Rand Fishkin on SEO, AI Overviews, and the Future of Content Marketing

0:00Everyone knows Ran Fishiskin, but if you don't know Ran Fishkin, you're you are one of the most legendary digital marketers there is. You started Moz in 2004 as just a community. Then you launched Mozro in 2007. Ma Moz is like one of the one of the main SEO tools and and you founded that and you taught so many people SEO

0:25and digital marketing, myself included. uh you're you're a huge name and and you have uh you have Spark Toro which is an audience insights tool and you started a a video game studio and you wrote a book. Yeah. Among among among a few other things but yes those are all those are all accurate. Um except I think I started Moz in 03. It was 03. Yeah. I can't get Wikipedia to change the that's what it was to be correct. That's what it was. Um, you know, and aside from just being like someone who's great at teaching SEO, you I just I think you are one of

1:06the most well-rounded marketers because when you were doing whiteboard Fridays, it wasn't just about SEO. It wasn't just no follow versus do follow. It was like here's how you should do social media marketing and how it affects everything else. Here's how you should do email marketing and little tricks that you can incorporate. Yeah man. I I mean I think that the the job of digital marketing is to be knowledgeable in a wide degree of pursuits and channels because I mean you know as well as anyone there there is no such thing as well the only thing that matters is what I do on this one channel

1:47or in this one tactic. Um, and I think unfortunately the SEO world, which I left gosh, like eight years ago now, seven years ago. Um, the SEO world does have a tendency to hyperfocus on just, you know, how high do I rank in Google and how much traffic am I getting from them as opposed to thinking holistically about how every marketing channel and tactic interacts with every other one and builds a brand over time and influences each other. Uh, and all the knock-on effects. And I think I think that part of my part of my job, right, if if my goal is to help people do better marketing, well, I better I

2:25better help people do better marketing of all kinds. You know, I I had this question later on, but it's an easy easy one to bring up. Like you you uh wrote on threads in 20 2015 the like biggest KPI was traffic and now in 2025 like no, not anymore. So I I want to ask you for your companies for Spark Toro and for Snack Bar Studio or or campaigns that you're doing, what are your KPIs? What are you paying attention to? Yeah, for most of social media marketing and the the small amount of effort that we put into uh search visibility, it's impressions, basically reaching more people. Um, and you know,

3:12granted there's ways to gify that and and mess with it, right? You could talk about politics and reach a huge audience, but obviously it's not going to be hyper relevant. Uh, but because our conversations are relevant to what we're doing, impressions is a great metric. It tells us how many people we're reaching and it shows us the relative success of different campaigns, different pieces of content, different posts, different tactics, all that kind of stuff. Um, and then you know that at the bottom of the funnel, it's conversions, right? It's people who sign up and pay for Spark Toro and how many people try it for free every week? How

3:49many people sign up for a paid account every week? That's the um that's the real goal. Yeah, that's the that's the biggest one. But so impressions. So, are you not um I know you're out of the SEO world, but I'm sure you think about this. Are are you not if you were an SEO, would you be concerned about like AI overviews? Yeah, especially AI overviews. I mean, you probably saw the 7% the 7% dip in Alphabet stock from the announcement uh made by it was made by uh by Apple's Apple's EddieQ and he said this has never happened in 22 years. It was it was Apple Safari browser declined

4:33for the or searches on Apple Safari browser declined for the first time last month. Never happened in 22 years. And that caused and he bl he blames AI for that. And then he's like, "We're exploring other options other than Google. We're exploring. We don't know if Google is the best." And so Alphabet shares sank 7%. It's like very relevant that you're this just happened. It's crazy that you're coming on now. So I'm

4:57sure you think about this stuff. Um, so if impressions are going up and clicks are maybe decreasing, but you're not concerned about you wouldn't if you were in the SEO world, you wouldn't be concerned about that. Well, depends what part of the SEO world, right? If you if you're working as an agency or consultant, there's there's tons of clients who can't be convinced that they should care about influence, not clicks. And that's really tough. if I were in house uh you know sort of running a company and and was in you know deep in SEO space I I think I wouldn't be worried um because my strong belief is it doesn't it

5:43doesn't particularly matter whether uh visibility in AI tools becomes more important or Bing versus Google or you know Safari maybe Safari switches to duck.go go. I I don't know like what whatever it doesn't matter, right? If you're um in-house, you can just adjust to the times and um make the right KPIs for your particular business. Y that's not a big deal. I guess I'm more so asking about zero click searches. So somebody somebody searches, they see they see their answer, they don't click into the result. there. Sure, the answer is attributed, but are they even paying attention to that attribution? Uh I yeah I don't think I

6:27don't think the goal becomes get traffic or even uh directly create branding there. I think the the real goal that you have when the answer is provided without a click is give the answer that is most beneficial to two parties. one, your audience, right? The people who are paying attention so that you can continue to be that answer. And two, your business. So, for example, you know, um there's an AI overview that fires when you search for what is audience research. And good news for us, Amanda wrote a blog post that basically Google has just taken uh that content and put it at the top and made it the AI

7:11overview. It doesn't it doesn't really mention Spark Toro. That's fine. Doesn't matter to us. What it does is it nudges people to understand what audience research entails. And so that if they decide, oh, okay, I do need to do that. I need to invest in audience research. I've got to understand where my customers are participating online, what they're paying attention to, their behaviors and demographics, that's great because long-term that nudges them towards Sparkurro, which is one of the only tools on the internet that answers that for them. So, right, like it's it's not about getting the traffic. It's not even about creating the brand impression right there. It's just about being the

7:49answer that nudges people in the direction you're trying to nudge them. So, you're saying if Amanda, your VP of marketing, didn't write that, people who are searching that wouldn't be nudged towards Spark Toro. You know what? To be honest, I can't say that for sure because we do a lot of content creation on the web and the AI overviews doesn't just pull from one source usually, right? right? It looks across the web. But what I can say is that uh for years we've been talking about audience research and one of the words most commonly associated with audience research that you will find across the internet is spark Toro. And so I believe

