Description
In this candid conversation, hosts Matt and Stephanie sit down with Rand Fishkin, CEO of SparkToro and former CEO of Moz, to explore the evolving landscape of audience research and ethical AI implementation. From challenging traditional venture capital models to building "chill work" companies, Rand shares invaluable insights on creating sustainable businesses without sacrificing personal values. Whether you're a marketer seeking better audience understanding, an entrepreneur navigating the startup world, or simply curious about the future of consumer insights, this episode offers a refreshing perspective on doing business differently in today's tech-driven world.
Rand - 00:00:00:
I think at the early stage, the thing that I generally recommend is rather than go and try and collect a bunch of data and then see where it applies, I would list out the problems you actually want to solve. Let's make sure that users can easily understand how to do every single thing in the product with no support. And as a result, anytime someone writes in to support, we take very careful note and we're like, oh, did three people write in about the same thing? Then we're fixing that. That's a thing that's broken. We need to address that.
Stephanie - 00:00:29:
Hello, fellow insight seekers. Welcome to The Curiosity Current, a podcast that's all about navigating the exciting world of market research. I'm Stephanie Vance.
Matt - 00:00:39:
I'm Matt Mahan. Join us as we explore the ever-shifting landscape of consumer behavior and what it means for brands like yours.
Stephanie - 00:00:46:
Each episode will get swept up in the trends and challenges facing researchers today, riding the current of curiosity towards new discoveries and deeper understanding.
Matt - 00:00:57:
Along the way, we'll tap into the brains of industry leaders, decode real-world data, and explore the tech that's shaping the future of research.
Stephanie - 00:01:04:
So whether you're a seasoned pro or just getting your feet wet, we're excited to have you on board.
Matt - 00:01:10:
So with that, let's jump right in.
Stephanie - 00:01:12:
Today on The Curiosity Current, we are joined by Rand Fishkin, co-founder and CEO of SparkToro, a company that's revolutionizing audience research and empowering marketers to understand their target audiences without relying on more traditional, expensive methods.
Matt - 00:01:31:
Rand is also the co-founder of Snackbar Studio and the former CEO of Moz, where he played a pivotal role in shaping the world of SEO tools. He's an entrepreneur, author of Lost and Founder, and a frequent keynote speaker, known for his transparent and candid approach to the startup world.
Stephanie - 00:01:49:
In today's conversation, we're going to dive into Rand's journey from Moz to SparkToro. We're going to explore how technology is transforming the landscape of audience research and discuss the future of consumer insights in an increasingly data-driven world.
Matt - 00:02:06:
Rand, we're so excited to have you here. Welcome to The Curiosity Current.
Stephanie - 00:02:09:
Yes, welcome.
Rand - 00:02:10:
Oh, Matt, Stephanie, thank you. Great to be here.
Matt - 00:02:13:
Rand, you've had, to put it lightly, a really interesting career trajectory. I think we wanted to start there. So you founded this company, Moz. You co-founded SparkToro. You've started an indie video game company called Snackbar Studio. How does all this fit together? Can you take us back to the beginning and share what has motivated this sort of evolution, the evolution of Rand?
