What happens to news when AI answers questions before headlines are even read? In this episode, Jon Clark and Joe DeVita talk with John Shehata—founder of NewzDash and GDdash, former SEO lead at Disney and Conde Nast—about the existential crisis facing digital publishers today.
https://page2pod.com - In episode #81 of the Page 2 Podcast, Jon Clark and Joe DeVita sit down with John Shehata—former SEO lead at Disney & Conde Nast and creator of NewzDash/GDdash—to unpack how AI and Google Discover are rewriting the rules of news distribution.
John shares his unique journey from developer to news SEO expert and SaaS founder, diving into how AI, LLMs, and Google Discover are upending everything we know about news visibility.
He reveals why keywords are becoming less relevant, how Google’s personalization and categorization engines actually work, and what publishers must do to survive—and thrive—in this rapidly evolving ecosystem.
Tune in for actionable strategies on paywalls, content refreshing, newsletter growth, and building AI-resilient newsrooms.
🐍 In This Episode
This is a must-listen for newsrooms, digital marketers, and SEO professionals navigating the shifting landscape of search, traffic, and monetization.
📎 Mentioned Links & Resources:
S4_EP081
===
[00:00:00] Jon: What happens to news when clicks disappear? When AI answers the questions before you've even seen the headline? Today we're talking with John Shahada, CEO of NewzDash and GDdash, one of the world's top experts on how audiences find news online. John has led SEO and audience growth at companies like Conde Nast and Disney.
Today he's building tools for the next chapter of news discovery one where traditional search is no longer the front door. This episode is about the existential crossroads facing digital publishers. We impact the collapse of ad-based traffic models. Why Google Discover has quietly overtaken search as the top traffic source for news and how AI overviews and LLMs are reshaping what visibility even means.
John explains how Google's personalization engines work, why content entities and categories matter more than keywords and what publishers can do to survive and maybe thrive in this new reality. John is building platforms that help publishers adapt, but he's also navigating a moment of deep uncertainty where long help strategies may no [00:01:00] longer work and new models haven't quite arrived.
Whether you're in news or traditional SEO, I think you're really gonna this one. Here's John Shahada.
John, welcome to the page two podcast. I'm joined as always by my business partner, Joe DeVita, moving Traffic Media. We're excited to talk news today, but I thought maybe we'd start with just the evolution of your career. I think the work at Conde Nast was incredible, helped solidify your expertise in the news field, and then moved into more of, I.
A product fo focused role. I'd love to hear a little bit more about just your career trajectory into more of a SaaS business, right? Or,
[00:01:47] John: yeah. John and Joe thanks for having me, guys. I'm very excited actually, to be on your podcast. So how much time do you have for the career?
[00:01:56] Jon: As much time as you need.
[00:01:58] John: Okay. I think I shared [00:02:00] this with few people, but I was almost a pharmacist, right? In another life. And then I think I woke up one day I grew up in Egypt, moved here to the states long time ago, and I and I woke up one day. I was like, I really hate pharmacy. I hate chemistry.
Why? Why am I doing this? And so I shifted college education to like computer science and software engineering and stuff like this. And I was like, so happy, so excited. Then I started working back in the days in an agency developing, products for them, like writing code and stuff.
It was a marketing agency. I was like, wow, this is exciting. What are they doing? I was like, and this is, I think back in the days of Alta Vista cus and all these kind of good stuff. You guys don't look old to know Alta Vista, but it's, older and then slowly is like moved slowly into the content side of things.
And dropped my coding. Every now and then, and I got my very first opportunity work in news and I'm very passionate about news. I don't know how [00:03:00] many years ago was a company called Advanced Digital, which actually the company that owns. Conde Nast. So it was the Penguin company, and these are all local news sites.
So I was this like young kid going around, I think they own like over 35 local news sites, like the Star Ledger plane dealer, like really big news newspapers and a lot of news sites like nj.com, al.com, mardi gras.com, NOLA and so on. So I'm like, I'm learning this exciting world of news and it was like the best thing about news.
It's it's amazing. Space between PPC and SEO. So if you think about SEO, right? It takes month to see the results and PPC, you put something and immediate results. News. SEO is like real time SEO, so it has the instant satisfaction of you work in something and you can see the results immediately, right?
Like when we're talking about news, but at the same time you work in an organic, so it's it was very good [00:04:00] world between the two. So I worked with them for 5, 6, 7 years. I remember back then Twitter and Facebook just started and I said, this social thing gonna be big. I even asked the owners to, to.
Speak was this new kid who did Facebook Mark Zuckerberg. It was like, what is the heck? And then after that, I built the very first social pla team, I think in the nation working on local news. So I did SEO and social for them, and I enjoyed it a lot. And then I moved to Disney. And in Disney I started as SEO and then moved to SEO and social I think this was like, back then there was no such a thing as audience development or audience growth.
So this was the closest to it. And we started with A, B, C News, ESPN. Then we went to Disney channels. We worked on building editorial teams for a, b, c entertainment. I worked in Oscars at the Academy Awards for I don't know, five, six years. And it was I. Very satisfying. And I work with [00:05:00] all the TV programs like, good Morning America, what would you do?
Shark Tank and stuff like this. So it's it was a good introduction, not just to hardcore news, but also TV and how we can take a program like Shark Tank and like what type of questions that people are asking and how to enrich this program. It was also a very good start for tV networks don't like to talk about themselves. So actually getting a, b, c entertainment to start talking about their shows, like Lost, if you guys remember, A Lost and all these great shows, right? And then after a while five, six years, I can't remember I moved back to Conde Nast, which is I think I only worked like three, four companies in my whole career.
