Dark Social and Attribution Challenges with Rand Fishkin

James Lawrence: I'm here today with Rand Fishkin. Rand, welcome to the Pod. 

Rand Fishkin: Yeah, thanks for having me. James Good to be here. 

James Lawrence: It's great to have you on here. I think most marketers in Australia know of Rand, but if you're in a very small percentage that don't; Rand is Co-Founder and CEO of Spark Toro, makers of very, very fine audience research software. In 2004, Rand created and launched Moz Blog, which morphed into Moz, almost became the most popular community in content resource for search marketers around the world. At the time you were the CEO of Moz you grew up to 130 employees, $30 million of revenue, over 30 million visitors per year. You've been pretty much a key player in the industry for the last 20 years. I've been fortunate enough to have been at a few conferences around the world where you've kind of taken center stage and informed marketers as to the big trends that we need to be aware of. So have to say your content is always in my top 2 or 3 thought leaders around the world. When I started the pod, I said, Rand's the guy I want to get on here. So it's it's great to have you here. Thank you for your time. 

Rand Fishkin: Oh, thrilled to be joining you, James. 

James Lawrence: I thought I'd start with a very general question, which is maybe a harder one to answer, but just macro landscape at the moment. It's obviously been such a big kind of 12 months in digital; generative AI, changes with Google... I mean, what are the big things that you kind of feel as marketers we need to be across? 

Rand Fishkin: Well, to be honest, I say this to a lot of marketers, which is what's happening in the world is generally less important than what's happening in your world. So if you are, you know, an e-commerce B2C marketer who's very reliant on advertising channels like TikTok and Instagram and direct consumer purchases, low purchase price, low consideration process; I'm not sure that there's a ton that's changed. I would say certainly there's some attribution weirdness that's going to be going on with with GA4, whose default models are going to over attribute certain advertising channels. Which is which is frustrating. There's some big changes around privacy, third party cookies going away next year and changes to how the ad platforms are going to operate. That will mean that you have less 1 to 1 matching of which consumers saw which ad and made which purchase. Instead you'll get these sort of federated lists of groups of people who, you know, this group of people experience this these ads and then they had these behaviors. But we can't tell you which one had which one ad view. So that's going to be changing. But I'm not sure it modifies a ton of what your process should be. 

Rand Fishkin: I think that broadly speaking, these big changes around privacy and advertising, third party cookies, what I've been calling the death of attribution, although maybe it never really existed in the accuracy form that we think it did. Yes, that stuff will have an impact on almost every marketer, but not, not even every. If you're doing SEO and keyword research and content marketing. It's probably not going to change a whole lot for you, right? Maybe it changes a bunch for your VP of marketing or your client and how they choose to invest, but for you less so. So I think this is one of the problems with being a marketer in the modern era is that you consume a lot of content that says, Oh my God, the sky is falling, everything's changing. And then you have to figure out, is my sky falling? Are all these changes actually going to affect me? And if so, how? And what do I need to do? Or can I sort of consider it as background knowledge, keep doing what I'm doing day to day, only change my strategy and my tactics in response to things that actually change my results. 

James Lawrence: I think that's a good perspective and I think it's easy as a marketer to kind of scroll through your LinkedIn feed and you couldn't possibly try to keep up with all the things that may or may not matter, right? I think it's a good segue into the piece published recently around ChatGPT and kind of usage of, I think for us down under here, I guess we'll be looking at it in two ways, right? Like the way that us marketers use tools like ChatGPT and Bard to help us do our job. But then what impact of users turning to products like that to find answers that they once would have turned to Google search for? Maybe we could talk a little bit about the piece that you did recently and kind of the usage of ChatGPT and what it's being used for, and is it kind of cannibalising Google? 

Rand Fishkin: No gosh, James, that is a missed opportunity. I'm going to make myself a note right now on my whiteboard. So go ask the data folks if they can show me the last two years of number of searches performed on Google. Yeah, and we can compare that against the traffic to OpenAI and the number of prompts that people put in, because I bet I would be willing to bet that Google has seen virtually no impact at all. Yeah, someone did do a study of whether Bing's traffic, right? Bing of course uses more of ChatGPT OpenAI. They have a partnership. I think being Microsoft is partly funding OpenAI, right? Etcetera. And the growth of Bing's market share was less than a 10th of a percent and the shrinkage of Google's was less than a 10th of a percent, at least in the United States. Which suggests that almost no one from a consumer standpoint is going to generative AI to solve problems that they used to go to Google for. But that would be a fascinating question. 

James Lawrence: Yeah, we were doing a webinar a few months back for an Australian audience and couldn't find anything - like very little data out there to suggest that that was the case. 