Rand - 00:02:37:
Oh, gosh. When I was a little boy, a little kid, I was deeply into video games. I loved them, was obsessed with that. And that is what got me into software, actually. My parents gave me an old hand-me-down computer, you know, giant box, weighed as much as I did. Did not have nearly enough RAM to play the video games that I wanted to play. I don't know if people know this, but back in the day with MS-DOS, DOS was like the operating system pre-Windows, you could manually rewrite the files that loaded the boot system of the computer to fool it into thinking that it had more memory than it really had. Now, there was no internet on which to discover this information. And so you had to like learn it sort of word of mouth from, luckily I had a friend whose grandfather worked at Microsoft. And so this is how I figured this out. And so then I learned to edit, you know, whatever, configuration systems, settings. That is how I learned programming initially. When the internet came out in 1994, five, six, you know, it was taking off. My mom was a marketing consultant who helped local small businesses in the Seattle area with things like yellow page ads and logos and letterheads. And she was like, hey, Rand, I've got this software for you. I think it was Microsoft FrontPage was the first thing she brought home in like 95. And I learned how to use that and started making web pages for her clients. And then when I went to college, I found that much more engaging than going to class. And so I became a college dropout who was a web designer and developer in my early years. And then we were terrible at it. Like we were so, Matt, Stephanie, I can't, let me tell you. You know how a consulting business is supposed to work? Low overhead, low expenses, decent profit margin, easy to get off the ground. No, our expenses were high. We were not bringing in clients. We were going deep into debt. Like we were terrible at this. I had at one point half a million dollars of credit card debt because pre 2008, they used to give so much credit to people who deserve no credit at all. Like me. Long story short, which folks can read about in Lost and Founder, is that eventually through SEO, we managed to turn the agency, the little consulting business, mom and son consulting business around. And then in 2007, accidentally tripped our way into building software and launching that. And we were one of the first SEO tool companies in the SaaS world, which we didn't even know what SaaS was at the time. Raised a bunch of venture capital. We didn't know what venture capital was at the time.
Stephanie - 00:05:28:
Totally, yeah.
Rand - 00:05:28:
Just generally made all the mistakes that an entrepreneur can make. But also we were riding this crazy wave, right? Google is growing so big. SEO is getting huge. And so, yeah, Moz has turned into a $50 million a year software as a service business with a couple hundred employees and a CEO who didn't really know what he was doing. Although, hot tip. I don't think CEOs know what they're doing. You know what they're really good at is networking and seeming like they know what they're doing.
Stephanie - 00:05:57:
Rand, that's going to be the 20 seconds.
Rand - 00:05:57:
Anyway, and I continue to fill that role today.
Stephanie - 00:06:00:
Right there.
Matt - 00:06:02:
Right. Yeah. Yeah. That's the lead in. That's the lead in line right there. CEOs don't know what they're doing.
Rand - 00:06:08:
Look, man, we're all just pretending. This is why I'm so opposed to CEOs making like 500 times the pay of their employees. They're not working harder. They're not better than you. They're not smarter than you. They're not more talented than you. We're just doing a job like anyone else. We just happen to have leverage with the board of directors to be able to convince people that they have to pay us more.
Stephanie - 00:06:28:
I think that's so true, especially founder CEOs. I love that about founder CEOs. There's so much. I know this is a weird word to use, but like humanity in that role. Do you know what I mean? Like it is a we have a founder CEO and his ability to relate to everyone in our company is so much higher than any CEO relationship I've ever had the privilege to work under. Like there's something about our founder CEO. It's a special, unique kind of setup to be working in. For sure.
Rand - 00:06:59:
Absolutely right, Stephanie. I think it can be a really beautiful thing when we embrace our humanity and our ability to relate to people. And man, it can also go exactly the opposite way, right? Where they just think that their shit doesn't stink and they take these twisted routes to egomaniacal domination. We could talk about that a whole lot. Like I think my journey has really been one of barely avoiding that. There was a time when I can look back and see myself in my early thirties. Moz was... Skyrocketing and blowing up. And I thought I was the cat's meow and that I could do no wrong. And if it hadn't been for some humbling mistakes and, you know, a path that made me a more thoughtful person, I, I don't know. I don't want to imagine who that rant could have been, but it might've been ugly. So Moz, the first company I founded, the one with my mom, it sold a few years ago, 2021, 22, something like that. I wasn't on the board of directors anymore. I wasn't at the company anymore, but I still owned my wife and I still owned a huge percent of the company, like 18, 19% or something. And we got this giant check in the mail, right? Like the company was not a successful exit by any means. Like the venture capitalists were not happy, but it's still buying a $50 million company is expensive.
Matt - 00:08:16:
There was utility there.
Rand - 00:08:17:
There was a number with a lot of zeros after it. And I was pretty freaked out that I'd turn into an asshole. Had like serious trauma, about, oh no, I'm going to be a terrible person. To which, I think the reflexive response would be. Well then, just give it away, and we did give away, a substantive portion of it. Well, certainly not all, and found a way to, sort of be comfortable with like, hey, it's okay too, we just splurged on a bunch of fancy clothes for my little brother. Because I want him to look nice, for some stuff. Yeah, we sent our nieces and cousins and some Fred's kids to college, right? Like that feels awesome, that's very, very fulfilling. Their parents couldn't have paid for it. So it's like, there's cool stuff you can do, but also you don't have to be a Turd Burger.
