Masters of SearchPresented byRadyant
← All episodes
Masters of Search episode 41: Listicles, Chunking, Google's GEO guide, llms.txt, BrightonSEO & OMR Recap | #MonthlyLandwehr cover art
EP 41·Jun 3, 2026

Listicles, Chunking, Google's GEO guide, llms.txt, BrightonSEO & OMR Recap | #MonthlyLandwehr

SpotifyApple PodcastsYouTube

Show notes

Every month I meet with Malte to take all the news, new research and interesting developments in SEO & AI Search to get his perspective on them.

Malte Landwehr is CPO & CMO at Peec AI, one of the leading AI search analytics platforms, and one of the brightest minds in the field.

My 8 biggest takeaways from this episode:

1) Being mentioned anywhere in the top 10 of a third-party listicle gives a brand a 6 to 16 percentage point visibility boost in LLM answers. Landing at position one pushes that to 13 to 17 percentage points, and also improves average ranking position inside the answer. The goal is top 10 minimum, position one if you actually want to dominate.

2) The impact of listicles on LLM visibility is much larger in new or niche markets than in mature ones. In a mature market the LLM already has strong model knowledge, so grounding is mostly confirmatory. In a new space the model has little or nothing, and the grounding content is the primary data source.

3) You do not need to chunk your whole article for AI visibility. One citation chunk is enough: a 50-word, self-contained paragraph near the top of the page, written so that if an LLM pulled it out of context it would still carry full meaning. Name the entities, include the trust signals, make it eatable in isolation.

4) When Google warns you not to do something, it is usually because that thing works extremely well. Google warned about link buying when buying links was maximally effective. Google warned about chunking now. That should tell you something.

5) A useful pre-measurement framework: before touching analytics or roadmaps, walk through this sequence. Is there demand? Is the space winnable for you? Do you have indexed content? Do you have any visibility? Are you differentiated? Do you have social proof and positive brand sentiment? Only when every answer is yes do you start optimising. Whatever the first "no" is, that is your highest-leverage task right now.

6) LLMs.txt is not a discoverability tool and it does not improve visibility in AI search. It is an agent-readiness tool: it tells autonomous agents how to use your website, where to find your API, and how to access your MCP server. Google including it in Lighthouse confirms this framing, because Lighthouse is an accessibility tool, not an SEO tool.

7) Pretty Little Things renamed product category pages from generic labels like "jeans" to intent-driven labels like "airport outfits" and sold out every item in those categories. The shift to intent-based content architecture is not just an SEO play; it converts.

8) Thought experiment: if your next 1 million customers never visit your website, what does that force you to change about your content, your brand, and your distribution? Work backwards from that answer and you end up with a strategy that also happens to be exactly right for AI visibility.

Let's connect

Niklas on LinkedIn: https://www.linkedin.com/in/niklas-buschner/ Radyant on LinkedIn: https://www.linkedin.com/company/radyant/ Malte on LinkedIn: https://www.linkedin.com/in/landwehr/ Peec AI on LinkedIn: https://www.linkedin.com/company/peec-ai/