And back then I moved in I built on the audience dev team for the US brands. So this is self wired, new Yorker, Bon Appetit, Epicurious and and all of those vanity Fair. So we built SEO teams, we built social media teams, and I think we did very good job. So they assigned me [00:06:00] to take on newsletters, and for me it was like newsletters.
It's this is really cool. And I didn't know much about email campaigns, but. I think when you apply the same kind of principles of audience growth, it doesn't matter what channel. And this is what I was trying to approach testing and learning and, all of that. So I did that and then I got promoted to the global VP of audience development strategy.
So I was managing all the SEO social newsletters audio podcast and stuff like that. And I had a small innovation team, like we were building a lot of products and also the all organic partnerships. So for seven years or eight years. And I think around COVID, it was very clear to me.
I had this entrepreneurial thing right when I was in college. This is how I paid my, for my college. I built a software that sells phone cards online. Back then, you had to buy the phone cards and scratch it to call, if you have a family overseas or something. So that was like a [00:07:00] really fun, and I was like, I always wanted to go back, but my career was actually doing well.
So I said during COVID when I start like, staying with my kids at home, it's I love this, I enjoy this because my commute was too long. So I start planning it and in 2022 I quit. And that's what I do. Full-time SaaS, just building products that I like for things I'm passionate about, like news, discover AI now and so on.
Sorry, it's a long answer, but it's
[00:07:29] Jon: it's perfect 'cause I think there's so many directions to dive into. On maybe a place to start is just looking back across your career. You've benefited from seeing these big sort of shifts in the industry, right? There was, first like really big algorithm updates that just wiped out sites than it was.
Mobile, then it was supposed to be voice. Now it's ai. Is there anything, that is fundamentally different about [00:08:00] this environment today compared to maybe some of those other, new technologies that were supposed to be, so disruptive? I don't think they ever fully came to fruition in the way that they were being forecasted.
I think ai, you could argue is maybe a little bit different.
[00:08:15] John: Actually, I totally agree with you. I think it's different from everything else that we saw before. I feel like every other disruption that came along to the industry was somehow additive. Right? Or enhancement or something that you can work with, like from mobile to voice to all the algorithm updates to the algorithm updates were happening on search, right?
So everything is even with social, if you think social at a certain point of time Facebook was actually sending more traffic than Google at a certain point of time and when, right. I was like, that's it. The dead of search. Like everyone gonna sit on Facebook and just find whatever stories they feed them.
Right. So we went through this, but I feel now this is different. Definitely. And there is a [00:09:00] whole argument online. I was like, is it the same? Is it different? What to call it? I'm not gonna get into this, but what I feel is this is the moment where everyone was going to blockbuster to rent v s tapes.
And people are, and Netflix just started streaming, right? And everyone is arguing this is a fad. This is it will not interrupt the industry. I feel it's the same thing. I feel, I'm not saying SEO is dead by any means, right? I think SEO will continue, people will continue to search. How will they search and where will they search, right?
And I think the same kind of principles of deciphering algorithms and understanding what works for our audience and how will we gain visibility. I think the same principles apply. But I think it's a pivotal moment in our industry. Not for every industry. Definitely evergreen content, top of the funnel, content, information or content will be impacted more than anything else.
But I wonder how news will be consumed, right? Not only in the next couple of years, [00:10:00] but when you start thinking about like Gen Alpha and Gen Z, right? It's like, where do they consume news? How are they gonna get news? How LLMs will surface news. Will Google still keep news unharmed from all the AI summaries and overviews and AI mode and stuff?
I think there are so many unknowns, but definitely the one thing that I know is like we are moving towards this direction and very quickly and very fast. Ca
[00:10:26] Joe: can we spend a little bit more time there? I really love the perspective we, and if you have a passion for journalism, you're always gonna wanna create some news.
But if you are a business type what are the ins? What are the incentive, what's the incentive today to start a news organization?
[00:10:42] John: I'm not sure about starting a news organization. I think it's, I think there is. There are gonna happen. A lot of consolidation will happen, right?
Where smaller and medium user organization will be consolidated in a bigger organizations especially that they have better revenue [00:11:00] diversification, right? Because like display ads and display revenue is gone. It's gone forever, right? Pennies, it's it's, it keep, it's keep giving cheaper and cheaper for the last 10 years, right?
So if you don't have a very clear, view of how you're gonna diversify your traffic or if you're not part of a bigger organization that probably be making their money elsewhere, not online or display ad or you don't have something like, hey people subscribe because my content, they can't find it anywhere else, and stuff like that.
So I think there are gonna be a lot of consolidations. I think local news will remain because you cannot find, this news was like bigger news organizations, the BBCs and CNN of the world, right? But there is a lot of content in between where people don't care much to subscribe or pay for. Right. It was like, hey, celebrity gossip and celebrity news and stuff like this, right?
So there need to be another business model. So I'm not sure about starting news. I think news [00:12:00] will evolve. If you remember the music industry was Napster and lemon, lime and all the stuff. It came to a point where it was like the whole industry well threatened and then they found, right, a new platform where okay.
People don't wanna buy a whole album in a cd. So I feel with news, and I said that to actually, I did a lot of editorial trainings and sessions with publishers strategies and so on. I said, what matters is the content, regardless, what is the medium? Yesterday was in new news on paper today was it's online tomorrow.
It could be a hologram view in your room, right? Or so that's what matters is the substance itself. I'm not sure if I answered your question, Joe, but it's
[00:12:45] Joe: I I think the answer that I heard there is from Napster to, it was like Apple Music, when you could buy one song at a time.
Yeah. That was the big transition. That was the business model shift. Yeah. That giving it away for free to artists getting [00:13:00] some portion of that revenue. So that hasn't been figured out yet with ai.
