AI isn’t killing news SEO — but it is forcing publishers to rethink where visibility, traffic, and value really come from. In this episode, Shahzad Abbas breaks down what billions of clicks reveal about Google Discover, AI search, and the future of publisher growth.
https://page2pod.com - In this episode of the Page 2 Podcast, Jon Clark sits down with Shahzad Abbas, VP of AI and Search Strategy at Define Media Group, to explore how AI is transforming the economics of news publishing, search visibility, and audience growth.
Shahzad shares insights from a massive analysis of billions of clicks, revealing why breaking news is still thriving, how Google Discover is overtaking traditional search for many publishers, and why that traffic may be less valuable than it appears. The conversation digs into the changing role of SEO, the rise of answer engines, AI-driven search behavior, and what publishers need to do to stay visible as the search landscape evolves.
📰 In This Episode
• Why breaking news still performs well in an AI-driven search world
• How Google Discover is changing traffic patterns for publishers
• Why Discover traffic can be high-volume but lower-value
• What AI search means for newsrooms, publishers, and SEO teams
• How publishers should rethink visibility beyond traditional Google Search
• The role of Search Console data in understanding real performance
• Why collaboration, experimentation, and adaptability matter in modern SEO
• Shahzad’s perspective on leadership, consulting, and building strategy in fast-changing environments
This episode is a valuable listen for publishers, SEO professionals, content strategists, and media leaders trying to understand what visibility looks like in the age of AI and answer engines.
Subscribe to the Page 2 Podcast for more conversations with the people shaping the future of search, SEO, content strategy, and digital visibility.
What do you think will matter more for publishers in the next few years: Google Search, Google Discover, or AI answer engines? Drop your thoughts in the comments.
🛠️ Tools & Resources Mentioned
• Shahzad Abbas on LinkedIn → https://www.linkedin.com/in/shahzadabbas/
• AI Overviews Study → https://www.definemg.com/breaking-news-thrives-in-the-age-of-ai/
• Meat Loaf will help you fix your website → https://www.tvguide.com/news/meat-loaf-seo-1004749/
• Define Media Group → https://www.definemg.com/
Jon Clark (00:00) Welcome to the Page 2 Podcast where we uncover the strategies, systems, and tactical decisions that move brands beyond Page 2 and into real visibility across search and answer engines. Today's guest is Shahzad Abbas, VP of AI and Search Strategy at Define Media Group, someone who has spent his career sitting right at the intersection of language and technology. He doesn't spend much time in the spotlight, but he probably should. He's built a reputation quietly behind the scenes as one of the sharpest minds in news SEO and AI strategy. Honestly, getting an hour of his time felt like a bit of a gift, and what we're trying to unpack in this episode is how AI is reshaping the economics of news publishing. Shahzad walks through a massive study analyzing billions of clicks that reveals breaking news is thriving. But the surprising finding is where that traffic is coming from and why platforms like Discover are overtaking traditional search, even as they deliver less valuable audiences. He also gets into the mechanics behind it all. How headlines, images, and editorial decisions now directly influence visibility across Discover. Shahzad is someone who deeply values language, authorship, and intent. Yet he's also advising companies on how to strategically use AI-generated content. That push and pull between scale and substance, automation and authenticity is really at the core of everything we discuss. I think you're going to really enjoy this conversation, not just because of the insights, but because of how thoughtfully Shahzad approaches the work. You can hear it in every answer. He's not just reacting to change, he's trying to understand it at a fundamental level. If you learned something new today, take a second to subscribe to the Page 2 Podcast. Leave us a rating or review and let us know what resonated. We'd love to hear your thoughts. All right, let's get into it
Jon Clark (01:42) Welcome to Episode 117 of the Page 2 Podcast. I'm your host, Jon Clark, and I'm joined as always by my partner at Moving Traffic Media, Joe DeVita.
Joe (01:50) Hi.
Jon Clark (01:51) We're excited to dig into all things news SEO with the VP of AI and Search Strategy at Define Media Group, Shahzad Abbas. Welcome to the show.
Shahzad Abbas (01:59) Hey, thanks for having me.
Joe (02:00) We want to talk a lot about a lot of things that you have written. Jon and I have spent a lot of time reading, trying to read everything you've written to prepare for this interview. But it became really clear to us, you have a passion for writing, you have a passion for languages, and you've studied English at Cambridge. I wonder if you could just start by telling us where that inspiration comes from. What are you reading now? What have you read that has given you so much inspiration?
Shahzad Abbas (02:27) Yeah, thank you. I appreciate the background. What's interesting is, I did study English. It was a passion of mine. I came into it a little bit late. But once I did in college, it really just sparked my interest. I just thought this was just the thing for me. So I was on a track to become a PhD. That's why I actually went to Cambridge for my graduate studies. I decided that it wasn't exactly what I wanted to do. This is back around the late 90s in 2000. Then I came back, the internet boom was going on. I had some ability to go into, I had some technical background. I had played around with HTML back in those days in the late 90s and so understood how to spin up a website. So I went into the technical field. Initially, I was more on the front-end development side. But then I became more of an SEO specialist as that field began to grow and people became aware of it. The reason I'm bringing up this background as a parallel to your question, Joe, is because when I saw the ability for people to marry language with technology, which is effectively what SEO is, it felt like an amazing confluence of the things that I've been working on and the things that I was passionate about. So that's where the English language, semantics, understanding intent, that and SEO kind of married really well. And I just fell into it and thought, yeah, this is exactly a nice confluence of everything that I'm really passionate about. So that's sort of my origin story, as it were.
