Discover how Query Fan Out is revolutionizing SEO as Jon Clark and Joe DeVito interview Noah Learner—Director of Innovation at Sterling Sky and founder of the SEO Community. Learn how AI Mode, ROSP (Return on Search Page), and structured data experiments are reshaping search strategy and revenue models.
https://page2pod.com - Jon Clark and Joe DeVito dive deep with Noah Learner—Director of Innovation at Sterling Sky and founder of the SEO Community—unpacking groundbreaking concepts like Google’s Query Fan Out technique, the ROSP (Return on Search Page) metric, and how AI Mode will reshape search monetization. Discover actionable insights on structured data experiments, future content strategies, and ways to stay ahead in an AI-first landscape.
Discover how these changes impact your SEO strategy, from technical optimizations to structured data and user engagement.
🎯 In this episode:
• What "Query Fan Out" is—and why it's a game-changer for SEO
• The introduction of "Return on Search Page (ROSP)" and why it matters to Google
• Why rank tracking is becoming obsolete—and what's replacing it
• How AI overviews and synthetic queries impact content strategies
• The critical role of structured data and schema in AI-driven SEO
• How Noah builds automated data pipelines using APIs for SEO insights
• Why testing and experimentation is essential for future-proofing SEO strategies
• The growing role of local SEO and video content in the age of AI
• Behind-the-scenes insights into managing the SEO Community Slack group
Noah Learner offers a blueprint for navigating the complexities of modern SEO, emphasizing the importance of innovation, automation, and continuous testing.
🎧 Listen & Subscribe:
Follow and subscribe to the Page 2 Podcast wherever you get your podcasts. Don’t forget to leave a review and share with your team.
📎 Mentioned Links & Resources:
Jon Clark (0:1.964)
Welcome to season four, episode 78 of the page two podcast. I'm Jon Clark. And as always, I'm joined by my cohost and partner in crime at moving traffic media, Joe DeVita. Today's guest is someone who needs no introduction in technical SEO circles and SEO circles, but we're going to give him one anyway. Uh, you've likely seen him sharing insights at major conferences like MozCon search love, uh, or perhaps you know him by his formidable nickname, the Kraken.
uh Noah Lerner is the current director of innovation at Sterling Sky Incorporated and the founder of the vibrant Slack group, the SEO community, with a focus on making our industry more connected and collaborative. He's a relentless builder. He's regularly pushing the boundaries of what's possible with data pipelines, APIs, and AI, and the mind behind some indispensable tools like Branch Explorer and Post-Omatic. uh But before we jump into our conversation with Noah,
Just a quick reminder, if you enjoyed today's episode, please take a moment to subscribe to the show and leave a rating and a review wherever you're listening. Really helps us continue to deliver great content. Noah, welcome to the show.
Noah (1:12.677)
It is fantastic to be here with you today.
Jon Clark (1:16.922)
Well, I know there's been so much discussion around Google I.O., perhaps too much at this point, but you had some really great thoughts uh as you had a chance to sort of reflect on the conversations that you were having um both with other SEOs, but also Google engineers. And I thought one of the concepts that you sort of introduced was this idea around Rosp or return on search page as a
maybe a metric to help define or sort of inform why Google was making some of these decisions. I'd love for you to sort of break that down for the audience and talk a little bit more about that.
Noah (1:58.643)
So basically I came back from Google I.O. and I felt like everyone that I went with was smarter than me. They all had these like amazing insights. Yeah, like and I wrote an article about it and I shared the names of all the SEOs that were present at the conference specifically so people could get a holistic view of what happened because I noticed as I was hanging out with my friends, I keyed in on specific small
Jon Clark (2:7.214)
It's a good group to be around, right?
Noah (2:28.627)
tunnel visiony things and other people had wider lenses and other very tight niches and I just felt like everybody needed to see what everyone else was saying to get a picture of it. So that's the first thing is surround sound is a thing and I think it's super important to get the insights of everybody versus just me or just my king even though his insights were critical. OK, so Ross, what is it? I sat down and I.
thought to myself, uh what the heck happened? What did I just witness? And I started to like take notes like crazy. And then I started to think critically about what happened. And there's all kinds of different critical thinking methodologies you can use. And one of them is like, put yourself in someone else's shoes, right? And Cindy Crum always talks about what Google is gonna do in the future. And I asked her in a podcast, like, hey, how do you do this so often? Because she's led the industry
in understanding things like entities, Fraggles, all kinds of stuff, right? And she said, well, I always put myself in Google's shoes and I try and understand from, you know, like to make a business case as if I'm Google, how can we actually make money? And I put myself in that kind of mindset and I started to think about how Google could make money with search. And for me, it's ad revenue, right? um
It's ad clicks, real simple. um So when I started to think about how they could make money, it was all about maximizing exposure to ad clicks and revenue generating clicks and building interfaces that allow them to accomplish that. And I started to look at the different metrics that maybe would matter to me if I was a product manager, things like time on page, actual revenue clicks.
clicks to conversions, et cetera, the cost to actually produce a SERP, what are the LLM costs to produce an AI mode response? What are the storage costs to store all that information? What are the costs to host anything? So, ROSP is return on a search page. And what that means is Google has specific costs to produce that specific page for that specific query. And then how much
Noah (4:56.357)
Revenue or they generating in relationship to that to that cost and I thought of it just like back in the day when I worked on For lots and lots of bike shops and I would always build these marketing flywheels that incorporated search and paid and I always targeted as high a return on ad spend or row ass as I could so I took that like row ass mentality to think about To kind of switch my perspective around and to think like okay, so if I was Google
What would I want to do? I would want to make as much money as I can from every single search page that I produce. But that can't be the only thing, right? Like uh to accomplish that, I also have to mean different things to different stakeholders along the way. Like users have to have an experience that they actually enjoy. The results that I produce have to be good enough. I have to produce an interface that makes people want to click.
