AI Mode vs. Top stories: Where news SEO wins (and loses) in 2026

Um, we're talking about AI and uh the whole theme today has been about AI and I'm going to talk specifically about AI in Google search and how that has impacted news publishers. So, a lot of that stuff you'll probably know about that. Um, but I'm going to reiterate a little bit as well to sort of set the scene in a set of context. Um, AI overviews is of course the big place where AI is integrated into Google search. We have these summarized answers at the top of a search result. They are created using Google Gemini large language model multi-step reasoning capabilities not 100% accurate. I don't know how many people have mentioned hallucinations so far before, but I still regularly see them in AI overviews and I'm more than mildly annoyed because they are presented with a lot of confidence at the top of a search result, but that confident confidence isn't necessarily warranted because you know we call it AI and they sort of become standard parliament that hey this is a AI tool but they're not really intelligent in a meaningful way. They're word predictors. They're very good word predictors but they're still word predictors. They don't think, they don't reason, they just predict the most

likely next word, most likely next sentence. And they have some recurring u um algorithms that makes them make them they're only smart in the sense that a dictionary is smart. They don't create new knowledge. They don't create new insights. They just regurgitate what they ingest. So that's also why they often get things wrong because they just predict words. And sometimes the words that are most likely to follow aren't necessarily accurate in any way, shape or form and not reflective of the reality that we live in. You cannot prevent hallucinations. I even asked Google and got an a overview of whether it's possible to prevent illucinations with others. The answer is no. It's just inherent in the technology. And that's something we do do need to keep in mind whenever we start relying more and more on AI to present us with the information. And especially when it comes to things like news stories and news events, if you get an AI summarized or AI generated answer, it might not actually be entirely accurate of what actually happened in the world and what the news websites are reporting. Of

course, uh, search engines like Google use grounding mechanisms, uh, where they perform searches in the background, summarize those to feed into the generated answer. This is, of course, by now should be all familiar to to most of the audience. Um, they're quite good at attributing sources. So, you get a source there often uh in the directly in the A overview list on the side. If you expand the A overview, you get a lot more sources, one for every particular section of text and content and then a full list there on the the right hand side or if you're on mobile in the form of a carousel at the bottom of an AI overview. On average, most A overviews would have five or six sources there. Very very rare when an AI overview does not show any sources at all. So the average number of site sources is about six, but keep in mind that less than half of all those siter links are actually ranked on the first page of the search result. Uh they cast a bit of a broader net uh than they do uh with classic search. Um available now in over 200 countries, 40 languages, there's a whole list of them on a Google support

document. I think pretty much almost everywhere where Google is uh active as a search engine, there will also be AI overviews present with very few exceptions. And the exceptions generally are for legal reasons where the particular nation's um information office is perhaps a bit fiercer against Google's integration of AI. This is the interesting bit. When there's a top stories box at the top of a Google search, so it's it's what Google would classify as a news query, we tend not to see an AI overview. There are some exceptions and those exceptions are generally when the news is what we say relatively old and by relatively old I actually mean a few hours like 6 to 8 hours old and there's nothing new that has emerged in that 6 to8 hours. So it's basically a one-off news event within a few hours Google will be able to then summarize that news event in an AI overview and you might get an AI overview on the same page as a top stories box. I've seen this happen in

different contexts. Sometimes the AI overview is below the top stories box, a bit further down the search result. Sometimes it's it's just above the top stories box. Um, so it does occasionally happen that you see news boxes as well as AI overviews, but generally for nearly all search results where there is a news box, there will not be an AI overview. And I think Google does this on purpose. It's a very deliberate decision that Google has made to avoid what they call hard news when it comes to AI summaries. And that's because until the AI has ingested all the content related to that news story, they cannot accurately summarize what is happening. And that ingestion is a bit of a delayed process. It's not real time. It's not as fast as classic indexing, that's for sure. So Google says, "Right, we're just going to show top stories and then maybe later down the line if there's nothing new happening, we can show an a overview summary of that news story." But initially when you publish an article on

the news website and you you aiming for top stories visibility, which is where most news publishers would get the majority of the Google traffic from, you are not competing with an AI overview. Of course, you are competing with AI overviews for anything that's not direct news for explainers, for background articles, for evergreen articles that you have. You are competing with AI overviews. And of course, news publishers also publish this sort of content in very large scales and therefore they are feeling the impact that AI overviews have on their traffic. Um, there's some tools out there you can use to monitor how often a overviews appear for your classic searches. I quite like what Cisco is now doing with their AI monitoring tools and they also see which particular queries trio and AI overview and whether or not you're cited as a source for those AI overviews. So that gives you some data at least to look at where the impact is in terms of the A overview potentially stealing your clicks, stealing your traffic or using you as as a cited source. When it comes