8:31that's why Google or Chat GBT or Perplexity or any of the other tools uh sees a question about audience research and looks for resources or data that is inclusive of or comes from Spark Toro because we've created that brand connection through language across the internet for the last six seven years. That's really interesting. something right when ChachiBT came out like this was twe I so it's crazy we we launched on product we launched a product on product hunt the same day that that chat GPT first launched I think it was December 2nd 2022 they launched on product hunt the same day as us and for most of the day we were number one and

9:19then through the sec but like slowly throughout the day we see this like this thing with a weird name chat GPT was just all these random letters thrown together in this like weird icon. It was rising up the ranks of Product Hunt. We were number one and I'm like whatever. We're number one. We're number one. Then it's then it's number five. Then it's

9:35number three. Then it's number two. What's happening? Then it's number one. And it's it just like blew everyone away. Chachi PT. And that night, which was the launch of our product, I ended up staying up until 4:00 a.m. playing around with Chachi PT being like, "This is in this is insane. This is going to

9:50change everything." And and then something that I couldn't stop thought thinking about is like doing SEO like how does this change how does this change publishers what's the incentive for people to write top offunnel content and especially now with with everything that we're seeing and I don't know if that's something that you've thought about. I'm sure it I'm sure it is. I'm very curious what I mean I think the incentive in my opinion the incentive has dropped massively if we're talking about being a um a publisher who makes money either on subscription revenue or on advertising revenue. But the incentive is still just as strong if your goal with publishing

10:35is to influence an audience to have a preference for your product or service. Right? So if you're a whatever consulting agency in any field or you make a product of any kind, consumer or B2B, you have a strong incentive to create content both on your own site but also across the web on social media platforms on the websites that your audience is likely to pay attention to, which is also therefore the uh website that any AI tool is likely to use as the reference resource in your field. That incentive is essentially create the content that you would want a an LLM or a search engine to pick up and

11:24show to people so that they will develop a preference or for your product or a nudge in the direction of the change that you're trying to make. You know what? May I I I had this thought as you were saying that maybe maybe like a lot of publishers or or even just I don't know like B2B SAS companies will be like, "Oh, there's no incentive anymore. We're going to stop doing top of funnel and only do bottom ofunnel and then you come in and you fill the gap and you do top off ofunnel

11:53and you just write all your content." This is what audience research is. And the best tool for that, it's Spark Toro. And then everyone Yeah. Everyone who looks it up, they're like, "Oh, Spark Spark Toro." This is why it's important. Edward, you've got it, man. That's exactly. Look, you know, in 2012 or something, right? I said exactly the same thing uh probably on stage plenty of times or on calls where I'd be like, "Yeah, yeah, SEO is BS. Ignore it. Don't do it. That way I can do it and rank number one, right?" And it's the same it's the same thing now. It's like, "Yeah, don't sure don't invest in

12:25content. Don't invest in social media marketing. Don't invest in publishing. Don't invest in, you know, visibility in Google and AI tools. That's fine, right? Like those of us who do are going to receive outsiz benefit from it. This method of marketing is so effective, I had to make sure it wasn't against Google's rules before I kept doing it. It's a form of SEO I call compact keywords. Whereas most SEO focuses on putting up articles to answer questions, how, what, when, compact keywords focuses on putting up dozens of pages that sell to searchers who are actually looking to buy. These pages rank on Google and convert so much better than

13:02normal that when I discovered this years ago, I couldn't believe this was allowed. It's less work, too. The average compact keywords page is only 415 words. Compact Keywords is a 13-hour deep course on getting sales with SEO. A customer recently said, "Each lesson is dense with information. You're giving years worth of experience boiled down into 15 to 30 minute lessons with no filler or fluff. I feel like I'm gaining a new superpower. Compact Keywords is about setting up an SEO funnel that brings you sales for years and years and years. It works with AI. It's less work than traditional SEO and it makes way more money. You can get it now at

13:46compactkeywords.com. Back to the podcast. I want to see since you since you mentioned all of those channels. So I want to ask do you think this was this was a question I went for a walk right before this and I just like walks are the best for thinking and this question hit me and I'm like I really want Ran's answer from this. So, do you think marketing is easier now or harder now than it was when you started MA? And you started MOA in 2003. And if you started MA in 2025, what would you do? I would not start Moz in 2025. Let's um that that space is just so incredibly

14:23crowded and I don't like crowded spaces. But um I think marketing now is massively easier in one way which is that there's so many more channels and they are much better understood and there's far more information about them and people paying attention to them and um reports and data being produced about how these marketing channels work. In 2003, Henry there was like there was so little like you know you say hey Ma's built this like great reputation in whatever from 03 to to 2012 13 like those 10 years the the bar was so low for us to create just any content marketing that's why I wonder if

15:11it was easier because you could do anything. Look, I think I think it was easier to stand out because the field was not crowded, but it was harder to do marketing because there was just not information about how it worked. And um very frankly, the number of channels and tactics that you could use to reach tens or hundreds of thousands of people or millions like barely existed. Didn't really exist. Internet adoption was so low back then really if you wanted to reach a million people which you you know not I I don't want to say easily can today but over the course of a year if you invest consistently in YouTube or

15:52threads or Reddit or LinkedIn or Tik Tok or Facebook like you can reach a million people in your audience if you sort of do a good job easily. That was not possible in 2003. There weren't even like there weren't a million people in the in the field like looking on the internet in 03. So yeah, I mean big big difference. Way more reachable audience today than ever before and reachable with our kind of skill sets. So yeah. Okay. So Snack Bar Studio, you're you're a co-founder in it. It's an indie game studio. How have you how have you done marketing for it? Like I want because I want to learn I want to know

16:32how Rand Fishkin like how Rand Fishkin operates with different marketing campaigns. You start companies, how are you marketing them? Because again I want to repeat this like I really think you are one of the greatest marketing minds and I'm not just like I'm I'm not just trying to like I have learned so much from you and there's a reason for that and I'm pretty good at picking out different marketing people. So I'm really curious how you do marketing campaigns and what you do for new sites, what how you build them up, what you do for for new for new companies, all that stuff. Yeah. Well, okay. So, in answer