Joe (04:04) And I know you speak multiple languages, too. So I guess that even helps you further to break down language into different components that if you only spoke one language, you don't even really think about. But maybe can we just on the personal side, like what have you read that inspired this love for language?
Shahzad Abbas (04:22) And that's again, a great question. It's not a question I get asked very often. So I really appreciate the digging in. So I'm of South Asian origin, right? I'm from Pakistan. When I went to school, so I was at UPenn and I studied English there before I went to Cambridge. When I fell into English, what I really loved was post-colonialism. And so what that means is effectively, post-colonial literature is effectively the literature of the South Asian and other colonial countries that were fighting against colonial imperialism, effectively. And so at that time, especially a little bit before and after colonialism had ended, there was just this amazing flowering of literature that spoke to the impact and both the benefits and destruction of colonialism on culture, right? How colonialist powers basically impacted how cultures wound up manifesting themselves after they left. What does it leave behind in its wake? And so for me as a South Asian, there was just this amazing flowering of Indian literature at that time. So from Salman Rushdie to Arundhati Roy to V.S. Naipaul, that's a little bit earlier. But effectively, all of those authors were really just becoming really well known and really valuable. And again, it spoke to me just because of my background. And so that's really what led me to not only become an English major, but also focus on that specific topic. The master's degree that I wound up getting through Penn was focused on post-colonial literature. That's what I studied at Cambridge as well.
Joe (05:55) We're going to try to make this transition into news, and I'll ask one more question about the types of things you like to read. It sounds like it's a mix of fiction and nonfiction, but are you just reading all the time? Like every day you try to read? What is your system for reading in a normal week? I'm sure you're reading a ton all the time.
Shahzad Abbas (06:16) Yeah, I love to read. It's really important for me. So there are a couple of different, I think, almost categories of reading that I have. So one is I'm always reading up on trade news. I want to make sure that I'm up on search news. Now it's even more just AI news. And then I've also got sort of a category where I'm always trying to read a book. And typically it is a fiction book, non-fiction book, fiction book, non-fiction book. So I try to trade off each one at a time. Now, as far as the SEO stuff goes, it's the typical trade news organizations, right? And we all know and love those. And so I do that. I think Substack is a big part of my news ingestion now. And so I have sort of a curated Substack that I look at, again, mostly focused on AI, because I'm really interested in understanding how AI is going to affect us and affect the world. And not just from a specifically SEO standpoint, obviously that's important as well, right? Like AI Overviews and AI moats and how the LLMs are taking away or not taking away from Google's search market, that's all fine. But what I'm also interested in is just AI in general and the impact that AI in general is going to have on our lives. I think it's the biggest technology since potentially ever. And so I just want to understand, I'm just interested in understanding all aspects of it, and so I'm always following people that are saying really interesting and deep and meaningful things about AI. Now, when it comes to books that I'm reading right now, so I have a book club that I'm a part of with friends from my UPenn days. And so right now I'm reading A Passage to India, for example. So again, that's a book by Forster. It has to do with colonialism, post-colonialism. I'm not always reading those types of books, right? Like, for example, another book that we read recently was The Body Keeps the Score. And so that is about trauma and understanding how trauma actually manifests in the body and not just mentally. I think there's a traditional bifurcation between mind and the body. And I think what this author is saying is that that's actually true in a sense, but also we can think about it in different ways because the manifestation of trauma actually impacts you in different ways, including the way that your body responds. So again, I'm reading all sorts of different things at all times, but I think it is important for people, in order to become well-rounded, to read different types of books. And as a humanities major, I'm going to say that there has been nothing more impactful to me in terms of my ability to read, understand, and just love language than constantly reading it, reading not only fiction, not only business books or industry books, but also nonfiction. It's really super important because again, it just gives me a perspective and thoughts and understanding of the world that is beyond what is my specific narrow view.
Jon Clark (09:05) Gosh, there's so much. I kind of want to make a separate podcast where we only talk about this kind of stuff, but let me try to make a loose transition. So you published a great AI Overviews study based on the billions of click data points that you have in your cohort. And there's a call out at the bottom of the blog post that you published mentioning that you used Claude for graph creation and some of the data analysis, but you elected not to use Claude for any of the actual writing. Talk to us a little bit about that. How are you thinking about AI's creation of content as it relates to your love and appreciation for the written word? Is there a conflict there for you at all?
