct, but probably is LM traffic as well. But for now, all we want to say is that LLM traffic uh is continuing to accelerate. And I'm assuming most of your websites look something similar to this. Maybe not that extreme, but something similar. Give me a thumbs up in the comments if you're seeing the same trend. It's going to be bad if I get a lot of thumbs downs. Okay. So, what LLM are these coming from? We did a study classic age. We we did a study 82,000 websites and chat GBT dominates that with 83% of all AI chatbot traffic. It's huge market dominance. Now obviously I think this will change over time as perplexity claude um etc start becoming more common. ChatBT was first to market, but you can't deny they have a very strong foothold and it would be very hard for any of them to knock them off. Um, Perplexit obviously accounts for 10% and, you know, co-pilot um is in there as well. But this concentration really means you you can focus on, you know, three or four big tools. Um, and I guess within that there's a range of AI usage. Juno, you've got maybe your mom and dad who are just using it to answer basic questions and then within that you've got maybe people that are very techsavvy that are setting up um their own custom GPTs to do product research or whatever. So there's a lot of gray within that but the sort of highlight is chat GBT is a monster in this space already. Now how do we know this? And the the problem is attribution is difficult, right? Attribution is a fuzzy area. Um, many AI platforms do not pass referral information properly. Traffic from Perplexity or Chat GBT desktop apps often just shows up as direct. Um, web- based ones, maybe Chat GBT, um, Perplexity, Claude, they'll give you the referral um, element. But what that means is there's a lot of unknown. There's a lot of stuff that's just showing up as direct. And we don't have numbers for the sheer volume of searches that are going into chat GBT. We just know that if you like our traffic's going down, our impressions are going up, and there's something missing in the middle, and it looks like this is the culprit. So, what does that mean? That means your AI traffic is being vastly underestimated. So, it's not all bad news when you're, you know, in despair and you see that your search is declining, but your tra but your direct is sort of stable and starting to grow. that is likely attributed to people maybe seeing your brands in an AI overview or somewhere else and then going direct or simply you being an a response um in chat JBT and somebody giving it a click. So it's not all bad news. This comes with challenges and opportunities. But what we need to know is measurement. So I guess one of the things I wanted to address just because it was interesting in the number is when you hear about all these brands I think there's an idea and everyone knows that chat GBT GBT chat GBT um perplexity all these big data models they train on the internet so obviously if you have a lot more mentions and you're a big website it's going to favor you that's true but what I wanted to touch on is that it's also an opportunities for smaller uh websites so you can see here that actually um traffic by source LLMs represent um a greater amount of traffic for smaller websites than they do for bigger websites. Now obviously there's some scale in that that makes it relative but what it means is they are showing up for these websites and they are delivering traffic on a meaningful scale. So what it means that if you are a smaller website and you're small in your game, there is an opportunity for you to win. This is the early days of a whole new search method and so laying the foundation now is going to be important. You don't have to be a huge website to dominate this space but it all but it helps. Let's be realistic. Okay. So in short to summarize that we are looking at AI overviews reducing click-through rate and your traffic and AI tools and citing um AI tools citing websites and brands boosting your traffic. But we have some problem with attribution. So the big question, how do we measure this? And that's where we come into Ahref's AI visibility tools. That was a bit of a a needless slide there, but here we go. Okay, the first one we're going to talk about um is brand mentions in AI responses for brand radar. So the first tool I'm going to go to here is brand radar. Now I'm going to touch on the fact that you have four five data sets here. If you are on a light plan, you'll get access to search demand and web uh web visibility, AI overviews, chatbt and perplexity are paid add-ons, but you can do a lot with this alone. So let me just load this up. Okay. So, as we said, I think before we go into AI overview, something I want to touch about on web visibility. Um, your mentions, you're looking for your mentions across AI overviews. And how do you work that out? Well, AI tools train themselves on the web's database, right? So, if you don't have good exposure on the web, aka this in particular index here, you're not going to expose in the other. So, what we're going to look at here is your um competitive uh score. or how often are you mentioned? And then we're going to look at are you mentioned in the right terms. So if somebody in the chat would like to give me a brand, this is always an experiment. If it's a niche brand, it might be hard, but give me give me a brand that we will search for web visibility. Okay, what we got? No takers. can be any brand. Last last time we did this, somebody mentioned WordPress. Live it up. Alert media. Patagonia revolve. I don't know