Jon Clark (5:37.272)
Right.
Noah (5:55.217)
I have to be, I have to make that experience delightful for users. You know, like there's all kinds of user engagement metrics that have to be in specific ranges. Like I want to make sure that people click X percent of the time where X is on a rain, like think of it as a curve of, of interactions that people could have. want to target a specific range on that curve. And so I kind of felt like if I was Google,
Jon Clark (6:14.274)
Mm-hmm.
Noah (6:22.897)
what I would do is try and maximize my revenue for every single search query or every single bucket of search queries or even category of queries. And at the same time, I'd know that a successful search experience is one where I'm making as much money as possible while also having my engagement metrics living inside specific ranges. So that's kinda how I was thinking about
Jon Clark (6:46.638)
I mean, it makes a ton of sense. Obviously, revenue is the primary driver behind Google's business model, right? And so I would imagine they're doing tons of calculations similar to that to understand, you know, ad placement, even the integration of ads, how they've become less and less more defined in search results, right? So now it's very difficult to distinguish between the two. um You know, they are now including
ads in AIO overviews. um Do you, I guess based on that methodology, how
How would you think about them integrating ads into the AI mode experience? Have you given any thought into that?
Noah (7:30.291)
uh Yeah, yeah, and I talked about that in the article too. Like, let's say you're owning a product and you want to maximize revenue, you have to start with a baseline, right? And so what's happening now, they're launching AI mode and it's a pretty clean interface. There aren't a ton of rich results inside there, right? So they're developing an understanding of what the baseline metrics are for how people click.
Jon Clark (7:51.808)
Right.
Noah (7:59.911)
you know, pixels from the top inside of an AI mode experience. And they also have a lot of historical knowledge around how people interact with web pages based on pixels from the top and also inside of different rich result experiences. And so they're going to take all of that historical knowledge and all of the knowledge that they're gleaning now with all the U S users who are using AI mode. And they're going to be able to understand like, Hey,
we expect someone to click X percent of the time if they're in position one or X pixels from the top. And then they're gonna start experiment, like if I was Google, I'd develop baseline insights into how people interact with the interface. And then I would start testing like crazy. Just like we as SEOs always think about testing like crazy, right? And so I would test, uh I would test,
placement of different rich results over time, and then glean insights from that and continually test and test and test. And so I think what we're gonna see is they're gonna start throwing in ads, they're gonna start throwing in local packs, they're gonna start throwing in different types of rich results and flight trackers and hotel booking widgets and all that kind of stuff so that they can monetize as much as possible for every query, every bucket of queries, every category.
of buckets. Does that make sense?
Joe DeVita (9:32.548)
Yeah, on the subject of measurement, um it feels like a pretty wide consensus that
Jon Clark (9:32.780)
Yeah, completely.
Joe DeVita (9:43.600)
organic search traffic is falling for just about every client and agency. But the position from Google, it sounds like from last week was you may be receiving less organic search traffic, but the traffic that you're still receiving should be of a higher quality. I got that sense just from some summaries of people who were at IO.
And I was listening to a conversation you had with Joy Hawkins a couple of days ago. And I kind of felt like your advice to agencies and clients was, we're going to have to think about measurement a little bit differently. um maybe I heard this wrong. correct me if I heard it wrong. But we're going to have to monitor over the coming months the dip in organic search traffic. We'll have to also try to understand where
are other traffic patterns changing. Maybe it's advertising, maybe it's social referral, wherever your different traffic sources come from, they may be changing because of this impact on organic search. But hopefully because the quality of the traffic coming from organic search has improved, there won't be a big drop in conversions. So maybe your conversion rate goes up from organic search in the months to come.
So I just, did I hear that right? Is it more about like, just you gotta pay closer attention to how all your other traffic sources relate to organic search and finding some correlations? Is your business still succeeding? If your business is still succeeding, there's gotta be a correlation somewhere, some other source of traffic. Did I hear that right or no?
Noah (11:26.043)
Yeah, you did, but it, but it also is really critically important to think about like what bucket of websites are we talking about? Like we were talking about Barry Adams earlier today. He dropped in an amazing post on LinkedIn about how basically news publishers are going to be losing a dramatic amount of clicks and clicks determine their revenue. Not, you know, the click is the conversion for them.
Because once people visit the site, then they get revenue based on ad impressions, right? And so, ah but for other buckets of websites, think about any kind of small business website or any kind of B2B site, just non-publisher sites, I think the thing to measure is the conversion action. And conversion actions are different depending on the website. Is it a form fill? Is it a phone call? Is it a text message? Is it a chat?