[00:13:06] John: It's coming along because a lot of LLM models are starting their publisher programs where publishers actually can if you think about Chad GBT, they have like their publisher program where you can join and then deselect who they wanna work with.
So there are some kind of direct deals happening. I think we are gonna move to a point where there is a revenue token revenue system. Where, do you know what, if we use your news in our training model or if we are grounding the search and retrieving live news from your site, right? We're using certain type of token and we will pay you every time we use your content or we even have it cashed or in the training model and we use it to surface.
So I think we're gonna move there eventually, and I think there will. There will companies that will emerge to be like ad networks and stuff like this will be like the middle, like interface between the news and [00:14:00] the LLM platforms. I think eventually we'll get there. I'm not sure how fast, but otherwise news organizations will be hurt tremendously if this revenue model doesn't evolve.
[00:14:12] Joe: I got one. Jon, I know you're chomping at
[00:14:14] Jon: the No, go ahead. I have one that build on that, but go ahead. Yeah.
[00:14:20] Joe: If you're a news organization today before that reality is this, for that tokenized tokenize re revenue sharing, do you block your content?
[00:14:27] John: I don't think you should block your content. I think you should.
Unless you are blocking your content or suing LLMs as a lever because you are I. Out there, and they really need your content. So they will do whatever to get your content. But if you are a small to medium site or business, if you block your content you're the one who lose. It's like blocking content from Google.
Back in the days Spain tried it news organizations, other countries, and then after two, three years they came back and say, you know what? Actually we need this. So I think we need LLMs in most [00:15:00] cases.
[00:15:00] Jon: So as usual, Joe asked exactly what I was gonna ask. But maybe to build on that if you we work with a local news publisher as well.
They have a whole bunch of local news organizations and they're, it's scary to the thread of having your content siphoned away and traffic siphoned away. And a lot of these news organizations are already under traffic pressure. And so I. In some cases, the default is well, we're just not gonna give that content away.
We're gonna try to hold on to, I think our point of view and discussing with them is exactly what you said, which is the longer you block it, the further behind essentially you get. And then from there on you're constantly playing catch up. Do you have any advice for I guess, organizations that are, really struggling with that assuming they're not suing right.
Like that would of course change the decision almost automatically. But, how do you approach organizations that are struggling with that decision?
[00:15:56] John: Sure. So I wrote an article about a week or two [00:16:00] ago. It's called the publisher's Survival Guide. And it's 11 different things to do and I'm happy to share it with you guys after the session.
I think it's very helpful. Took me like a month to write based on my experience. So I think there's so many things, but one of the things I, three things I mentioned is diversification of revenue. That is diversification of traffic and diversification of audience Diversification should be, if it didn't start already, it should start now, right?
It is one of the things I talk about also is like the building direct relationship with your audience. Publishers has been always dependent on other platforms to send them the traffic, the Facebook, the Google, and stuff like this. And they ignored things like building their own newsletters, right?
It's was not sexy, didn't work, whatever. And knowing that newsletters audience actually the most likely segment to subscribe to a pizza new subscription, it's the highest, it's the most loyal. These are people who said, do you know what? I like [00:17:00] your content so much. I'm signing up for your newsletter.
So building these kind of direct relationships apps don't work for everyone. Some new organization might have apps. There like a much higher friction than signing up for a newsletter and need some kind of daily habitation. I call apps as like the, our, day-to-day our times, bookmarks of the times.
Do you remember when we had bookmarks in the browser? Thousands of them. And then we end up clicking the same three bookmarks that we use every day. It's very much with apps, so you need to use it like every day for people actually to download it and use it. So diversification of your traffic.
Are there any other sources that I can get traffic from? Or is like mainly Google, Facebook, discover has been like a big sending the biggest traffic for news publishers above everything else, even above Google. Search. Most of the publishers now is like about 70%, 65% to 70% of their traffic from Google Discover is, there is an opportunity there.
And so on, right? Is [00:18:00] your audience young or old? Can you diversify this type of audience? Can you diversify your content? I think there are so many that can be done. Also looking into consolidation. I think I mentioned it. I think a lot of news organization will start consolidating because there is power on coming together, right?
Either consolidation as business entities, right? Or consolidation in forming unified fronts, right? To approach LMS with one United Voice, right? One single publisher is different from having a hundred speaking on all of their behalf.
[00:18:34] Jon: And I would guess you would have a little bit more bargaining power when you go to the table to try to get revenue for every time it's used, right?
[00:18:42] John: Yes, absolutely. Yeah.
[00:18:43] Jon: Yeah. That makes a ton of sense.
[00:18:45] John: And the strong, yeah, the strong carries the weak, right? Is so there is a strength in, in like unity and coming, like speaking with one voice.
[00:18:52] Jon: Yeah. That makes a lot of sense. I read, I was reading another blog post. We'll link to it in the show notes around some of the analysis that you did [00:19:00] around, the different verticals within news that were actually being impacted by AI overviews.
And one of the things I really like about your studies is the depth of data is significant, like millions of impressions or, and it's very different from some of the other studies that are like, oh, we looked at a hundred thousand. You know what I mean? It's just the context is much more meaningful in my opinion.
And I thought one of the really interesting things was, at the time of the study, I think it was less than 4% of general news queries were actually trigger triggering an a IO. And so I don't know if you have a contrarian point of view on how disruptive it is currently. I think future or forward looking is I.
I would guess it's gonna grow, right? But, how is, how disruptive is it really based on the data that you're seeing?