what Elgato is. I know what Patagonia is. Campfire trees. Bathify. Okay, I'm gonna I'm gonna pick one and roll the dice. Let's go for Patagonia. We can do another one when we do AI overviews. Um, so first of all, we're going to enter our brand here. And this is, as I'd like to remind you, in web visibility. So, what we're looking to establish here is across all of the web, well, within our index, however, it's 17 billion pages, how many brand mentions do we get? Now, this is nothing without some comparative data. So let's look at let's use the suggest more to generate some competitors. Let's get rid of Marmmont because the m the animals show up park tricks. Okay. So what this is going to show us without a graph for some reason is mentions over time. So here we can see the spike. We see Northace at a massive spike. Sometimes this is when when they do a press release or sometimes it's actually just when they release a new part of the website. we we crawl it uh more. Um but so North Face is in strong first position, Patagonia in second. And what we're going to do is we're going to put that as we're going to put that as cumulative to give us a better idea of the competitive score. So here we can see when you're looking at your website, okay, particularly when you're looking at your own niche competitors across the training data that LLMs have, how well or how often do I appear? It's a good starting point. If you're really low in this, it's a problem. you probably need to start producing more more content and we'll talk about meaningful content later. Now, the second one you want to know is because of the way LLM's train, they have this thing called semantic uh relativeness or similarity, semantic similarity. If your competitors show up in a list and you don't, that's a problem because what you want is when somebody says to um chatbt, hey, give me the five best outdoor jackets. And if it's mentioning Arctrics, uh, Mountain Hardware, because it recognizes Patagonia and as, um, semantically similar, it should pro prompt that as well. So, what we want to do then is find times where it mentions, let's just put down, yeah, let's put all of these. Mentions all of these, but does not contain PA. Now, what you would do here, now this is the web, this is not AI prompts, but you'd be looking at the type of content that you're being mentioned in here, and you'd either produce your own or you reach out to them because it's really important that you appear. See, there's there's only 34 results out of thousands. So, actually, the scale of this isn't that bad. You're Patagon, you're pretty well set up. But if it was the other case where this was thousands, it would suggest that there's a problem and it's not you're not setting yourself up um for victory in the AI overviews. Okay, so let's get rid of that. Another thing we'd want to talk about is um maybe the category. So let's say the Patagonia. Again, this is to do with training the LLMs and um semantic similarity. So what we've established now here is the scope of our mentions, how big we are competitively, where we are in the market. We understand are we mentioned in the same topics as our competitors um as particularly important ones. And now we want to say okay are we mentioned in the same in the right categories. So let's say for here I don't know maybe like let's say I'm trying to think of something a bit more specific because they will show up for lots of this hiking and maybe you would go into the filters here where it contains Patagonia and it contains hiking which will not be a hard ask I would imagine. Now again you have to imagine you've you've um got a more specific uh topic. So maybe a hrefs and something new for us like um AI overviews where you notice that actually the correlation between us and AI overviews in the tool is not very high that's a problem. If it was very high we're doing well. So that basically what we're what we're trying to do with this looking at web mentions is we're trying to evaluate um the foundations that we're building on you know because AI and um and LLMs are going to use this to generate their answers particularly new ones. Okay. So next we will move on to AI overviews. Right. Would somebody else like to suggest a new brand? I feel like Patagonia was a bit easy. so much on it, but we don't want to go too niche for this live demo. Don't want me me scrambling around trying to get the data to work. It's Megan. I love Thank you. Thank you. We love it too. The web visibility is a really unique feature to us and it's really cool because it is what LLMs are trained on. So, if you can analyze that successfully, then you can set one one more up the funnel and then look at AI overviews and start to optimize for that. But you can't really get one without the other. Nintendo, thank you, Matt. Nice simple, although I don't think their their latest console is doing particularly well. Atari, Capcom, that's a bit odd. Let's Let's uh let's go for something a little bit. We will put in Capcom as well. So again looking at so now we're on to AI overviews and this is the next step up. So we're looking to measure we know we're being mentioned but we just can't track it but how much are we being mentioned? So we want to know the impact again here if we are Nintendo we can see that compared to Xbox and PlayStation actually we're pretty well mentioned but we are I put Capcom in there but um we are the third runner in this in this particular race. And then we can go down and we can start to see the type of um prompts and responses that are coming up in