You know, any of that stuff is where I'd focus. And I'd incorporate a couple different things. I'd look at, we want to know where we, you know, I think about conversion actions and then all the stuff that leads up to that as leading indicators. What's happening with impressions, clicks, click through rate, and position. That's coming from Search Console, right? And then,
Jon Clark (12:50.318)
Mm-hmm.
Noah (12:50.961)
let's go wider and bigger and think about GA or Google Analytics 4 and think about, hey, what's happening in my organic channel? Am I seeing overall traffic up or down? Am I seeing time on page different, i.e., hey, it's higher quality, people are engaging more fully with my content when they land there, and that's a leading indicator to get us to the actual conversion action on the website that we're tracking with GA4. And I'd look at
that conversion rate based on, on, on that organic channel. I don't know. I haven't looked at this enough. Like I got back from the conference and it was like, I had a bunch of time to put together an article. And then I just had a bunch of projects that I had to dive in full time on this week because of recording two podcasts, which is not like a normal week for me. So it's like trying to fit it all in, but I think it's critically important to like.
Jon Clark (13:41.326)
Thank
Noah (13:50.003)
You know, Google is saying, and I, and I took this note down as I was like reading, reading their documentation, like, Hey, we're sending you higher quality traffic. I'm like, okay, great. How are you measuring that? You know, how can we measure that?
Jon Clark (14:3.310)
Yeah, I actually wanted to ask about that because there's been a lot of comments coming out of Google from their spokespeople around how traffic's actually increasing and people love AI overviews. But if you look at industry studies or you just talk, right, like we're having a conversation here, I would guess there's been many a time where we don't like that result.
Did anyone grill any of the engineers at IO about this? did they try to get to, like, where is this actual data that you keep referencing? Or is it sort of a black box?
Noah (14:44.243)
Um, it wasn't, it was, uh, that was one of the funnier parts of the discussion, think. Um, to my mind anyway, was one the funnier parts. I don't remember who told them that clicks in CTR were down like crazy. And Google engineers at that point in the conversation, if my memory serves me correctly, leaned in.
and the look on their face was one of surprise. And I think I told them that, I don't remember, but.
I said, are you surprised? Really? And then I started cackling like really loud and other people started laughing at that point because it was, it was such a disconnect. was like, how could you not know the clicks are down? You know, at different parts in the conversation, uh, engineers said things like we measure hundreds and thousands and millions of different things. Okay. So that's.
I'm holding this in my head on one side and then on the other side, it's like, we don't know the clicks are down. And it's like, wait a minute, that's like one of the four, like the four metrics that we get to see in search console, right? We get clicks, impressions, CTR and position. So like that was wild. And then after the discussion, I was talking with another SEO named JC Schrenard and he was saying, Hey man, they look at this stuff differently than we do.
Jon Clark (15:55.768)
Right.
Noah (16:21.191)
We look at it from the website perspective of like, hey, my site is getting less clicks for this query on this page. And Google looks at the success of a search result page as like, does this entire SERP get more clicks? Does this entire SERP have more impressions? Does this entire SERP drive more revenue? Potentially they might look at that for.
from their perspective. So that moment and that serious disconnect was probably, he thought was caused by that kind of like completely different perspective. And yeah, it was wild. It was just a wild part of the conversation.
Jon Clark (17:9.986)
Yeah, it's sort of interesting to think about it more deeply because the focus of Google as of late has been to keep you in the search experience. I theoretically, right, until you click on an ad and they make some revenue. So it makes sense that maybe the metrics that they're looking at are different. Like maybe a key click metric for them is that I've clicked on a couple of different.
options that have gotten me to a search results page that ultimately I clicked out to an ad, which, you know, maybe in their mind they translate as a more valuable click and therefore it's sending you more valuable traffic. I can see how they could extrapolate that. It's definitely not, I think, apples to apples on what we're accustomed to reporting against, but I guess I can see how that would make sense. One thing you said earlier,
Um, Was around, you know, the, the sort of like the fan out technique. And you talked a little bit about this in some of your, your recent articles as well. wanted to double tap into this as well, because you sort of showed, um, an example on a recent webinar where you, sort of maybe guess that that, that those were those sort of fan out queries that Google was implementing. Have you thought any more deeply around how to
sort of pull those out and sort of, yeah.
Noah (18:36.615)
Yeah, so I have and I've played so and anybody can do this too. Like any, I think most SEOs, if you have not experimented with dev tools in Chrome, I feel like you're missing out on a lot of visibility into how the web works period, right? But like if you open up Google and you go to AI mode and you conduct a search, open up dev tools, go to the network tab, clear out the all of the stuff so that
it's clean, there are no rows of information, conduct a search, and then you'll see row by row all the stuff that happens after you conduct a search. And you're gonna see different rows. Some rows will be like a JavaScript file, some rows will be a CSS file, some rows will be an image, and some rows will be like an API request. And uh I found one that...
appeared to me to be the query going off to Google and I talked about that in the webinar with Joy. And what comes back is HTML that gets written via Ajax into the AI mode experience. And there's a whole bunch of divs, um know, like HTML divs inside the search result page that are made visible, have text written to them, and then are hidden once that Ajax code gets uh
pushed back in and I forget the exact parameter for it, but we could, I guess we could do that and I could show you, but the other thing is Mike referenced another parameter which was, I it's like list threads. And if you open up the network tab, you can filter by things and you just type in the word list and you'll see that request right away. And my understanding of what that is, is that it's,
uh that that's just building the history, the history of conversations inside AI mode. And his takeaway on it was different. um I get the sense that he believes that that's another way that Google is personalizing the results to you. And it's another way of offering context. I don't know that I agree with it. And the reason why is that, let's say I'm asking it about like, what's the best, and I don't drink anymore, but what's the best high ABV,
Noah (20:59.283)
hazy IPA and then I ask it about building a data pipeline, that other search in my history really doesn't offer any kind of context, I would argue for that next conversation. Yeah, but that's what's returning in the list threads response. So I don't know. My sense is that I don't see how to reverse engineer the query pan out.