Stephanie - 00:08:59:
You don't have to be a Turd Burger. It's true. So how did that then affect how you approached SparkToro?
Rand - 00:09:07:
In a ton of ways. When Casey, my co-founder at SparkToro, and I conceived of the business, we both knew that we did not want to run a big company. We didn't even want to run a midsize company. We wanted to keep it really small and work with lots of contractors and external people, of course. But what's nice is in this era, you can do that. So we're running a north of a million dollar a year SaaS business with three people. And we have an approach to work that we call chill work, which is essentially that we want to build a business that does not require, although sometimes we often put in more than this, but does not require more than 20 hours of work in a week for many of us. That's been like a guiding philosophy in how we build and design all sorts of things. Everything from, hey, let's make sure that users can easily understand how to do every single thing in the product with no support. And as a result, anytime someone writes into support, we take very careful note and we're like, oh, did three people write in about the same thing? Then we're fixing that. That's a thing that's broken. We need to address that. But as a result, the product is really, really easy to use. People who have no idea about audience research software can log in for the first time and figure it out really quick. There's tons of other things that flow from that. We are not willing to do sales processes. If you write in, this is one of the few things that we don't fix. If you write in and ask us to do a demo, we're the team from P&G and we're spending $50 billion a year on advertising. We would love to take a demo with SparkToro and we write back and say, sorry, we're a tiny team of three. We don't do demos. Here are some videos you can watch. The response is often, you got to be kidding me. Who wouldn't do a demo for us? We're like, hey, that's who we are. We're a little team. We'll answer your questions on email. No problem.
Matt - 00:11:00:
If it's something that's so user-centric in its design and so lovely to use, you want that to be the proof in the pudding, right? You want the person's first experience to be immersive. It's more immersive than getting in and getting your hands dirty.
Rand - 00:11:13:
Should be low friction. So, you know, that speaks to a lot of the things that we do with the business in terms of cost efficiency and savings and how we build software, how we design things, what we expect from our team, what we don't expect. It's super different. It changed how we raised money. We raised actually from a lot of folks like yourself, lots of just industry people who were interested in what we were doing and thought, oh, that's a great idea. I'd love to be an investor in that. And they put in between $25,000 and $100,000 each. And a few years ago, we actually paid everybody back their initial investment. And now we do profit sharing pro rata, which I don't know why that doesn't exist in the tech world. It seems deeply weird to me. And so this is exactly the same way that Snackbar Studio is funded and a number of other businesses. I've actually convinced it's the right word. But we open source the documents so that anyone can raise money using our funding structure. So you can go online and search for SparkToro funding. There's a LLC docs that you can just fill in the blanks with, save a bunch of money with your lawyers. Snackbar Studio is a C-Corp. So if you're doing C-Corp instead and you want to do it that way, those docs are also open sourced. So, yeah, we're just trying to also give back, pave this path for other people to follow.
Stephanie - 00:12:29:
It almost feels like flipping boundaries on their head. There are boundaries that we normally don't have in business that you're like, I'm not willing to compromise on these boundaries, which is our personal time and our being there. But then there are other boundaries that are normally held quite tightly in business that you're like, those don't need to be held. We should be sharing over here. Super interesting approach.
Rand - 00:12:51:
Exactly. I think you could probably tell from listening to me even for, you know, whatever, 10 minutes that- I'm quite a contrarian. I try to hold these two opposing views in my head, which is I hate guys who say, well, actually to everything. And also there's a lot of existing power structures and ways that things are done where I say, well, actually, does it have to be done that way? Can't we do it some other way?
Stephanie - 00:13:15:
I love it.
Matt - 00:13:16:
Contrarian is a little bit of a loaded term, right? I mean, I think it's come to mean someone who challenges an idea for personal gain, but there's some value in just calling out dogma for what it is.