17:01to that last question, it's different for every different company. Uh, and Snack Bar Studio is going to be extremely different from Moz or Spark Tour or anything I've done before, right? So, so Moz was very SEO and content flywheel, right? Like we produce content, we put it on our website, then we try and rank better for that content because we were in the SEO field. It made us look good. When we ranked well in Google, and when people found us, they thought we were an authority, and then they came to the site and subscribed, and then they signed up for the free trial, blah blah blah blah

17:29blah. With Spark Toro, I recognized that the zero-click problem was such a huge huge issue and I was trying to build something outside of the SEO space. And so we went heavily into really three things. One is press and PR like trying to get the sources of influence that our audience already read and subscribed to to cover us, to write about us, to talk about us favorably, getting speakers on stages at conferences to talk about us and show screenshots of what we were doing. The second one was uh social media marketing. So at at the time it was very heavily Twitter and LinkedIn, a little bit of YouTube and Reddit, a few

18:07other channels. Now now that Twitter is mostly dead and mostly chat bots, um it is uh a little more threads, blue sky, LinkedIn mixed in there. And uh the third one is is for sure like a a content email play. So we do we have a very popular newsletter. I think about 40 50,000 subscribers to that newsletter. Every time Amanda sends that out over the next like few days, you can see a spike in people forwarding that and and coming to our website and signing up for the um the free and paid versions of the product. So that those email newsletter subscriptions are just huge. I still think email is a great

18:47channel. For Snack Bar Studio, I've done almost no marketing. The only marketing I've done so far is about the company itself rather than the game and that is intentional. Um, so I did a big datadriven study about um, games. I I actually did it with a couple of software engineering friends here in Seattle. We looked at I think about 80,000 games that have come out on Steam in the last 12 15 years. Uh, and we basically looked at how well they performed. Many many games as you're probably aware announce their sales. You know the publisher will say like you know on their Twitter on their Discord

19:29or on YouTube, hey we crossed a million sales, we crossed 50,000 sales. We sold 5,000 copies of this game. We only sold 200 copies of this game. Like you know whatever it is. And so we basically indexed those announcements, built up a huge repository of them, built a machine learning model to um correlate announcements with metrics, publicly available metrics we could find like the score that the game received from Metacritic, the rating on Steam, the number of reviews per week on average, all these kinds of metrics, and then built a regression model that that had a nice correlation with actual sales. And then we modeled and looked at all the

20:09games that that sort of, you know, uh, fit particular paradigms and the tags and genres they were in. And that's what helped me select cooking, mid-century Italy for the setting. Um, uh, sort of chill action RPG as the as the style of the game. A bunch of different things about the, uh, structure of what the game was going to be. And then we looked at how marketing sort of affected these things. And and the discovery was marketing more than a year before a game comes out is not just ineffective, it can work negatively against you. Um, and that's because basically the people who would be most excited about the game,

20:54like it's in the perfect genre for them. It's like exactly the thing they want to see, they get tired of waiting. And then when the game comes out, they're sort of they're like over it, right? They're like, "Oh man, that old thing like whatever." So, we have done no marketing for Snack Bar, the the video game itself. Um, yet because we're still about 2 years away from launch, we think maybe 18 months, but that is going to happen. You're not going to see much of anything on the marketing front until between 9 and 12 months before the game comes out in early access. Um, and then

21:29I, you know, I'm happy to talk about like what I think is going to work in that sector, but I think that's probably less useful than having me back on in two years and I can tell you what did or didn't work. I'm s I'm super interested in gaming. So, we me and my friends, we

21:43made the first video game in crypto. This was in 2018. We we made Pokemon on the blockchain. It was really successful and it was it was it was crazy. I learned so much about the game industry. We had we had this amazing uh this this amazing musician who did our OST. It's the OST is still on Spotify. It's like yeah, it was it was so great. Um and okay, but I'm very I'm curious about your data driven approach because to me

22:08that's so cool. So aside from like the okay don't start don't start talking about it before like before uh one year like before that what were you learning about like different announcements and announcing games as well was it like a style of writing or words that worked phrases? Yeah. Uh so let's see. The the generally most successful uh forms of marketing are when you show off what would I call it? um unique features, especially visuals and and things that sort of surprise people and attract them in emotional ways to uh not I would say not a game's unfortunately narrative or um its sort

23:04of uh visuals outside of game play. Gamers, especially over the last like four or five years, have become extremely immunized to sexy trailers that don't actually show gameplay. And they really want to see what is it like to play the game. GTA 6. Yeah. Yeah. Yeah. I mean, look, if you're one of these big studios and you have a giant title that already has millions of fans and, you know, 100 million people who are familiar with the game universe, the rules are different. No, no, but I'm saying to Mor that was a big complaint of the recent trailer was that they don't show gameplay. Yeah. Yeah. And I

23:42think I think that's a a really fair complaint, but also that complaint exists because millions of people watch that trailer. If you're an indie studio with no history, you can't get millions of people to complain about your trailer. Like, nobody's even going to watch it. So the things that we've seen that are really successful are extremely short form video visual content that shows gameplay that uh is surprising, emotionally triggering. Um that that emotion could be, you know, for horror games it's often like fear, surprise. Um for uh a for some narrative games it's um the text content and how interesting that can be. Uh, a recent game that that

24:26did a really nice job with this is called Esoteric Eb. They did a great job of showing off their like weird unique writing and how it fit with with a particular genre that's that's sort of popular in that space right now. Um, you can see this with a a little bit with the visual novel style games. And for games like ours, it's a lot of what games like Hades or Call to the Lamb or um you know those Nolita or uh even even Vampire Survivor whatever like those type of games show off this sort of sexy crazy um unique visual action in the game play itself and that's what that's what we've

25:08seen be successful. I I also would recommend anybody who's deeply interested in this topic uh check out a newsletter called Game Discover Co. from Simon Carlos. I I actually consulted with him a couple years ago on the the project that I'm talking about and on the game. Um I think his his insights and his data on this is is absolutely topnotch. How did you by the way Hades

25:31was excellent. That was my COVID game. That was like I was sick with CO. I just laid in bed playing Hades. How did you get the the data to to like figure out what works and what doesn't? And is this the type of thing that you can recreate for any brand? Um, yes. Was it very manual or was it