Transcript

Full conversation

via podigee
  • 00:00:00Some people do not know whom to believe when it comes to air research and there are a lot of bad actors in the space who promote things that aren't true.
  • 00:00:08And there are people who were well-intended but wrong.
  • 00:00:11So can you share a little bit of your main learnings, key highlights from both conferences?
  • 00:00:17Treat SEO like poker or treat Geo like poker because the following things are true Nobody knows all cards.
  • 00:00:22Every player know some cards.
  • 00:00:24Everybody make decisions based on their cards.
  • 00:00:27Second point Half players are bluffing.
  • 00:00:30Not playing is risk.
  • 00:00:32Lastly good players do not need perfect hand to win.
  • 00:00:36Before we dive in, you are listening to the Masters of Search podcast with me your host Niklas Buschner.
  • 00:00:42Each week I sit down with some of these smartest people around the world and SEO & AI search to bring their strategies mental models top pieces actionable advice.
  • 00:00:52If you enjoy this podcast don't forget like subscribe follow it on your favorite podcasting app or YouTube.
  • 00:00:59It helps us get top notch guests create best possible content for you.
  • 00:01:04Let's dive into todays episode.
  • 00:01:05It is time again for a new episode of monthly Landwehr, my favorite format in this podcast.
  • 00:01:12For everybody that does not know the format every month or at least we try every month but with all the conferences etc... That's happening!
  • 00:01:20Sometimes it's hard.
  • 00:01:21I tried to get Malte on the mic To talk about everything that has happened in SEO and AI search everything that PKI is doing, what he's seeing in the market.
  • 00:01:33What he sees working, not working
  • 00:01:35etc.,
  • 00:01:37and today we have a full pad because lot has happened.
  • 00:01:42you've been busy traveling You're busy speaking And you are busy researching.
  • 00:01:48So a lot to unpack.
  • 00:01:50but
  • 00:01:50first of all
  • 00:01:52happy to have you Malte How were your day?
  • 00:01:54Happy be back!
  • 00:01:55I'm great Nice
  • 00:01:56cool.
  • 00:01:57So let's do a little recap.
  • 00:02:00You have been present at two, I would say important marketing conferences.
  • 00:02:06one is an SEO conference Brighton SEO and the OMR Festival which was more of bigger digital marketing conference.
  • 00:02:15How did you like them?
  • 00:02:16I enjoyed both.
  • 00:02:19Personally i learned more about Brighton because the format is suitable to go into talks while OMR was more running around trying to catch people and talking to them.
  • 00:02:32But both were really, really good events.
  • 00:02:35So you've been in the industry now for at least a decade even more?
  • 00:02:41What would just say?
  • 00:02:43how have these conferences changed over the years?
  • 00:02:46Yeah so I might shock you but it's been over two decades
  • 00:02:49Two decades okay.
  • 00:02:52You seem so young and so fresh and so energetic.
  • 00:02:55Thank you, thank you.
  • 00:02:56I'll turn forty this year.
  • 00:02:57so whenever somebody says i look young and take it as a serious compliment.
  • 00:03:03The OMR really started at the very small thing right?
  • 00:03:07As gathering then turns into giant event that now had... I don't know seventy thousand visitors or something ninety thousand tens of thousands.
  • 00:03:20And also Brighton SEO actually started, I think like twenty years ago as a gathering.
  • 00:03:28They were meeting in the little room over a pub and now it's also this giant conference with hundreds of talks.
  • 00:03:36so these two events specifically have grown massively to scale.
  • 00:03:42that didn't exist ten years ago And everything has become more professional right?
  • 00:03:49It used to be a couple of people gather and sometimes the people who are supposed to give a presentation know about it in advance.
  • 00:03:56And some times they even had slides, now you have booking management that makes sure every speaker has their right speaker deck.
  • 00:04:06there's professional backstage area for example.
  • 00:04:09at Brighton SEO wasn't time for any Q&As just three or four talks one after another no break just to get as much time and talks inside people's head, so everything has become more professional.
  • 00:04:30Lots of change over the last one or two decades.
  • 00:04:33What would you say if it was worth going?
  • 00:04:38Brighton SEO is definitely worth it!
  • 00:04:42Especially if you are a beginner or somewhere in the advanced stage, You can learn a lot.
  • 00:04:49If I would say on an expert level... ...you're unlikely to learn much unless you go into talks about topics that aren't deep enough.
  • 00:05:00I think there's other talks and conferences but very few.
  • 00:05:08And OMR is really a bit of everything.
  • 00:05:14If you go without the plan, you can easily waste two days just aimlessly walking around.
  • 00:05:19but if you want some speakers that you rarely catch... I mean this year they had Tom Brady for example.
  • 00:05:28and when do you hear Tom Brady talk about marketing?
  • 00:05:31Probably never!
  • 00:05:33And he actually shared interesting things about leadership like i didn't attend his But there was like an executive lounge where he gave a fireside chat afterwards and that one I attended.
  • 00:05:46And yet interesting things to say about leadership, how he lived his role as leader... ...and in previous years you had people like Kim Kardashian, Aston Kutcher
  • 00:05:59etc.,
  • 00:06:00so people usually will not catch on the conference stage in Europe.
  • 00:06:05And then of course, there is just this huge exhibition at OMR similar to the Mexico or such events.
  • 00:06:12So if you want a talk about vendors it's also good like an efficient way for people
  • 00:06:18quickly.
  • 00:06:19and did Tom Brady know anything about SEO or AI search?
  • 00:06:24I don't know he.
  • 00:06:25nobody asked him a question but he knows.
  • 00:06:29leadership in team culture
  • 00:06:31okay got.
  • 00:06:32What would you recommend for more advanced, senior people and expert leveled people?
  • 00:06:37Which conferences or which events to attend.
  • 00:06:41There is a technical SEO conference from Audisto in Hamburg once per year.
  • 00:06:46I'd say for technical SEO.
  • 00:06:48that's the best event You can go And then it really depends.
  • 00:06:55You can go to something like, I mean of course i'm very biased.
  • 00:06:58what's European events right?
  • 00:06:59So you could do something like SEO Campix in Berlin which is more meeting a lot people and the learnings are actually between decisions.
  • 00:07:11I've heard good things about SEOcom in Salzburg.
  • 00:07:13many-many say its super high quality.