Google's AI overview Street Fighter 6 sales. And again, what we can do here is the exact same thing um as we did before. We want to see where is PlayStation, Xbox, let's get rid of Capcom. Where are those two mentioned, but Nintendo is not? Does not contain. Let's go for all. Now this is particularly useful in AI overviews because again um we're looking to work out a semantic similarity. Why are we not appearing? What are the prompts that are showing up? Now in this case it might be yeah okay who is Xbox or PS3 release things that we don't want to but it might be more um niche things like Monster Hunter crossplatform. If that is cross crossplatform we should probably be appearing there. Again, you're going to have to use your imagination to make this relevant to your company. Okay. One other thing I'd like to touch on. Let's say, so in here, you don't just need to search brands. You can search topics as well. So, let's say um adventure travel is your topic. You know, whatever you want to relate to your business. Suggest more. What else we got? Backpacking, ecoourism. Let's search these. This is really good way to work out what is the most popular term that's being mentioned, particularly if you've got a range of different things in your company and people express things slightly differently. Um, you can pick one. Also, when you have this, it's really interesting to look at the kind of questions that prompt Reddit or how many they have. So, here we're trying to analyze again the mentions um impact, how how strong are they? um what are the top ranking um topics within um AI overviews? And now here we want to see is how many link to Reddit. Now what this can tell you is if out of the 8,000 4,000 have have a Reddit link. It would suggest pretty strongly that you want to have a strong place on Reddit. Now that doesn't mean I would say you go in and pretend to be a normal user and then just plug your company. No, it's a terrible idea. But have a have a have a Reddit have a subreddit of your own that's branded. uh invite questions, answer them honestly, and try and grow a bit of a presence because clearly it's very helpful and that will affect how people talk about you. But obviously, as you know, Reddit can go very badly when done uh poorly. Okay. Um I should mention that you can do these for any country. Um if you put in Spain, um let's say if you put in a brand, you'll just get the language um off that. So let's say you put in Nike for example, you'll get Spanish language uh results. Um but you can sometimes get a popular English English result as well in Spain. But yeah, you can filter by any country to see your data set. Um and just for fun, let's go to perplexity and I will take one more web flow. Okay, I see one more web flow. Let's see what's happening with web flow. What's being suggested? Figma, Shopify, Square, F. Okay. Yeah. Let's see what they've got. And you see the data is a little bit thinner here because it's brand new. But again, what we're seeing is the country. What is what is a CMS? And what's really cool here, and I think this is something that uh will probably develop into a feature at some point is looking at how your brand is mentioned. Again, we're talking about LLMs, as we know, are just prediction machines. They're just looking at what's the next predictable work. So we want to look at what are the words that are surrounding um that are surrounding your brand. Okay, in some cases it can be a description. Um this is not a great example included, webex. Okay, so many platforms offer um storage included like like Wix. So you know that if somebody's going to ask a question about platforms with storage included, Wix will show up. And a it's a it's a way case of also documenting the way you're described. That's going to basically build what we would call your AI positioning. Okay, so we have covered web visibility, the training data, AI overviews, how you're showing up in Google, and then a little bit about chat GBT and perplexity. Um, I should mention you can on AI overviews dictate uh whether this is just in the overview or the keyword. And you can also dictate whether the term shows up in the content, the URL or the header we can. Okay, Patrick. Great. Any plans to bundle AI options together in Brown Radar? I know that we're still in discussions as to how we package this. It's early days. Um, but I don't have any concrete information to tell you right now about how we're going to bundle this um, at the moment. So, do we have any immediate questions on AI overviews, web perplexity, chatb? [Music] Nope. And nothing popping up here. Let's move on to a brand new feature that we have released called AI citations. There's not really too much to it at this point, but it's a really great um way to get a snapshot of your competitor's presence in AI or your your own. So, if you have a verified website, even on a free plan, you will see this for your verified website. Um other than that, it's available for paid plans. And if you actually want to see the chat GBT information, you have to have the chat GBT um add-on. Um so for example, you're seeing here citations. And this is an important uh distinction. These are not mentions. These are citations. What does that mean? It means that uh your website is being linked as a source or your website entirely is being linked um as the answer. Um, so it's a little bit different, a a little bit um if you like more concrete than just a mention. So we're moving sort of down the funnel. We've discussed all the things you can do with mentions and now we're