[00:19:47] John: So I'm not sure which which study we did. We did one last year and we did one a week ago, so a week or two ago. So let me open it here. So there are so much numbers, but I think the good news is [00:20:00] comparing last year to this year, we found that AI overviews, specifically AI overviews hasn't increased that much.
I think last year when we did it was like 3.9. This year is like 3.5, 3.6. So it's very constant, which is good news. But there are some verticals like you mentioned it's much higher than others. Like I think with health, specifically health, it's the highest. Right, followed by technology and science.
But when you come to top trends or like top breaking news or world news or sports news, we haven't seen that visibility of AI overview. But the interesting thing is we start noticing that I think the difference is last year we're only measuring on what's happening right now, right?
This is trending. Is there an AI overview or not? This year is like we start measuring how long AI overviews, if any, to start it. And one of the things, yeah, one of the things that we noticed is like for sports events within six hours, seven hours [00:21:00] Google will tell you, Hey, this is the recap of the game right.
This is what happened, stuff like this. Right. And I feel like there is some kind of a connection with like search demand or search volume or I. Content production, or it could be a mix of all of them. Right? And we like, when we see what, like demand starts, like slowing down the AI overviews, like start rising up and then if demand comes up, the AI view.
So it, it was a little bit different. So we're still testing, we're in a very early stages of testing, but I think health. Is the most impacted. So this is one thing in the previous study I analyzed like more on the evergreen side of things as well. Evergreen and top of the funnel kind of content and how to and service content and all of this gonna impact be impacted like in a huge way.
I think if you think about affiliate content is hey, these are the 10 best sunglasses for men, no need for metal man anymore, which is devastating for a lot of news publishers. Or the 10 password managers [00:22:00] for 2025, right? It's you can do this study in comparison and I think el and m will reach a point where like it won't need some of that content.
It just, it can actually analyze it based on the specs and everything else based on UGC. If you go to Reddit for example, right? You'll find a plethora of reviews. Actually firsthand reviews. I use this product that's horrible. Or I use this product because I couldn't find any other product that worked for very, my specific edge case.
That's why Reddit has been like Google has been using them and have a direct deal. That's why Reddit actually is more visible or impacting results that happen in certain l lms, and so on. Yes, I
[00:22:42] Joe: think I heard two pieces of advice there, and I maybe if I got this, if I get this wrong, please correct me.
But the one piece of advice may be that general content that you used to find some success with. In terms of SEO will become less important. Small publishers, small news organizations probably [00:23:00] should focus on now is their unique voice. What can they say differently from everyone else? The other thing I heard you say was around breaking news that's around six or so hours.
There's only gonna be one person rewarded for the breaking news in an AI output. Right. So is speed the advice there? Like you've got as an organization, you've gotta figure out how to get faster?
[00:23:20] John: Yeah so with breaking news, what we found is there is, AI overviews doesn't appear much, but in a certain type of news like events or games or stuff like this, they will start appearing.
And we're very early on, so like we're tracking for example, all what's happening, the craziness that's happening around the world and different regions and we wanna see what happen, like when will AI overview. But the one thing for sure is like. If the user search Google in a different way, AI overviews will appear.
So if you, for example, with everything that's happening now, if you if you search for Iran, Israel, right, [00:24:00] you'll see top news and you'll see probably like anywhere from 10 to 30 news articles in Google and stuff like this. But let's say you ask what is the latest with Iranis Israel? Now you are sending a very clear signal to Google is I don't wanna see, I don't wanna read the news.
I just tell me what's going on. Right? So the query makes a huge difference. Will people start figuring this out and say, do you know what? I don't wanna read the news. I just wanna see what's happening now with LLMs, right? Most of the LLMs right now tap into the deals that they have with news publishers.
So if open AI or TGBT, right, just oh, I have a deal with Reuters, BBC and CNN, and you ask me something like this, I will service. New news from there, because I already have all their feeds and they write so much content and they're like all over the place. But if you write, if you search for something very specific, and I'm trying to find something extremely specific. The what happened in the game yesterday between team one and team two the [00:25:00] injury or something, right? And let's say they don't have it, it's a very specific or a very niche topic. Then they ground their search and go to the web and try to find results and they give you the sources.
So I think there are still opportunities. The problem is people don't click much in change GBT, right? So the problem is the economy of like clicks. So that's the issue. They have to pay publishers in a different way because people don't click much unless they wanna finish the task with a purchase.
So I, I did all my research. Should I buy this phone or this phone? Radiation quality camera, all of this. Now I know exactly this is the phone for me. And now that even that is coming to l LMS where now you can like purchase from LLM, right? And you can buy the phone. So
[00:25:46] Jon: yes. Really interesting.
I think one, one area of news that at least so far doesn't appear to be as disrupted is discover. And for a long time that was a black box in terms of how to even get [00:26:00] content surfaced there. Folks like yourself, Lily Ray, others have done a great job of pulling back that bail a little bit.
Your talk at SEO week I think was really insightful as well. And I wanted to dig in there a little bit. So one of the. I guess technologies that you've built was G dash sort or GD dash or G Dash? Yes. It doesn't
[00:26:20] John: roll out the time very easily. It meant to be like Google Discover, so GD.
Dash, got
[00:26:25] Jon: it. Yeah. So maybe give a, 10,000 foot view of what that product does, and then I'd love to dig into Google Discover a little bit just in terms of some tips and tactics there.
[00:26:34] John: Sure. Absolutely. I've been building products for so long. When I left Conde Nast I start thinking about what can help news publishers gain more visibility, right?
And so we built our very first product, which is NewzDash. It was actually something I built when I left Disney. And before I joined Conde Nast, it was like how we can track every single trend who is ranking for these trends. So this was back in 2014. If they don't [00:27:00] rank for these important trends, what they can do and stuff like this.