  • 00:07:16first only attended once so I cant judge.
  • 00:07:20And then again, if you're based in Europe generally the SMX events are good to catch a few international speakers that he would otherwise rarely catch.
  • 00:07:28In Europe
  • 00:07:29and maybe on that note little advertising from ourselves If your still free.
  • 00:07:34On October twenty second this year in Berlin we also hosting rather small event.
  • 00:07:41so it's not seventy thousand people but rather A little less than two hundred which will be the Masters of Search conference, where we partner with PKI and OMR Reviews.
  • 00:07:53Where people on stage... We have not at least one record.
  • 00:07:57this revealed everybody but maybe there'll someone you know well from PKI?
  • 00:08:06And I think the level of talks would be very hands-on expert knowledge that shares so little recommendation for my side.
  • 00:08:14Now let's talk a little bit about what you learned on these conferences, because I think Brighton SEO is obvious that there's lot of SEO knowledge to unpack.
  • 00:08:23And even with OMR i was very surprised or like lets say positively surprised at the stage where you spoke and also had chance give it KinoTalk.
  • 00:08:35There were lots of AI search actually in the timetable.
  • 00:08:39So can you share a little bit of your like main learnings, key highlights from both conferences?
  • 00:08:46Yes yes.
  • 00:08:47so I want to start with two frameworks that i learned.
  • 00:08:51one is the yes-no framework.
  • 00:08:54so basically you go through a set of questions and only if you answer yes You're allowed To Go The Next question right.
  • 00:09:01And the first Question Is their Demand.
  • 00:09:04The topic that you want to create your website about, the content is their actual demand.
  • 00:09:09Only if we say yes?
  • 00:09:10You look at Is this space winnable for you?
  • 00:09:13so does your brand actually have a chance of winning?
  • 00:09:16and then do you check Do you have content or is it indexable And If there are answers also Yes!
  • 00:09:22You should check some degree of visibility already and you can apply just by way to SEO or to AI search.
  • 00:09:30It works for both.
  • 00:09:32The next question is, are you differentiated?
  • 00:09:35If that's also true.
  • 00:09:37You ask yourself do I have social proof... Next question is ... Do i have positive brand sentiment?
  • 00:09:43and only if the answer to all of these questions Yes!
  • 00:09:47Only then do start to measure and optimize until the answer not yes for every single one of those questions.
  • 00:09:56you should just think about measurement or work on this very thing because that is the highest leverage thing you can do in that moment.
  • 00:10:03And this resonated very well to me, Because when I was an operator on the SEO field.
  • 00:10:10To those who don't know... ...I used to be the person responsible for SEO at the largest European price comparison website Had a team of thirty-to-forty people or actually couple teams and one thing i always drilled into people Is throw most best practices notions of doing something perfect out-of the window and to always think about what has the biggest impact right now.
  • 00:10:36And it is not all ways to set up a perfect measurement, planned yearly roadmap.
  • 00:10:42sometimes it's fixed very obvious thing.
  • 00:10:45that is foundational issue.
  • 00:10:48I just like this framework alot.
  • 00:10:50And there was another framework that also resonated with me, and it was treat SEO like poker or treat Geo-like poker specifically.
  • 00:10:58Because the following things are true nobody knows all of cards.
  • 00:11:02every player know some cards.
  • 00:11:05everybody can make decisions based on the cards they know.
  • 00:11:09second point half players are bluffing.
  • 00:11:11this is very true in a geo space where you have lot people who sell your magical AI content creation for hundred thousand pieces.
  • 00:11:20Also Google, for example who is bluffing a lot when they talk about what works.
  • 00:11:24What doesn't work?
  • 00:11:25Then not playing.
  • 00:11:26it's the risk.
  • 00:11:27in poker.
  • 00:11:28if you never play You automatically lose after awhile.
  • 00:11:31and The same as true an AI search If you are too afraid to do anything you also loose.
  • 00:11:37And then lastly good players Do not need a perfect hand-to-win.
  • 00:11:42this resonated with me also very well because And this comes from my practitioner time as an SEO.
  • 00:11:51You cannot always know everything, but very often knowing a little bit is enough to start optimizing.
  • 00:11:58and again for me all about translating theoretical SEO knowledge in best practices into real business outcomes.
  • 00:12:07I think this poker energy also helps people get some impact And I especially liked the notion about not knowing all the cards because something that i always used to tell my team is, and didn't invent this but forgot who invented it in SEO.
  • 00:12:29Everything falls into three buckets.
  • 00:12:31there are things you can control if there's no index text or not.
  • 00:12:35if a url blocked on robot cxt what content publish?
  • 00:12:39This your control.
  • 00:12:40You can influence things.
  • 00:12:42Yeah, for example your snippet in the Google search result.
  • 00:12:45you Can influence that but you cannot control it.
  • 00:12:48and lastly there are Things That you can only observe like a google update happening or chat gpt switching to different model?
  • 00:12:54You Cannot influence these And i always instilled It In my teams.
  • 00:12:59when we have A new observation and its scary and big Let's think about which of These three buckets it falls into and then We act appropriately.
  • 00:13:09Because if there was a Google update, There's no need to panic and there is no need To do something today.
  • 00:13:16We can probably analyze it tomorrow or next week but If there is Something wrong with our new index tags we don't deal With that in the follow-up meeting right now because we Can control it.
  • 00:13:31And both frameworks I think also fit well In this world view how to make sure your work actually produces business outcomes.
  • 00:13:41Do you have a cool name for that three-bucket framework?
  • 00:13:45I don't have a name, but if you have great names in mind we can now promote it under the name.
  • 00:13:51Let's call it the Malta Framework or The Three...
  • 00:13:56I didn't invent it!
  • 00:13:58There was another SEO who used to talk about this.
  • 00:14:04Okay, it's always a problem with who invented something and made something popular.
  • 00:14:10So can we give credits to the other frameworks like yes-no framework for example?