Rand Fishkin: So one of the articles I linked to in the ChatGPT piece was from, I think, a substack writer who noted this research around, you know, being versus Google. Yeah, ugly numbers from ChatGPT and Microsoft reveal that demand is already shrinking. So Bing's market share hasn't grown at all. This is a net article. Net actually owns my old company Moz now. And yeah, as of mid-August Bing had not grown at all. Yeah. Oh they're down 0.01%. 0.04% from January to July. 

James Lawrence: It's an interesting interplay, isn't it? And like because we're a little bit different down here, Bing has always had a larger cut through in the States than it does here. We're generally like 96-97% Google, which is the numbers are same, a little bit elusive regardless of market. 

Rand Fishkin: I think one of the smart things that Google did early on was, this is my memory, is like 2004-2006; Google really prioritised getting better results in even smaller markets, right? You know, Australian market, New Zealand, South Korea, which is a relatively big market, right? It's not a top ten countries. And Microsoft, first with what was it MSN search and then live search and then Bing. They really they tried to, like, get everything right in the United States and Canada and then go from there. And seeing as they never caught up here, they just never, never grew their market share anywhere else either. 

James Lawrence: It's really, really interesting. And I guess the rollout or integration now of generative AI into the actual SERPs is kind of the big player there, right? 

Rand Fishkin: I mean, I think it was one of their hopes. Here's the way I would think about it. If I were Microsoft, which is I’d kind of be laughing all the way to the bank, right? Because sure, my search share didn't grow, but I now have a technology and a brand. And this partnership with OpenAI have funded them. I would be feeling excellent about the fact that that is turning into a probably tens of billions of dollars a year product. And you know, if you're Microsoft, I think the time to worry about; ‘hey, are we winning or losing in search’ was 15 years ago. I think right now the question is how are we doing on market cap? How are we doing on technology as a whole? I'd be thinking much more strategically about ChatGPT versus; oh, no, I'm not winning. As a marketer, though, we care a lot about whether people are using Google or Bing or whether they're going to ChatGPT and getting answers. And here's the way I would think about that. Generative AI will make its way into some search results. 

Rand Fishkin: In some ways, yeah, maybe it will make searchers incredibly happier and more satisfied with their results, in which case Google will probably adopt it whole hog and it'll replace a ton of what we now call rich snippets or instant answers. Or maybe it will be a small piece, you know, sort of like how Google News shows up in 0.85% of search results or recipe cards show up in a lot of recipe type searches. Whatever it is, Right? So it'll be a niche inside of search that some marketers have to think about a bunch because their content applies to that and some don't. Again, I go back a little bit to the; does it affect you right now? If not, the only thing I would really be thinking about is can I get my brand name next to more mentions of the words and phrases I want to be associated with all around the web because that's the sort of equivalent of keyword research and link building for a generative AI solution, right? If you and I go to ChatGPT now and we say; hey, what are the leading CRM tools it's going to give us Salesforce and and HubSpot. And that's because so much of the content it was trained on has those brand names next to all the words, right? Because it is just a predictive text. Yeah. Or predictive token system. So if I'm trying to compete with, you know, folks in the CRM tool space, I want my CRM to be mentioned. Anytime the word CRM tool is mentioned. Yeah, get how do I get in? The New York Times? How do I get in to Wikipedia? How do I make sure anybody who's talking about it on Reddit is talking about it? How do I make sure anybody who's talking about on Twitter or on LinkedIn or Facebook or, you know, text that shows up in YouTube transcripts and comments, all of that stuff which I think from our view, feels like an extension of what good marketing has been over the last 5 to 10 years anyway, right?

James Lawrence: Which is for those listening and not watching, Randy's sipping, sipping honey lemon tea and pointing at -

Rand Fishkin: And making crazy eyes. Yes, of course. 

James Lawrence: But that's right. Google's kind of scoured the web for the last 20 years, and AI is kind of scouring text and databases and content, and it's like, what are the pointers in the right direction for the kind of the leader in this space? And looking at high quality references versus low quality references. Yeah I think that's absolutely right. I think the other question to ask is do people go to ChatGPT and search for answers to questions that lead them directly to brands they would buy from. 

Rand Fishkin: And we don't have that answer today. But I looked at thousands and thousands of ChatGPT prompts and I saw very little of that. Honestly, I don't think I saw, out of all the ones I analyzed, maybe 1 or 2 could I classify as, oh, that's the type of search I would expect someone to do at Google. And then the results that are returned by ChatGPT actually led someone to a brand that they might buy from that they wouldn't have considered previously. It's a task engine, right? It's do this thing for me, not, I am curious about where I should go to sort of buy a solution to a problem. 

James Lawrence: Yeah, that's it. Referencing an article that he published in late August on the Spark blog. And the actual thing was a huge data set, right? 