And we have like amazing roster of clients like BBC Reuters N-B-C-C-B-S, even like niche clients Dodgeville Germany to media group. Netflix FanDuel like we can, okay, so then Google Discover start becoming like a really major traffic driver for news publishers. And the funny part is like when Google started Google Discover, they came to us came to me at Condé Nast and say, Hey, we wanna try with some of your some of your brands.
So back in the days, Google Discover was driven by RSS feed. So we give them an RSS feed. And so I was always like, oh, Google finally found the missing puzzle for social because they tried social so many times and they failed miserably in, in many of these attempts. And I didn't know back in the days that Discover will be that huge.
But now Discover is sending, as I mentioned before, 70% of Google traffic. So it's way above search. [00:28:00] So we build GDdash. And the idea behind that is if historically keywords are the foundation of search. You optimize and search for keywords. Now there is no keyword in discover. What would you do?
So the foundation in Google Discover is mainly based on two fundamental elements, entities and category. And this is how Google actually look at content before it matches with user interests. Okay, Jon and Joe are interested in LLMs and AI and stuff like this. So they built for each user a portfolio, right?
Of interests based on what you read, what you visit, what you search for, and stuff like this. So I. Google knows all your interests. Now they look at a piece of content that John Shahata wrote about SEO and stuff like that. I was like, and they analyze that piece of content. Okay, this belongs to marketing, right?
Or belongs to like entertainment. In under entertainment. It belongs to music news. And under music news, probably it belongs to rock and [00:29:00] roll, for example. So this is categorization, right? And Google has about 2000 categories. And then they look at entities and entities are real life things. People, locations organizations and so on and so forth.
And they say, okay, this article os it's talking about Taylor Swift's, it's talking about the Grammys. It happened in this location and so on, and other celebrities were there. So they have all these entities. So when they take an article or content and they categorize it based on categories and entities, then they match it to the user interests.
I'm trying to make it in a very simple way, but much more complicated than that. So we said, okay, we can't understand because Discover is highly personalized. I cannot figure out who is the personas of the millions of people visiting your sites, but I can. Actually understand your content the same way Google understand your content.
So that's what GDdash was built for is okay, let me analyze your content and categorize it in categories and entities the same way Google does. And then figure out, okay, when you write about this entity, when you write [00:30:00] about this category, that actually what resonate much higher than anything else with the audience based on their interest.
So we start showing people like we had a client who. As a business news client, very serious business news. And then I was like, oh, we continue writing about business and it definitely, it's the highest portion, but we found out that whenever they write about animals or pets or animal services and stuff, that the ROI of these articles were 200 times 200 fools.
The news, right? I was like, Hey guys, Google loves that content from your site. The audience resonate and engage with it. I'm not saying you're gonna drop the business news, that's absolutely not. But if business news is like for scale and this is your core product, this is works very well, right? So maybe an editorial strategy, can we increase that content by maybe another 10%?
Stuff like this. So this is GD Gas, and then it evolved and more other things are analyzing your evergreen content, analyzing your search traffic, and now we're adding [00:31:00] AI visibility to it as well.
[00:31:01] Jon: I was going back through your slides from SEO week the other day, and two things popped out at me based on other conversations we've been having and the release of AI mode.
If you think about what Discover is, it seems like there's some parallels with AI mode, right? Because it's highly personalized based on, what you're reading or what you interact with. It's less of a search result and more of a summary, which is maybe a little bit different than discover, but still somewhat of a component there.
Is there, is that a fair comparison of maybe. Technology or the insights from Discover that Google's sort of carrying over to what the AI mode experience is? Or am I making a stretch there?
[00:31:44] John: I think, I remember when I spoke about it, I said, if you wanna understand personalization more, you differently need to understand how Discover is working.
And to understand this kind of personalization, you have to go back into your content to understand how is it personalized to all these different segments [00:32:00] out there. And then you start categorizing your content the way Google categorize it. So this is like understanding the general concepts, and I think with AI mode, right?
And LLMs, it's, there is no keyword tracking, right? It's ev there is a different answer for everyone, but there is a recognition of your content based on the passages, right? So it's not a URL ranking. The, this whole URL notion is gone, right? It's like the passage, right? This passage talks about this entities and enter these kind of questions.
This is how we view the content, or this is if we analyze your piece of content that you think you wrote about business, but in reality it's not actually after analyzing it. The relevancy is very low or very slow, or it's not deep enough and stuff like this. In a normal analysis, I don't think LLMs will care about it, but there are so many other layers, right?
Like you talk about personalization you talk about your history with LLMs because personalization alone and then [00:33:00] understanding the memory, like they have memory now. It's okay, this is what you like, so I'm not gonna, even if this site is very good, but you don't this is not your preference.
I'm gonna show you your preference, what matches your lifestyle and everything I know about you. Right? There is a higher brand reputation. Branding will be extremely important when it comes to LLMs and stuff like this. So there is differently, some similarities, but I think AI in general, or LLMs and AI mode and overview is a whole different science in itself.
[00:33:30] Jon: The other thing that I thought was interesting as it relates to AI mode is the. Recommendation for Google Discover around like building out adjacent topics like around the main topic and therefore you're able to, get more visibility in Google Discover. It's, it feels a little bit like an approach for query fan out, which is the same where you're, you might still be targeting a keyword in this case, but there's all these sort of adjacent topics around it that as Google is, synthetically creating these [00:34:00] searches to build the response, you need content along that path in order to have it surfaced.
Right. So I thought that was an interesting, I dunno, loose relationship between what AI mode might become and. How to rank in Google. Google Discover. Go ahead Joe.