to optimizing for those A overviews, um we can we can talk about this often. People generally don't click that much from an AI overview to cited sources, but you still ideally want to be part of it, at least get a chance to get a click. I quite like that a overview actually cast a bit of a broader net than regular search results. It's not just the 10 blue links. Several different studies, this one from zip tie, for example, shows that a lot of the cited sources come from outside the traditional search results and they tend to focus on on clusters of content. You know, you have a cited source here for pain relief related to tooth extraction. you get sent to a cluster of content, a particular paragraph of content on a web page that that talks about this. Similarly, best polish delicacies. I got this screenshot because several months ago I was in Poland for an event and I wanted to know what kind of food I should try when I was there. You click on any of these cited sources. You get sent to listical pages specifically to the section of that listical that talks about that particular delicacy skipping the rest of the page. I find it quite interesting. This one is another

example. Inlay souls pain relief. again you get sent to a paragraph of content related to that particular aspect of the query. So this leads us to the the the chunking optimization tactic. I'm a little bit ambivalent about this to be entirely honest. I don't think we should necessarily restructure all all our web pages just to tap into these fairly feal citation mechanisms. But if you want to if you want to be shown to do busy work to your bosses to to your clients from hey we're targeting LLMs we're targeting optimization for AI overviews this is one thing you can do it's probably not going to hurt your classic SEO it might help your efforts to get into an LLM as a cited source but you know as we see from the data the clicks aren't really there and the classic tactics all still work you know you want to be visible in a overviews you got to do your SEO properly it's as simple as that and that's That's a dead horse I'm not particularly going to be beating today. The question is of course what's the impact on traffic both for news websites and general websites.

Google says there's no direct impact. Um although they are a bit clever in how they phrase it because they basically say total organic click volume from Google search websites has been relatively stable and click quality has increased yada yada yada. They also say that any third-party studies that show otherwise are based on flawed methodologies, isolated examples, or traffic changes that occurred prior to the roll out of AI features. So basically, they're saying, "What are you talking about? AI overviews don't lose you any traffic." No, it's all fine. It's great. We're just going to send you more traffic and it's going to be better traffic. And all those third party studies, they're wrong. you know, third party studies like this one from similar app that shows the amount of zero click search is going up quite spectacularly since the introduction of AI overviews or this particular study from an that shows the click-through rate on non-branded queries go down quite a lot after a overviews has launched or this one from a refs which shows an average reduction of 34 and a.5% clicks after a

overviews were introduced or this particular study from uh data from authoritas published by press cazette which show a significant ificantly lower clickthrough rates on pages that show an AI overview or this one from Pew Research, which guess what shows the exact [ __ ] same thing. So yeah, Google is trying to gaslight us. Basically, AI overviews cost us traffic. It's as simple as that. users do not click from air overviews to cited sources as often as they click from a classic the list of 10 blue links which makes perfect sense because again a overview provide the answer more or less provided they're not hallucinating directly in the search results. So why would you want to click through to recite the source unless you're not entirely sure that the answer is correct. So for users, air overviews are very good. They're very useful and they don't necessarily need users to click through to to decide the source. So of course they cost you traffic and no matter what Google says, it's just the reality we live in. And we see that everywhere all over the world. In fact,

we have Google search console data showing this exact same thing. The famous crocodile graphs where you know we see the impressions go up or stable because we get cited as as a source which is an impression but we don't get the clicks. people don't click. As simple as that. So yeah, they're being deliberately deceptive and but you know, they hold all the cards really and there's not an awful lot we can do about that. So yeah, a overviews cost us a bit of traffic. However, it it can vary quite a lot depending on the context of the AI overview. The data I've seen from search console profiles from a lot of my publishing clients is sometimes fairly minimal, especially if it's a very news focused publishing client. That shows maybe a small decrease in traffic to some publishers that focus more on not necessarily hard news topics, but more maybe on entertainment news, celebrity news or just uh you know lifestyle stories which aren't really classified as news as such. they get a lot bigger hits from AI overviews. They lose a lot