Rand Fishkin: So we looked at we looked at millions of visits or millions of devices that visited OpenAI and chat GPT specifically. And we also manually analysed thousands. I say manually, I manually analysed hundreds, and then we had thousands and thousands of full prompt series. So between 1 and 5 or 6 prompts that people had entered into a ChatGPT session and we were able to see what those prompts were and the full text of them. And then we used a classification system, ChatGPT itself actually inside Google Sheets to classify about 4100 of these 7000 ish prompts. We threw out a bunch of them that were, you know, weird short one word. Like poking around. People would put weird things in like mustache, mustache etc. And, you know, those kinds of prompts are not particularly useful, but the ones that appeared to be real prompts, which was, you know, more than half of them, we analyse those, we classified them into what was their purpose, what was the person trying to accomplish here, and then produce this big report that gives a breakdown of all of those. And you can see many of the common words and phrases that people use in their ChatGPT prompts like ‘create’ and ‘learn’ and ‘list’ and ‘build’ and a lot of stuff around programming, write JavaScript, etc.  

James Lawrence: Which kind of lends itself to your point earlier around we don't really know where Google and Bing take the integration of generative AI into the SERPs, but it's not entirely implausible that lots of it remains unchanged and that certain queries will lend themselves more to this kind of content. I've got the data in front of me and 30% of the queries you analysed were programming related, 23% education related, 20% content related. I presume that's creation of content. So that's 50. That's almost 75%. Right? Then you got 13% sales and marketing, 6.7% personal and other 3% political and not for profit. So it's there. What's the best enterprise software to do or who's the best agency to do Y or, you know, where can I buy certain types of fashion? 

Rand Fishkin: So those types of queries would fall under two subcategories in education. One is personal education and one is professional education. And then there was sort of education in terms of school, right? So I'm doing homework or I'm trying to put together a lesson plan for my class. That one was the largest in the education segment. And then personal and professional education, which is where you would classify probably most of the prompts that one might use to replace Google. You know, we're talking about sub 10% of all ChatGPT users. 

James Lawrence: Cool. I suspect that trolling through 8000 prompts would make for some pretty interesting bedtime reading. I think it's really good, a good perspective to control what you can as a marketer. Let's not get too bogged down in the big shifting, you know, tectonic plates. Let's, you know, worry about our patch, which I think leans into the next area. I'd love to chat with you around Dark Social because I think that is an area I don't think lots of us understand. And I think if we do and if we can arm marketers with a better knowledge of that, it helps to make arguments within businesses and to explain actually what's happening out there for a brand. Can you kind of just talk about the concept? There's a really good piece on the Spark blog around it, which I'd urge everyone to kind of read. Can we explain? Because I think people just don't understand that this is a reality. 

Rand Fishkin: Sure. So the dry, boring version of this is all the social platforms over the last ten years have to some extent either prioritised linked lists, native content, whatever, a post on Twitter, post on LinkedIn, a post on Facebook post on Reddit that contains no link to a third party website. Right? No link to the open web. And they have also all pulled back on their logging of analytics attribution. So the ability to see that a visitor from LinkedIn came from LinkedIn has gone in about 30 or 40% of all visits that LinkedIn sends out from its platform. 

James Lawrence: How does that actually happen? So literally, 100 people visit my website from LinkedIn. And when I look into GA4, it's showing referral. 

Rand Fishkin: Which will classify as direct type in. Here's the real story in my opinion, which is these platforms are not dumb. They didn't do any of this by accident. Here's what happened. The leadership sat in a room and they said; hey, we noticed that a lot of people are getting a lot of value from… let's say we're LinkedIn. A lot of people are getting a lot of value from LinkedIn and they're posting links to their website. We're sending hundreds of millions of visits every year off of LinkedIn. Why are we doing that? We could be keeping more of those and if we kept more of them, marketers would learn over time that the only reach people on LinkedIn is to advertise with LinkedIn. Also, when we do send traffic out, which should be less often, let's not let people know that we sent that traffic. That way they don't go invest more in organic LinkedIn, they invest more in paid LinkedIn because paid is the only way that we'll show them the real attribution. This sounds nefarious and malicious. It's not nefarious and malicious. It's capitalism, right? It's a bunch of people sitting in a room and going; Well, how do we make more money? Well, does this hurt people? Hey, just some marketers. Screw them. Give me the money. 

James Lawrence: Follow the money. Follow the money. 

Rand Fishkin: I mean, yeah, that's the story of big tech over the last 15 years, right? That progressively every platform Google, Facebook, Apple, Amazon, Reddit, LinkedIn, Twitter, TikTok; every one of them starts out more marketer friendly, consumer friendly. And then over time, I think Cory Doctorow has actually written an outstanding piece about this called ‘The Enshittification of the Internet’. And he basically proposes that every platform, every major platform that has success by being outstanding to its users. And then over time, it realises it can trade user value for advertiser value. How do we be awesome to the marketers and the advertisers who are going to fund our platform and make us lots of money? Because we have to show that to our investors so that we can go public. And then after you go public and they realise growth is slowing, they start being shitty to those advertisers and marketers as well so that they can keep all the value for themselves.