25:49like Yeah, it it was a little manual. Yeah. So, I would say I think most of our most of our data collection came from just googling around for games. So, we built an automated system that essentially, you know, for each of the 80,000 titles or whatever, we we filtered out the ones that basically had almost no reviews, which which brought the list down to like 5,000 or something. And then for each of those, we did a a Google search, not manually, but like an automated Google search for a variety of um things. So, that the name might be something like Hades Game Sales or Hades Game Sales announcement

26:27or that kind of thing. And then we would look for hey did the publisher announce the number of game sales on a particular date. Then we would record that in a database here's the date here's the game here's the you know ID and here's the number of sales that they announced passed on a certain day and then we would look at the okay how many reviews did they have on that day as oo you know uh leading up to that time etc etc. So,

26:50it's buildable. Anybody can build it. The data is all public. It's on the internet or in Steam. Steam has a very nice API that you can call for this stuff. Um, and it's very crawable, too. The application to other businesses and other sectors is quite challenging. But I would say one of the things that that we use at Spark Turo, right, Spark Toro does exactly this. We we try and look at people who visit this website also visit these websites, also listen to these podcasts, also subscribe to these subreddits, also listen to these, you know, uh um YouTube channels, podcasts, whatever. And those pieces of data come

27:27from clickstream data. Clickstream data is not it's sort of technically public, but you have to buy it from a provider. Um we use Datos which uh was bought by Semrush a couple years ago and Datos essentially gives us a you know clickstream for 10 million plus devices. Uh so this is like you know phones and desktops and laptops and then every URL that is visited in a browser they capture they aggregate those they anonymize it they remove any PII and then they essentially sell the aggregation. So, they're not telling us,

28:04oh, Edward visited these websites. They're telling us, hey, here's, you know, 10,000 people visited or 10,000 people in the panel visited this site and of those 10,000, 3,000 visited this site and 2,000 visited this site and 500 visited this site and etc., etc. So, that's that's essentially how we build our data set uh behind Spark Toro. And it it kind of works the same way that this game project that I was talking about, right? So like we're trying to figure out what's successful in games marketing but Spark Tour is trying to figure out what does an audience do who visits a particular site or who has a

28:43you know particular word in their profile or any way you can describe an audience. We try and provide that same kind of data. So for snack bar your main data you were just getting it from dos and from uh and from steam. Sorry. Uh, we didn't use Datos for the Snack Bar project. We used Datos for Spark Toro. Okay. How are How are you How are you crawling and and getting all of that for

29:10for Snack? Uh, for Snack bot. Yep. Exactly. Okay. Yeah. Just built a bot. Um, I think we used one of the APIs that crawls Google to get the automated Google searches. So there's like uh SER API, SER, a few other ones. There's I think there's five or six different competitors in that space that crawl Google for you. Were you only looking at the at the first announcements for each game or were you looking at like a bunch

29:37of different announcements? Yeah. Yeah. We would we I think we had on average five or six announcements per game where we knew sales. Um because many you know if a game is using sales to sort of promote themselves they usually do it many times right they'll be like Assassin's Creed Mirage crossed 500,000 crossed a million crossed 2 million cross 5 million crossed 10 million like you know Ubisoft likes to share those numbers because they they create um both demand in the market and also help their stock and make people excited about the company that kind of thing. And you said

30:13I think you said I think you said this. It was just Twitter that you were looking at for announcements. That was No. Oh, it was it was a bunch. No. And announcements come from all over. We We found a healthy number of announcements on Masttodon. There's It turns out there's a lot of indie game developers who like publish their stats on

30:29Mastadon. What What year was this? We beginning of last year. Oh, okay. Okay. I see. Yeah. Yeah. Sure. Maybe maybe two years ago, actually. Dude, I I gotta say that is your approach to like figuring out what works with games and like that is so brilliant. That's why that's why that's why you're my favorite. That's that's like No, for real. Like I mean, but this is, you know, this is like my playbook, right? Like I I like to make big data studies and then like look at correlations and I don't know, I just opinions are fine. Like I don't have any problem with opinions. And there's

31:04people who are more successful than me who are fully opinion based. Like they they don't use data. But I just like to have the numbers, right? To be able to prove to myself like, "Hey, before I make a, you know, make a judgment, let me make sure that it's it's backed up by something." There's a rare, very, very small amount of people who have the most incredible intuition and who are like basically oracles. You can ask them any question and they I have like I have three friends like that. My friend Ram,

31:31my friend Solomon, and my friend Yuri. And they are just it they are right about everything. It's it's it's crazy. I am not. And I am the type of person who needs to learn from these from those people or from or from data. And I am like proven wrong all the time. All all the time. And uh I mean there's in the history of humanity there's no one who's right about everything. It just like it doesn't happen otherwise. No no they're not right. I'm exaggerating the stock market, right? Yeah. Yeah. No, I'm completely exaggerating. But they are like it their intuition is just it's shockingly it's it's shockingly good. I

32:07I don't want to discount intuition like like we were talking about, right? I I I want to be clear that there are people who don't use my approach to what you know, whatever they're building, whatever they're creating. They use an intuition based approach. Some of them have been extremely successful, far more successful, you know, on paper, whatever, than than I've been uh with my companies. And I think I think that's great. I just like the data driven approach. I think one of the reasons I like it is because it's justifiable. Like something that I always hated, Edward, I hate I'll tell you like even as a kid I didn't like it

32:38when my parents did this which was the like you know because I said so I wanted a reason right like you want logic andre doesn't want to be told no I mean I'm happy to be told no I think for for example why should I not go 80 miles an hour on the highway because the risk of death every mile hour over 60 goes up so dramatically like it's just ridiculous I want to be told no like Speed limits are a great thing. Tell me about that. Vaccines, awesome. Like, oh my god, tell me to get a damn vaccine and tell me which one and tell me when