looking um at citations. It's really cool. This just takes you to the corresponding report which again is a great way to look at the keywords they are uh scoring for. You can hear see all the different links here that um in this case HubSpot are uh scoring for. And again, I would look at how they're being mentioned and then you can play about with the competitors um once you're in there. Okay, just reading through AI overviews are not exportable yet I don't believe. No, they are not. Okay, so that covers our AI AI citations um block within site explorer. Again, all you need is a verified website and it will show you this as an overview. Um but let's go to our next one as we are almost at half an hour. Um traffic from links shared by AI, traffic from AI bots, web analytics. Okay, so we know that they are sending traffic as we mentioned with the citations. There are um LLMs um sending traffic directly to your website. How do we track this? Now, we have a lovely web analytics uh tool. Um it's free to use. Um it's much lighter uh script than Google. Um it's also the data is a lot faster. I think we return data in like five minutes I think it is. And whereas Google Analytics takes about 24 to 48 hours to um process it. So it's a lot faster and we deliberately built it to be a little bit simpler to be honest with you. Um but we're constantly working on it. We've just added funnels so you can do quite a lot with it. But today we are talking about AI search. Now to find this in our tool is wonderfully simple. There's just an AI search channel source. I think you I you can do this in um Google Analytics with some configuration, but here you know you just have it set out and you click here and you'll see your sources and it'll break it down by the main um LLMs that are providing channels uh sorry providing traffic and you can see the pages they're driving it to. So if we you know pick let's just say chatbt what are the pages that are doing really well homepage and then a few um specific tools or blogs which is pretty cool. Uh let's go back because what also you can do it's not so useful at the moment but you can compare let's just share compare to search and direct you can compare your AI search uh to your direct um and your search so you can see when this is coming up over time and you can also get some information about bounce rate visit duration I think Google or released um where Google staff released a thing saying that uh a visitors coming from AI sources or overviews or or chat GBT um are higher quality. Um certainly the data doesn't support it at this stage, but we don't really have a lot to go off. But what this does mean is over time you can track that and particularly if you're seeing your search go down and you're seeing your AI search come up. Um that's a really good way to show that your work is not in vain and it's being offset. Um, any questions on the traffic sources? Timothy, sorry. What are LLMs? LMS are large lang I see. Oh, there you go. Large language models. It's an AI. It's what chat GBT and everything is. It's how they give you answers. It's just another word for uh an AI chat. Some pricing tweaks. Okay. Any other questions about web analytics before we move on? I can see there's a lot of people uh getting getting in there with the questions. Okay, moving on. So, how do you track your own keywords? How are they impacted? And particularly, what about keywords that aren't in our index? So, if you got super specific keywords, they're almost definitely not going to be in brand radar yet because that index is relatively new. Um, so we recommend doing that in Rank Tracker. Let me pull up the correct tab. Here we go. So, this is our rank tracker tool. Uh, most of you that have used HS for a while will be very familiar with it. This is where um you're putting in all your keywords um and you're tracking their position over time. Now, what else is cool is if you're seeing a drop in your traffic, it could be related to um the amount of AI overviews in your key keywords or important keywords or all your keywords. So, what you can do here is you can go to SER features, which is at the bottom of the overview, and you click on AI overviews, and it's going to tell you, oh, I've just undone that. It's going to tell you of all the keywords that you're tracking, how many of them have an AI overview in their SER, and then how many of them do you own. So, I guess whether you own the SER or whether you not, you're going to see a traffic drop, but if you don't own the SER, that's much worse. Um, so here you can see that we are currently at seven total and uh four owned. And you also see that they're not constant. Google is experimenting. So um if you go to the keyword list here, this is a little thing that says this tells you the fe the type of SER uh feature. You can see there's an AI overview. Here you can see we actually lost an AI overview. A bit of context, this is our help articles. So um it's not like normal competition you'd see in a SER uh but it serves for the purpose of this demonstration. So if you were to click on this, we will automatically compare uh with a previous SER and what you can see there used to be an AI overview that we held um and now we well we actually it's the same article that would have um appeared in position two and the AI overview we lost it but now we just have position one. This is more interesting obviously in the case where um you lose the AI overview and a and a competitor then takes it and you start to look at um their content and what they've done right. Um, so that covers