Rand - 00:13:28:
That is wonderfully said, Matt. It's genius because you're absolutely right. The brand of contrarian is terrible right now. I think that it's often used by not to generalize, but by reply guys, right? Well, actually guys. And I never want to be that. I don't think that's what we're doing. I think what we're trying to say is, what if we question the way things have been done and try to find ways to do them differently that are caring and empathetic and thoughtful and transparent? That's a different kind of contrarianism.
Matt - 00:14:01:
I love that.
Stephanie - 00:14:02:
I was going to move on to the next question. Although, Rand, I will say, do you want to talk about Lost and Founder a little bit?
Rand - 00:14:07:
Sure. Yeah. I mean, here's what I'd say in brief. Lost and Founder, I'm very proud of it, which I don't say about a lot of things. I'm very proud of it because I know that it has helped a ton of entrepreneurs to avoid or think more carefully about raising money from venture capital and about how they build their businesses and scale them and what really matters in their companies and their lives. If those are things that someone in your life is worried about, Lost and Founder might be a good read for them. And I would also say that I wrote it. Do you guys know when you're going through a traumatic experience and it's really hard to see the forest for the trees? I find this a lot with people who are going through a romantic breakup, right? And they think, oh, that partner, my ex-partner is so terrible in all these different ways. And then five or 10 years on, they look back at it and they say, they're not terrible. I'm not terrible. It just wasn't a good fit. It wasn't the right match. And I do look back now, seven years, I guess, since the publication of Lost and Founder, and think there were a lot of things in there that just weren't a good match. I was not the right person for that asset class and for that type of a business experience. And I want to do things differently and I want to pave a different path. But that doesn't necessarily make everyone who invests in companies from the venture background evil. And it doesn't necessarily make every founder who says yes to venture a bad person. Correlation is not causation. Statistics represent a wide range of experiences. I think if I had to raise venture today, I probably could do an okay job of it and maybe just be more thoughtful about the kind of company I want to build. But it certainly encourages a lot of bad behavior.
Matt - 00:15:54:
Lost and Founder was your therapy, your way of reconciling this. Sounds like you were experiencing was you're going through a transformation where there are implications for your own work, but you want to do something intentionally different from the way the status quo is approaching Venture or whatever elements you want to talk about. And that caused some dissonance, right? That caused some issues that were challenging to work through. Writing the book then was a form of therapy for you. It sounds like it was a way of keeping score, going back with maybe a more objective viewpoint, right? Understanding what happened and getting it on paper.
Rand - 00:16:36:
Always with the goal. I think this is the best thing that my editor, Niki Papadopoulos from Penguin, did for me was, and for readers, is she always made it audience serving, reader serving. She's like, great, you wrote all this stuff. That's very cathartic for you. I'm happy for you, sort of. But here's what we're going to put in the book. This is what makes this useful to somebody else. I think that's why Lost and Founder did as well as it did and continues to be on lots of nice lists. You know, I have like an alert set up for it and I get pinged every time someone says something nice about it. Yeah.
Stephanie - 00:17:13:
I think that's one of the beautiful things about entrepreneurs in general is that they tend to be hungry for things to relate to and things to help them understand the world in which either because everything is shifting all the time in entrepreneurship. It's a good audience to have for a book because they're they are hungry for knowledge. And I will say I have read it and, but I didn't read it when I was in that world. But it is so easy for me personally to relate to because I've worked for a couple of startups in my past. And I'm also my partner worked at an accelerator for a long time when we first met. So just that is pure immersion in that world of people who are in it, just hardcore. They're in it. They're trying to figure it out. They're trying to make those moves and highly relatable, very, very relatable stuff. So I enjoyed it quite a bit.
Rand - 00:18:06:
Only wish is that the humility that early stage first-time entrepreneurs come to that world with maintained as they became wealthy and powerful and important and built these important technologies and ran these powerful companies. I was talking with another founder friend recently about how few of our mutual friends who had had extreme amounts of success, right? Like where their net worth went above $100 million or something like that. How few of those people stayed grounded, normal, fun to spend any time with, you know, thoughtful, charitable, equitable, just all the things that you would want from a human being seem to go away the richer and more powerful you can. Anyway, this is probably why I had all this trauma when I got that big check, which wasn't even close to that number, but-
Stephanie - 00:19:02:
It definitely puts into sharp focus the moral quandary of money.