Shahzad Abbas (09:49) It was important for me personally to write this blog post with my own thoughts, with my own language, with sort of my own creative hand, and to develop the throughline of what I'm trying to say and making sure that there is no intermediary between what I'm trying to say and who I'm trying to say it to. Because if you use AI, what's happening is it's taking, to some degree, its own interpretation of what you're using and then providing that to the audience. And that's fine in some ways. I've of course used AI to summarize content, to provide briefs, to even use as part of what I'm providing to my clients. I think everybody's doing that. In this case specifically, I wanted to be very clear to the audience that this was only our work, that this was only my work, that every single word that was written in here was actually written, quote unquote, the old-fashioned way, because again, I did not want there to be any buffer between what I wanted to say, how I wanted to say it, and how I wanted the audience to react to it. Otherwise, you're giving up a little bit of your real intent and meaning because it's just going to use different language to get to what you're trying to say.
Jon Clark (11:12) Got it. Let's maybe circle back to the AI Overviews study. I'm sure you work with publishers who are embracing AI-written content a bit, I don't know, maybe even in some cases, like creating AI personas that sort of push out articles. How are you advising clients around that? I mean, we've seen some websites who push out tons of AI-generated content. You might argue it's more slop than real news, but we've seen them get hammered in recent updates. You know, there's also the challenges of hallucinations and to your point, the tone is no longer in sort of an editorial tone in some cases if you give up too much of that control. So how are you guys thinking about that in terms of advising clients on sort of all your competition is doing it. So we need to figure out a way to embrace it, but how do we do it smartly and with guardrails around it that allow us to maintain an editorial, ethical component?
Shahzad Abbas (12:09) Carefully and judiciously. So what I mean by that is everybody is resource constrained, and so we're not coming into this blind or idealistic. There are plenty of publishers we have that want to write more but don't have the resources to be able to write that content. And so what we're trying to do is identify those gaps in which they don't have, for example, an actual writer on that specific topic or beat and let them know that, okay, maybe this is a way that you can break into this. And as it potentially succeeds, if it does, then maybe you put somebody real on it. But it can be a starter as far as the foray goes into topics that they're not covering, and then allow that to be a bellwether about whether you actually invest more resources into that area. So that's the way that we're trying to approach it. We're not advocates for massive scale. It just doesn't work. It hasn't proven to work now. And I think even if it did work, again, I've seen the studies and I'm sure you guys have where initially it works and then it gets hammered. And so we've seen these cycles before when it comes to Google. Google will eventually catch up and it's always trying to provide its point of view of the most helpful content available. Google has not said that AI is verboten by any means, but at the same time, anybody who's actually seen AI content at scale that isn't carefully editorially curated can spot it from a mile away. And it doesn't have, at this point, I believe, the same level of quality that a human editorial view can provide. And so from that perspective, I think it is a dangerous game you're playing if you're scaling out AI.
Jon Clark (13:54) Got it. Let's maybe circle back to the AI Overviews study. Maybe set the stage for folks who haven't seen the study. What was it? I think we all have heard or have maybe even experienced the outcome of it, but was there a different hypothesis or data point that you expected to find from the study as you started it? What was the goal when you started analyzing this data?
Shahzad Abbas (14:15) Yeah, so let me just, I think, set the stage for the audience if they haven't had a chance to read the study. What we tried to do in this study, as I wrote a blog post, and I wanted to understand if there was a difference in performance for breaking news. So again, what I mean by breaking news is news that appears in the Top Stories carousel of search results. That Top Stories is effectively what I call breaking news. It's news of the moment. It is news that is potentially even changing throughout the course of minutes or definitely hours, especially in the course of a day or two days or three days. Think about something like the Iran War, right? That's constantly changing. Anything that you're searching for right now is going to be very different from yesterday or two days ago. So that is an example of breaking news. And so what I wanted to understand is, is there a difference in the ability for publishers to perform well in breaking news versus editorial and evergreen content? We had sensed this happening because we have a pretty robust dashboard across all our clients that we are keeping track of. And so when we're looking at breaking news sections of their editorial content, we're seeing that they're actually performing well versus more lifestyle or editorial or health content where that is really suffering. And so we had an idea that this was actually happening. And that's what was the genesis of the study. What I wanted to sort of do was validate this hypothesis I had that breaking news is a moat around which AI and AI Overviews and the LLMs have not actually bridged. And I wanted to understand, if that's the case, how much of an impact it's actually had. And then the third thing that I wanted to do was break out web search versus Discover and understand if there's a difference between the two. Those are the reasons why I went into this research to begin with. Now, what I found as far as the research goes was a couple of things that surprised me and a couple of things that I expected. So what I expected was that breaking news would be protected from AI. Why is this? It's protected because AI right now, as it stands, the technology itself is obviously grounded on training data and then it needs the RAG system, so the Retrieval-Augmented Generation, to actually grab web search and bring it into the generative summary that we all see in AI Overviews or any other breaking news chat. That is going to be somewhat of a limitation. Google has 25 years of a durable system to get the latest news and provide it to its users. So it knows how to do this already. The LLMs are not built to actually grab the latest information. Initially, when the LLMs were, before the RAG systems actually came out and web search was integrated, they would say, hey, this training data stops at a few months ago or whatever. So from that perspective, I expected that breaking news would be protected. It turns out it is. So what we saw was, so we did a study, it was like a six-month period, and we saw that breaking news was up like 103%. It was crazy, right? And so it was doing really, really well. Evergreen content and so we were able to break and segment across our entire portfolio of sites that were in this study breaking news content versus non-breaking news content. So evergreen content, lifestyle content, health content, etc. And we saw that that was going down. So there's a clear breaking news is this way, it's going up, and then evergreen content is going down. That's very clear. What we were able to understand is that breaking news is doing well in Top Stories, but not as well as we thought. So I think that was the surprising finding, which is that the majority of the driver of breaking news is actually not the Top Stories carousel. Now, if we're thinking about breaking news in terms of just overall search performance and how it's measuring against the market per se, it's doing really well, right? It's up like 2%. So breaking news in web search is up 2% over this time period, 2 or 3%, something like that. That is well overperforming against the general search market, just everything, which is down like 40%. So if you're a stock and you put money into breaking news, you're doing okay. But if you had put in stock into Discover breaking news, that's where all the growth is, because it's up like 140%, whatever the number was. And so that's what surprised me. I didn't expect to find that because I always sort of assumed that Discover is more of an evergreen lifestyle entertainment channel. And it turns out it's not. There's opportunity here for Discover to actually be more of a breaking news channel. And this may just be because users are habituated to Discover being a feed that is not just for health and entertainment. They don't really care what they're seeing in terms of the differentiation, or as if they wouldn't click on breaking news just because it's in the Discover feed. If they see it and it's of interest to them, they are going to click on it regardless of where they find it. And I think that's what's happening, is that those folks who have personalized for, let's say, a national news publisher like the New York Times, if they get information that's of interest to them that is breaking news content, then they're just going to click on it wherever they find it. And I think that's what's happening.