James Lawrence: Enshittification. 

Rand Fishkin: And enshittification of the internet happens in big tech. I know shocking. It's you know, it's one of those things where you read the piece from the document and you're like, wow, that's so insightful and very well written and incredibly obvious. 

James Lawrence: How do you suggest or recommend that the in-house marketer who their reports are saying the sales are coming from Google ads and from SEO? You know, as a humble marketer where you've kind of got vested interest in the organisation and stakeholders, we're all marketers. We know how to do it. You're not doing a good job. How do you push back on this and kind of advocate for the marketing that is actually working? But it's difficult with data to prove that, because we've been preconditioned over the last 20 years to know data is everything and data doesn't lie. And I'll make all my decisions based on data and I can see you rolling your eyes. And we're trying to wage war against that as well. Data can help, but a lot of it's meaningless or fake or doesn't tell the full story. So practical ways to kind of actually try to push back on it. 

Rand Fishkin: I'm not arguing here entirely against being data informed. I think data driven is a little more dangerous because it removes all creativity and removes all sort of serendipitous opportunity, right? You'll never find a new channel. You'll never find a next big opportunity. If you are only studying what worked in the past. I would say two things. One, it is marketer's job to tell the same story that you and I are telling today to their boss, team, clients, that is your obligation. Marketers you have to story tell and say; hey, this is what's happening in the world and here's why it's happening. You can understand why Google and Facebook and all these companies have an incentive to do this. 

Rand Fishkin: Here's what some smart companies have done. They have accidentally or intentionally turned off ads for for a set of time. Right. 60 days and 90 day ad shut off. They didn't shut off all their ads. Maybe they tested shutting off half the ads, 30% of the ad spend. And then they tried to see; hey, did we see the same number of conversions a month later, 60 days later, 90 days later, six months later? Boy, if we did, that means we were wasting that money on advertising. That means Google was lying to us about whether that assisted conversion was really assisted or whether they just had enough user data to put our ad in the user's journey between learned about us and bought from us. In which case we have a lot more advertising budget to pull back on and redistribute into channels that can help more people learn about us who will buy from us. Yeah. That story should not be hard to tell. It's a powerful one, right? And it's clever. It's thoughtful. It's also data informed. It's saying; hey, I'm not telling you exactly which channels aren't performing. I'm saying we should test and the only way to test is to pull back, right? You have to shut it off and then see if you keep getting those conversions. And this is what happens time and time again in a ton of these stories that I've seen. It is we shut off a bunch of our ads or even all of our ads, and we saw that we still got 90% of the conversions. Something is very wrong. 

James Lawrence: I don't know if it was any of your pieces. It might have been. But the analogy being that certain segments of advertising are very much handing out pamphlets or putting a billboard at the front of your retail store and kind of claiming that that was the the thing that drove the person as they went into your car park, to come in to purchase anyway. I think that there's two analogies going on here, right? One is some ads are presenting exactly the right kind of product to the right kind of person. At the time of purchase. You could think of a Google shopping ad for a product someone's just searched for and you have exactly that product and you have the best price and also a brand people have heard of and have a positive association with. And you have it in stock, you've got free delivery and most of your competitors don't. Should you buy that ad? Probably. Probably you should buy that ad if you don't buy that ad, well, the Google shopping users even see, you know, they probably won't. 

James Lawrence: This is also similar to I'm selling skincare products on Instagram and have a great little short ad video. And my ad video gets lots of views and gets a surprising number of clicks on Instagram. And I can see that the purchase happens after the video view and after the click in the same session. Yeah, well, okay, that's probably driving a lot of my revenue. Those two examples are completely different to; I'm selling software as a service to chemical engineers at plants across North America, and it's a six month to twelve month sort of life cycle of people learning about us and our brand and our product and then and getting into our sales funnel. 

James Lawrence: And, you know, my marketing qualified leads are coming through from these channels and they're turning into sales like these. In those cases, you can shut off all advertising and see. We kept getting leads for a long time. In fact, we saw very little. What's going on? Oh, what's really going on is that it was our brand, our PR, our content, our organic social, our event presence, our sales people, all those things, our email marketing. Those were way out of performing advertising. But what Google and Facebook and Apple were really good at was making sure that chemical engineers who matched our customer profile saw our ads somewhere as they bounced around the web knowing that some of those people would were going to buy from us eventually anyway, or get onto our list. And so really all they're doing is knowing lots about the behavior of Internet users and injecting themselves into the process at some point.