Jon Clark (21:11.406)
Right, because there's no relationship between. Yeah.
Noah (21:28.165)
in API requests. My sense is that, you know, Mike has this tool called Q4ia that I haven't played a ton with, but I definitely think it's rad and I definitely think people should play with it. em I mean, if we go back to first principles thinking, like we always start with like, what do we know? uh We know that Google is saying that they're using Gemini 2.5 to produce AI mode results.
Jon Clark (21:34.862)
Yeah, I was going to ask about that.
Noah (21:57.499)
And so a part of like reverse engineering, how query fan out works is going to involve using Gemini 2.5 to understand like how those queries are built. And then I think it's a lot of testing and experimentation. Like going from a query to break the query down into the entities that are involved and then fanning out to under like understanding what bucket it goes in and then asking questions of Google's different knowledge graphs.
based on what that entity means.
But beyond that, it's like total black box land. think we're all, I mean, the thing that was interesting to me was that when I was talking to the other technical-ish people at the conference, that's all we wanted to talk about was query fan out. And Google has three patents that talk about, uh one is like stateful search or stateful chat um experience.
And then there's another two patents and I believe Mike King's talked about this both on LinkedIn and in his articles. And I think investigating those patents is probably going to be the most useful way to learn more about it. um The question, I mean, the bottom line question is like, we perceive as search practitioners that things just got super complex.
and that they got very, very technical. And the question is, is that the case? Right? Or are we doing the same work that we've always done? And I think, and I want to cite Mike again, because I think the dude's brilliant, like vector embeddings and looking at web pages as chunks of text. And I want to cite Cindy Crum too, because she got us all smarter back in the day.
Jon Clark (23:39.715)
Right.
Noah (24:2.909)
thinking about Fraggles, which are the little passages that if you perform a search and that result takes you to a middle of the page that's all highlighted, at least for me, it's like a light lavender highlight, purple, whatever. That is Google saying this little passage of text is the best answer to the query. And so it makes a ton of sense that Google is probably using that same type of technique, so we need to do our vector embeddings around
Jon Clark (24:15.704)
Mm-hmm.
Noah (24:32.541)
passage lengths. And then as we're building content, I think uh to compete and win, um we're probably going to need to focus on the core things that are important for every single business. Like what product or service is the most important and what buyer journey around that product or service is going to make us money and then optimize to win that battle. Like that feels like the core thing because the work
that's going to be necessary to win an AMI mode, AI mode is huge. To get one of eight results or 10 results or five results, there are a lot of people producing content to win for every product and service. And so the amount of effort that's going to take to like optimize every single passage on every single page of every single buyer journey.
That's a really difficult task. And so that's why I feel like you start with the journeys that are going to make you money. Optimize and test for those. Probably start with bottom of funnel and work your way up. Um, and just test, test, test. And if people don't have testing in there, the way that they roll either as an individual or as an agency or as an in-house team, it's going to be very difficult to compete and win. Um,
Like there's this killer's book called, I'm trying to remember the ex CEO of Procter and Gamble wrote it called like, it might just be called strategy, but the core takeaways are where to play and how to Right. And um for me, the where to play is, are the buyer journeys that we care about.
And how to win is going to be gleaned through a whole hell of a lot of testing. And uh yeah.
Jon Clark (26:34.434)
So, I mean, it's sort of interesting, but you could almost imagine applying a ROS type of metric to your own pages in order to better identify which pages you really want to put that effort into of vectorizing, uh you know, chunking that content up and then sort of measuring that performance over time. um Yeah, super interesting. You talked a lot about the sort of the journey and optimizing content sort of around that journey.
Noah (26:52.264)
Yeah.
Jon Clark (27:4.563)
And you mentioned a little bit around how YouTube and local were basically like non-existent at Google IO. I'm curious your thoughts around how those might play into the future or how you could maybe see those changing as Google maybe shifts their focus to them once AI mode is more prominent. Like any thoughts there? I know you do a lot in the local space on the bike side and things like that. So.
Noah (27:12.499)
Mmm.
Jon Clark (27:32.590)
And from a product perspective, I think you would bring a unique perspective there too.
Noah (27:36.403)
Yeah, great question. uh My head's swirling a little bit, so bear with me for a sec. So I work at the best local SEO agency on the planet. Having said that, that's Sterling Sky. Having said that, I'm like the guy behind the scenes who builds all our tools. I build data pipelines for Search Console, data pipelines for Google Business Profile, uh LSA data, ads data, different rank tracking stuff.
Jon Clark (27:42.316)
Yeah.