Joe (19:59) Can I ask you to make a distinction between breaking news and Discover? For a reader to see breaking news, they've got to show some interest in a news subject. But with Discover, you're in a Google app and you've got this it's really passive, you have this personalized feed of stories that kind of endlessly scrolls and hopefully you click on something eventually. But those are two really different behaviors. And my question is, well, if you agree with that, yes or no would be great, and there's a different value in the people who click through. Someone who's interested in a certain news story, they see some breaking news content, they click into it, they read it. Someone who's passively scrolling a personalized Discover feed eventually clicks on something. They didn't have an intent when they started and they were just convinced to click through to something. The value of that traffic has got to be different. They're both driving traffic to the site, but I would imagine one is converting at a higher rate than the other, if the conversion goal is a new subscription or something. Not really a question there, but I guess agree or disagree: if your goal is new subscriptions, it seems like you want to focus on trying to be the breaking news outlet rather than the Discover outlet.
Shahzad Abbas (21:19) Yeah, so I agree exactly with what you're saying. So again, for everybody who's listening, what's happening is web search is down like 40% or so. Discover is going the other way, so it's like 30% up. And so it's actually crossed in our view for the first time. For the first time, we've seen more traffic from Discover than from web search, which I think is actually a big deal, given that historically, Google was meant to be a medium through which publishers were found. And that is still the case to some degree. It's diminished for certain. And now a secondary product, Discover, is actually sending more clicks through our news publisher-heavy portfolio than web search, which is, again, I think that's somewhat astounding. Now, as far as the value of Discover versus web traffic goes, and I'm going to say specifically web search organic traffic, you're absolutely right, Joe. When we talk to our clients about conversion rates, about engagement rates, about any of the metrics that actually really matter to them, they always say that Discover does not convert as well, that the value of the Discover click is not equal to the value of the organic click across the KPIs that matter. So put another way, let's say web search and Discover are effectively equal as traffic drivers now. If the value of the organic traffic is like a dollar, maybe the value of the Discover traffic is like 75 cents or something like that. There's more to make up in terms of value. You have to get more clicks from Discover in order to make up the value that was lost in web search. And so you're relying on a traffic source that is a little bit more black-boxy than even web search. Discover is more difficult to game and figure out and understand what actually moves the needle, and that traffic source provides less value to you than the traffic source that you're trying to replace, which is web search.
Jon Clark (23:27) I definitely want to come back to how to rank better in Discover, but I'm curious what all this data means for local news publishers. I think they've always probably been at a disadvantage for Top Stories anyway compared to the national publishers. They probably have a better chance in Discover, but if that traffic value, as you've described it, is less valuable traffic, what does that mean for local publishers? How do they compete in this space? And maybe to build on that, I don't know if you can say, but were local publishers part of that cohort of data, or were you looking primarily at national publishers?