[00:34:16] Joe: Yeah, it was, it's the beginning of a content strategy, right? Yeah. Like under how that query Fanout is gonna use your page.
Can you talk maybe would love to hear if you have any kind of new thoughts on how to structure the content on a page. Are you thinking about that a little bit differently now?
[00:34:33] John: Yeah. So I think there's topic clustering. There is query fan out and there is adjacent content and they're like they go around in likes, like similar orbits, but not the same kind of a thing.
So the adjacent content strategy I was talking about in Google Discover is let's say Tom Cruise is trending because of Mission Impossible. You don't write entertainment content, right? But you have a site like Architectural Digest, which I used to [00:35:00] manage, right? And we have celebrity houses and homes, pictures from their homes.
And because we went there, interviewed them, so I used to do this when I was a ConEst, right? It was like, we will push that content. I was like. Where is Tom Cruise hiding when he's not on the set of Mission Impossible? Right? So something like that, right? Or where, what film locations for, game of Thrones or whatever.
Here you are capitalizing on a big trend, even that you don't, you're not writing specifically about Mission Impossible and Tom Cruise, but you have content. And what we have seen that most of the time this content rise with the whole wave because search engines love to diversify and give different angles and stuff like that.
So that was the idea. Adjacent content, right? Government shut down. WebMD will write something about they're not talking specifically about the government shutdown, but they might write something about like, how will this impact your health? So stuff like that queer fan Out is a very amazing concept, right?
It's like, where. [00:36:00] LMS and Google and others trying to figure out all the different searches that go, like in Google you write what is the best SEO software for news, and hopefully NewzDash will show up like number one. But in LLMs and AI overviews okay, let's think about all the other queries that may go along with this.
What is the best technical SEO software? What is the best content? SEO software for news? What is the best, whatever? So they trying to figure out all these different queries around this topic, and then they grabbing each one, sensi, sensitizing all the information, and then providing you with Hey, if you are thinking about content, this is, could be great ones.
If you're thinking about real time, these are great ones. If you're thinking about technical. So it's fanning out, right? It's oh you're. Looking for running shoes, and then they look at all the different things. If you build topic clusters right about this, right? It is let's say you're running [00:37:00] shoes and best running shoes for women, for men, for flat foot, for whatever.
And you're thinking about the cluster and all the different sub topics. In a way you are optimizing for that. So it's all like similar orbits a thing. Same kind of strategy and way of thinking but different applications in a way.
[00:37:17] Jon: I wanted to double click on the Tom Cruise example.
So in that scenario, right, you have the piece of content already. Maybe it's, I don't know, a year old, two years old, because you've done that interview in the past. So when Tom Cruise comes out with a new movie, what's the what are the tactics to get that piece of content pulled into discover? Is it actually refreshing the content?
Is it pushing social signals to it? Is there a strategy there or is it just a natural uplift because of all the interest that it gets pulled into discover on its own? Do you drive signals into it?
[00:37:53] John: Yeah, absolutely. Google tells you there is no way to optimize for Google Discover.
And from a technical point of view, they could be [00:38:00] like, correct, but there are actually 11 ways that you can influence your performance in Google Discover. Right? And it's interesting you say signals. For example, let's say you, this article is buried on the site and you bring it back to the homepage and do some updates and maybe change the URL if this is your.
SEO philosophy, right? And then maybe update the headline, update some of the images, right? And then do a push notification now. And then suddenly a lot of people came and visited the article. So you're sending a signal in through Chrome data. 'cause Google Discover and other Google products look at the usage of articles and stuff through the Chrome data that they have access to.
All of these are very strong signals for Google Discover, right? And I was like, oh there is a traction either audience traction or social traction about this article. So we have clients where. They write the same article actually every week the most popular comedy movies to watch on Netflix this week.[00:39:00]
Right? And they don't change the URL and every week they take the same article, refresh it, push it back to the homepage, do all the different things that I mentioned about to create traction. And we see it coming again in Google Discover. So it depends on what type of content and the situation.
Yes. But the, these are like the general elements that you would do for an article like that.
[00:39:22] Jon: Let's talk about URL. Refreshing. I feel like that's a, a hot topic. Not just for news, but even like we do big blog content audits and, look for content that has an evergreen nature to it, but hasn't been updated in five years.
In most cases we're not suggesting changing the URL itself, but more about, updating the content and, what sort of best practices do you apply or preach when you're going in and refreshing content like that? One thing is the date, right? Some people leave the original published date.
Some people leave last updated, which, we've seen some studies that show that you shouldn't do that. So there are tips like that, that are related to, [00:40:00] actual content refreshing that you've discovered discovered through your testing.
[00:40:04] John: Yeah I. Let's divide the content in a very simple way.
There are so many ways you can divide content, but let's say there is new content and there is evergreen content, okay? And there are so many variations of like colors, but news content. Let's say that we will not write a new article or refresh the content unless there's substantial change to the, there was a shooting and now we know the victims and now we know the shooter and there's more, I probably writing new content or like doing a new URL and stuff like this, right?
But if it's like a small TBO or stuff like this wouldn't change the URL but if the article developed into a second day, a third day and fourth day differently, you wanna create a new URL every day. So I'm gonna put news on the side now for Evergreen in the cones, I created a project called the Pine Tree, and the whole idea of the pine tree is they are evergreen.
They last forever, back in the days many years ago. You write a very good article, [00:41:00] it can send you traffic for 5, 6, 7, 8 years. You're good nowadays. The last study I did was like few years ago. You'll be lucky if you get like good, solid nine months to 12 months from a, an evergreen piece of content. So definitely because the Google now know the most recent updates about the topic you're writing about.