Rand - 00:19:07:
The modern age. You feel this in every sort of ecosystem, right? In media and influence spaces. You feel it in the world of technology. We certainly feel it, obviously, in the worlds of consumer insights and marketing realm. We feel it in politics, of course, and social issues and culture. Like, it's just, it bleeds all over everything. And it feels like the problem of our age. It bleeds into, like, climate change. You know, it's everything.
Stephanie - 00:19:34:
Let's talk about SparkToro a little bit. You built SparkToro to help marketers find insights without relying on surveys, focus groups, things like that. What I think of as primary research. There's an ongoing debate in the industry, I feel like, about the utility predictiveness, things like that of surveys versus transactional behavioral data. I would love to hear you talk a little bit, if you don't mind, about the strengths and weaknesses of each approach and how you think they actually maybe complement each other.
Rand - 00:20:04:
I do. You've probably read some of my stuff on this. I love surveys and interviews. I think they're super valuable for getting our soft insights and detail-level insights into particular problems. We use surveys and interviews all the time at Moz to improve our business. They're great. However, there are certain types of questions that people, human beings, are just terrible at answering. Biggest among these would be things like, hey, when you are planning a vacation to Italy, what websites do you visit in that process? People are really bad at answering that question. They can't really do it. I'm not saying they're lying. They're not lying. It's not that the information is necessarily wholly inaccurate. It's just incomplete. And if you were to survey or interview 100 people, you would not get nearly the data that gives you correct information about what an average person does and in what order these websites are visited and which ones are most popular in that sector. That is not a good use case for surveys and interviews. It just isn't. Oh, I'll tell you one of my favorite stories and examples on this. So early on in SparkToro's life, first six months or nine months after we launched, there was a they were direct to consumer. And they're also sold in like specialty grocery stores. They made a specialty food product and they were doing quite well. This is like during the pandemic, home chefs are just going crazy for this product. They're going to ship. And they had on their website, I noticed they had signed up for a SparkToro account. They also had on their website a post-purchase survey where they asked people, how did you hear about us? In my opinion, is one of the worst use cases for a survey.
Stephanie - 00:21:53:
All the time. No, I just want to interject.
Rand - 00:21:56:
They're very common, right? Like these are super popular, the post-purchase survey. There's good questions to ask in there, by the way. I'm not saying throw it out entirely, but a bad question to ask is, how did you hear about us? The reason it's a bad question is because almost certainly they heard about you from 17 places. But they're only going to remember or think about the top one, right? Or whatever. And a ton of people are going to mark whatever's first on the list or the most sort of popular thing or the thing that they think will get them through the process the fastest, the top of mind thing. So I asked them in order to dissuade them from doing this, I asked this company, hey, would you guys be willing to try and experiment with me? Would you add, they had one of the dropdowns was an influencer. You heard about us from an influencer. And then they had a secondary dropdown where they, you know, where they listed the influencers who had mentioned their brand. And I said, hey, would you just do me a favor and add these two famous food celebrities, like food world celebrities, into that dropdown?
Matt - 00:23:01:
Were they famous, Rand? Were they truly famous?
Rand - 00:23:03:
They were quite famous. They were like people you would see on Food Network and like all that kind of stuff, like real chefs.
Matt - 00:23:08:
I thought it was a red herring play.
Rand - 00:23:10:
It was a red herring.
Stephanie - 00:23:11:
Did they promote the product? It sounds like a no.
Rand - 00:23:14:
Yeah. They had never, ever in their whole lives mentioned the product. It wasn't Martha Stewart, but like Martha Stewart. And oh my God, almost everyone who chose the influence, I think it was like 85% chose one of those two very famous people.
Stephanie - 00:23:26:
Everyone was like, Gordon Ramsay told me about this.
Matt - 00:23:30:
So you're proving out the causation is more driven by top of mind awareness in general.
Stephanie - 00:23:35:
Familiarity, things like that.