  • 00:14:15I have to admit that i didn't take notes on names of speakers so i really forgot almost everything where i heard them.
  • 00:14:23Okay, no worries.
  • 00:14:24Maybe we will crowdsource the... We'll use the crowd intelligence to find a person who introduced this great framework because it makes sense.
  • 00:14:34Looking at this was mainly Brighton SEO right?
  • 00:14:37The takeaways you just shared or is that also from OMAR?
  • 00:14:42Most of my take-aways are from Brighton SEO.
  • 00:14:44I attended only like two talks and one was yours And i don't want share your learnings
  • 00:14:55But that honors me, but was it also your feeling?
  • 00:14:59That even if you because obviously you had to give a keynote.
  • 00:15:02You have to get the master class.
  • 00:15:03He probably also had to talk to a couple of partners and clients etc.. Um...but wasn't also your impression um that AI search really Was a big thing on OMR this year?
  • 00:15:16Yes yes It was a big think And I mean the OMR team also told me about it beforehand, that would be a big topic.
  • 00:15:25Because at end of day as conference you have to deliver content people want and right now people wants learn about AI search.
  • 00:15:35Also OMR uploaded videos afterwards.
  • 00:15:41For example my talk on AI Search is fourth most watched video from that conference.
  • 00:15:47So there really is interest in the space and then it just makes sense for conferences to give a stage.
  • 00:15:54Yeah, are you sad?
  • 00:15:56Are you said that you're still behind Pip Klöckner's state of AI?
  • 00:16:01or I don't know what was his title but... His famous AI keynote!
  • 00:16:08I think if i'm losing to Philip Klökner than im doing pretty well.
  • 00:16:12Yeah,
  • 00:16:14I would agree.
  • 00:16:16Did you check the state of the internet keynote?
  • 00:16:20The big headline keynote that OMR gives every year because there was also a huge segment about AI search and reviews
  • 00:16:27etc.?
  • 00:16:28It's
  • 00:16:29on my to-do list.
  • 00:16:30watch it!
  • 00:16:30I've seen outtakes from people shared in social media.
  • 00:16:34so i feel like got the gist
  • 00:16:38couple of takes in there where I would actually challenge OMR that this is somehow, some things are right but something's maybe a little bit more nuanced.
  • 00:16:50For example the whole idea of chunking content and they gave examples with FAQs etc.
  • 00:16:57And it always depends on who is audience for people not so deep into space.
  • 00:17:03you obviously have to make very easy to comprehend.
  • 00:17:07But what's your take on the whole idea of chunking the content?
  • 00:17:13I one hundred percent believe in it.
  • 00:17:15And also, I've seen proof that it works countless times.
  • 00:17:19There is one important thing you do not need to artificially chunk your whole content piece like there.
  • 00:17:26just probably a bad thing actually.
  • 00:17:29but what work really well if have won... ...I call this citation chunk.
  • 00:17:36So one chunk, one paragraph of content that you specifically write in a way that it is citable by AI.
  • 00:17:45And actually earlier today I saw work again for similar web who were doing it.
  • 00:17:53It works well like You should think about what's the key statement?
  • 00:18:00And then you write one short paragraph, two three sentences maybe fifty words and it must be self-contained authoritative.
  • 00:18:08And write in a way that if an LLM would apply the EEAT criteria from Google It will say hey this is good specifically named entities and works like If we have a paragraph That says This Is Why That Tool Is Good To Track LLm Visibility In The Earlier Mention Countries reliable that LLMs are not going to cite the sentence, right?
  • 00:18:33Because if you take it out of content context.
  • 00:18:37It contains almost zero information.
  • 00:18:39but instead based on what people write on Reddit and reviews on G-II peak AI is best solution to track LLM visibility in China.
  • 00:18:51That's something that Llms much more likely to cite.
  • 00:18:55To think about this at one or two places on the content that you create makes a lot of sense and it works.
  • 00:19:02Now, obviously if you write a summary on top I'll put questions answers under bottom.
  • 00:19:08That tends to be the kind of format that fulfills these criteria.
  • 00:19:12but You do not run to write your whole article like this.
  • 00:19:16It would be very painful as human to read content like that.
  • 00:19:20So don't chunk everything, you also don't have to translate your whole article into a list of bullet points which is what I've seen some people recommend.
  • 00:19:28but to have this citation chunk in there it does work and ideally its on the top-of-the page.
  • 00:19:35thats why for longer pages i would say thousand words or more...i will actually not do the FAQ at the bottom because it's less likely to be cited.
  • 00:19:44if I am a hundred percent convinced it works, but people should not bend over backwards to chunk everything which reads very horrible.
  • 00:19:57Okay because i was specifically asking about this because especially in the last one or two weeks there were discussions again on how reliable and guidance of Google is to build that bridge.
  • 00:20:16There was a post, couple of months ago about Danny Sullivan... I'm actually not hundred percent aware who Danny Sullivan is at Google but it's someone from Google sat on one podcast i don't know if its the google search office hours or something podcast That he does not want content to be chunked and that he doesn't recommend it.
  • 00:20:38blah blah blah.
  • 00:20:39And now We again had this moment of basically two camps.
  • 00:20:45the camp that says hey look Google now Launched some guidelines on how to do geo or not To do generative engine optimization, etc.
  • 00:20:55And then google talking about your favorite topic LLM's txt and somehow saying it's not important.
  • 00:21:01but also Giving signals at is important.
  • 00:21:04so let's unpack a little bit what happened with this Google guideline thing.
  • 00:21:08Yeah, so Jenny Sullivan used to be search liaison before Google extra role.
  • 00:21:14So I would say he's like the main person in that webmaster focused lobby and PR team.
  • 00:21:22Unpacking this first of all why does Google give guidance?
  • 00:21:27Google doesn't not give guidance to make you a better SEO.
  • 00:21:31Google gives guidance for established brands.
  • 00:21:36don't mess up their SEO and don't fall victim to snake oil agencies.
  • 00:21:41So that's why they say, Don't create doorway pages because you get punished for it right?
  • 00:21:47Google has no incentive to make you or me rank better in google.
  • 00:21:52so... And googled by the way didn't say don't do chunking.