Rand Fishkin: The most obvious one of these, of course, in the last example I'll give, is paid search for branded terms. Yeah, right. So somebody, you know, somebody searches for Spark Toro…

James Lawrence: They're looking for you. 

Rand Fishkin: Yeah. I mean obviously right. If we buy the paid ad, they're going to have heard about us in lots of other places. And if we didn't buy the paid ad, we rank organically for it anyway. Maybe some percent of all the people who search for it are going to click on the organic listing for Spark Toro. But Google's kind of like; well, we can sort of play the Mafia game of like, oh, it would be a shame if one of your competitors was to come in here up top in front of you. That's essentially what they're saying with our paid branded ad. That you don't buy the branded search term for your brand. We're going to sell it to somebody else. 

James Lawrence: Literally yesterday with a client and that exact principle. Which was organic conversions that trailed off. And we were like, yeah, but you guys fired up branded ads at exactly the same moment and they're like there's no, in this instance, not even any competitor bidding. It was like, Guys, you're spending money to pay and you're the same person anyway. 

Rand Fishkin: Here's the only reason, the only logical reason I've heard why you should buy branded paid ads in that case. And that is; if the money is largely meaningless to your bottom line and margin. We're paying $0.50 a click for those and you know, the average whatever purchase price is $5,000 over two years or something. And we just want to know all the variations of branded keywords that people are using. Which Google refuses to provide organically, but they'll give us all of them with the paid version/

James Lawrence: We had some work in the hotel space where it worked as well, where basically Booking.com, Expedia, the OTAs was siphoning off if you're the direct hotel and that otherwise because they could sell your inventory take their 15% to 25% and we actually did the extreme test you're advocating which was we turned it all off and bookings went down, bookings to the OTAs went up and we went back and kind of bought direct. But yeah, I agree with you. I think in the vast bulk of instances, unless you're spending a very small amount of money to kind of control that narrative, it's a hard play in terms of the social data. I just wanted to kind of share it. Percentage of visitors attributed to direct visits from these platforms. We talk like the research you did TikTok, 100% slack, 100% discord. 

Rand Fishkin: So basically, if you're getting any traffic from those platforms, you're getting absolutely nothing in UGA saying they're coming from their Facebook, and Instagram was at 75%. Anything being shared there, which is also potentially much higher quality traffic, right? You've got things being shared, referrals, people who are so bought in really depends on the platform, right? For some people, TikTok traffic is very valuable to them if the demographic is the right match. And for some people it's useless. Like funny story; Spark Toro had a couple times where we like went viral on TikTok. Some creator in the marketing world did videos about us and got hundreds of thousands of views, sent a ton of traffic to Spark Toro. Oh, my God. The worst quality traffic you've ever seen in your life. Like, just honestly terrifying people who clearly had no idea what we were or what how to use it. They couldn't even figure out an email sign up. They were like sending us really terrible messages. Everybody was signing up with like, fuckyou@gmail.com. This is sort of why did you bother? Why, why even enter that email? Like you're not going to get the confirmation code for it.

James Lawrence: Right?

Rand Fishkin: I mean, unless you really own that email address. 

James Lawrence: Strong, strong, strong email. 

Rand Fishkin: But I'm just saying, right, that it's different for every different provider. And what that research shows is that dark social is usually somewhere between 20 to 100% of all the traffic that you might see from a particular network. And if you are looking in your referral data to try and figure out which channels are valuable to us, you're playing the game wrong because those channels that underreport the traffic they actually send you and the channels where you don't participate, guess what? They're not going to look good. 

James Lawrence: It's fascinating. I would encourage all listeners to the pod to read the article and then see how it applies for the business you're marketing. You mentioned Spark Toro, and I'd love to to jump into that now. So was it 2018 that you started? Why did you launch Spark Toro and then what is it? Talk us through it. For those marketers that aren't yet aware. 

Rand Fishkin: See, whenever someone says, Why did you start Spark Toro? I always have that Homer Simpson moment. Where is brain says; ‘Don't say revenge, don't say revenge, Don't say revenge’. And then Homer, Of course; revenge. That's it. I'm out of here. So for folks who might not know the story, I had a contentious last few years at Moz after I stepped down from the CEO role. I think Spark Toro was in part a personal mission that I had to prove mostly to myself. But also to a small number of people on my board of directors who maybe didn't listen to me at my previous company and that maybe I still knew what I was talking about and could build a successful software product for marketers. And Spark Toro is something I strongly believe marketers need, which is essentially to be able to understand not just the channels, but the individual sources of influence across channels and on channels that are often ignored that reach their audiences.