33:14because can save my life and prevent me from getting shingles, which is like the worst, most painful thing. Friend of ours just got it. Um, anyway, so like I love the rules. I absolutely, but I also want justification. Like, show me the numbers, right? I want to see that 10,000 people took this clinical trial. They took the shingles vaccine. when they got it, their reported pain was so much lower and that how long it lasted was lower, blah blah blah blah blah. I don't want to see the like, hey, rub this, you know, compound from a beetle

33:42on your chest. Like, whoa, whoa, whoa. I'm not I'm not into that. Yeah. Yeah, I'm Yeah, I get that completely. You want to be Yeah. Just there has to be reason behind it. I'm the same I'm the same way. I'm I'm like happy proven wrong. Yeah. Oh, yeah. Yeah, I love I love being proven wrong, right? like you want to you want to show me some new data just you know the thing that uh I think is really interesting for marketers especially one of the things that's fascinating is like you might be listening to this conversation between the two of us and like yeah data is

34:15really convincing I like data I you know these strong correlation studies with real numbers statistically significant like I understand stats and math blah blah blah blah blah blah that resonates with me you're in the minority if that message resonates with you the way it resonates with you and you're in the minority. Most people are not convinced by stats. For for example, uh I was having this conversation with some friends the other day and they were like, "Oh man, Chicago is really

34:43dangerous." I I had this like, is it? Let's let's go look it up. Well, statistically, you know what? Seattle in the 80s and 90s, every day of those two decades was like five times more dangerous than Chicago is today. and you grew up here and like you don't think Seattle was dangerous when you were a kid, right? So why does Chicago seem dangerous? But there like but I know this guy and this guy was a police officer in Chicago and he said this terrible thing happened. That's an anecdote. It's a story. It it might be a true story. I I I think it is a true

35:16story, but in the grand scheme of things that does not make Chicago statistically dangerous, right? Highway driving way more dangerous than Chicago. And and I think this is a really hard message for people to gro and understand. And as a marketer, you should be careful about learning what your audience resonates with. It might not be data. Like you can use data to inform your own opinion if that's what resonates with you. But if you're making a message, you know, that's going to appear on your landing page or homepage, the email that people are going to open, the content that you're going to create, the social media

35:52post, data and statistics often is not it. anecdotes, stories, emotions, things that like trigger people's fear, surprise, happiness, joy, um even anger. Those things often are more effective at changing minds and nudging behavior than data is, which is infuriating. And like I hate that about humanity, but also we live in the real world, so let's recognize it. Well, I make this show for people who like data, so you're you're in the you're in the right you're in the right place. But I think that's a it's a good lesson for all of them, right? To be like not everyone is like you. Oh yeah. Oh yeah. I mean I know so many

36:34people who are not and it's uh there are a lot but but then if you're when you're obsessed with data then you also have to be obsessed with questioning the source of the data and going really deep into where it came from. Yeah. Yeah. Yeah. I mean like I'll tell you um good example of this. So Spark Toro I mentioned uses Datos which is like our clickstream data provider but occasionally we see something funny in the data and we're like this doesn't doesn't feel right like something's off here and and this is why I love digging into it and I like doing big studies based on it because occasionally we'll

37:08find something funny and I'm like oh what's going on here? Oh okay got it. Datos has this like particular bias. How do we control for that bias and like present the data accurately? Um you know it's just how panels are built. panels have biases. I mean, a good example of this that almost everyone's familiar with is um elections, right? So when when you have a a political election, you know, the United States had an election last year. It was very well covered worldwide and um most folks are familiar with the fact that like the polling data, which you know, polls are usually based on a few thousand people

37:47and the polls were statistically uh quite quite accurate, but they were off by around 1.5% 2%. right in in favor of um the Democratic candidate here in the US uh Kla Harris. And then you know come election day it turned out like hey that 1.5% 2% that the polls were off swung the election in the direction of Trump even though like you know sort of statistically it looked like Kla was going to win from your polling Trump won. And this is the same thing that happened in 2016, the first time Trump won, right? Like polls statistically undercounted him, I think that time was like 3 or 4%. Um, and it's a really good

38:28thing to recognize like that doesn't mean polls are wrong, right? Polls were correct. Like polls got it so close, two or 3%. That's incredible. Like that's so damn close. That's that's the kind of data reliability that I would love to have in all my products, right? But when it when it when you're talking about a really close like you know 47% to 48% type of thing obviously means the difference between one candidate and another. And so I like to be aware of those what those differences are and where there might be polling error or or statistical sampling error. And so we spend a lot of time in the raw numbers

39:09that Datos gives us trying to trying to identify what those are so we can control for it. Yeah. Yeah. Really important to look at that stuff. I want to I want to go um back to like what your really So this is just a personal question. What got you so interested in SEO when you started when you started Moz and and and like and then grew it and then what got you really interested in in audience data because now you're doing Spark Toro? Yeah. uh with Moz. It was uh basically we were running a a web design consultancy, you know, back in like 99 to 0203. And our customers started asking

39:51us, our clients were like, "Hey, when you build the website, like we want to show up in at the time it was like Dogpile, Infosek, Glycos, you know, um MSN search was big back then, Yahoo, and and this nent one, Google, right, that was that was climbing the charts. Um, and we basically hired subcontractors to do SEO for our clients. But then we we were so broke. We were running out of money. We couldn't pay our subcontractors. And so I had to start doing the work myself. And I just found I had a a prediliction for it, right? I I like doing the work. It was

40:25interesting to me. I I like the like sort of game of trying to get, you know, rankings and links and do the right content, do the right keywords, like that matching the algorithm. And that was really fun for me. Um, and I found I was good at explaining it to other people. Probably even better than I was at actually doing the work was I was really good at explaining how SEO worked to other people. And and that's what built the SEO MA blog and then the the Moz subscription, you know, sort of followed from that. Um with Spark Toro, my frustration, Edward, was that um even even the last few years I was at

41:01Moz, I found that there were many businesses that I was in contact with who essentially like they couldn't realistically build their business exclusively through ranking in Google. And I think that's happened, as we talked about earlier on the show, that's happened to more and more companies, right? where it's like, hey, you can't just rely on Google. That that can't be your only channel. Uh especially with the rise of zero click