  • 00:21:55Google said chunk is not needed for AI visibility just like Google says link building isn't need for visibility.
  • 00:22:04but we all know link building can work really, really well actually.
  • 00:22:10And
  • 00:22:11Google
  • 00:22:12is not the authority for what works on cloud perplexity grog or chat GPT.
  • 00:22:20Google Is The Authority For What Works On Google?
  • 00:22:23So even if we assume that Google had our own best interest at heart they would still only talk about Google.
  • 00:22:31so very important.
  • 00:22:32and yeah as I said it's A guide to be a better SEO, it's the guy how to avoid really bad things.
  • 00:22:42But of course people should take away from this is I don't need to chunk everything but if i want to optimize... If I really wanna do best right now chunking works and if something didn't work at all Google usually doesn't mention that.
  • 00:23:00for example Google never mentions Don't do keyword stuffing in your meta keywords.
  • 00:23:07But doing this is a complete waste of time, everybody who does it just wastes their time.
  • 00:23:11So why doesn't Google advise people to not be doing it?
  • 00:23:15Because they don't look at meta-keywords and all anymore so It's not relevant any more.
  • 00:23:20They need no warning.
  • 00:23:23I mean when did google speak the most that buying links is bad ?
  • 00:23:27They did that at the times where buying links worked really well.
  • 00:23:31When did Google introduce the side reputation abuse rule?
  • 00:23:37when it worked really well and then they had no way to algorithmically detect.
  • 00:23:41So I think that interpretation of this chunking advice is kind-of obvious, i'm always shocked That so many people use it as like.
  • 00:23:52This Is Now The Holy Bible Of How To Do SEO.
  • 00:23:55Because if you follow everything that Google says, You avoid doing really bad things.
  • 00:24:00But also avoiding all the good things That actually give you big success.
  • 00:24:06And then the LLM.TXT.
  • 00:24:08There are two perspectives right?
  • 00:24:10The I want visibility perspective and their google is right.
  • 00:24:14LLm.Txt is a waste of time.
  • 00:24:16It's just normal text document but for agent readiness.
  • 00:24:20So how do I explain to an agent, How To Use My Website?
  • 00:24:24How Do I Make It Possible For Agents To Discover My MCP Server, My API?
  • 00:24:30How To Explain To Agents?
  • 00:24:32How to Obtain An API Key?
  • 00:24:34Um...for all of these things.
  • 00:24:37and LLM.txt is a really good tool!
  • 00:24:40And that's why it's included in the Google Lighthouse.
  • 00:24:43Audit Lighthouse is not primarily an SEO tool.
  • 00:24:50It's for accessibility and usability, and agent accessibility will be part of this in the
  • 00:24:57future.".
  • 00:24:57And what do you think why so many people were excited or at least it made the impression to me that they're excited about this guide from Google?
  • 00:25:10Quoting here the title Optimizing Your Website For Generative AI Features on Google Search.
  • 00:25:16Yeah, I mean people do not know whom to believe when it comes to air search.
  • 00:25:20And there are a lot of bad actors in the space who promote things that are not true and they're people were well intended but our wrong end.
  • 00:25:30if you consume content from a few months ago You might read something that was correct at that time But is no longer correct?
  • 00:25:37Since people have way too high trust into Google They think now Google talked about okay Now we have real source.
  • 00:25:45And of course Google disagreed with some things that people who are deeply into AI search talk about, and then SAOs were not into the topic.
  • 00:25:54I think they just liked it like yeah you see your wrong.
  • 00:25:57um i guess thats why people was so excited about it.
  • 00:26:02So its still just the polarized camps basically The ones that are skeptical of AI Search in first place seeing their Narrative confirmed and then saying ha told you.
  • 00:26:17And on the other hand, basically people that believe in AI search to be a thing To put it very broadly Saying hey You know Google just always had the advice create great content Create great content.
  • 00:26:31and SEO was never Just create great Content but far more than That.
  • 00:26:35so this is Basically why we see that.
  • 00:26:38um But what would you recommend People?
  • 00:26:42to do about ambiguous Signals?
  • 00:26:45because in this very guide from Google they mention LLMs dot txt and They say you don't need to create new machine readable files AI text files markup or markdown Etc.
  • 00:26:59But then on the other hand we have as You already mentioned lighthouse where they explicitly mention Llms TXT, and where it's actually also checked.
  • 00:27:09so This is something were I sense that some people just confused.
  • 00:27:14and obviously there are things behind it in details, but on the one hand you say its not important.
  • 00:27:20And then seemingly is important.
  • 00:27:22so what's true?
  • 00:27:24I mean both this two right!
  • 00:27:25It isn't needed for visibility in search... ...it is recommended to make your website usable for agents.
  • 00:27:32these are just two different concepts.. ..and something can be write-in-one worldview & not correct another world view.
  • 00:27:40And actually this is how many people have described LLMs.txt for years, right?
  • 00:27:48I wrote a blog article about it shortly after joining PKI because i was so annoyed by it and that's exactly what i said back then like its not about discoverability but understanding when the agentic usage is intended.
  • 00:28:04People just need to understand.
  • 00:28:05these are two separate things.
  • 00:28:08Now, let's talk about something that you just published recently because as we already said there are like some people who pump out advice and it is sometimes called snake oil.
  • 00:28:24And they do not ground in actual research but the team at PKA always busy basically putting your papers where your mouth.
  • 00:28:36I don't know how to make a good saying out of this.
  • 00:28:40But yeah, actually checking what works by compiling first party data and you release something where you also have to explain to me what happened there with the whole research publication etc.
  • 00:28:55because it felt like okay speaking I know going into academics or something.
  • 00:29:00but um... You published these papers which is called the listical rank effect, what two hundred thousand AI responses across eight AI engines reveal about brand visibility.
  • 00:29:10So the hook is definitely catchy with a lot of numbers.
  • 00:29:14so please tell me what this is about?