tracking your own keywords that have AI overviews to understand the impact particularly if you've got, you know, keywords that are contributing, you know, maybe some in some cases five, six% of your entire traffic if you got really important ones and then all of a sudden an AI overview pops up um that can be responsible for a big dive in traffic. But what if you have really longtail niche keywords um that just don't show up in um brand radar? Quite simple. you add them to um keyword tracker. Um you can do so for example maybe we had something super specific like how to use brand radar I think it was the one constant suggested just add the keyword here and then depending on what plan you're on either in a few days or in a week depending on how um how often they update um that um keyword would appear here and you'd be able to track and the same whether it's an AI overview or not. um really useful for longtail keywords specific to you that are maybe very close to purchase as well. So anything that any sort of intents that you know around your business that are people looking for a solution uh but aren't obviously um obvious to the um to the standard buyer. Okay, that covers rank tracker. Do we have any questions? I think I see Michelle is on fire in the chat covering most of the questions which is great. Okay, we will move on. so let's recap what we've covered. You're getting mentions but less traffic. Okay, you can track its impact by looking at web visibility. Do you have the foundations good enough to actually show up um and in chat in AI chats? If so, where are you featuring in the right kind of articles? Are you featuring next to the right kind of terms? If that's a yes, then you move on to AI overviews. Okay, we're in AI overviews. Let's get specific. Uh what are they saying about us? What brands are we showing up next to? Are we appearing for the right uh topics? Then you do a topic search. Are the are the topics that we are betting on? Let's say it's um I don't know uh summer hikes. Are summer hikes trending or no, it's actually autumn walks. you know, you can start to build a picture of the trending um topics as well as your mentions. Then off the back of that, we went and looked at, okay, so we've covered mentions pretty clearly. We're then looking at uh the traffic to our website. Pretty simple with um our web analytics tool. Again, I'll mention it's free. Um and even there, you can sort of re reverse engineer um what's working by looking at the pages that are getting uh traffic for that. We know it's not perfect because of referral issues, but it's pretty close at the moment or as close as we're going to get uh for any of the competition out there. Uh and then rank tracker is the way that you um are seeing AI um overviews impacted on keywords that are very important to you and ones that maybe we don't have in our index or brand um brand radar. Okay, so that leads us on to the big million-dollar question. How do you get mentioned more by AI? Now, this is not the main topic of this conversation. Um, but we will give you some snippets of findings and then send on some information after. So, first thing, sounds obvious, don't block LLM crawlers, but you'd be amazed. Now, obviously, this is a personal choice. Sometimes there's a lot of strain on your site and you you have to make a decision. Um in fact we see that uh chat GBT and Claude a CCB bot are getting blocked by nearly 6% of top traffic sites. That that may not sound like much but it's a massive shift particularly how publishers work. And there's this obviously idea that you want to gate your content so that chatbt isn't giving it away for free. And there are circumstances where that might be correct. But for most of us, we are kind of, let's say, putting up blockers for oursel. So blocking these crawlers, I always like to think it's like closing your shop door because you're worried about window shoppers. You know, they're not going to buy now, but they might show up in the future. What you're doing by blocking LM crawlers is you're well, first of all, they've already crawled your site. They know the basics, but you're blocking them from new information. And so if you uh change your information about your brand, your product or you have got great content, you're not letting them in. So really, you're um stopping yourself answering the questions uh that might be useful to your users and being included um in the short list. And I think it was a really good product marketer that said that the main problem brands are going to have now isn't um fighting for who's got the better features. it's showing up in the initial short list because so many people are going to do their initial spec on um on AI tools. So, it's not um it's it's not a perfect system, but it is an opportunity. By letting AI in there, you're giving you're improving your visibility and you're giving it access to your fresh content, which is obviously your most recent product and um positioning. Um okay. Um, but there is obviously a little bit of LLMs kind of ignore your blocks anyway. Uh, we wrote an article on that. I think Michelle can share it. Um, they're not the the most adherent to uh your LM LMM.txt. So, don't block crawlers if they're not breaking your site, that is. Next up, okay, here we looked at what correlated to showing up in Perplexity Chat GBTI overviews. um what factors were highly influential. Um it might not surprise you but it's about creating deep comprehensive content. So look at I mean look at the correlation of the data for chatbt AI overviews and it's very