more traffic. Basically, the good news is that AI does seem to generate more search queries on Google. We see a correlation this on 23 to 24 with usage of AI and usage of Google. And I think and I think a lot of other people think have jumped to this conclusion that heavy use of AI chatbots like Chat TPT encourages people to use Google more to doublech checkck what the chat bot has been telling them. And so we might see a bit of an offset of the loss of traffic by increased search volumes. However, I think the loss of traffic is still very clear and very measurable for most websites out there. This continues in 2025. The latest earnings call uh Sund said that AO will drive meaningful query growth. So hey, more queries, more searches and incremental total query growth for search. So Google's still growing as a search engine. However, that I do strongly believe and this is sort of reflected in the way that Google themselves talk about this that Google doesn't send as much clicks as a percentage of the total search volume to

the wider web. So, AO views and news. Top stories still win. News does have a shorter shelf life. You got a smaller window of opportunity for news stories if you're targeting top stories because at some stage you're going to be replaced by an AO overview. Evergreen content you just lose. Most evergreen search queries will have an A overview at the top. Um, Google traffic in general is down and for most news publishers, if they're just maintaining traffic, they're probably winning. Having said that, it is still by far the largest single traffic source for any news website out there. And if you want to jump on the AI bandwagon and stop investing in AI, I'm I'm pretty sure that all your competitors are going to be absolutely delighted by that because you're just making it easier for them to steal your clicks away from you. You have to invest in SEO. You have to keep investing in the channels. Yes, it's going to be harder to get the traffic, but you know, ignoring it is not going to save you. You got to just work harder at it. Now, I want to talk about about AI mode. AI mode is an interesting one.

Of course, get clicks, dive deep into AI mode directly from an AI overview. Now, Google is actively encouraging us to use AI mode. You go to that, you get an interesting summary page, which is, you know, I'm an AI skeptic. I think nobody will be surprised by that. I sometimes do go to AI mode because it does give me useful information, especially if it's more of a a detailed question where I know to get all the answers I need, I might have to visit multiple websites. And in AI mode, I don't have to do that. I don't have to engage with multiple websites, click away their cookie consent forms, their tolerate their their email newsletter overlays, and get bombarded by irrelevant ads. AI mode will just give me all that same information without any of the friction of having to engage with websites. And I'm saying this as an AI skeptic. You know, more importantly, AI mode is not immune for news. This is an interesting one, a Ukraine pre-deal. Of course, you get a top stories box on classic search. You click on the AI mode tab, you get a summary of everything related to the news there. So, you can use AI mode to

get the latest news and you don't need to engage with news websites. Again, like with AI overviews, AI mode will show site sources, in this case, news websites there. And the question then is, do people actually click on those sources? And the answer is yes, no, they don't. Uh, there was an interesting fairly small sample size study done not too long ago by Kevin Indick from Growth Memo. But I do find the data interesting because it sort of correlates with what I'm seeing myself. People just don't click. They just they just don't click. When you're in AI mode, you tend to stay in AI mode. You don't need to click to websites. And that's also partially because the AI mode answers are good. I think they're better than AI overviews, less prone to hallucinations. And I think that's because they tend to summarize what they scrape off the web rather than auto to generate an answer and then try to to ground it. And you can just answer ask it follow-up questions and stay in AI mode. So the outbound clicks generally are just very very low. And if there is an outbound click, it tends to be commercially incentivized click that somebody wants to buy something which of course is not relevant for news websites. News websites just don't get

any clicks. um widely available as well since October in the pretty much the same countries that a reviews are active. News like I said is not excluded. The good news is that adoption rates are fairly low. Uh 75 million daily users is what Google's giving us. But that means it's a fairly tiny percentage of the two billion daily users that Google sees. Google is trying to encourage users to use AI mode with that dive deeper in AI mode button under an AI overview and also by making it a prominent feature on the standard google.com landing page. However, if you press the enter button and you type in a search, which most people will do, you get the classic search result. So, they're not making it the default search just yet. And in fact, I don't think they will, but that's a different rant which I will not go into too much today. I think it's too risky for Google to make AI mode the default at this stage of the AI wars. And then we have the web guides. This is

a quite an interesting one where they cluster a search result um into different aspects of the search. So you're basically combining all the best of AI overviews and AI mode and classic search in one search result. The cluster of articles will organized by subtopics and you get a short summary at the top of each cluster followed by classic blue links. and they again they don't seem to trigger when there's a top stories box. I have not been able to get a webcast search result for news items. Um which I think is quite interesting as well. I had an example not too long ago where I did a query where I wanted to know why Graham Wry left Monster Rugby. He was a rugby coach. Monster Rugby is one of the the Irish the four Irish provincial rugby teams. And last last season Gra suddenly departed mid-season as head coach despite the fact he had a pretty successful season so far. And then I I got a web guide search result on my mobile. So of course as typical SEO I immediately took a screenshot of that. So you get the quick matches at the top which is classic search and then you can expand that to get more classic search result and below that you got the web