41:28searches and all this kind of stuff. Like, you got to be in more places. And there was there was no way I could help these companies understand where their audiences actually were and prove it to them. Um and I was, you know, we were quickly able to see how survey data was terrible for this. Like, you can't ask

41:45people, "How did you hear about us?" It's like, it's the worst question. nobody can answer you consistently or honestly or or fully. Um they don't even know how they found you. Um and so yeah, I wanted this I wanted this product that could help solve that problem and um that was the genesis of Spark Toro. What I you're like you should do other types of marketing. That was so so as I said like that was one of the things that you that you taught when you were still working at when you were still at MOS and that was like such a valuable lesson which is be a well-rounded marketer

42:24because if you want SEO results that's also really going to improve your SEO results. There you go. That's that's exactly I mean look Google I think if you and I were designing a search engine one of the most foolish things that we could say would be hey let's only ever pay attention to exactly the ranking factors that sort of SEOs would try and uh manipulate to rank in our search engine that's just plain dumb right um and now you can see you know with with large language models and and the rise of AI that they are using just this vast variety of signals that is essentially

43:04how commonly referenced are you as the sort of brand or authority on a topic uh not just across the web but in books in magazines and radio and television and sort of the offline world and what people are talking about what they're saying on social media e even just what are people typing into their phones and like emailing each other and right like all these different sources are being used as um training data for large language models. I think uh is it Facebook is in a lawsuit right now because it it trained its LLM on um on books without paying for them, including Lost and Founder, by the way, including

43:46my book and Geraldine's book. Wow. Wow. Yeah. And it's it's one of those frustrating things where you're like, "What? You you get to do that for free? You don't have to pay anybody. You can just like steal it essentially." And they're like, and they have like, you know, the DOJ, I think, has an email chain between Facebook exacts that's like, hey, this is super illegal. And they're like,

44:04yeah, just do it anyway. Wow. Wow. It gives you this sense that like if you are trying to do marketing effectively through one particular channel and you are not paying attention to the fact that influence is created in a massively diverse set of channels and sources, you're going to have a bad time. I think the best the best brands, the best marketers, the best entrepreneurs realize I need to be present in the places where my audience pays attention and and do the the right work to be in those places rather than, you know, hey, I'm really good at Twitter. I'm just going to stick to Twitter. That's a tough one. And and

44:48Spark Turo shows shows those places where your audience is. you I mean look because I I think it was insane it was insanely frustrating to me that there wasn't a like an inexpensive product where I could just go and say here's who my audience is tell me where they are like tell me what they pay attention to that you know I I Casey and I had this idea in 2018 that it's like it can't be that hard like that data must be available somehow it is quite difficult to get after after you know, working on

45:23this problem for the last seven years. It's hard, but it is solvable and I think that's a a superpower that Spark Toro can deliver and I don't know, I hope I hope it's something more people invest in. So, like you said, stories get people to take action, stories get people to move. What are some like crazy customer success stories for Spark Toro where customers like they just like getting this audience insight data helped them so much and how did it help them and what results did they get and where were they before? I'm sure you

45:54have like tons of these. Yeah, a few. Um, okay. One of the most fun ones, uh, this is from a few years ago, but a um a TV show contacted us, like basically the um whatever the the marketing product team for this particular television show. I I'm not allowed to share which one it is, so they didn't give me permission for that, but it was a a TV show that essentially was getting cancelled on one of the major streaming services, and they were trying to get

46:22picked up by other streaming services. And so they essentially used Spark Toro data to get demographic and behavioral information about people who watched that show and subscribed to kind of the um that whatever social channels that like that show was producing. Um, and then they use that data in their presentation pitch to pitch and win a contract with another streaming service because they were able to show that that new streaming service that they were reaching an audience that that streaming service really wanted to reach that they didn't have subscribers in yet. Um, the show was like a was kind of sci-fi uh little like like historical sci-fi, but

47:06um that essentially they didn't have that audience. this this streaming service that picked them up. And so it was a big win for that show to be able to get uncancled and use Spark Toro data to make the pitch. And I just I love that story. Like it's uh it's so fun. I'll um yeah, probably probably keep that one in mind for for the rest of my life. I just I I love that we helped a

47:30show get uncancled. And and you use it I mean you use it you use it to launch Spark Toro. You you were like Oh, sure. Yeah, of course. Yeah. Yeah. Yeah. I mean, I think a lot of people use us for specifically for launches either not just their company, but oftentimes like a product launch or even a piece of content, right? So, a lot of agencies will use us to um make a pitch for, hey, here's where we're going to promote our content. Here's the topics that we're going to focus on, right? Because Spark Tour gives topics of interest, search terms, that kind of stuff. Um, and then here's how we're

48:04going to position it to reach these audiences because we know that this audience pays attention to these kinds of YouTube channels, these kinds of podcasts, these kinds of subreddits, these kinds of social accounts. And so, we're going to use the same language and emotions and like style that's been successful for other content launches with these creators when we launch our own thing. Actually that was so before I yeah you before I say so it was very clear how was how it would be useful if you're doing like outbound sales or if you're getting onboarded onto a company you want you want to understand that or you're an agency all that stuff but you

48:39know I love content marketing one of the reasons I love content marketing is because you made so much content about content marketing and I I guess I wanted to hear more about how Spark Toro is used for content marketing and I guess you're saying like Spark Toro shows the content in your niche that is successful So you can just use the style and the themes and the topics. Yeah. So we we do a few things on this front, right? So like you can see specific pages that perform well with a particular audience um especially in search. So you know if you if you go to Sparkurro and you

49:14search for an audience like I don't know landscape designers in Canada or something, right? and we'll we'll show like hey these pages come up most frequently uh in all the search terms that landscape designers search for and so you can get a sense of like oh right like this type of content tends to be successful then you can also go over to the topic section and you can get um basically topics of interest to those people what I like doing I like creating stuff that people haven't invested in before so I looking at that list of sort of topics of interest that landscape designers in Canada might have and then

49:56the ones that have been successful and be like hang on there's a hole here like this this is missing this is no one's no one's doing this one let's make let's make some of that content I think it also pays to probably invest in the the sort of highest affinity topics for an audience um but yeah that can that can super help and then once you make it Spark Toro also helps with where should

