Garrett Sussman: John is the CEO and founder of NewzDash, an award-winning real time news SEO software and g dash, Google Discover Analytics and Optimization. He’s the founder of Nest the News and editorials SEO Summit, and former global VP of Audience Development at Conde Nast. Leading SEO, social and email strategy for 16 iconic brands including am I gonna read them all? No. I’m not gonna read them all. But Vogue, GQ, The New Yorker, and more. John’s the contributing author of The Art of SEO, USA Today, Top 10 SEO Experts. Jeez, what don’t you write in, John? Okay. So he’s written in some things. Former SEO lead at Disney, ABC. He’s trained thousands of journalists, built multiple SEO tools, and continues shaping the future of editorial SEO across the globe. He knows a thing or two about the publishing industry. Presenting, unlocking Google Discover code, analyzing 200+ million articles. Please welcome John Shehata.

John Shehata: Hello, guys. It’s amazing to be here. I just want to say thank you so much. I know it’s the hardest to come to a session after lunch. It’s like this trade off, do I go take a nap or come actually listen to this? So I also want to thank the team for the new chairs. Amazing. I can tell like my back from yesterday to today. 

So today we’re going to talk about Google Discover. And Google Discover is the largest source of traffic for content creators and news publishers nowadays. The largest. It surpasses everything. It surpasses Google Search itself. So today what I’m gonna talk about. And the other interesting fact about Google Discover, Google Discover 80% of the feed is completely personalized. So if you think about the memory layer and ChatGPT and all the stuff, this can actually tell you a lot of stuff that actually can help you with all the LLMs optimization. So why should you care? How do you get in Google Discover? How to optimize? And few other reasons there. Now I want a fair warning, I have 104 slides and I have 30 minutes so that’s why I go in 2x speed, but don’t worry, you’re gonna have all the slides, I write all my notes, so don’t worry about that. Everything is in my slides. Okay. This is some of the companies I worked on. So fun, I even optimized Reddit at a certain point of time and I had access to older Google search consoles, that was fun. I own NewzDash which is like real time SEO tool for news publishers and we have Nest which is like amazing small virtual conference. We have about 800 people a year that just talk a geek about SEO and news. And I have a small tool called GD Dash that analyzes your Google Discover traffic. Okay.

So Google Discover. You guys know Google Discover. It’s like Google’s social feed or you might want to think about Google Discover. It’s like this person who pretends to be your friend who comes to your house uninvited all the times, he always brings gifts to you, right, you don’t know if they’re coming tomorrow or not, they keep coming day over day and then suddenly they stop. And then you don’t know, did you upset them? Did you make them angry? And then they come and show up again. That’s really Google Discover. It’s a mystery in itself, you don’t know why they come, why they go, and this is what we’re gonna talk about. So this is Google’s version of Facebook. 

Google Discover is all based on the foundation of matching entities and content classifiers against user interests. Okay? And I’m gonna talk a lot about that because this is very important to understand if you want to understand personalization in different other tools. When we talk about user interests and a lot of other tools are using this, what is user interest or how Google and other memory layers build a portfolio about you? They understand your activity, right? What content do you read? What’s your search history? Hey, you’re looking for a house, you start looking for mortgage rates, you stop looking for mortgage rates, it means you might got something already so they instead of giving you like, oh you should be reading this about mortgage rates since you stopped searching for this like evergreen topic, they no longer give you this. So they understand the evolution of your search history. They understand what do you read in Google like on Discover in other apps, in Android. Actually Google has access to many of the other apps you have in Android and they can see what you’re doing on Instagram, what do you like in YouTube and other places. So they see all of this. They understand your location and there are three different modes of location. Where is your work location? Where is your home location? And where is other locations? Vacation and stuff like this. And it was like, if you’re on vacation, what type of content should I give you? Versus home or whatever. Your device and explicit signals when you tell Google Discover, hey I like this content, I like this website, and so on.