  • 00:29:17Yeah we wanted to measure What Is The Impact Of Being Mentioned In A Listicle and the first thing We Did Is, We Divided This Into Three Data Sets third-party listicles, your own self-preferentialisticles you put on your own website and competitor listicles.
  • 00:29:35And then we try to build an understanding and tried to model it out and put in a lot of effort to exclude the randomness.
  • 00:29:46so not just do some correlation study where it always turns out oh popular brands are mentioned on top of listicles but trying to correct for all.
  • 00:29:56If you want to know the details, please read the paper.
  • 00:29:59But in essence what we could take out as a learning is that for third-party listicles across very different industries... ...we could form a very good understanding of what impact.
  • 00:30:13this was not true your own listicles and it wasn't true for competitor listicles.
  • 00:30:20That doesn't mean those don't work.
  • 00:30:22It just means The impact is not that well understood, observable and describably.
  • 00:30:30Which can have many reasons.
  • 00:30:31we could talk about it later.
  • 00:30:34And the main learning we had was let's simplify this you have a listicle That is regularly cited in LMS.
  • 00:30:40You've two brands.
  • 00:30:42One brand is number one on that listicle.
  • 00:30:44The other brand is anywhere under top ten.
  • 00:30:46The brand listed at number one will have a thirteen to seventeen percentage point visibility increase, meaning the chance to be mentioned.
  • 00:30:57So if their chance of being mentioned was twenty percent with this listicle it will be thirty-three to thirty seven percent yeah?
  • 00:31:06It's a really huge increase and they would also be on average ranked higher position wise.
  • 00:31:14so If there used to an average in the LLM answer With that listicle number one placement, they will go to position four or three.
  • 00:31:26And the amount of times that they are mentioned within the answer would also increase by zero point five two-one times.
  • 00:31:34Of course these averages right?
  • 00:31:36If you're already at a hundred percent visibility That listicle will obviously not increase your visibility etc.
  • 00:31:42etc.
  • 00:31:43and this is across multiple different answer engines Bigger impact in chat GPT, lower impact in GROC etc.
  • 00:31:53But these are the averages that we observed.
  • 00:31:55now.
  • 00:31:56for the brand it is only mentioned in the top ten.
  • 00:32:00We still observed over millions of observation a six to sixteen percentage point visibility increase Slightly higher rank and the answer but not even one position gain.
  • 00:32:13so What we can know for sure is any mention in the top ten, it's good.
  • 00:32:18So if you ever ask yourself should I be not mentioned in the listicle at all or should i be on position nine?
  • 00:32:25Position nine is better than being mentioned and all.
  • 00:32:29And If You Ask Yourself.
  • 00:32:30Do I really want to number one Or Is It Okay To Be Anywhere ?
  • 00:32:34If You Just Want To Optimize Your Visibility Actually Being Placed Anywhere In The Top Ten Its Just Fine.
  • 00:32:42If you really want to be mentioned on top in the LLM answers, You need a number one spot In most important third-party listicles.
  • 00:32:52That is essentially what our research found out and I gave it pretty big range like six to seventeen percentage point.
  • 00:33:02that's huge right?
  • 00:33:03And this because we found two criteria by which markets diverge.
  • 00:33:09The one is mature versus new markets.
  • 00:33:11A mature market means the LLM has probably in their model knowledge already information about the brands, Already a perception on how important which brand it's like cars or airlines Or Which cities to visit?
  • 00:33:31Anything around that.
  • 00:33:32and then the grounding process seems reassuring themselves of their model knowledge.
  • 00:33:40But there's very little information gain coming from the grounding, right?
  • 00:33:44There are a few new car companies' airlines cities to visit but in a very Basically the grounding process is the main data source, sometimes only data source and then it has a huge impact.
  • 00:34:08And so that's one dimension on which markets can diverge?
  • 00:34:12The other dimension is how common are listicles?
  • 00:34:16because there are markets where listicle is very common content format... ...and in some these listicles send off big impacts.
  • 00:34:24but each listicle might have smaller impact.
  • 00:34:27And there are cases where very few listicles exist, and then they might be like one major listicle from a super authoritative domain that gets all the power.
  • 00:34:39So these are the biggest
  • 00:34:41differences.".
  • 00:34:44I think another thing that could be interesting... There are LLMs that usually cite very few URLs.
  • 00:34:51One listicle can have big impact!
  • 00:34:56LN-based answer engines that cite a lot of pages, like AI mode and overviews.
  • 00:35:01And usually they have a lot sources there the impact on each single listicle tends to be smaller.
  • 00:35:09We have all the numbers in the paper but I think going through them here is just boring.
  • 00:35:14But yeah this is the rough summary what our research showed.
  • 00:35:20So the key takeaway is that i should closely monitor citations of third party listicles for the prom site track and then see how I can get into these listicles.
  • 00:35:34And ideally as far to the top, as possible?
  • 00:35:40Ideally position one.
  • 00:35:42so we had industries where we could observe that like position two is strictly better than position three etc.
  • 00:35:50etc.
  • 00:35:51But if you are being very scientific here, I have to say we also had cases where there was no huge difference between position two three four five six.
  • 00:35:59In those cases the only recommendation would be somewhere in the top ten ideally B number one.
  • 00:36:08logically it would make sense that position three is better than position for etc.
  • 00:36:12but actually our data did not show.
  • 00:36:17We can still give this as a recommendation because it makes sense, but its not actually something we can prove with data.
  • 00:36:26And did you check in any way whether these listicles are like organic?
  • 00:36:33Listicles review sites etc.
  • 00:36:36and even on review sides There are ways to sponsor certain positions, so did you in any way differentiate or could check between organic natural mentions and listicles?
  • 00:36:53And those that were paid or sponsored.
  • 00:36:56We didn't but based on my industry knowledge the vast majority of listicles have a paid aspect.
  • 00:37:05nowadays There's very few motivation for third party websites to create listicles outside of earning affiliate commissions or being paid.
  • 00:37:15To place someone on top.