Shahzad Abbas (24:04) Jon, so you hit the nail on the head. This is actually my next step. I think I'm going to potentially extend this out to local as well. We have a fairly large local cohort across the nation. I think it's 200-plus local publishers in our portfolio. What we're seeing is up and down. It's unclear. So right now I'm actually gathering data. If you notice, I think the latest Google updates did mention local specifically. And so what I'm trying to understand is whether that actually trickled down into value and increased performance across these surfaces. It's a great question. I will say that just almost anecdotally, our local publishers at a general level did seem to be fairly resilient when it came to these traffic losses. Now, that doesn't mean that they didn't follow the same trends. They did. But I think my sense is that Google is trying to showcase local publishers if it's able to triangulate where you're coming from. And so like, if you're typing in a local search result, let's say, so I'm in New Jersey right now, there's a congressional election up for grabs. It's NJ-12. So Bonnie Watson Coleman, our rep, is retiring. There are 12 people trying to vie for her seat. So it's a massive scrum. And so there's a lot of information out there about that. And so if I'm typing in information related to that, I am going to get Top Stories results. There was just, like, I think yesterday or the day before, a debate that occurred in Princeton. And so that came up in my Top Stories. And so I saw that. And so I think Google is trying to make a concerted effort to showcase local stories. I think local stories tend to be higher engagement anyway. If it does sort of bubble up, I'm seeing this in Google News. I'm seeing this in other places as well. And so I think the opportunity is there for local to be able to be resilient and perform well. But they have to understand what their lane is. Because Google is always assigning us a lane, right? Eventually, what you want to do is understand your lane, excel in your lane, expand your lane carefully in different ways. But effectively, like, we are seeing some signs of local publishers doing really well. And so I would say on that front, not all hope is lost.
Jon Clark (26:25) Got it. Well, let's talk about Discover ranking factors, if you can call them that. We had Lily Ray on the show two weeks ago and dug into this a little bit. But from your perspective, there's been some great research in the space. Google says there aren't any specific elements, but for example, image sizing has been shown to have some impact there. If you start working with a client, are there two or three tactics that seem to provide a better opportunity to at least be included, or are there other factors that you're looking toward?
Shahzad Abbas (26:56) For sure. There are definitely opportunities to improve your Discover performance. We've worked with clients where that has actually occurred. So much of it has to do, so the first is just making sure their technical specs are up to date, right? Like, making sure that those are table stakes. The most important one is 1,200-pixel width. So Discover has a requirement that says all original images you're publishing need to be 1,200 width. We've found multiple times where a client did not meet those requirements and effectively were gated out of Discover. Now, one of the frustrating things about Discover is that unlike Google Search Console, which has a very comprehensive suite of tools that can allow you to understand things like indexing and site maps and crawl stats, there is no such thing for Discover. So there's no tool you can look at for Discover from Google search that says, "Okay, these are the problems that we're finding, fix them." And so you have to kind of back into this rather than just being able to get it straight from the horse's mouth, as it were. So from that perspective, we have to make sure that, so we've got now our list of things that we check for from a technical perspective. I think the most important one is just the 1,200-pixel width requirement. But there are others as well, and so we just want to make sure that is met. Once that's met, once the technical requirements are met and we can ensure that you're effectively eligible, you meet the eligibility requirements to actually get into Discover, the next thing is really dialing in the headlines and the types of images you're using. Those are effectively the most important things. With respect to headlines, so much of this has to do with entities, framing, and also tone of the headline. So we've got a robust Discover analysis workflow and process in which we take a look at editorial content and compare it against our baselines and then compare it to, let's say, for example, if a site has tanked in Discover compared to what it was doing previously, in order to understand where the gaps are and how to improve. And so, for example, one of the things we find is that you may think that people are attuned to clickbait, right? So by clickbait, I mean something like you're leaving significant portions of the information out. If I'm talking about a sports entity and I say the name of the team, but then I leave out the name of the player, then all of a sudden you're leaving something out that is so grounding and important for Google Search and probably the other LLMs as well, is that you're leaving a gap. And I think you're leaving a gap not just for the LLMs to parse or Search to parse, but you're also leaving a gap for the user. And I'm seeing that more and more, the idea of clickbait, which seemed to be all the rage, is now gone by the wayside, effectively. Like, it doesn't work. You have to actually name the thing. You have to have stakes involved. You have to be able to make sure that the way that you position this content is actually relevant to the user. So headline framing is actually one of the most important things. And then obviously, the image that goes along with it. And so we have image training that we do. So we do have a lot of technical back-end processes that help us make the recommendations based on real data. And so when we're going through this Discover analysis that I'm talking about, one of the things we do is not only look at headlines, but also images. We can do image visualization comparisons now at scale using AI and then trying to understand, for example, is it a profile image? Is it a landscape image? Is it an action shot? Is it whatever, right? And then segment those and provide a statistically valid set of results that lead to higher Discover performance. So for example, it usually winds up being multiple things. It is: if you use an action shot and include two entities and you frame it as a question, those perform better than if you do these other three things. And so what we're trying to do is come at this, it's a very data-oriented approach in order to help people understand how to construct headlines that follow this general formula. Obviously, you're talking about writing, and so some of this is going to be intuitive and some of this you can't always follow, but we're trying to give them an idea of how to frame this information and also give them an idea of definite don'ts, because we wind up coming up with lists often that are clear negatives. And so more than even just the yes, affirmatively do these things, it is "Don't do these things," because we are seeing real suppression when it comes to Discover performance when you are including an image like this or a headline like that. So those are some of the things that we're doing to try to understand how to develop a Discover analysis that is meaningful and actionable to our clients.
Jon Clark (31:55) Super interesting.
Joe (31:55) I have a question about breaking news too. I know you and the team at Define do a lot of technical audit and recommendations for your clients, but I also know you guys do a lot of training with publishers. I'm curious: there is a technical foundation that you need to get right to be served for a breaking news story in a breaking news carousel. Can you tell us a little bit about the training that you give to a publisher? A lot of it has to do with editorial publishing, I assume, and the timing of publishing. Can you talk a little bit about some of the common things you use when you're training a publisher, like the editorial decisions they need to make to be considered in a breaking news carousel?