How does it work? So there is two things that the foundation of Google Discover, entities and content classifiers. I like to call content classifiers categories for simplicity, right? And then they match this against user interest with what I just spoke about right now. You guys remember that story? Oscars, right? I optimized Oscars with the Academy Awards for like six, seven years so I’m gonna use that example. They never invited me to the red carpet. They put us in a hotel where I like watch Oscars on TV. But anyway, so if you think of this incident, right, you think of entities are Will Smith, Chris Rock, Oscars, Academy Awards, these are the entities, right? So entities are real time things like people, places, events, organizations, and so on. These are entities. And then when you speak about categories or as Google like to call them, content classifiers, right? It’s like a broad section of topics together, right? So for this specific story it can go under like arts and entertainment, film and TV industry, celebrities and entertainment, other like movies, gossip, and so on, and there different relevancy or salience for all of these kind of categories. So this is the foundation. Google looks at your content, classify it into entities and categories, and then looks at user interests and do the matching. Right? If you try to understand every single user through a personalized feed, there is no way you can do that. Same thing with charity and other.

So what you try to do is reverse it in a different way. You try to look to a group of users of a similar interest that the search or LLMs or whatever is rewarding you with that traffic. And this is the best approach to optimize for these kind of platforms. When you look at the content that appears in Google Discover and the traffic, we have one of the tools we have, we have GDDGD dash, and we have about 500 publishers. Right? And the majority of the content, 80% of the traffic that you get is only focused mainly focused on organizations, people, and locations. So for every piece of content you write, you need to think about these three entities and how they fit into your content. So why should you care, right? It’s like it brings the highest amount of traffic, that’s okay, that’s fine. This is a recent study we just did like a few weeks ago, and you can see like if this is a Google pie of 100%, Google Discover year after year keep growing. Okay? So right now on average it’s about 65% for content publishers. I just worked with someone a couple of days ago, 99% of their traffic is Discover which is very scary. Google search as a share of the pie keeps shrinking. And I feel like Google is telling us like, look at this amazing thing Discover. I’m gonna keep you giving you more traffic. And on the other side they keep like taking traffic away from search. Right? And then all the other platforms. The other amazing thing which you guys have this data in Google Search Console, Google Discover has almost four times, four folds the traffic, the CTR you get from web. So you can see CTR here, it’s about 8.9 or I can’t see the numbers here, 8.9 CTR for Google Discover. And look at the the web, 2.1. It’s amazing, it’s huge, right? Eight times for every, the ROI for Google Discover.

Another thing also, so this is some drama, a couple of days ago I found a hack in Google Search Console because they said Google Discover is coming to desktop and I found that the data already exists in Google Search Console so I put the information out in LinkedIn and I’m so naive or I don’t know. I reach out to the Google team, someone specifically, I said, oh this is awesome, thank you for having it in Google Search Console. When’s it gonna come to the API? Within one hour they removed it from Google Search Console. So but it has been on on desktop for almost at least 16 months of data. I already found it there. 

Okay. So before I before I start, right, this is not a strategy. This is not a long term strategy. If you are focused and addicted to Google Discover, you’re gonna get hurt badly. Okay? If this is the main channel you’re getting the traffic from. Okay? So this should be a supplemental thing. This is like a short term strategy, you squeeze as much from Google Discover as you can but this should not because I have seen publishers who switch their approach to Google Discover and like thin content, clicky headlines and stuff like this, and within few months they lost all their traffic. So don’t do that.

Okay. Google has a lot of like hidden messages in their guidelines, but let me decipher this for you. I believe this is the approach or this is the process that goes when your content get in Google Discover. The first step, you must be indexed in Google. Right? So Google crawlers need to find you and index you. You get automatically become eligible to be in Google Discover. Then they look at, is this a new URL? Did we see it before? Does it match our Google Discover policies, high resolution, large images, and all the stuff. Then you get in a sandbox. Okay? That’s why when you publish something, you don’t get Google Discover immediately. There is some kind of like a lag time where they start like classifying your content into another level or deeper level. What are the entities? What are the content classifiers? And so on. And then they gather sufficient audience data. So Google has all the Chrome data about your content. Who is visiting your content? And when they see social and audience traction on that piece of content, it even pushes like makes it more eligible for like Google Discover. And then you appear in Google Discover and they match your content with the user interest. This is type of content that you should not be in Google Discover, right? So especially the very first one, you see a lot of sports sites that they have these graphic or adult oriented kind of like content based on the golf course and stuff like this. Usually it doesn’t work well in Google Discover, and there is a lot of other stuff as well. 