Jon Clark (27:54.606)
You
Noah (28:5.093)
And then I combine it to help our team do the stuff that they do to compete and win. like all the local SEO testing is driven by strategy that our team designs out. Enjoy leads. She'll have specific questions and she'll bring a question to me and I help her answer the question. I help her get the information that she needs to understand something to make a strategic decision one way or another that's going to drive how we operate.
But um in terms of like what I see the crystal ball of the future for local being, I still feel that like that concept of clean results, understand baseline actions, throw in rich results over time. I know that local stuff is making its way into AI mode already. We did a personal injury search the other day in AI mode, and then we switched to all.
And what was funny was that we did a personal injury. don't remember where the personal injury search was. It might've been Los Angeles or Miami or something in AI mode. got a local pack and wait, it wasn't a person. Hold on a second. I don't remember what it was. Maybe it was like best car under. Was it the best car under 30,000 with a turbo? think. Yeah. We, we, yeah.
Jon Clark (29:16.278)
Really? Okay.
Joe DeVita (29:21.393)
Atlanta, it was in Atlanta.
Joe DeVita (29:28.338)
You too. We saw.
Noah (29:31.251)
So it was the best car under $30,000 with a turbo yielded a local map and local results under that, I believe, in AI mode. And we switched over to all, which is the standard way we search. And there was no map. There was no local pack. So I think we're going to see, I think the hard part for Google is that on the Google business profile side, just to get down to that little niche, like they moved all of the
management tooling into the front end of the SERPs so that if I'm managing my Google business profile, I have to do that from the search result page. So there's a lot of infrastructure stuff they're gonna have to deal with to incorporate both the tooling for the management side and into the search result side. it's unknown to me, but I know that it'll get there. That was one, you asked about local. What was the other, there were two things.
Joe DeVita (30:28.082)
You too. You too.
Jon Clark (30:28.631)
YouTube, like video.
Noah (30:30.011)
YouTube, yeah. I think YouTube is the same kind of thing. I think YouTube is going to find its way in there. No problem. And the reason why is that they talk ad nauseum about multimodal. People are searching in new, harder, more interesting ways. They're taking pictures of things with tools like Gemini Live or with their phone or the Google app on their phone. And they're saying, what is this? What does it say? You know, like it might be.
uh stuff that's in a different language, translate it, or um show me what type of tree that is, or what type of butterfly, whatever. So they're doing video search, they're doing image search. I don't know about video yet, but um they're doing image search, and they're doing text and voice, and so it totally makes sense that video is gonna be represented in the SERPs.
Jon Clark (31:20.640)
sense.
Noah (31:20.689)
Now, the bigger question is how do we help our us, our businesses, plus whoever we're working for, if we're in an agency environment, how do we get them the same levels of visibility and the same levels of leads or success, whatever their success metrics are?
Noah (31:46.341)
I keep thinking about getting outside of the search engine and I start thinking a lot about video content. I personally do not have TikTok on any of my devices, but there's a huge amount of money to be made in TikTok. I don't quite know how to navigate it, but I, a personal level, do not feel comfortable with any of my data going outside of the United States. And on the YouTube side,
I know that video is incredibly important and people interact with video in ways that they can't through text. um Whether it's learning how to cook something, whether it's learning. uh For me, I was buying pans last year and I was watching all these like video reviews of specific pans and they would say like, this doesn't do this one thing well. And they'd show you it scarring or burning eggs or
whatever. um And it just, I know that they just dropped Veo3, which is an AI video production tool. I still think that there's something incredibly powerful about humans on video. And you guys believe in this too, right? Like you're producing a podcast because you believe that it's going to help your agency appear to be
Jon Clark (33:10.092)
Right.
Noah (33:15.151)
an amazing resource and industry thought leaders and all that stuff. Yeah, and including video. And you know that it works. And we at Sterling Sky know that video works. And we invest a ton of energy into producing content, both that helps the industry at large, but also helps us drive leads for our business. And um I think it's one of the most strategic things that Joy does.
Jon Clark (33:19.572)
including video.
Noah (33:44.441)
in a huge variety of ways that she's strategic. Like people don't give her enough credit, I think. Like she's effing brilliant and she wouldn't really dig me using the effing, but she would love the brilliant. yeah.
Joe DeVita (34:0.283)
In addition to TikTok and YouTube, LinkedIn prefers video too. It seems like the feed is, my feeds are full of video. And maybe we just double down on content, because that's one of the things that brands do have control of, the types of content that they develop. I think we're all getting comfortable with the next year or two.
Noah (34:5.170)
Yeah.
Joe DeVita (34:24.530)
things are gonna be in flux for clients and how they organize their SEO strategies and investments. But the one thing that I think is pretty clear is you should, your content strategy should shift a little bit toward producing the types of content that you or your company is uniquely qualified to put out to the world. So whereas in the past we might suggest a different kind of hub and smoke, hub and smoke approach.
to creating content around central themes. Now the focus maybe leans more toward, you're a dentist in Atlanta or you're a car dealership in Missouri or whatever. Like what are you specifically qualified to answer? Don't try to answer big general questions just because you wanna optimize for.
dentist in Atlanta or a car dealer in Missouri. What are you specifically qualified to answer? think that's a little bit of a shift in content strategy, I think.