Shahzad Abbas (32:39) Absolutely. Training is a core part of our service offering. We've been doing it forever. A lot of the training that we do is especially geared towards editorial teams. It is about making sure that the editorial teams understand that we're working for them and not against them. So, we're not trying to antagonizingly give them recommendations that they don't want to do. I think that's really important. We have to come at the editorial team in a way that feels like we're allies rather than adversaries. That's really important. Mechanistically, the way that we approach this is typically along fundamental SEO best practices. It's literally title tags, headlines, subheads, and providing that information in a way that, again, shows the data-driven approach. Everything that we do is very heavily data-driven, so any presentation that we ever do will show a set of title tags that are really well-dialed in, that have entities that meet the minimum standard title tag length, and then well-formed meta descriptions, H1s, and subheads. And we present that to them, and then we present the opposite. Then we show them the performance. Invariably, the performance disparity is so great that it's very clear that working on this approach will allow the editor to get their content out to the world. That's what they want, right? So from that perspective, that's what we're trying to serve is a way for those editorial teams to get their writing out to more of their audience. That discovery zone is just greater. So when we're developing our editorial guidelines, A, we want to make it as simple as possible. We do not want to overcomplicate things. Editors are not SEOs; that's not what they're meant to be, especially when we're talking about journalists. They have lots of other jobs to do that are very important. And so we're not sitting there trying to dial them in on doing the minutia of keyword research. We're trying to give them the tools to understand intuitively how to actually write in the way that is very natural to them, versus an SEO-ified or AI-ified approach to headline writing or content writing.
Jon Clark (34:53) I'm glad Joe brought us back to breaking news because I had a couple of questions here. My wife used to work at the New York Times and she would always tell me the impact of updating a title as something new was sort of added to a post or something like that. I was always amazed by that. Granted, she's working on probably one of the strongest domains across the internet, so that helps. But as it relates to Discover, that strategy is not something you would employ, right? You want more of a static title there, or is that headline or title update applicable to Discover as well if you start to see traffic dipping?
Shahzad Abbas (35:29) Yeah, don't sleep on title tags. Title tags are so basic but so valuable. You're 100% right on that. When it comes to Discover, that's actually a great question. I am not sure of the impact that updating a title or even a headline would have on Discover performance after it's actually been indexed. I have a very strong understanding of how that appears in breaking news. If you update your title tag or headline, Google is coming back every few minutes to actually re-index content, especially if it's in the News sitemap, and especially if it's gotten pinged to actually say, "Hey, there's been a significant update to the story." It wants to present the latest story to its users. So 100% that is impactful. How that surfaces to Discover is actually an interesting question. I don't know the answer to that question simply because I don't know whether that actually gets juiced up a little bit in Discover for a news piece that is of the moment, or how that actually impacts on a data basis. It might be something I might look at.
Jon Clark (36:34) It's kind of interesting too, right? Because Discover just had its first official algorithm update. So I'm wondering if they're starting to separate those out and maybe there is an opportunity there. It sounds like more to be learned before we can make an explicit decision, but it's kind of interesting to think about.
Shahzad Abbas (36:48) We're definitely seeing somewhat of a separation between the Discover systems and the Search systems. I think Glenn Gabe was the first to point this out a little while ago, and I would concur with that statement, which is that the systems that had been effectively... I think Discover was sort of a child of the organic search index, and so it was built upon that. Then over time, I think it's just evolved, especially now that it's sending even more traffic often to publishers than web search; it's becoming its own entity in its own right. It's moving away from the nest in a way. From that perspective, it requires its own unique algorithms, assessments, and signals. I'm not surprised that some of those are going to live in Discover only versus Search. Having a core update that is Discover only speaks to the viability of that statement.
Jon Clark (37:34) I have maybe a two-part question around breaking news. Is it better to be published first or is it better to be published best as it relates to getting pulled into the Top Stories carousel? And I guess the second part of that question is the published date. I've seen a lot of case studies, and I think we've seen it in our own client data, where it's better to only have the single date of when it was last updated, not published and then a recent update. Do you have any thoughts around the impact of having multiple dates on a page versus one?