Can you optimize for Google Discover? So Google official statement, you cannot. Right? And if you look at it from an engineering perspective, yes, there is like everyone gets a completely different feed. Right? So how can you optimize for that? But in reality, in my opinion, I was the second site that got in Google Discover in 2018 when Google actually used to power Google Discover through RSS feeds. Okay. This is the original Google Discover. So working on Google Discover for that long of a time, I believe there are 11 different ways you can optimize for Google Discover. I’m not gonna like go over every one but you can have every one in the slides but let’s talk about some of them.

The first one is, remember the key is categories and entities. So you need to understand what categories and entities that actually drive traffic to your site. Right? And how would you do that is by analyzing your site. Right? So you can see actually Google Discover give us a lot of hints. These article about machine learning, these articles about mortgage rates. So you think of these as like display entities. Right? If you say. So how we can gather that information to match it against our users? I’m gonna So we built a tool for that but I want to give you how we did it so you can do it your own. This is not like selling the tools or anything. The first thing is you need to connect to Google Search Console. Right? There is a lot of tools. There is – you can connect it to your Google Sheets. Right? So you connect through the APIs. You can use BigQuery free. Second thing is you need to get the content for each of these pages. So if you have the CMS already, right, you have all that content in the system. So now you have connection to Google Search Console, you get the traffic of Google Discover Daily, then you get the content of each page, and then you utilize NLP. NLP natural language processing. Google has one, it’s good for like maybe 20 countries, Other countries may have their own specific NLPs. The NLP simply takes your content and say this piece of content is about these entities, these categories, and this is like the relevance of these entities and categories to that content. Once you do that, you start grouping things based on impressions. And impressions are more important than clicks. If you have been paying attention for everything we’re seeing, clicks keeping going down. But are you there? Are you visible? And impressions, this is Google decision to show more or less of your content. Clicks reflect on your headlines and images. Are you doing a good job or not? So you’re gonna measure everything against impressions, not against clicks. 

Okay. So this is how we do it in GD Dash after we do the four steps that I showed you. And then you start discovering what is the cluster where your site actually getting all the traffic or impressions. And for this site you can see like very food and drinks are very big, but why they are writing about sports. So this start giving you ideas like are you off topic? And because Google, a lot of like the updates recently they’re doing is like, why are writing about this stuff? That’s not your core, right? So this can give you these kind of ideas. On the second example, sports, news, and arts and entertainment, so I call arts and entertainment a striking distance kind of category. You have some visibility, it needs to go like grow a little bit. Once you do this – so I’m gonna give you this, how how would you use that. Right? So this is a news site, a very prominent news site in SEO that’s using the tool. And you can see that when they write about news, the very last row, they do 218 million. Right? And they wrote 1,500 articles about news. Right? This is this is the core product that they do. But they also write about pets, completely two categories. For every article they write about pets, they get 9,000 impressions. For every article they write about news, they get three impressions. So I’m not saying you’re gonna stop writing about news, but you understand that the ROI on pets is three times the ROI in news. So this is a conversation with editorial teams. What’s our strategy for that? Right? Same thing here. 

This is entities. When they write about the Russian conflict, right, they get a 110,000 impressions for each article they write. But when they write about the golden retrievers, who doesn’t love – I’m I’m a dog guy, I think someone asked a couple of days ago, they get 400,000 impressions for every article. So it’s four times. Analyzing your content and understanding what your content stands for, categories and entities, and understand how Google view your content and which content they reward with more impressions is very important so you can work on your editorial strategies and work with your editorial teams. You need to deliver timely content, right? It’s like trends are very important in Google Discover, right? So understanding what’s trending out there. So we see a lot of times when something trend and it hits Google Discover around the same time, you get the spike of the interest. 

There’s two types of trends. There’s like, you know, evergreen trends, like everyone’s talking about whipped coffee during COVID. I found that in TikTok. We asked the Bon Appetit team to write something about it and we got like, I don’t know, 15 million impressions for just that one article. So there is this evergreen. But there’s also the news trends and what’s trending right now and stuff like this. So these are two types. One thing that we do like to analyze because we have user panels like let’s say we have like five million users in US that we track. They give us access to their Google Discover feeds so we see what they see. And we like to track what’s trending actually in Google Discover. So this is another way of looking at it.