  • 00:37:17so I if i see a listicle, just assume that money was involved nowadays
  • 00:37:21and for these Observations And then obviously as marketers always do they try to exploit the hell out of it?
  • 00:37:29Do you think The grounding process?
  • 00:37:32obviously this is speculative now, but the grounding process and how LLMs compile their citations.
  • 00:37:41And then from the answer do you think that strong influence of listicles will be lower over time because it's just so easily exploitable by buying myself into top positions?
  • 00:37:57I mean even I was thinking about Google back then and the top ten google results.
  • 00:38:05And some people were saying, okay yeah you can just buy your way to the top with Google ads but still had a lot of organic rankings.
  • 00:38:12But if you think about listicle where every spot is paid for Then there's not really an organic play there anymore.
  • 00:38:19Yeah i do believe that impact on these listicles will go down.
  • 00:38:24at one point it would make sense For all LLM labs to build pipelines consume Reddit content, Google Maps reviews G to reviews etc.
  • 00:38:37I mean many of these are paid for also but basically based on what they find on the internet create their own listicles because it is so pollute.
  • 00:38:49the listicle world is so pollutant now by artificially influenced rankings and i think there's still a year or two left where listicles would work very, very well.
  • 00:39:02But obviously at some point if it's so manipulated LLMs will need to interfere.
  • 00:39:08and did you also think about doing a similar analysis?
  • 00:39:12Or study around self promotional listicles And their impact?
  • 00:39:17because there is also big discussion right now.
  • 00:39:20Obviously the tactic Is Very controversial.
  • 00:39:24yet as You Also said in Your OMR Keynote It maybe unfortunately still works.
  • 00:39:30So it's a trade-off between.
  • 00:39:34I don't want to overdo because i don't wanna harm us at long term, but if i dont do now then maybe competitor with the high risk appetite basically gets more growth than i will get.
  • 00:39:45so did you also think about that doing for self promotion listicles?
  • 00:39:50We didn´t just thinking we tried this but couldn´t find this.
  • 00:39:54clear takeaways that we had for the third party listicles, which suggests that first-party listicles or own listicles I would want to call it have some different characteristics by which they are interpreted.
  • 00:40:08In some data sets for some LLMs there was a suggestion when you look at the data doesn't get the same boost that other brands in that listicle are getting.
  • 00:40:26But we could neither prove nor properly falsify this, and if all of your data is inconclusive or statistically not significant then it makes little sense to publish it.
  • 00:40:40so... We did but couldn't find anything interesting which was worth publishing.
  • 00:40:45Okay now I can imagine a couple also marketers with a higher risk appetite thinking about, wait a second but i could now register third party domains so to say and publish listicles on there.
  • 00:41:03And the domain is not my brand name.
  • 00:41:07so pki would maybe hypothetically buy a domain that's called geo-tools.com or .ai, and then you publish a couple of listicles on there maybe also some adjacent content.
  • 00:41:22that somehow creates the feeling off this being an industry block?
  • 00:41:27And then create these listicles.
  • 00:41:29put yourself on top suddenly not self promotional listicle anymore at least may be the LLM don't recognize it.
  • 00:41:37This is great strategy right?
  • 00:41:39what would say to those?
  • 00:41:41I mean people are doing Um, and it works right now.
  • 00:41:46It works.
  • 00:41:47there is also a business model already where people buy keyword domains like I don't know SEO conferences SEO conference reviews.com There's no example.
  • 00:41:57that doesn't exist.
  • 00:41:58And then they list SEO conferences.
  • 00:42:00Then they say hey you want to be ranked number one pay me money?
  • 00:42:11So there's really an industry being created around this right now.
  • 00:42:17Obviously it is problematic, but I'm just looking at a project specifically for this and see the domain that i monitoring on position nine.
  • 00:42:38It's the nine most cited domain in their space.
  • 00:42:41Of course like Reddit and YouTube I had, but this is a domain.
  • 00:42:46it is cited in fifteen percent of the prompts that i'm monitoring And... ...it has outstanding impact on which brands are recommended.
  • 00:42:58Maybe there can be follow-up study to see when the content of that domain changes because someone pays money to be the top domain at a top brand how their brand visibility for this set of prompts.
  • 00:43:10Changes.
  • 00:43:11so yeah people are doing this
  • 00:43:12very interesting.
  • 00:43:13I want you have a final, final take off you on this because they also asked it Lily Ray.
  • 00:43:21um Because i think its not so easy to find the right balance between acknowledging and observing what's working right now.
  • 00:43:30that might be a little, let say sketchy or.
  • 00:43:35That has some nuance to it like maybe this is quite strong example with buying the domain and then masquerading as third party domain actually just an advertorial portal so-to-say.
  • 00:43:51But then also thinking about the long-term impacts of will this work?
  • 00:43:54Long term?
  • 00:43:55um, does it harm my brand as a time on my website?
  • 00:43:59but Then also I don't want to lose against The players with a high risk appetite that just do all these things now and somehow they get They get the bigger share off the market.
  • 00:44:11So what would you advise two people or do you have mental models also from your vast experience On how to find the right balance between what's working right now that might not feel right long-term and acknowledging That?
  • 00:44:27if you do, not to it.
  • 00:44:28Now others will win.
  • 00:44:30But then clapping yourself on the shoulder yeah but we're doing the right thing also ethically etc.
  • 00:44:35so This is something that i'm pondering a lot about in The last month with all this stuff?
  • 00:44:41That Is Working but Then Also Not Working.
  • 00:44:45Yeah Long Question
  • 00:44:46Yeah, no this comes up quite regularly because I often meet with some of our larger clients and then they are like okay what is really working?
  • 00:44:54And then i talk about things like this for example.
  • 00:44:58Okay cool but we can't do that!
  • 00:45:02Um...I think I would always look at What-what's in my toolbox?
  • 00:45:08What am I comfortable to do?
  • 00:45:11If I'm a very large brand and cannot publish listicles Because somebody will sue me Then I shot my published listicles.