Rand Fishkin: If it turns out that if you want to sell an indie video game in North America to an audience who really loves crafting and strategy games, you need to be in front of these creators on Twitch and YouTube and TikTok and Instagram. Who? Who are they? Which ones? How would you find out? Gosh, what could you do? You could survey and interview your audience. That's really tough. How do you find people who play games they don't really publicly post about them? You might be able to do some steam review data mining and but it's pretty tough. How do you figure out where dentists in California are getting their dental equipment? Sources of information, right. What newsletters or events or conferences or trade publications do they read and pay attention to or their particular Facebook groups they're part of, or all of that kind of stuff? It's doesn't fit under the classic definition of dark social. What it fits under is hard to identify sources of influence. 

Rand Fishkin: I keep the word influencer out of my mouth because influencer has come to sort of mean half naked person on a beach with a six pack who's like; here, you should buy this blender and you can have six pack abs and a beach like me. That's not what I mean. What I really mean is all the niche sources of influence that could potentially influence a person and a group of people. And that's what Spark Toro is designed to help people do. It's been nice. We did not launch until 2020. So this is our third year of operation as an actual product. But yeah, it's helped tens of thousands of marketers already. That's that's pretty exciting. It's brilliant. The whole influencer being a dirty word. It just shouldn't be. 

James Lawrence: And I think us as marketers need to challenge it, kind of rise up against that because it is like digital marketing is becoming old school marketing, right? And I think a tool like yours helps businesses, particularly those that might not have huge research budgets that 

some of the big corporate ones. You can pay to have big bodies of research done to help find, you know, the dental this or the gaming influencer in a certain market so hard if you're the vast majority of marketers that are operating in a business. 

Rand Fishkin: You know it's so funny. I was thinking about this the other day with, do you remember the television show Mad Men? That was on maybe ten years ago. 

James Lawrence: Yeah. So classic, right? 

Rand Fishkin: Jon Hamm plays this director of an advertising firm in the 1950s and then into the 60s in the United States, in New York. And, you know, the clothing is great. The scriptwriting is great. The dialogue is great. It's more than a little misogynistic, which probably the advertising world certainly was, too. But the interesting part that I was thinking about is there's a scene where it shows a client, you know, potential client for the ad agency coming into the firm. They're all wearing suits and whatever. And one of the guys who works at Sterling Draper, you know, the fictional name of the company that's supposed to represent one of the big New York ad agencies, pulls a binder off the shelf and the binder is filled with information about where you can go to reach certain demographics. And so I think the the advertiser in this case, the client is a car company and they talk about the car that they're making and who they want to reach with it. You know, they look and they say, okay, according to the research, we need to be on this channel. We need to be on this time slot. We want to be on this these three shows. We want to be in these two magazines. 

Rand Fishkin: I think they mentioned like Time magazine and NBC in the major cities. That research came from giant large scale survey data. These big survey firms would ask a million Americans like, hey, where do you what do you watch and read and listen to? Who are you? Blah, blah, blah. So if you wanted to reach a dentist, you'd go into an ad agency, they'd grab the binder and they'd tell you, okay, great, you're going to be advertising in these three papers. There's no binder, right? In the Internet era, all that data is right on the web. Yeah, we could find it, right? We have things like clickstream data sources and social following graphs, APIs and search results. We could collate all that data together, amalgamated, and then build a binder. And that's essentially what Spark Toro is. Yeah, it's sort of the ad agency binder for the 21st century. 

James Lawrence: There's a blog article surely on the Sparks our website for that. You know, you Don Draper. There's something in there Rand.

Rand Fishkin: Just much less drinking. No, no terrible behavior towards women. 

James Lawrence: But you know, the comparison between the two. Where do you think it plays really strong, the actual software, like who's it? An absolute no brainer for…

Rand Fishkin: If your brand or company sells to a certain set of professionals, right? So like B2B use cases or even B2C that targets an interest group that describes themselves publicly online, for example, tabletop role playing game enthusiasts. All of these people are not shy about putting the RPG hashtag in there, whatever, Mastodon and Twitter and LinkedIn and Facebook and Instagram profiles. And so it's very easy to identify them. Dentists, orthodontists, interior designers, chemical engineers, podcasters, right. Anything where they publicly describe who they are. And that's your perfect audience. It gets much more challenging and we often don't do a great job when your target is well. Pretty much everyone. If you're selling car insurance or homes, I'm trying to sell real estate in Brisbane. Yeah, well, everyone's kind of a home buyer potentially. I mean, obviously there's demographics around it, age groups and income groups, but Spark Toro, that data is not publicly available on the web and we don't try and whatever, you know, what Facebook does, right, where they buy data from banks to get that sort of private data and then apply that to profiles so that you can do targeting. We don't do that and we never will. We think that's unethical, even though it's not technically illegal in every country, but anything that's publicly available, data that's on public profiles, we anonymise and aggregate. And that's why it works so well for B2B. 