One concept I’m gonna throw it out there. Hey John, Tom Cruise is trending because of Mission Impossible. I don’t have any content on Tom Cruise, that’s not our thing. So there is a concept I came up with, it’s called adjacent content. Adjacent content, so I used to have one of the sites, one of my sites, Conde Nast was Architectural Digest. Right? Has nothing to do with entertainment. But guess what? We had a story on every celebrity around the world and pictures in their houses and where do you live and stuff like this. So every time a celebrity trended, I will take the main story about this celebrity, we push it to the homepage, where is Tom Cruise hiding when he’s not filming Mission Impossible? And guess what? We ride the wave with the trends. So thinking about adjacent trends, it doesn’t have to be exactly on Tom Cruise and Mission Impossible, but if Tom Cruise is trending and you have something adjacent to that, you can capitalize on the moment. So just an idea to leave you with.

Titles are the most important thing in Google Discover. And Google is telling you, oh, don’t do clicky, baity headlines, but this is official Google guidelines. But the reality, they do work. Okay. These are the most popular headlines and trafficked articles in Google Discover. But you cannot be outrageous. You cannot promise something that you will not deliver. Right? So it works and here are some of the examples. What type of headlines work? Curiosity, sensational, click worthy content, list studies and so on. Fear of controversy, reference to well known individuals, celebrities, entities focused headlines, incorporation of trending topics and so on. This is my favorite, firsthand headlines.

I’m 63 years old, I worked hard my entire life but I just got fired after announcing my retirement. Is this even legal? This article remained in Google Discover for almost like a month. Right? And the funny part, this is a Yahoo article. That guy didn’t write the story. They quoted the guy from the story, put it as a headline and it was doing very very well. People connect to these kind of headlines. So first hand headlines work really, really amazing in Google Discover. Objectives, rare, surprising, whatever secret to, not prepared for limited time. So these are the trigger words that make people click on your headlines. 

Google Discover is an amazing way to do – on your site you can do A/B testing. So when you have like a channel page linking to an article, you can use scripts to try different headlines on your site and see which one actually get higher click through rate and start utilizing these in your OG title. So this is a very thing. Now we’re talking about AI. So take all the data you collected through Google Search Console and entities and stuff like this, throw it all in your own model and start asking this model, what works best for me? Right? Which headlines? What kind of patterns of headlines that work for me? What type of stories and so on? Number of words, we see much higher number of words in Google Discover. But this is not to say, hey, start increasing the your titles or stuff like this, we’re seeing longer titles and headlines work in Google Discover. This is very interesting. Are you guys familiar with OG titles? Yes? Okay. Good. Okay. So OG title, I start analyzing different countries. When OG title is different from the headline and the title, 48% of the time US, sorry, 34% of the time in US, Google will go with the OG title. Definitely utilize OG titles in your CMS if you’re not doing that. In Germany, it was so weird. It was almost like 61%. Okay? So it’s every country is different. Diversify.

It says AMP top story but it’s not AMP. This is Google saying like it’s news or stories. Right? Videos, it shrunk from 4% last year. A lot of videos were very visible last year, much less this year, 1%. This is the six different types in Google Discover that you will see over there. Now there’s two different types that you should be aware of. The news headlines, these are non-personalized. And usually it’s about 20% of the feed. Right? And usually they stay from 30 minutes to a couple of hours. Right? But almost everyone sees them, but for a shorter period of time. Classic articles headlines, these are all the personalized ones. Right? 80% of the feed or more. And less people see it because it’s very personalized but they get you a lot of traffic.

Instagram, YouTube are very very popular. The number one website in Google Discover is YouTube. So Phil was talking about it, That’s another reason he should add to his list. This is the number one site in Google Discover is YouTube. Quality visuals, large images, high resolution is very important. One thing we noticed after we analyzed the top 100,000 images for the top 100,000 articles, the majority of them are people. Close-up people, right? So this works very very well. When people like the reader can connect to the person on the image, usually works very well. Different colors work better than others. I’m not saying this is causation, but there is high correlation there. Backgrounds and all the stuff, you guys all of this. Adult content works very well. Hey John, is there adult content in Google Discover? Not really. It’s like, hey, a new study shows that users of Pornhub do do so and so and so. Right? So usually Google Discover minimized the amount of content that appears there, but no surprise, that content has the high CTR for some reason, 13% versus news is like 2%. Who cares about news, right? Shopping, the highest shopping category that works in Google Discover is swab meets outdoor markets, flea markets, and stuff like this. Antiques is the lowest. When it comes to news, sports news is the highest, 10% CTR versus like business news and stuff like that. Sports, I was I was amazed that Australian football has much higher, 14% for whatever reason, versus American football that only has 10%t. Now I believe there’s social buzz has is a factor in Google Discover which means that when they look at your social traction, right, did this story get a lot of traction. One of the things, I was in a conference, Ona, and they they were talking to the person who made the deal on Reddit with Google, and he said they don’t only send them the content, they send them also the velocity of up votes and down votes in Reddit. So they know which content in Reddit is actually gaining momentum, which one is not. They have that information from Twitter, from Facebook public pages, and other places as well. So they can tell it’s like Kraut Tangle of old days, which content is actually doing well on other social platforms.