50:19I promote it to whom and how, right? Like what what subreddit should I be participating in and sort of replying to people on? What um YouTube channel should I be pitching? Which podcast should I be trying to go on? What press sources should I reach out to for some like PR coverage? Uh where should I write a guest post or like submit an article to? um what websites where if I could, you know, maybe I should put my piece of content on this website rather than my own website or in addition to my own website to get additional um coverage and attention and rank in the LLMs and get an AI overview and all

50:57that kind of stuff. So, yeah, it's it's really um I think it solves a bunch of problems that classic like SEO focused tools would would not help you with um and aren't designed to. Yeah. Yeah. when you when you used that data to launch Spark Toro, were you using were using a PR agency armed with that data or were you

51:17doing the outreach yourself? I found that this is not true for everybody, but like because I'm a person who's known in the marketing field, when I reach out to a publication in marketing universe, the the response rate is much higher, right? And so that's kind of a a cheat code that I use to be able to get lots of coverage for Spark Toro because I I wanted to ask that makes a lot of sense because I wanted to ask if like someone doesn't have a huge budget but they're like oh my god this is super cool I want to do this like should they be using Spark

51:48Toro and how if they don't if they're like if they're bootstrapped Yeah. Yeah. Yeah. Yeah. I mean, I I think the answer is almost always it pays to at least have the data, like to know here's where our audience is. Even if the pitches are really hard and you're not having success early on, people are turning you down, it's okay, right? Like you can over time grow that audience, get better at the pitching, hire a PR firm if you want, hire a marketing person who's going to do those pitches for you. Um, participate in places where you where there's no gatekeeping. Social media is especially great for this. Um, YouTube

52:30comments are pretty good for this to be honest. I don't know if you've played around, but there's lots of sectors where like the YouTube video is on one topic and then you'll see like a highly upvoted comment that has some additional useful stuff from someone and you know it's not clear always like hey go buy our product like that you know that tends to get you banned. Same with Reddit. But helpful consistency and participation in communities where people pay attention often gets you followed, shared, liked, trusted. Uh

53:03that's that's worked really well for me. I I participate in a lot of Reddits, like in a lot of subreddits because um and I I almost never refer to Spark Toro like or Snackbar, right? I'm just trying to like hype people up and help them and Yeah. Yeah. Are you commenting as your name as Ran Fishkin? So then Okay, that makes sense. Yeah. Yeah. You can't you can't be anonymous. You have to like you have people need to be able to Google be or actually what people do on Reddit is they is they pin something to their personal profile and then someone goes who is this? They click on on the

53:37Redditor and then they see that pin right there on Reddit. I and I guess I guess it would kind of be the same thing on YouTube. Who was this person who keeps commenting? Oh, look at this. Look at their look at their channel trailer. It's this this thing. And same thing is true on Twitter, on LinkedIn, on uh Facebook, on threads, on blue sky, you know, all all those places. People have the same behavior. They like see someone a few times and they're like, "Who who is this? This this person sounds smart

54:05on this topic." Click. Yeah. Yeah. Yeah. That makes a lot of sense. I want to go back I want to go back to SEO. I I I'd be like so remiss if if I didn't mention this. So I was I guess I was like one of the first people to cover when you did the Google API leak which was basically like a one year exactly ago. I think it was in May. Oh yeah. Yeah. So So you dropped that man. That was Let me tell you what. Well, let me tell you what happened. So crazy. So like uh I was I was on I was on Twitter and Twitter had

54:37it I had just started doing like AI summaries of things that were trending and and I look at that to get news to make content on and like I see I see Google is in there just in the general trending. Oh my god the biggest Google leak of all time is there and I just took a screenshot of my laptop with my phone and put it as my IG story and then that started going viral and I'm like I need to like people want to hear about this. So, I literally just and my podcast editing skills were not what they are now. Like for short episodes

55:09now, there's like tons of tons of cutaways and graphics and all of this stuff. Back then, I I wasn't doing any of that. I was just recording myself like speaking or reading something. So, I literally just read your article on the Google API leak and that me just sitting in a chair reading it going, "Oh my god." like reacting that alone got like 30,000 views and I made so much content on it after and everyone was so interesting and it was so what you showed about how about how uh clicks in the SERs are are used and engagement is used and you always predicted that and I

55:46this is something that I said in the in the podcast over and over again which is like if you've been doing SEO for a while it's very clear that engagement is important engagement on all levels is important. It's like very clear. You know, I had on uh Bill Hartzer who was a pretty well-known SEO. I I was a couple months ago and something that I asked him. I'm like, "What do you think is more important? Is it is it backlinks or engagement?" This is like one of my favorite favorite questions because you will see a lot of SEOs who want to die

56:17on the hill and say that it's backlinks. But I think I totally I I Yeah. Still you still you you still see that and I and I'm it's so clear to me that it's like you could get all these backlinks get get to I don't want to but if you have my understanding is during the department of justice trial right was it you know they they had like googlers um submitting sort of testimony right where they they swear to tell the truth and the whole truth and and they can be thrown in jail and fined right if they

56:48lie and the the the Google person right? The Google spokesperson, I can't remember who it was, one of the engineers that they subpoenaed said that uh engagement essentially the the look at clicks and and um bounce bounces back to the SER and all that kind of stuff. Basically, that was the number one signal, right? And that uh I I want to say links were in the top five, but not the top three. So yeah, uh I can't remember exactly what the testimony, but but engagement was the number one signal according to sworn in subpoenaed testimony. Could that person be lying? I

57:23guess so, but it seems unlikely, right? I mean, just given both of our experience, like it's it's just so clear that it's engagement. It's like you make a better page and the number one position could have all these links, but if your page is outperforming, if your retention is higher, if people are clicking through to other pages and they're not pogo sticking, they're not going back to the SERs, they're not going back to the search engine results page, like that's the best thing. I I just can't I cannot imagine a world in which you design an algorithm that says, "Hey, if this thing has more links, I don't care whether