Noah (35:30.493)
I agree with that, but there's also like weird content pockets still. Like I was thinking about, um there's this bike ride near me where you can climb this hill called Flagstaff Road. You can climb 3000 feet and four and a half miles. It's really hard, right? And to do something that hard, generally you need to set up your bike.
so that it has like a hill climbing setup. Imagine trying to take your bike and to transform it so that it's easier to pedal going up super steep stuff. Your bike comes with a specific set of gears and chain rings, which are the things with all the teeth in the front that are connected to your pedals. The size of those things determines both the chain rings in the front and the gears in the back.
Those two things combined determine how hard it is to pedal up a mountain. Now imagine trying to solve the problem of how do I buy a hill climbing drivetrain upgrade. That content does not exist on the internet. And that drivetrain, if I was to upgrade because of the way my bike is set up, it would cost me well over $1,300 to upgrade my bike.
because there's no upgrade path for my specific drive train that doesn't include an electronic upgrade. So I have to upgrade from manual shifting on my speeds to electronic shifting. And that means that once I make the shift to electronic, bam, 1300 bucks. And ah so I'm still thinking through this. I mean, my mind is stuck.
I feel stuck in a couple different ways. The first is there are still content pockets. Is it worth trying to get those? Does that align with our business goals? Should we build things that we see we can win? The second thing is what does the landscape look like in a place like, let's pretend we're in Houston, Texas, and I want to get business as a painter in Texas.
Noah (37:54.173)
There are probably two to 500 painting companies in Texas, in Houston, right? How could I possibly hope to win when AI mode is only rewarding the top X number? Like, so that I have a really hard time thinking through. And then um just generally, I'm thinking about page structure a lot and buyer journey structure. Like, is it one big page?
that has tons of internal linking in the page, thus making it super easy for Google to understand how it's all connected and relevant. Because I am seeing tons of listicles. If you search for anything best of uh inside of AI mode, there's tons of listicles. And what's wild about that is you'll say best widget under y price. And it'll return a listicle.
that has some results from that listicle that meet the conditions and other products in that listicle that do not meet the conditions and all of those products will get returned in the response. So like, is it building big long pieces of content or is it building hub and spoke stuff that has awesome internal linking? Like those are the three things that I'm thinking personally a lot about. uh
I can see what I think Mike's doing, right? You look at that big article they wrote, it's 12,000 words. That's an ultimate guide to, right? And if you look at each different section, he's treating it like passage ranking for each individual section, and each individual section is talking about the specific questions that someone would ask around that topic, and it's answering each of the individual questions around that topic.
So I think that's a great piece of experimentation to pay attention to if you want to see how someone's testing that's smart. That's a great example.
Joe DeVita (40:5.790)
Fuck.
Jon Clark (40:6.486)
In fairness, Mike's articles have always been quite long, but yeah, I totally get that. uh That is definitely an interesting one to cover. I think the other interesting thing about that article is it's one of the newer topics. um And so he's almost like first to market on a lot of the conversation in such a post.
Noah (40:9.917)
Totally. Oh yeah, yeah, yeah.
Jon Clark (40:35.254)
I think that positions that content really, really well. uh
Joe DeVita (40:41.610)
Can I ask Noah if I know the information is really fresh? It's just from last week. But if you think about the being in flux for a year or two as Google figures this out, the transition to more AI mode, maybe as, but. m
Do you slow down your SEO investment on big projects? If you had a recommendation to build some crazy calculator or some big development investment, do you slow down on that and focus on things you can control a little bit less with less money, like the content and structured data? think maybe we can lead into that too.
Noah (41:25.523)
Great question.
Noah (41:31.123)
Honestly, I don't know. I mean, if I'm being totally honest with you, I don't know. And, um...
I think we are collaborating with our clients to come up with the answer to that question in different ways, depending on the client. Like we're doing a ton of different projects that would fall into that larger, um larger bucket, but clients that rely on a ton of like top of funnel traffic, like we're probably going to be pausing how we think about top of funnel stuff, you know?
tons and tons of blog content that falls into that commodity type content. That's stuff that I think is becoming clear that we're gonna do less. And again, I wanna be super clear about this. I do not drive strategy for our agency, period. Like, that's not my, this is not my bag. I'm the one giving them the information to make decisions. So I wanna be clear about that.
Jon Clark (42:10.582)
and
Noah (42:37.681)
I'm not speaking on sterling sky when I say that because they're smarter about strategy than I am based on what they look at all day long. You know, um, so I just want to be clear about that. Um, I think the calculator things still are driving traffic and they're still driving conversions. I mean, like that's still w and we have tons of sites, like think about any personal injury attorney. I don't know if you have ever worked with any, but
Jon Clark (42:51.350)
And so the.
Noah (43:7.165)
calculators are super efficient at driving traffic. The problem with trip with calculators is that every state has different laws around all the stuff that someone would build a calculator for. And not only that, a lot of it is up to the discretion of the judge. And so like, it's this weird thing that you can't just like, cause we actually thought about building calculators across all 50 states for all different types of attorneys and all that. And it was like,
Whoa, this is huge. Like we can't get it. It was like, oh my God, hey, where am I going to get the data? Right? Like, how am I going to understand what the laws are in all the states and how much of this is up to discretion and what are the ranges of settlements look like and all that stuff. But em I think those are still smart investments.
Jon Clark (43:39.214)
This is
Noah (43:58.267)
I hope that's helpful.