guide search result which are just different sub questions basically like you had the official statements the analysis the fan reaction and all of those with a few cited links which you could expand individually to get more source links and as a user I found this quite useful. However, at the same time, I also found myself less inclined to click through two sources as I would have been in a classic search result because a lot of it was already captured in the summaries. But if I wanted to deeper dive into, for example, what the fans thought about the the coach leaving, I could go to sources quite easily. So, I found this quite an elegant solution. I found it a nice little middle ground of AI and classic search. And officially, it's still an experiment. However, unofficially, like I said, I've been seeing these and I'm using an iPhone and Safari, and I've been seeing these pop up quite regularly in Google searches that I do on my mobile. Um, because it's still officially a lab feature, we don't really know what the traffic impact is,

but I don't expect it to have as big of an impact as a overview or AI mode. I do think there will be some impact because like I said the summaries are quite good at providing details that you might might be the only thing that you need. So you necessarily need to click through but it does seem to be formatted in a way that that you know the summaries are quite short and if you want to read the detailed stuff you will need to go to those web pages. So I think you know it's it's a nice integration. I I suspect this might be very similar to what the new default search is going to be like in the near future. I think both Google is happy with this because it still allows for accurate advertising targeting and providing links to web publishers and search results so that they can get the regulator software back uh while still being seen to be using AI in their search results. And I think users quite like it. Me as again the AI skeptic user, I quite liked it. So I was quite happy with with the web guide results and from a user point of view, I

was really pleased with the kind of experience that it gives me. Google has a completely uh complete documentation of course on AI features and your website. So keep an eye on that for the latest developments. In a nutshell, we are seeing a replacement of classic search with AI enabled search features. AI overviews are an evolution. Basically, featured snippets on steroids. If you're a hard news publisher, you're pretty safe. But your peripheral content that is not necessarily hard news will suffer from AI overviews. Circumvent that by building direct brand loyalty if you can. And there are ways to increase your citation probability. AI mode is potentially a real traffic killer. I don't think Google wants to kill the open web. I genuinely don't think they want to. So that's also why I don't think they're going to roll out AI mode as a default experience because if they do that, it would would kill the open web. They just would. And I don't think it's in Google's best interests to do that. Um, web guide is the nice middle ground. I see that sort of as the future of search where we will get uh that nice middle ground of having AI be part of

the search result while still offering very clear links to sources presented in such a way that users can very easily find exactly the kind of information that they're looking for and it would mean slightly less powerful impact on on traffic as well. So no surviving the AI first era which we're now arriving into for news publishers it's very much against trying to im immunize yourself against whatever the tech giants are up to and that means you have to have loyal audiences that go directly to you as a website that circumvent the platforms or type your brand name in the platform in which case they don't have a choice but to show your website and it also means you have to understand what your audience wants from you. So talk to your audience. You know, there's one of those so scary things. A good friend of mine, Andy Jarvis, he runs a marketing consultancy here in Northern Ireland, Global Clients. His mantra is talk to your customers. And I don't think enough websites or businesses do that. They don't talk enough to their customers or they just broadcast to their customers and don't listen to what their customers have to say. And I think, you know, if you want to survive, you need to understand what your customers want from

you. And that means listening to your customer. Multimodal content, of course, uh it's not just about classic articles, text articles anymore. News hasn't been about that for a very long time. So, get your videos out there, get your podcasts out there, get your newsletters out there. Do whatever your audience wants you to do when it comes to formatting your news in the way that they like to consume it. And yes, brand awareness, brand power. Um, if you are a commodity publisher, as in your target audience wouldn't necessarily miss you if you were gone because they can just go to another website, you're going to have a bad time in the near future. I'll give an example. I work with several Formula 1 news websites and some of these I have to actively look at their logo to remind myself which of my client websites I'm actually on cuz they look similar. They have similar content. They have similar stories and there's nothing really that sets them apart. If half those Formula 1 news websites were to disappear tomorrow, nobody would really care except maybe the people who work for them. And the audience who wants to read that content, they'll be fine. They're just going to read a different website.