58:01people like it or not. Show it at the top." You'd be insane to design an algorithm like that. Insane. One of one of the things that I love about SEO is that like when you understand Google's algorithm, you understand all the other algorithms. Like you understand the Tik Tok algorithm, you understand the Instagram algorithm, you understand YouTube, you understand all of it. Like it's Yeah. I I mean I think Yeah. I think Tik Tok is a unique one in a way because Well, and social is a little bit different. Like the one thing I'd say that's really different from social and search is search's goal is to get people

58:40the answer as fast as possible because Google knows that the faster they give at least you know accurate information the more likely it is someone will perform another search and come back to them and trust them and you know search again and again and again. Whereas Tik Tok's goal is just raw engagement. Like how long can we keep people on the app scrolling through more things coming

59:04back more frequently create addiction? And and addiction is different from answering a query. And so the optimization for addiction, which is what Facebook and YouTube and Reddit and Tik Tok and all these other ones do, is just it's a different incentive structure. And so I think that that's the that's the core difference between those two. And I I I I do think there's real differences in how they design what shows at the top of a social feed versus what ranks at the top of a search engine because of those incentives. Yeah, that makes a lot of sense. But it's it's I I just I'm trying to on the spot

59:43think of like all the correlations that I see where it's like uh I mean practice the practice of being able to look at you know I think um NBA players have this right where they like when they when they play a game of like street basketball or something they I can't remember the name of it but but NBA players have this thing where they they like can clock someone's skill immediately like in a few seconds, they can say they can like accurately estimate how how great is this person at like, you know, running from the break, how good are they at like getting around a certain kind of

1:00:18defense, how good are they a three-point shooter, a free throw shooter, like everything, right? Like they can just clock that instantly. And I think professional SEOs, I had this certainly back in my day, could like look at a set of Google search results and be like, that's why this ranks here, that's why this ranks here, that's why this ranks here, that's why, you know, and and I think the same is true of social media marketing. You can sort of look at a feed and be like, I got it. I know exactly what's going on, right? And be you might be a little off, but you get

1:00:46this like sixth sense, the the the like Spider-Man Spidey sense of why an algorithm works. And I I think you're right that applies, you know, if you learn that in SEO, you can apply it to social. If you learn it in social, I see it less frequently for whatever reason applied to search. I don't know why. I have some theories, but I think they would offend people, so I can't say those theories. But if you look at like I mean if you look at the type of people that a lot of SEOs are and you see it in the comments on social media, you see it

1:01:15on you like it's like it is crazy to me some of the things that people don't understand in SEO that to us are are so are so uh what's common sense are are so like of so of course obvious um within you know what with engagement the way that I I guess I would relate it is it's it's uh p okay so like um pogo sticking in in Google going back to the search results is the same thing as swiping up and looking for another video. That's like that's that's how I relate it. And not pogo sticking is the same thing as actually watching a video through yeah

1:01:54backlinking is the same thing as a share because a share is the strongest factor in a in a social I think I think now uh most of the social platforms have switched from sharing to commenting. So, if you look at the um you know when Elon bought Twitter, he like shared the the ranking system for how Twitter shows things in the feed. And basically a a comment was worth I can't remember if it was it was something like 33 times the weight of a of a retweet. So like if someone comments on a tweet, that is the strongest signal by far. That was that was old Twitter or that's the new the

1:02:31new Twitter. Okay. Yeah. Um, and uh, LinkedIn it works the same way. We've done a ton of testing. Someone sharing your post, don't get me wrong, very kind, like great, especially if they use a comment, uh, or, you know, if they they sort of repost it to their own feed, but not nearly as strong in terms of reaching more people for

1:02:54impressions as if they leave a comment. I wonder why that is. Like if I'm thinking about because what I do all the time is with any platform I think about things from the perspective of the person designing the algorithm. I love algorithms. I make them myself. Like they're very interesting to me. I wonder I like wonder why a comment is more valuable than a share. I think the only um the only explanation my wife is going to sneak in here because we're supposed to be done 15 minutes ago. Sorry. Hi. Uh

1:03:24we'll end we'll end in a second. Yeah. Do I need to run to something? No, I had a question for you. Okay, keep going. All right. Um, this the last one. Last one. Last one and then we'll wrap. Last one and then we'll wrap. Thank you. Thanks for your time, Ren. Bye. Bye. Uh, sorry, Edward, can you remind me what was the what was the question? Why are why do you think comments are more

1:03:45important than shares? Oh, yeah. Yeah. Um because uh a if you're designing the algorithm and you see that a comment is a stronger predictor of someone coming back to the app, opening it again, subscribing to the channel, um like like engaging more versus a share, then you're going to bias your uh algorithm to showing things that

1:04:13will incite or induce comments more. more so than sharing. That is my that is my theory and I would suspect quite strongly especially for the micro blogging platforms. So like threads or blue sky or Twitter or mastadon um although mastadon and blue sky are are less algorithmic. Uh I would guess that a lot of the repeat engagement like their signal that someone is going to start using the app and then come back to the app happens from did they comment. That makes a lot of sense because people I people could be sharing because they're angry, but I think they're probably less likely. They'll still comment if they're angry, but

1:04:54maybe it's they're less likely to comment. Retweet is like a, "Oh, this is cool. Other people should see it." Comment is like, "You've hooked me." Ah, yeah. Yeah. Yeah. Rand, thank you so much for your time. You're like, you're an inspiration to me. I learned so much from you and it's absolute pleasure to have you on the show. Everyone, check out Spark Toro. Is there any anywhere else that you want people to go and and find your stuff, find you? I mean, if people are particularly interested in uh in Snackbar and seeing what we're up to, snackarstudio.com um is where you can go. There's there's an email list you can sign up for, but

1:05:26we won't send any emails for a long time. Uh and I'm most active on LinkedIn and Blue Sky, so you can find me in those places. Edward, thank you so much for having me. I really appreciate it, man. Thank you. Stick around after this ends because a video will just have to upload. And uh yeah, so for everybody watching or listening, this is episode

1:05:45674 of my daily digital marketing show. 674. I'm almost losing count days in a row doing the show since I started. Haven't missed a single day. And everybody, I will talk to you again tomorrow. Bye now.