James Lawrence: That's always the challenge with B2B, right? Like obviously LinkedIn can give you some certain targeting, which helps, but it's really challenging to in those more niche areas.

Rand Fishkin: A few years ago you remember James, when Facebook audience insights existed. Yeah, like 2010 to 2016t. Donald Trump gets elected and it's the Cambridge Analytica scandal. And so Facebook like pulls all audience insights out of your organic Facebook pages. But in Facebook audience insights on your page, it used to tell you like, oh, people who follow your page also have are interested in these brands, these companies, these pages, these public groups. That was awesome. Twitter used to do the same thing. And Twitter analytics, they pulled back on it, right? So all these companies have all this data. They're using it to show your ads and sort of get you as an advertiser. But it's your data. You built that audience. You're the reason people are coming to your Facebook page, your Twitter account, your LinkedIn account, your pay, whatever, your subreddit, and they're hiding it from you so that you'll buy ads from them. And I think that sucks. I think you should get that data. And this is one of the reasons that Spark Toro is also a very inexpensive product is because we want to help a lot of like small and medium sized marketers. We don't have any like enterprise support for like, I feel a little bad. There's a very large one of the big five ad agencies in the world has been trying to get a Spark Toro subscription. But they need you to like jump through a bunch of hoops and provide them with whatever support and all these guarantees and security compliance checks. And I keep writing to them and being like, we're not going to do it. 

James Lawrence: You've designed this business to allow you to live the life you want to cook the nice meals you want to quickly. 

Rand Fishkin: Like, I'm not I'm not going to do it. Sorry, man. Omnicom or whoever, just like you go away, you can buy this data from other people. You can build it yourself. 

James Lawrence: What? This is a free trial. So anyone out there sign up? Free trial? 

Rand Fishkin: No, no, no, there's not no free trial. We'll do you one better, though. We have a forever free version, so you never need to pay. You don't have to put in a credit card. You can just have a free account that you use. I think we have more than 100,000 marketers who use the free account every month. Just completely free. So go play around, do your research, and if you're using it and you're like, gosh, I actually really want more from this, like, want to see more rows of data, more kinds of data, great. Then you can sign up and it starts at 50 bucks a month. 

James Lawrence: Yeah, it's fantastic. And where's the product going? Where do you see it kind of heading in the next few years?

Rand Fishkin: We're about to come out with a V2 of the product. Don't know if it'll be out by the time this pod launches, but the V2 will be a little bit of a different interface. And on the back end it is using some new kinds of data sources, including the same clickstream data set from data that we used for the ChatGPT analysis. So if you're kind of like, wow, how did they get all that data about what people visited and how much traffic it got and where they went after and before, That's how we can get that data. So I have been a huge fan of Eli Goodmans, who runs data for a long time. He and I worked together at Jumpshot back when did those Google Zero click studies. At the end of my tenure and then into the start of my Spark one. And then we have a large amount of public profiles from LinkedIn, which is great. And we sort of marry those two together with a third data source and that is search data. So for the first time, so many people have been asking us; it's great to see like which sources of influence, which websites, which social accounts are popular with my audience, I also want to see which keywords are popular with my audience, what are people searching for? And so now we'll be able to offer audience data around keywords as well. So that's where it's going and then I think to be honest, I don't love to predict out. We really like being responsive. So I think we're going to launch this, see what people love, what they don't love, what they use, what they don't use, and then double down on the things that people are asking for and want. 

James Lawrence: So love it. It sounds like Ryan's revenge tour is in full flight. 

Rand Fishkin: It's funny, you know, I say revenge, but the interesting thing is the longer the time frame is that you move away from some painful incident, the less sharp those edges feel. Moz is sort of faded to me into; well, that was that was sort of a lovely learning experience. And I made a lot of friends and got to have wonderful connections to people like yourself. And a lot of marketers know who I am because of that experience. So I don't know, maybe I don't have to carry a grudge about it. It's in the past. I was young. Hopefully, I'll make fewer of those mistakes going forward. I think, you know, Lost and Founder, the book I wrote, is kind of a nice literal bookend right to that journey and tries to tell me and other people hopefully, what not to do in the future. 

James Lawrence: You've definitely added to the industry and helped a lot of people make sense of things that are potentially very, very complex. And it's hard to make the complex simple which you do, which is which is probably a better legacy than

Rand Fishkin: Oh, my company was worth $1 billion, right? 

James Lawrence: It'd be nice to have both, though. 