they don't have sufficiently distinctive features or if they have distinctive features, they're not emphasizing them enough. They're not showcasing them enough. You have to showcase what makes you unique, what makes you different, and use that to try and build that loyal audience and bind your readers to you. Don't be a commodity publisher that just writes what everybody else is writing because you're probably going to disappear in the next two to five years. I have a newsletter. I very occasionally blather about this stuff. I write maybe once every six to eight weeks. So if you're want to stroke my ego and subscribe, I'd be very grateful. Um, and that is all that I have. I wonder if you have any questions or if this rant was sufficient to get people to shut up, Barry. So I took from this entire presentation that you firmly believe that uh Google's right in the increasing queries and increasing traffic to websites, right? That's what I took from this. It's always great to be able to see um and and and see your picture and your viewpoint of news publishers. There's a number of strategies that news publishers are not adopting yet. And

we've talked about this on our show a number of different times. I want to throw a couple things. You slammed AI overviews for quoting out ofate stats like English speaker examples for newsrooms. What practical rules would you add around date stamps or data sourcing or corrections so L&M's stop latching on to stale and decontextualized numbers? You can't prevent LM from doing that. Um they will make stuff up if they don't know the direct answer. Um and if if they if you provide them with the answer, then they're just going to show it and not send you any traffic. And for publishers that are focused on acquiring traffic, why would you want to do that? Again, I'm a bit amused sometimes at the rush that the SEO industry has to optimize for AI search. I'm like, why the [ __ ] would you? It's not going to earn you traffic. Just you're helping the LLMs becoming more accurate and and becoming much better at providing answers. People don't click on those site sources nearly as much as as maybe you want them to. And yes, we'll talk about this top of funnel stuff and and

brand awareness, all of that. Great. >> What's the ROI on that? How do you measure that? And people are proclaiming the death of search and the death of SEO. Like, oh no, it's all about LM. It's all about AI. Oh, I'm sorry, but when I look at my traffic graph for my clients, that's a load of [ __ ] [ __ ] >> Winning by by a significant margin. So, please just keep investing in classic SEO. >> Abs. Well, you did say AI isn't the root cause of the crisis. It just accelerates weak strategies, right? >> Abs. Absolutely. Absolutely. I think a lot of publishers have to be honest with themselves and and look at the data and say yeah Google traffic has been on the decline for a while and that is not necessarily Google's fault. I do think Google is part of the blame, but part of it is just publishers have been chasing after the cheap clicks for too long and again not doing those things that I talked about earlier, not building those loyal audiences, not not emphasizing the the the unique USB as as a brand and AI just hastened that decline and everybody's panicking now. I'm like, yeah, you should have been changing course 10 years ago and if you're only doing it now because you feel the AI

heat, you you've been having your head in the sand for too long now. One of our listeners here says it's just because the bosses are obsessed with AI search. It does impact destination. It does and and we've talked about it is that the I mode um that's your newsfeed and you don't have to go to to the destination, right? So it's the interesting thing about AI mode though, Aaron, is you generally don't like it. It tends to be one of those cases where people dip in and very few people stay there. They just go back to classic search. users don't really want AI mode, which is again one of the reasons why I don't think Google wants to make it the default experience. >> Well, just wait until they plaster all the ads around there, then they'll be fine. Right. [laughter] >> Well, is there a kind of a non-trafficiire to editorial planning so leadership doesn't just stare at page views? I mean, we talk about brand awareness, we talk about visibility. I think there is something to be able to track news items

in the AI mode because that is a a venue that they're going to be affecting, right? Whether or not they've got the traffic itself. >> What are we talking about when it comes down to that that non-traic KPI? >> Yeah, I mean this has always been a tough one because this is the case for all of the the surfaces that Google has be that classic search, AI mode, even image search, the news tab. It's about impressions versus clicks of course because most publishers still care about the clicks but what what I find as an interesting correlary to that is quite can you measure impressions in an AI system like AI mode and then look at brand searches off the back of that does it increase brand awareness which then can be reflected in brand searches and direct visits to your website >> and it's interesting to build models around that to see if if increased citation in AI answers does actually lead to more brand traffic either direct traffic or brand search I think that's an interesting question to to ask and and maybe try and solve because that is really the only way where we can see a

positive impact from a citation in an AI answer. >> Yeah. And that's still brand search is at least impressions on the site, right? and that that growth, but you also have to have all players on the team communicating and coordinating because if if something goes off in rogue and starts creating some structure that is affecting AI and the and and and analytics is not trained or the models aren't prepared for that, then it's it's you have to be uniform and you have to be focused inside the organization. A couple questions from the audience here, Barry. From everything you you're seeing, when does traditional news SEO still beat AIdriven answers in 2026? And when does it just not stand a chance? That's from Ryan Cooper. >> Yeah, good question. Um, basically, classic news SEO wins in the moment of urgency, in the moment of news creation, when it is a fresh story, because the there's no way the AI can provide you with an answer at that particular