So your topic can easily be out of date very quickly. Right, or outdated. Sorry, very quickly. So my advice is you need to create an editorial calendar for your most important evergreen pieces. Where you come back, right? And say, okay, we need to refresh that. Now, if it's still driving traffic and ranking will and all that stuff, I will keep it as is.
But I think what I build in Condesa now in GDdash is like, if the traffic drops below 25% of the maximum traffic point for that piece of article, we need to re-go and reinvestigate what can be done? Are there new data? Are there better [00:42:00] competitors writing content about this? Right? What should we do?
And I think you should test with dates. Every site is different, right? If you are writing something like, and Google showed that the, this content was written four years ago. Maybe people don't wanna click on that because it's it seems like outdated and I need fresh data. So I think is like the best approach is to test.
I think there is no magic answer for everyone, but it's like testing. So identify which content you need to refresh. That's step number one. Step number two, do you like analysis of the landscape or other types of content are appearing? New data that were not available for us? Stuff like this, we need to go more in depth stuff like this.
And then you go refresh the content. For URLs, I like to keep maintain the same URLs for Evergreen and not change them. Because you wanna maintain that history. Another part very quickly we did in Evergreen is we consolidated content. So for every piece of content that we refreshed, we consolidated five to 10 [00:43:00] pieces of content.
Because after a while when you work on the same site for many years. You will definitely find that you wrote about the same content. Probably. I dunno, in some of our magazines were wrote about the same exact topic 3000 times and they didn't differ that much from each other. Right. How many times you can do nail design was like a whole different angle.
Right. By consolidating you giving better chances for stronger content to drive links and traffic and become more in depth versus diluting your content in a hundred different ways.
[00:43:32] Jon: In that scenario where you're collapsing, I don't know, a hundred or 3000 pieces of content, how do you land on the URL to choose?
Is it a net new and you redirect everything into it? Or are you choosing like the. The best performing URL from that cluster and then moving everything into that URL.
[00:43:50] John: I don't know when I will ever stop saying it depends, but it really depends, right? In case, right case. My, my first goal is like, what is the [00:44:00] best article that have history, have strong incoming links, authority links and have and had the decent, strong traffic, right?
So maybe we can utilize this primary content. And again, you need to think about copyrights. Can we modify that piece of content or not? Do we have the rights to it? All that stuff. So there's so many different elements, or maybe all of that content is not good. You need to have a fresh start, but I always go with option one first. If there is something that I have that sets that criteria, I will start with it first.
[00:44:32] Jon: Got it. I know we're bumping up on time. How long do you have? Do we have you for another 10 minutes or so?
[00:44:38] John: Sure. 10 minutes is fine. Absolutely. Yeah.
[00:44:40] Jon: Okay. I'm enjoying.
I wanted to talk a little bit about paywall content. I feel like in the age of ai, we don't talk about or hear about, the challenges with paywall and news. And there's so many different ways to implement it. Do you have any tips on implementing paywalls of different sorts and types?
Is there a way to [00:45:00] tease the content, to still solicit the signup without causing a lot of friction, there's the SEO piece, but more from like the UX side of things. Yeah.
[00:45:09] John: Yeah. So the SEO piece, I think there is a lot of content out there about the schema and stuff that you can do and stuff like this.
I think I wanna talk more about the strategy and which touches on the ux, right? I. Does your content actually will be the content that people will subscribe to, right? So this is a big question. It's not every piece of content or every news organization have content that people actually are willing to pay money for, right?
So this is the biggest question. The second question is is every piece of content, let's say we assume that yes, you have the content that people wanna pay money for. Does it, is it every piece of content needs to be under behind the paywall? Maybe you're right about entertainment and that entertainment traffic that entertainment content drives you a lot of traffic, but it's not the core content that actually people wanna pay for.
So you leave the [00:46:00] entertainment traffic or entertainment content, like available and accessible and stuff like this. And then certain amount of like analysis or in-depth whatever on other elements of the world. This is like the content that you're actually people are paying for. What is the value exchange here?
I'm gonna give you my money. What are you giving me? It's oh, I'm removing all the ads from the site. It's much faster. You get like a crossroad a crossword puzzles. You get exclusive recipes. You get all these kind of things. There needs to be like some kind of a value exchange, right?
Because I think it's like you don't ask people for money immediately. I personally, when I work in Conde Nast is I converted people to newsletter subscriber. That's much less friction than Hey, give me $5 a month, right? And then you start building and nurturing this kind of relationship and.
It's a much better conversion and you have much better control in the environment versus Hey, they came to the web, [00:47:00] okay. Through the web or through social. Not every type of audience is a good subscriber or well convert to become a subscriber. So you need to have your data signs at or hire a team to understand the, or calculate the likelihood of that specific user to become a subscriber.
Okay. And people who are one and done, most likely, they will not, they only focus on the topic and they leave the site. But people that you see, like coming to your site once and twice and five times a month and stuff like this, so it's not one size fits all. And you play with sampling, right? It is okay, this user, I'm gonna give them three free articles, and then I'm gonna ask for their email.
And, but I wanna tell them, Hey, for this email, I'm giving you all of these things. And I was like, oh, this is a good deal, right? I'm gonna give you my email and you're gonna give me 10 articles. I love your site and I like to read it. And then you was like, Hey, do you wanna get this and this for that amount of money?
So it's a sequence. It's right of actions. [00:48:00] And you ask, you have to test. I think one of the things I did was, epicurious, right? So I wanted to convert more users to become email subscriber. And I was like, we had the generic interstitial and interstitials do work, and I know people hate it, but interstitials work if they're done right?