Jon Clark (43:58.754)
Yeah, I that was always the risk with AI, right? Like the nuance use cases of, like you just mentioned, right? Like very specific differences between state, even maybe tiny differences in some cases. uh There's just so much opportunity for error. Yeah.
Noah (44:16.977)
Oh, they're massive. They're not. Yeah, but that's the thing. They're not small. It's like. Like the amount of years that someone has to pay alimony like some states it might be capped at five. Some states it might be 20 like just imagine for bill, you know, a billionaire couple that breaks up. That's just like hugely impactful difference in terms of outcomes. But yeah, wild. um I felt like.
Jon Clark (44:40.481)
for
Noah (44:46.033)
I haven't been to a ton of these IOs. This is the first one I've attended. I was talking with Jordan Cooney from Previsible the other night, and he was talking about the fact that he'd been to six of them. so um having an opportunity to talk to with Google engineers is pretty special, and they do care about users. That was clear. um And they do care about producing a sensational product. That much is also clear. um
My takeaway was that it felt like a big deal. Like it felt like, it sounds super tripe, but it feels like we're entering an entirely new era and it's like the slate is clean and we're all starting again from scratch. And I think the agencies and organizations that use testing heavily, like every single thing that's in our SOPs has been tested. And um if you're not testing,
you're going to lose. It's just that simple. It doesn't how good you think your gut is. Like um it's just a whole new error and it's a new search surface. And um if you're not testing to learn how to win in it, you're not gonna win.
Jon Clark (46:3.106)
Yeah, yeah, it's a great takeaway.
Joe DeVita (46:3.582)
Can we talk a little bit about structured data? It feels like that may be the best place to experiment with in the short term. Do you feel like there's any advice you would give?
Noah (46:16.039)
I think it's a fertile area to test. think Google has offered lots of mixed messages on it. In documentation, they'll talk about how structured data and specific uh files that target uh large language models like an LLMs.txt file are not necessary to appear in AI mode. And then when you look at the documentation that they released the other day that talked about
ways to appear in AI search, they talk about structured data, which is the exact opposite of what other docs say. the gray area is where we test, right? Hey, they're saying two different things. And then there's the other thing of Google says X, okay, let's test it to confirm that that's true, whatever the X is, And um so I think it makes sense to test it. um
Jon Clark (46:51.182)
All
Jon Clark (47:15.672)
When the whole debate on LinkedIn around like, is schema useful? it not? I think one of the challenges with interpreting the results is more than likely if you're ranking well, you're probably using some sort of schema. So if you only evaluate pages that are included in AI overviews, right? The propensity for them to have some sort of schema is probably high. And so you almost sort of skew the tests.
just by nature of the results that you're evaluating. So I always thought that that was a, you know, something sort of tough to overcome when you're trying to decide before there's an official announcement from Google on yes or no. And so.
Noah (47:48.019)
Mm-hmm.
Noah (47:58.781)
Here's something wild. Have you ever done a correlation study where you basically take rank data for an industry? So you're basically thousands and thousands of queries and you look at as many different data points as you can in relation to rank. And we did this for dentists, I believe, and we collaborated with Places Scout. They gave us access to rank data.
for all kinds of different dental terms across the United States. And their rank data has a lot of different data points, lots and lots and lots of different stuff. So you can correlate reviews, review velocity, distance from the centroid, all these different things. And what I took away from that study was that very few things correlate very little to rank.
Like it was like, it was like everything correlated very, very, very slightly. And it was like, this story sucks. how do we tell this story? You you want to see like one thing relates perfectly to rank. I, and I think of a lot of this stuff as it's like, um, well, there's two ways to think about it. There's many ways to think about it, but one way is the allergy analogy where, Hey, I'm doing great. And.
Um, my health is dependent on when my coffee cup runneth over, you know, like, Hey, I got some dust, throw that in the cup. Hey, I got some, some, some grass. Hey, I got some dog hair. Hey, you got all this stuff. And all of a sudden when enough of those allergens rise to enough of a threshold, Hey, I, you know, I'm having an allergy attack or the, yay, my website ranks or my red website tanks. If it's the wrong signals, right?
That's one way to think about it. um
Jon Clark (50:1.762)
So is the correlation, oh sorry, go ahead.
Noah (50:1.915)
In the sec, yeah. Yeah, no, that's fine. We can move on. There's more ways to think, but it doesn't matter.
Jon Clark (50:8.302)
I was going to say, is the correlation then you actually need a lot of these things working together? Yeah. Yep.
Noah (50:13.927)
Yes, yes, 100%. And oh, the other way is this concept of the ABCs of ranking. And the ABCs of ranking are a reference to, uh if you look at the Google trial docs, uh something got released a couple weeks ago. m And they talk about the ABCs of ranking. One of the Google engineers talks about it. And the A are anchors or lengths. B is
the body, which is content, and Cs are click data, which are user metrics derived from Google search surfaces and from Google Chrome. And that those three sets of signals are manually tweaked by Google engineers in order to control how the algorithm works. Up until now, that's how it's worked. so like, when I think about ranking in the future, I wanna know,
how much to each of those three things matter. And it feels like the B signal, the body signal is being turned up big time, you know, and that's what vector embeddings are gonna matter.