instant. So, you need to be fairly quick with your story. Get it out there fairly quickly. Make sure it's optimized as much as you can and then do follow-up stories. Keep that story alive. Keep the top stories box alive. That's that's your winning opportunity. And like I said in the size, you have a smaller window of opportunity now. Previously, you had maybe two days in top stories and now I think it's maybe six eight hours. >> So, it's a smaller window of opportunity. So, you got to maximize it. But >> that is still when you win. the moment it becomes a a longer older story and news 4 hours is old, you have a higher risk of losing out to AI generated answers because they just they they catch up quickly now and they can provide summaries very very quickly. >> Another question from Mark Scott. How does Barry see news publishers diversifying away from traditional web traffic? What are the next battlegrounds for publishers? >> Yeah, this is an interesting one. Um, I'll be honest, I'm not entirely sure myself yet. Um it really depends on how news consumption will change in the next several years and even if it will

change. Um and news people don't stop reading the news. I mean news fatigue is a real thing but generally we like reading the news. We like to stay up to date. It's just how we consume the news that might change. I find my own habits changing a little bit where I find myself reading less content on the web and actually turning on the radio or doing like one of those uh quick news catchup podcasts that go out two or three times a day >> and just listen to that when I'm driving or something like that. >> That's a perfect example of multimodal where the publisher actually moves to the consumer's preference as opposed to just staying the web content and the pop stories. Right. >> Exactly. And I think also as a publisher, if you have an audience that's very visually focused, for example, a young audience that likes to spend time on Tik Tok and YouTube, maybe do like short form videos like a 30 seconds video on on a particular story that you >> Mhm. >> then publish on Tik Tok. And there's different ways you can monetize that sort of stuff as well with with uh maybe some sponsorship logos or ads or something. So it doesn't have to be just giving away the news for free. But yeah, publishers are ironically

generally very slow to change despite the fact that the news industry is is very fast moving as organizations. News publishers generally change very slowly and like trying to to get an oil tanker to change course but they will need to because user behaviors are changing and we need to follow that. Having said that, if you're a publisher who's not in a moment of of critical failure, as in, you know, you're okay. Maybe not where you want to be, but you're okay. You don't necessarily need to change rapidly. You, you know, becoming a first adopter is very rarely the best strategy. It's better sometimes to just wait and see how is this going to pan out. What is is the user and the consumer going to do? See who who else is doing that and let them make their mistakes and then try to follow in their footsteps without making those same mistakes. >> Very good. I got one last comment cuz I know we've had these conversations before. Um, you've been blunt with a lot of publishers blaming AI while shipping quote unquote shitty websites with shitty experiences. What's one UX or ad load threshold? And we we've had these discuss discussions offline where you'd

tell a publisher fix this before you ever think about AI mode strategy because user experience data will sink you either way. Are there any thresholds that you can share with with our audience right now? >> I had an interesting one recently where it was an Australian website and I loaded it on desktop without my ad blocker because I want to see what the experience is like. There were so many ads and so many overlays and pop-ups. I couldn't actually see the headline, but there was no content for me visually on on that first viewport. And I'm like, what are we doing here people? What are you doing? I mean, there's an official standard which is about 20% ad density. I think it's actually 30% ad density where some advertising purists to draw a line. I think that's way too much. I think that 30% even 20% is way too much. If 20% of your viewport is advertising, you [ __ ] up somewhere along the way. The challenge is to make ads integral to your content in such a way that it doesn't interrupt people engaging with your content. And that is something actually you need good user experience designers for. And if you do that, the

the actual percentage of advertising becomes almost entirely irrelevant because they just feel natural to the website. Uh and that's why I think programmatic advertising is probably one of the key problems here because you have no control over what the ad shows and how it is displayed and the website just >> and we're having render blocking issues all the time. >> And I think if if you're a publisher, if you're a big publisher, you can build direct relationship with advertisers and you definitely should. If you're smaller publisher, you need to band together with other smaller publishers to form a sort of uh inventory block and then get better deals from from from advertisers because you cannot rely on programmatic to save the day. It's just going to ruin your website no matter how you integrated. >> Yep. All right. I think uh I think we've filled out the swear quota for the uh for the conference today. So, >> no, no, no. Always counting, man. Thanks so much, Barry, and a great presentation.