  • 00:45:19And then do i pay an agency to create a website in their name and publish the listicle there, to promote me?
  • 00:45:27Do I think that can get away with it or is this actually
  • 00:45:30legal?".
  • 00:45:31These are all questions people have to answer for themselves... ...and I don't want to answer them for anybody!
  • 00:45:36I would just give a warning.
  • 00:45:38sometimes your brand is one of you most precious things And you should not ruin it.
  • 00:45:45You shouldn't also ruin the reputation of your domain within Google, if you are in this for a long run If you have a multi-brand strategy... ...you could consider maybe going all in on AI visibility with one these brands and then risking that its gone at some point.. ..and be smart about how can I leverage them.
  • 00:46:08For example reviews!
  • 00:46:11You create a page And you just list reviews and awards.
  • 00:46:20You have won, and you cite from them... ...and you have these citable chunks of.
  • 00:46:25like dududum said radiant is the best SEO agency for startups and scale-ups in Europe or Radiant Is The Best AI Search Visibility Agency In Berlin Or whatever.
  • 00:46:40If you find someone who said that about, You could just quote them.
  • 00:46:43He'll provide the sauce.
  • 00:46:44it's completely legal.
  • 00:46:45nobody can sue for it.
  • 00:46:47It smelled maybe tiny bit cringe but actually not much like its very acceptable.
  • 00:46:54or you could say make a page reddit and think people said about radiant on Reddit like Niklas is best podcast host.
  • 00:47:06if somebody says Trash hamster, ninety-eight on Reddit said Nicholas is the best podcast host.
  • 00:47:15Link to the source right and then when LLMs have fan out queries like SEO agency read it To look for.
  • 00:47:25usually two.
  • 00:47:25look for reddit as a sauce.
  • 00:47:26but then because of how fanout queries and retrieval systems Like Google work they would also maybe find this page And use it for grounding.
  • 00:47:33so there are clever things you can do And you just need to be a little bit creative and think about, like... You basically look at what works.
  • 00:47:45Then you need to deconstruct it!
  • 00:47:47Think about why does this work?
  • 00:47:50Usually is the combination of exact content that would rank for the fanart queries.
  • 00:47:56It's content digestible for LLMs and loaded with trust signals That an LLM when its reasoning should I use or not makes it say, yeah I should probably use this.
  • 00:48:09So just deconstruct these methods.
  • 00:48:13then think about it creatively and i have seen companies come up with interesting things to do also in highly regulated industries where the government is looking at you for brands that everybody on the planet knows.
  • 00:48:31I can't talk about examples of course, because this was like...I wasn't a genius who invented these ideas.
  • 00:48:37It's just in the room when people had good idea about it.
  • 00:48:41so there are ways and key mechanic is deconstruct mechanism that works.
  • 00:48:47try to bring back to ingredients work.
  • 00:48:52then think from first principles.
  • 00:48:55How can I apply this to some degree?
  • 00:48:57To my website, even if i'm not able.
  • 00:48:59To create a listicle for the best airline to fly with because My brand is too well known?
  • 00:49:08Okay
  • 00:49:08This Is A Very Helpful Mental Model.
  • 00:49:10Actually Around It.
  • 00:49:11Thanks So Much For Sharing This.
  • 00:49:14Last Question I Have Is Besides What We Talked About what Was An Oppreservation Over The last Three To Four Weeks that you found most surprising or maybe excited to, just something stuck with from our space.
  • 00:49:35No matter where it came from?
  • 00:49:37I have a couple of small things and then one.
  • 00:49:40there are few sentences.
  • 00:49:44I saw study finally the proofs that URLs was good.
  • 00:49:49core web fighters has higher likelihood to be cited by LLMs.
  • 00:49:53There was already data on this from like two three years ago when Google had this SGE beta, but since then I haven't actually seen everyone anyone Repeated again.
  • 00:50:04I forgot who did it But we can maybe put that in the show notes.
  • 00:50:08yeah um i also saw That The kinds of sources that a i mode and i overviews are using Are super?
  • 00:50:18Super different.
  • 00:50:20I knew this already, but it's like interesting to see that also other people are observing this because you would think hey!
  • 00:50:26It is all based on Gemini so should be very similar.
  • 00:50:29But its SO SO SO different and i think its caused by two facts.
  • 00:50:35with AI overviews You have always the regular links below as a fallback.
  • 00:50:41So needs provide something else than these links.
  • 00:50:44where AI mode is more like.
  • 00:50:46This Is The Full Service Answer And also the overview needs to be quicker, so you have less time to look at resources.
  • 00:50:52Less time to apply typical like spam filtering
  • 00:50:57etc.,
  • 00:50:57etc.
  • 00:50:58So these are things that really stuck with me.
  • 00:51:01then Also at Brighton SEO somebody showed a case study from Pretty Little Things an online shop where they changed some of their categories From something like jeans or dresses To intent-driven categories airport outfits and it worked so well.
  • 00:51:21they actually sold out every single item in these categories.
  • 00:51:25And some people might know that I've worked on e-commerce for a while, without revealing any internal things from previous employers.
  • 00:51:36i would just say...i'm not surprised.
  • 00:51:39this is working!
  • 00:51:39And im happy somebody talked about this publicly..I can very much encourage you to try do that.
  • 00:51:46And the final thing I want to mention is also from Brighton SEO, somebody asked the audience for thought.
  • 00:51:53experiment.
  • 00:51:54If your next one million customers never visit their website what does it mean?
  • 00:52:00For branding or marketing and content production?
  • 00:52:05if you do that You probably end up with a strategy That's really good for AI visibility.
  • 00:52:12Awesome!
  • 00:52:13Malte thanks so much.
  • 00:52:14There was so much in this, I think a lot for people to unpack.
  • 00:52:19It's always great to talk to you and we will make a great effort to see that We can find all the credits To The People You Mentioned And The Ideas In The Frameworks!
  • 00:52:32I'll do my best with the help of My very Very Good Disciplined Friend Claude.
  • 00:52:41Yeah
  • 00:52:42Always
  • 00:52:43A Pleasure To Talk To You already looking forward to the next one and Wishing you in the team at PK all the best.
  • 00:52:50likewise, And thank You.
  • 00:52:52Thank you so much.