Jon Clark (51:25.994)
I mean, in the other way, the C value may be getting turned down, right? Because if what we're seeing from clicks going out of search results in AI mode, et cetera, are decreasing, it would seem like the content piece or the B piece is getting turned up and the C piece getting turned down. um Yeah, interesting. um Well, I definitely wanted to...
Noah (51:41.373)
Mm-hmm.
Jon Clark (51:53.290)
Maybe do a hard pivot into the SEO community. em I've been I've been part of the group for quite some time. I'm more of a lurker to be honest than a contributor, but anytime I've come to the group with a tough question, there's always been plenty of folks who are willing to jump in and help me figure it out, which has been amazing. And you guys are coming up on your two year anniversary. Is that right?
Noah (52:16.125)
Yeah, yep. Two years in July. like, I'm like, oh my God. It's either been super fast or just like forever ago. It just feels so, it always feels like a lot. You know what I mean?
Jon Clark (52:17.282)
Yeah, yeah, so congrats on that. oh
Jon Clark (52:29.430)
Yeah, yeah. And it must be a ton to manage. Like, how do you organize your day around everything that's happening on that side, right? Like, the user engagement piece of it must be just overwhelming at times, and then also having a full-time job and other things. uh You how do you work through all that?
Noah (52:51.835)
Good question. um The answer is discipline. And when I struggle, it's due to struggles with discipline, right? But I think the thing that I try and do is I do community stuff early in the morning. So like 530 to 730 or something like that. I'm doing community stuff. And then throughout the day, if I get messages, I answer them in waves, usually like lunchtime or whatever.
Jon Clark (52:56.066)
You
Noah (53:23.079)
That's just my personal inputs and outputs, right? But like, how does the community operate? Well, I think the thing that makes it awesome, not the thing, but uh some things that I'm proud of are the fact that like on day one, we came up with a code of conduct. Like literally the first day we came up with a code of conduct. And that's morphed over time and updated and stuff like that. And I asked for help moderating the community pretty much on the first day.
And um I did that because I had built another community called agency automators that had about a thousand people in it. And it was really active and robust for a while. And then it slowly started to do the decline. And the reason it declined, I think was I just ran out of steam and I didn't want to have that happen again. So I decided to ask for help right away. And there are about 20 people that helped moderate the community.
And then there are a of folks behind the scenes that help us with different stuff, whether it's promotion, helping produce the campfire chats or the ask me anything events that happen. m Some people help me a lot in building the email template. And um I've sort of taken on the newsletter as uh something where I want to have a voice. But it's hard. mean, like you mentioned that question of like, how do you get it all done?
um I think the only way to do it is to like basically have a core set of priorities that you want out of life. Really goes that big, right? And what are the things that really matter to me? My wife, my kids, my dogs, um having a house that I can call my own, doing great work that I'm proud of, having an impact on others. Those are the lenses.
And so anytime I get an opportunity to do something, I look at it through that lens of, it going to help me accomplish any of those goals? Right? You know, having enough money to retire at a time of my choosing. That's another like core goal. If it doesn't meet those through that lens, I kind of need to say no. Um, and my ability to be laser focused like that is what helps me get a lot of stuff done.
Noah (55:45.287)
And when I stray from that is when I get just out of focus. And that happens, you know, like I had to cancel our campfire chat this week because I just felt overwhelmed. And it was like, you know, like I started a local SEO meetup, a Denver uh version of the SEO communities for a meetup. did one this past month and we had 30.
I don't know, 32, 34 people showed up and it was like, holy crap, this is great. This is making an impact. This is bringing people together, but it's also a lot of work. You know what I mean? So it's like everything like that. You have to look through a lot of different lenses.
Jon Clark (56:28.226)
Yeah, that's amazing. Um, I always find it hard to say no to things. Um, I've been, I've been trying to get better at it. Um, as our agency has grown, right? Like you really do have to prioritize if you want to, you know, get to those, those life goals. Um, so, uh, before we let you go, we, we've started a tradition on the show, sort of giving our listeners a little bit, something extra. We call it the page two podcast prediction.
Noah (56:33.875)
100%.
Jon Clark (56:57.966)
um So we'll basically ask the same question. um In 12 months from now, when you go to Google.com, what do you expect that experience to be?
Noah (57:10.717)
Mm.
Noah (57:14.109)
Based on how AI overviews rolled out, I think that AI mode is going to be the primary search surface for everybody everywhere.
Really. And what I mean by that is if you go to google.com or you open up Chrome and it opens with the search window and you produce a search, right now it takes you to the all tab. In a year that's going to go to AI mode. And all the rich results and ads and all that stuff that's missing right now is all going to be all over the place. And that whole concept that AJ Cohn talks about of like the inshitification of Google.
is going to be, you know, 2.0.
Jon Clark (57:57.890)
Yeah. Love it. Can't wait. ah Well, thanks Noah. This uh has been really insightful. Appreciate you taking the time. I know you're busy. ah
Noah (57:59.579)
Yay!
Noah (58:10.502)
Yeah, it was fun. Sorry I talked so much. I didn't give you enough like 30 second chucks.
Jon Clark (58:16.096)
No, this was fantastic. uh So before we let you go, if you enjoyed the show, please remember to subscribe, rate, and review. We'll see you next time. Bye bye.
Noah (58:25.637)
Excellent, thanks so much.