clear that word count and sentence count are really are really important. Now that doesn't just mean long words and count sentences. That means that AI tools like like depth. Okay. In reversely, the flesh score, am I probably pronouncing that incorrectly, which is a readability score, um, shows a negative correlation, which actually might suggest that more sophisticated content actually performs better, more academic content. No surprising. Um, and finally, domain domain rating. Um, it favors content from highly authoritative sources. No surprises there. I think the key insight is is that the basics don't really change. Google spent a long time trying to work out what was good content and AI isn't going to overwrite that anytime soon. So, writing deep authoritative content is still going to work. Um, interestingly, if you look at the bottom of that chart, you'll see things like uh back links and keyword count um are actually negatively correlated. Again, we would take this with a pinch of salt, but what it means is that it's not like a traditional um search engine result um where you know the top the one with a thousand back links is going to show up. It chat GBT um perplexity AI overviews, they're looking for different factors. That doesn't mean they're worthless. It just means that they are not um if you like at the top tier of um factors they are looking at when analyzing content. Now, this is a very early space. This is very early studies. This might change, but I think saying generating wellthoughtout, wellressearched content is going to be the future. Think less top tips, think more everything you need to know backed by data and experience. Those short articles or top tips, so those quick how-tos, chatbt can write that itself. What it can't do is write things from an author from from industry experts where you can cite them or leading or brand new stats. that's where you're going to add a lot of value. So, no hacks there, but um interesting to know. Um and I think this is summed up really well by our CEO, Tim Solu. So, he shared that we got 14,000 new users from Chat GBT. And the kicker is we didn't do anything special for Chat GBT. Now, I know you're saying AHFS has a big content team and we've worked hard it for a long time. That's true. But what we're seeing is that all that work we did in content really translated over well to this new AI era. Now there will be challenges we face but in creating great tools for marketers. Basically start focusing on the product. Why? Because word of mouth as it has been for the past 20 years is still the most powerful tool in any marketer's um arsenal. So create something that you're proud of that people will talk about and there you're already on to a winning start. We published tons of content um addressing different problems. We we wrote really in-depth studies and guides that people then talk about and link to again quality. And then probably finally as we educate our customers um and support them throughout it. So they then become experts and they then start talking about HS. Again, this is one of the ones in marketing that's always really hard to justify because marketing is sometimes a numbers game and you you know, how do you justify the the value of uh well, we're going to launch an academy that's going to teach um uh all our customers to be power users and it's going to take three months to build. Sure, I don't want to have that argument either, but if you can start laying the groundwork um for these things, it's really good. And I see that Michelle has mentioned Black Hat SEO. You got there ahead of me. So that was uh Tim's take. Moving on to um Ryan's take who is our director of content. His take is it's just it is just SEO. But I want to add a caveat to that that Michelle has just added. We are now much like the beginning of SEO the early days. There are tricks and hacks you can use to gify the system a little bit in your favor. um which we've got an article on called uh LLM is in its black hat era. However, what I would say is that when people built massive, you know, link um profiles on private blog networks um or they invested heavily in keyword stuffing, eventually all that content was devalued or those links were devalued when Google updated. So I would not be surprised if we get to a stage where the same things happens with chatbt and perplexity and they become more sophisticated. So by all means read about the hacks. Um enjoy them, maybe even try them, but do not expect them to be the source for long-term success. Like all things, it's about brand first and then quality content. You know, we do not want 10 things you can an SEO should know. We need an in-depth thing from an SEO who's been in the game for 10 years. That's the kind of twist. Um, and producing them over time will guarantee your success because quality content is quality content and it's going to be mentioned in other places. The distribution channels are going to work. So, that's within your control and something you should do. Now, that doesn't mean you shouldn't look at distribution channels and featuring on Reddit, etc. These are all tactics, but the core of if you like SEO and content still remains here. That's our take. Like I said, we actually have a um a YouTube um done by my boss Andre um on how to write high quality content that will rank in um AI chats. I think I've got it here. If Michelle if you could share that in the chat if you can find it um that's a really good one as well um to to look at and the caveat is to end end all this end all this dialogue this diet