I was speaking to someone and I was like, oh we have this tool that can trick Google in thinking that this article has a lot of traffic because this is part of social Google Discover algorithm? Is this article getting like a burst of traffic and stuff like this? And I said you don’t have to trick it. If you have push notifications and you have like a very strong article you want to push, just do push notifications. Google Discover is looking at Chrome data all the time to see which articles are getting like high velocity of traffic. Mobile friendliness, this is a mobile product so it does make sense. EAT, so hey you should definitely speak to Lily, she’s right here. And I linked to some of her stuff right there. So this is good E. It’s so hard when they add the second E. So EEAT is very important to Google Discover. You should get your knowledge panel for your brand. If you go to Google Knowledge Graph and you search for their brand, this is the one on the bottom, and I search for ESPN. So Google Knowledge Panel, it has about nine billion entities. They give a score for ESPN 37,000. When I went to and searched for Bleacher, they said 20,000. And if you look at the traffic data and visibility, ESPN has much higher visibility than Bleacher Report. That could be an indication of like, you know, it’s usually a score for a different thing, but it’s a good indication how well you’re doing.

You should do like person schema on all your author pages and so on and so forth. So do this at your own risk. Here is a black hat, gray hat kind of tactic. Google Discover only recognized new URLs. So a lot of publishers in Europe, what they do is like they push the article, it doesn’t get traffic or Google Discover traffic within a couple of hours. They re refresh the article again with another URL, change one word in the headline, swap the image. And a lot of times it works because Google Discover do not compare URLs. It’s not a good strategy for evergreen, but a really effective strategy for news. This is a live article from CNN actually where Google Discover was like indexing every live update that we were doing. So it was like, they don’t care. It’s like a new URL, we didn’t see it, we were gonna get it. But the perm link got the highest visibility. This is another site where they do the 10 best movies to watch this week from Netflix. So every time they refresh, it’s the same URL. Every time they refresh, they get a spike, so it can work also on the same URL, so you can definitely do that. 

Spy on your competition. So the best and simplest strategy when it comes to Google Discover, see what’s working and repeat. That’s really the simplest way in Google Discover, right? So you need to understand what’s working in Google Discover overall, right? You need to understand what are the most viral content in Google Discover, and then it’s like, oh, we can talk about this, we can talk about this. Taking into consideration that you have authority and your site is relevant to this topic, usually it works if you write about the same thing. You ride the wave, right, or do adjacent. YouTube is #1. You can see here from YouTube, right, for every hundred visitors or visits or clicks that YouTube gets in Google Discover, Yahoo will get 23. Right? So YouTube by far is the biggest if you’re doing any video optimization because you can see all the different types that YouTube is getting as a traffic and start mimicking this for your YouTube channel. You can figure out your category because categories change all the time. There was a big snowstorm so Google will discover will push weather all the way #1. Now politics is the biggest thing, so you understanding what Google Discover about every day or every hour is extremely important. This is a trick I used to do on Conde Nast all the time for seven years. I will see what’s working in one language for one brand, automatically translate it, have someone like read proof it, right? And then push it with a different language in other brands that we own. 76% of the time it works. It’s the same algorithm. And if you have authority on that topic, can translate and syndicate and send to other sites.

I think we have, we’re zero minutes, I listed here 11 different ways that you can lose your, 11 reasons that you lose your Google Discover, I’m not going to go through them. The one thing I want to just warn you about, unsafe content. So you can go to Google and Google can tell you how much of your content they consider to be unsafe by adding a filter and safe equals active and you can see, okay so how many articles I have in Google, let’s say 100,000, you add that filter and you say unsafe equals active, they say like I don’t know, 20,000. So this is an idea to give you how much of your content Google thinks that it’s unsafe in Google Discover. There’s another 20 slides there for you guys to enjoy. I hope this was useful to you. Thank you so much.