Rand Fishkin: Not the worst, right? I mean, here's the thing, though, James. Maybe you're personally connected to a lot of very high net worth individuals. Geraldine and I have come to know a good handful of people in our lives who have lots of money or way too much money. I don't think they're actually happier or better off. I think diminishing returns start somewhere in the low few millions of dollars and then they continue, and once you get to 10/20, the drop off is extreme. And I think unfortunately, a lot of very wealthy people struggle to search for meaning. Think that they're supposed to use their money to buy themselves happiness and also to immunise themselves against criticism. And it becomes really tough because a lot of the global narrative is about how rich people are destroying the world. And that is a true narrative. But I think it creates this psychology. I wish Moz had made us some more money. Sure. But I think $1 billion is the wrong amount. 

James Lawrence: That's good. I think it's good for you to have that perspective. And yeah, for sure. Definitely observing people that I know that have a lot of money. It doesn't bring happiness. I think happiness is human connection and a safe place to live. And good food on the table shared with good people. Then the time and the resources to do the things you like. 

Rand Fishkin: Money gets you to a point. Absolutely agree James. I'm hoping to write a sequel to Lost in Founder probably the next 18 months, maybe publishing in a couple of years and spend a whole chapter on the money stuff that nobody talks about because, you know, Moz did sell, right? I think it sold in 2021. That was a very strange, life changing amount of money for us, but also awkward, scary, sad in a lot of ways. 

James Lawrence: Really interesting. And I think that the thing is, most Australians would say that Americans talk much more openly about money than we do, 

Rand Fishkin: Which I think that's true. 

James Lawrence: I think that would be true.

Rand Fishkin: Americans don't, in my opinion, Americans don't talk about it nearly enough, especially higher earners, because I think many folks are ashamed. Right? They feel like and maybe some shame should accompany, when a vast majority of your country is living in terrible conditions and there's no social safety net, you should probably feel ashamed that you manage to be one of the few at the top. But also if you don't have those conversations, it's pretty tough to change things. 

James Lawrence: Yeah, this is not where I thought our podcast would go around. I love it though. It's great. Thank you for sharing. It kind of segues into the final part of the pod. Career advice for in-house marketers and young marketers out there. So I normally finish the pod with what's the one piece of career advice? But if you could just for a couple of minutes, just talk about what advice would you give to that marketer out there that is kind of starting out in their first role, the first couple of roles, like what's going to make for career success based on what you've seen? 

Rand Fishkin: Specialisation. I think specialisation is one of the keys to building not just a successful career, but a career where you are known for something and therefore recruited for that thing. And it's the demand that's in the small niches. When I say specialisation, I don't mean specialise in Google advertising. If you specialise in Google advertising for boutique hotels, yes. That's the kind of specialist level that I'm talking about. And then when people say, gosh, you know, I really need someone who's an expert at this, you'll be the person on everyone's lips. You'll be the the name that is associated with that. The other piece of advice is you have to do marketing for yourself. I know that lots of people don't love personal branding and they don't love personal content creation. And I'm not telling you you have to do a lot, but if you want to be successful in the marketing world, you need to be known by other people who might help your career recruit. You suggest you to people who are looking recommend your work to potential clients or your next employer. Even just connect you with people who you get along with. 

Rand Fishkin: My suggestion is you have two ways of doing that. One is content on the Internet. Social media content, a blog, video series, a podcast, write all those kinds of things to where you where you share your knowledge. You publish what you know and what you've learned, and you help people in those ways. And the second is personal communication. You can maybe do it through email or an email marketing list, which I'd really term in the first bucket of sort of content creation for the web or in person. And in person is weird because events in the digital marketing space have been shrinking dramatically even as the field has grown. Yeah, it's weird. Ten years ago there was probably 200 events worldwide for digital marketers that had more than 500 attendees. Now I don't think I could name 12. So it is it is shrunk dramatically and that sucks. It's a problem. But going to those events, you will find people who are excited to meet you. Who want to network and in a kind and friendly sense, and also learn from you if you get on stage or pitch content and get in front of organisers, even at local stuff that can make a difference too. So those are the two ways that you sort of get your knowledge out there. Just like anything else, right? You're a marketer, so you know that no great product just is made, launched and then people will just find it. That's not how anything works. Nothing works that way. Same thing with you and your career. You've got to let people who might help you, who might want what you do, know what you can do that's unique. And the way to be to get that uniqueness is to specialise. 

James Lawrence: It's great advice. Thanks for being so generous with your time today, Rand. Really appreciate it. I recommend to all the listeners. It's not a free trial. It's free forever. But even putting that to the side, sign up to Rand's content, as I said, to kick off the pod like when something hits my inbox from your end. I love it. It's thought provoking. It makes me think. It makes me question what we're doing for our clients. I'm glad the whiteboard is back. It's behind him. For those that are listening and can't see him. And I'm glad it's back and he's tapping it. It's real. It's not a zoom background. Thanks so much for being on the pod Rand. 

Rand Fishkin: My pleasure, James. Thanks for having me.