tribe is that AI traffic is still a small slice of the whole pie for now we look at these numbers search direct social you know they're 85 86% for search and direct alone LM's a 0.1 one. Now, this number is going to grow and obviously we have a lot of unknowns with um Google's pure AI um going to be coming in as well. Um but for the time being, it's not panic stations. It's double down in in in good quality content, high quality writing. Think about your presence and mentions and how you're being mentioned, but it's not panic stations yet. and you should be able to measure the impact and uh and grow your visibility on AI. Right? That um brings us to time on the large chunk of this um webinar. So we have about 10 minutes where I can field your questions. Let's see. All right, we'll have a look and see what's in here. Okay, folks. I don't Michelle, you have been on fire as I don't see any unanswered questions yet. See, Patrick Okay, folks. There was a little bit of a lag on the chat, but if you have any final questions, we can address them. I see you, Michelle. I think it's all just a bit hazy now. It's a lot of information. It's a lot of tools. Um, you know, even as a builder, there are some unknowns. But what we do know, um, is that particularly things like how your brand is mentioned in the words that surround your brand are very important. Um, and you want to control that as much as possible and put out content that basically, how would we say, presents your brand, your products in the light that you want, for the correct terms that you want. And then if it's very authoritative and very good, you'd hope that um others will pick it up and they will rewrite it. If not, you can invest in actual partnerships where you can get people to write about your brand and that that um content as well. Distribution is going to be key as well. One thing worth asking, how are you thinking internally about uh the personalization aspect? Patrick um assuming you're talking about personalization you are referring to the fact that chat tbt will take into account your previous um chats and give you serve you answers based on that. As far as I'm aware we're treating it um like Google in that we are the average user who has nothing and we will give you the average response. So the actual response itself is more of a bell weather than it is an exact because everyone will be different. I would add to that though is that most people are free. Most people have chat that use chatb no most chat GPT users are free. Therefore these results are probably a lot closer to what they see than a paid user and they are the bulk of the audience. Um I hope that answers your question. Uh, is it fair to say that AI works a little bit like traditional PR? Yeah, that yeah. Um, theory digital I think it's it's not dissimilar. Um, there's a lot of agencies pushing now digital PR. So providing um mentions in authoritative websites. I think the only difference is it's not just a mention, it's that you want to be mentioned with your specific positioning and then you want to get even narrower. Like if you have a product within your suite, you want that mentioned. So it's it's probably I would say it's a cross between partnerships and digital PR. But yeah, there's a there's a a very big crossover in that. Uh, let's see. Gordon McCllum, can you only get free user data, not paid GT data? Correct. Uh, Gordon, I would assume that's the case, but I will double check that with our devs. Um, but yeah, I would think that's a fair assumption to make. I don't think they're going to give us access to the individual queries of paid accounts. Um, but I will double check that and I can follow up with you. There you go. Michelle's on it. I'm assuming we're talking about Patrick the Oracle, the man who sees everything coming. If he's on it, don't worry. We'll have a fix very soon. Okay. Uh let's see. I'm just going through the rest of the question. Is there an API for HF's analytical tools to plug into a dashboard for example? Yes, we do have an API and we have an integration with uh Looker as well that you can build a Looker dashboard. Um I'll sure Michelle can send over our documents about what um endpoints we have and what data you can extract for now. I hope at some point there will be a search console tool for LLMs even if it's just yeah that would be very helpful particularly if we can integrate it with hrefs I think this is the way it's going that's what everyone wants to get but this is the early days and I think there's a little bit of panic with SEOs but if anything it's a big opportunity we are now again in in the wild west of search um so there's a lot of opportunities to lay some solid foundations um for further growth. great chat, guys. Really active in it. I'm going to see if we can save this. Some really good suggestions, some really good uh questions as well. If there is nothing else, we can wrap up. I'll give you 30 seconds to put in your last question. Anything pending? And it looks like we will finish. Oh, we got one in. That's Michelle. All great. Okay. Well, folks, we're going to wrap up and you will get five minutes back. Thank you so much for your questions and time. You've been amazing. Um, and yeah, if you have any questions um that we didn't address or any follow-ups, it's got your brain ticking. Uh, please feel free to reach out to us. You can find us on LinkedIn. Um, just search my name. Um, and yeah, we'll be happy to field um your questions and followup. So, once again, thank you very much and have a lovely evening and or day. Bye folks.