I was like, Hey, subscribe to our newsletter to get the latest recipes from our site. Oh, okay. It worked fine. But then I start figuring out Jon and Joe are coming to that specific page because they love Italian cuisine, and this is like a very specific recipe for Italian cuisine. So what if we address this right interest in them in our messaging?
Want more Italian cuisine recipes? Sign up here. We saw a 16% increase in subscriptions. From that, just adding these two words. So there is a lot of AB testing and stuff that can definitely like a whole convers conversion optimization that can be done around these pages.
[00:48:59] Joe: Sounds like [00:49:00] you would advise against using a standardized a.
Article meter. You need to be smarter about it. You need to understand the reader a little bit better and customize the meter to try and push them down that funnel toward a subscription. Is that what you're saying?
[00:49:15] John: Yes. I think you have to test a lot. You can start with a benchmark. Build your benchmark, right?
It's or your baseline, right? And from the baseline you start doing all these different tests. What if we move from one to three? What if we only apply this to search, but not social? What if we, give them three for the people that we have seen multiple times that coming in the last two, three weeks versus the people that we only saw once, all these kind of tests.
And you start like figuring out what type of audience that actually convert to us. What their characteristics. Oh, the love. I don't know. Taylor Swift. What if we mentioned Taylor Swift in our messages or something exclusive and so on and so forth.
[00:49:56] Jon: So one last thing before we wrap up with our final question.
The Nest [00:50:00] Conference. It's one of the, I think the largest or the only news dedicated conference. So super important for people in the industry to, to join there. Do you have a date and and a location for that yet?
[00:50:11] John: Yes. Thank you so much for asking. So I'm very excited about this year. It's October 21st and 22nd this year.
It's virtual and that's why we have a lot of people like, and I think we. I can call our conference global because we have people from 55 countries that attend, and I think this is, this year is the largest year ever that we did. We have 20 speakers that are speaking for two full days. And everything is recorded.
So even if you can't attend certain segments of the conference, you can come and replay, like even with questions and answers and chats. So it's a really. Great platform. Yes. So it's in line. Mike King will be the keynote speaker this year. We're gonna talk, a lot of sessions are about AI from a building agents to building for news and building like, content that is AI [00:51:00] resilient and conceptually and big pictures.
But we'll continue also as like SEO for the newsroom and the technical aspects of news. We have Line Ray, we have Mike, we have Marshall Simmons. We have a huge group of people. Huge. Really amazing line lineup of speakers this year.
[00:51:17] Jon: Marshall Simmons is a, is an OG in the news space as well.
So it's gonna be fun to hear him speak. He doesn't speak often, so if you can get him on the lineup. Yeah, we
[00:51:25] John: him out, we got him out.
[00:51:26] Jon: Yeah. That's incredible. My wife, it's basically her mentor at the New York Times I always hear great things about him. It's had a chance to hang out with him a couple times.
He's an amazing guy. Yeah. Yeah. So to put you on the spot, have any discounts for the listeners?
[00:51:39] John: Absolutely. How about I send you the discount? You guys can have it there. So I'm gonna give you guys a special discount for 25% which is a little bit more than I think what we have right now is early bird.
So your users can use it
[00:51:52] Jon: When you're trying to. Fill it out, a subscription. Are there any, like, how do you land on the price point?
[00:51:59] John: So I wouldn't [00:52:00] claim that I am the, I'm the expert on that. There are people, all, what they do is go to market and how to, price your products and stuff like this.
I'm not the expert, but I would say is like how competitive your product versus others. Right. And what is the need for your product versus other is this is something, a product that I can find in a hundred different other places. So it's or this is, it's only here and I'm willing to pay $50, a hundred dollars a month versus now I just, Hey, do you know what, I'm gonna pay a couple of dollars because I hate the ads and I, they have this like little bit extra stuff.
So it really depends on the uniqueness of the product that you're offering and the how the need of the user that for that product.
[00:52:44] Jon: Sounds like a perfect opportunity for testing.
[00:52:46] Joe: Yeah. Always. Yeah.
[00:52:48] Jon: Well, I know this has been. Fantastic. Really enjoyed the conversation so far. Before we let you go, we have a prediction question that we like to ask.
If you go to, I don't know, Google Discover [00:53:00] in the next 12 months, what would you expect that experience to be? Or, how would you expect it to evolve with AI and everything else?
[00:53:08] John: That, that's a great question. I could be wrong, but I think Google might test AI overviews on Google Discover and we actually start seeing very low instances or small times where actually Google will introduce AI overview.
And I think the way they would do it is because it's a very highly personalized feed. The AI overview could be like the intro maybe to the feed. It was like, today, these are the things that matter that happened, and then you will see all your stories. I was like, you know when you send the newsletter and you write an intro and you have a list of all the stories below it, I think AI overview will come.
Definitely. I think it will be smarter and smarter. About what do they serve you? I doubt that they will disrupt that platform much because it seems like this is the last platform that publishers are getting much traffic from. [00:54:00] And I think Google wants publishers to be happy, right? It's Hey, look at Discover you're getting a lot of traffic here and don't worry about search.
Right. A thing. And you'll need
[00:54:09] Joe: another lawsuit.
[00:54:10] John: Yeah. And, it can make a lot of noise for them, but I have been wrong before, so I don't,
[00:54:17] Jon: Well, before we let you go let the audience know where they can find you. Social and we'll put all that in the show notes as well.
[00:54:23] John: Absolutely. I'm very active on LinkedIn more and more than ever before. And I'm also like on Twitter as well, so John Shahata and and Twitter. It's Jay Shahata.
[00:54:32] Jon: Alright, well thanks so much. This has been amazing and if you enjoyed the show, please remember to subscribe, rate and review. We'll see you next time.
Bye-bye.