Shahzad Abbas (38:09) Yeah, I'll take the second part first, which is the published date. That's been a big bugaboo with Google. We've seen lots of instances where Google is just pulling the wrong date from the page. There are lots of places where the date can actually live. It can live in the structured data, it can live on-page, it can live even on different sections of the page that the CMS actually pushes. We've seen lots of different versions of pages have multiple different dates and Google is actually pulling those in, surfacing them in search results, and then it being wrong, sometimes wrong by years. That can be super impactful to the viability of this as a search result. Even if it is evergreen content, if it's last updated in 2023 and it's 2026, are you going to click on that evergreen content versus something that was updated three months ago? You probably won't, because there's just more updated information that you want to get and it's showcasing it to you in search results very prominently. So that's a problem. I think you do have to be very careful and judicious about specifically which dates you actually provide. I think you're right, Jon, that you want to showcase only those dates that are clear, meaningful, and most impactful to the user, and not have two separate dates because we've seen time and again that Google chooses the wrong date or the date that you don't want it to show. Reducing that point of friction is extremely important. As far as the first part of your question goes, when it comes to breaking news, I think first is probably best. I think it aligns with the race to actually get a story broken and be the original source of that breaking news material. One is that you'll just get the links back from it. Two is that Google sees you as the first, and so I think that is also helpful. From that perspective, I think generally speaking, being first is super valuable. How that holds and whether or not eventually you settle into what is best is a different story. For example, if let's say the New York Times and CNN... let's say CNN has the initial break on the news, but over the course of the day or two days in which this specific news lifecycle is actually live in the search landscape, although you sort of have that head start, I think that it can be overcome. If you publish at a cadence, at quality, if you do all the SEO right, if you have a live blog or anything else that supports the ability for that news content to continue to surface in your top carousels, I think it can be overcome for sure. It's not as if you're always going to be behind; you can play catch-up as well.
Joe (40:52) I've got to shift gears a little bit. Shahzad, 15 years ago, you worked for TV Guide at about the same time social media was just becoming popular and disrupted companies creating, whose business model was built around creating and publishing content. You must have learned something from that experience working with TV Guide that now, in today's era of AI disruption, or at least with publishers' business models slightly being disrupted by AI, that maybe you're seeing something that you can relate to from the past? And I'm curious, with all the conversations you're having with some of our country's largest news publishers, is there something that they're missing or not really willing to accept with the reality in which they're going to have to figure out?
Shahzad Abbas (41:43) That's a great question, and I appreciate you bringing it back. History rhymes, right? We see a version of history in terms of these cycles repeat themselves. And it is always well worth looking back at historical cycles to recognize what we can understand from them and also understand what is different from this specific cycle. I was at another legacy site, it was TV Guide, not Reader's Digest, but the point is still the same. But the point is still extremely valid because at that time, I remember our GM, Facebook had just sort of come out publicly. This is like 2006, seven, eight, something along those lines. People were really starting to recognize Facebook as an important entity. My GM at TV Guide, he actually requested us to sign up for Facebook from his account to all of our accounts. Because he was like, "Sign up for this thing. It's important if you haven't already." And so that's how I actually got my first Facebook account, because he was like, "This is worth paying attention to." So shout out to Paul Greenberg for getting me out there first. When it comes to that disruption, I think there have been multiple disruptions. So social was a disruption. Google algorithm updates were a disruption, right? So the Panda-Penguin era was a disruption, where all that sort of backlink network work was kind of wiped away. There was the mobile stuff that happened in the mid-2015s was a disruption. And so there have definitely been these sort of large-scale cycles of how Google has changed and how consumption habits have changed. I think one of the things that I would say when it comes to AI that's really important to understand, the way that I like to think about this current cycle is by using a term that I think Rand Fishkin is actually trying to advocate for, which is this is an area of SEO where the S-E-O is Search Everywhere Optimization. And I think that has a lot of credibility and valence to it because the mechanism of information retrieval hasn't diminished, it's only increased. So Google Search is actually, as a product, continuing to increase in terms of how many people are searching for content on Google through the app or through google.com. When ChatGPT first came out, we thought, hey, this was actually going to cause Google to potentially decline in search market share. Maybe it has a little bit, just overall. But in any meaningful or material way, it's not true. And so what we're seeing right now is that effectively what's happened is AI is stacked on top of Search. So Search is still whatever it was, trillions of data points that it has. AI is just on top of it. And it's a mechanism that people are using to get back into Search, to go deeper into either validating or purchasing in terms of their journey and completing their journey. So I think that's important for everybody to understand: that there's a lot of hype around AI, and especially AI tracking tools. Google Search is not going away. That is very, very important to understand. Also, Google, from a distribution and hardware and market position standpoint, has a very strong position. Gemini is integrated within Google Search, it's a standalone app, and the distribution is something that nobody else can touch. So whatever growth areas you're seeing as far as OpenAI or Claude, the reality is that the full scope of data, if you think about Gemini plus Google, they're actually in the lead. And it's going to be really hard for anybody to dislodge them because, let's say even from a legal standpoint, the Justice Department did not say that Google Chrome and Android have to be separated from Google proper. And so now they have carte blanche. They can really put the pedal to the metal and allow themselves to go full bore on the integration, to full-scale vertical integration between AI and Search. So that's the one thing that I would say, is that it is a Search Everywhere Optimization world. All of these LLMs are growing. It is on top of, and not in replacement of, Google, and Google is very well-positioned. So it would not be prudent to think that Google is pass'e9. In fact, they are in a stronger position than they have ever been.
Jon Clark (45:58) Got it. Yeah, I think that's maybe a good place to wrap. Do you mind if we jump to some rapid fires?
Shahzad Abbas (46:02) Yeah, go ahead. Yeah, this is great.
Jon Clark (46:04) You mentioned Substack earlier. I'm sort of curious: if there was a news startup today, would you recommend Substack as a CMS to build on, or would you still think about a traditional CMS as the way to go for a news publisher? \
Shahzad Abbas (46:19) I'm old school. I want to keep my people with me. I don't want to give away my audience to anybody despite how valuable it can be. I think you can have Substack as a supplementary, but if I'm starting anything, I want to start my own and allow it to grow organically, endemically, and have it be owned by me and have that direct relationship. I don't want there to be an intermediary between myself and my audience.
Joe (46:42) If you could add one piece of data or view of data within Google Search Console to make your life a little easier, what would it be?
Shahzad Abbas (46:50) Top Stories. They've never had a Top Stories view. Secondarily, I'm going to expand it a little bit, but now AI Overviews. Please, we'd love to see what that is. But I think it's difficult for them to want to do that because it's going to show a stark reality.
Jon Clark (47:04) Maybe not a rapid-fire question, but your comment reminded me: are you integrating the citation data from Bing into your data analysis recently, or is it not expansive enough to make some decisions against? \
Shahzad Abbas (47:16) We haven't looked into it yet, but that's a great point, Jon. I think we can use it as sort of a proxy. It's sort of like the Nielsen model of a sample size that's probably good enough. So I think that's worth looking into. But I think the problem is that, frankly, there are not a lot of publishers that are looking at it actively or even have Bing Webmaster Tools. So we have to maybe push them, and even on us, maybe we have to recognize that they are innovating and trying to push the needle for higher standards. And that is something that we should look into as well.
Joe (47:49) Is there a KPI that your most sophisticated clients care about that your least sophisticated clients don't really care about yet?
Shahzad Abbas (47:59) Yeah. I think most of our clients are fairly well-attuned to revenue. Eventually, everybody is in business, right? And so they're trying to generate revenue, and Search is a revenue-driving channel. That's fundamentally what it is. And so we have to allow a clear view into how much it's actually contributing to that revenue. I think all of our clients are doing it to some degree. I think some are looking at it more heavily as a clear KPI than others. And so it's not necessarily that not everybody's looking at it. Everybody is. But I think there are some who are doing it in more sophisticated ways than others.
Jon Clark (48:38) I think one challenge with a news organization is figuring out what those SEO roles should be or those audience roles should be sort of in the newsroom. Is there, and I'm sure you get asked this question, especially in your training, even your consulting, is there a role that you're recommending today that maybe you didn't back in, I don't know, 2015, 2016, especially with the growth of AI and that sort of thing?
Shahzad Abbas (49:04) The folks we talk to that we have relationships with that are main points of contact within the publishing space, what we are trying to do is expand their remit to include AI. That's really, really important, not just for them or for us, but also as a representative of the organization to make sure they are understood as having the chops to weigh in on these important decisions that will affect how AI actually functions and how external AI bots are able to access the information that is within your site. That's really, really important. And I think that's what we're trying to function. Again, there's not any bureaucracy in my company, it's just three people. And so like, I literally changed my role in title to include AI. It was just because it was necessary and sort of a reality of the phase of this technological cycle that we're in. That's what I'm trying to do with our clients as well, is to make sure that, especially our points of contact are known for being leaders in the AI space because nothing actually prepares them more for being leaders in the AI space than being leaders in the Search space.
Jon Clark (50:18) Totally. That's one of the toughest things in those types of organizations is knowing who to go to. So yeah, it's a really, really smart approach. All right, last question. I'm really curious: what's the best piece of advice Marshall Simmons has ever given you?
Shahzad Abbas (50:31) Marshall is a boss who doesn't act like a boss. Like, he's amazing and frankly just changed my life because he effectively brought me over from TV Guide, which was a place I loved, but I had to commute to the city. And so we wound up, I've been remote since 2011. I don't know if there's one thing that he's ever said to me, but I would say that his approach has been so meaningful in the way that I'm trying to approach things, which is that he really does allow people to find their own path, allow people to recognize the things that are important to them, that they're passionate about and that they are good at, and just gives you a lot of rope to try to do it. So there's never something that I've posed to Marshall that he said no to. It was more always like, "Let's see if it works," kind of thing. And I think that's super valuable. And that, plus his extremely collaborative nature, has been extremely maturing for me because that has helped me in my own consultation and managerial roles. As I'm managing and consulting for other clients, it helps me understand how to approach these relationships in a very meaningful and productive way.
Jon Clark (51:44) Gosh, before we wrap, in our research, I found this amazing article on TV Guide that I think you wrote. It's "Meat Loaf Will Fix Your Website," back in 2009. It was an enjoyable read. I'll link to it in the show notes. But Shahzad, this has been really insightful. I appreciate you taking the time to be thoughtful with your responses. So if folks want to connect with you online, where can they find you?
Shahzad Abbas (51:56) Yeah, I'm on LinkedIn. I don't have Facebook. I don't have Instagram. I don't have TikTok. But I am active on LinkedIn, so you can find me there.
Jon Clark (52:14) Perfect. And we'll link to your profile in the show notes as well. Thanks again for being on the Page 2 Podcast. And for those who are listening in, if you enjoyed the show, please remember to subscribe, rate and review. We'll see you next week. Bye-bye.
Shahzad Abbas (52:26) Thank you guys.