Rand - 00:23:36:
It has nothing to do with did that person influence it, right? And they're not lying. They're just like, oh yeah, I've heard of that person. That probably was the person who brought me to you.
Stephanie - 00:23:47:
Exactly. Yeah.
Rand - 00:23:48:
Anyway, I loved it because once they put it in there, they realized, oh, this is a terrible idea. We should not be asking this. The other reason it's a terrible idea to ask, how did you hear about us? Is because if you invest your marketing dollars into places that are already sending you people who are buying, you will always overspend in places that are already working for you and never look at new channels and new opportunities. Yeah, blind spots. The whole thing is a blind spot. The very fundamental design of the question is blind spot. This is a great example of where surveys and interviews fall flat. There's tons of other places where they do. And there's tons of places where they're useful. For example, at SparkToro, last year, we hired Asia Orangio, who's a conversion and SaaS consultant. And she went and interviewed, I want to say 30 or 40 of our best customers and some people who job title-wise and work-wise looked like our best customers. And she asked them interview questions about how they use the product or didn't. And those findings, which essentially challenged a lot of our notions about what we thought people valued in SparkToro, those informed what we built this last six months, a year, launched in May and had our best three months in two years because of it. And surveys and interviews, crazy valuable. I love them.
Matt - 00:25:17:
What was the insight, if we can ask? I mean, don't diverge your strategy, but what was uncovered?
Rand - 00:25:21:
For folks who are listening who might not know what SparkToro does, a quick overview, like you might say, oh, I am a marketing professional and I want to help my business understand our audience better. So we target real estate agents in the United States and we make, you know, special kind of software that helps them track their processes. And so we're selling to real estate agencies and individual agents who own their own business. So you could search for SparkToro people whose profile contains real estate agents and you want to see which social networks are most popular with them. Might say, oh, yeah, you know what? We don't have a strategy at all for Threads and Threads is quite popular with real estate or YouTube is quite as popular or Pinterest is quite popular. Gosh, why are we putting energy into Twitter still? That's not popular with real estate agents anymore. Are real estate agents using AI tools? Like are they into ChatGPT, Perplexity, Gemini, like all these other? Oh, it looks like they are starting to adopt it. I was talking to my real estate agent a couple of nights ago because he lives across the street from us. No wonder you sold us a house in this neighborhood. And he was saying like, oh, yeah, I've been using ChatGPT for a lot of my presentations, you know, where I get the like data about a house and then I have it put together sort of structure around. Anyway, that could help tell you the answers to those things. So SparkToro does hear the websites that they visit. Here's the social networks they use. Here's the people they follow on social. Here's the topics they're interested in. These are the podcasts they listen to. Here's the YouTube channels they subscribe to. All that kind of stuff. What we thought people did was not what they did. We thought that people were using SparkToro, seeing, for example, a list of the YouTube channels that real estate agents subscribe to, and then going and pitching to be on those channels, you know, putting them into their YouTube advertising list, basically essentially checking those off or exporting them, making a list, and then using that to do their follow-up marketing opportunity. No. Turns out what they were doing was they were exporting that list, making a high-level graph, just a graph of use, and then putting that in a presentation to their director of marketing or their board, their client if they're an agency, and using that, our data to pitch people on, please give me budget to go invest in this. That was their use case. It's very high level, very strategic. They were not down in the weeds, like choosing the different thing. Maybe sometimes a little bit once they got the budget from their leadership team or client or whatever. But we were not serving that use case. Our best customers were like spending hours downloading and making graphs and blah, blah, blah. And so Asia said, dudes, you need to make your product visual. You need to make this show. Give them one click to screenshot and put it into the report, like make it a drag and drop over to PowerPoint or whatever. And we were like, okay, yeah, let's try that. So we released this new version. The visuals are like 85% of the upgrades that we did. And people went nuts for it.
Stephanie - 00:28:29:
That's such a nugget. I love that insight.
Matt - 00:28:32:
Perfect. It sounds like there was a little bit of an evolution of your ideal customer profile, really, as part of that. You thought that the people looking at the dashboards were the ones that were making the decisions on media spend. And it turned out these were, at least these were people who were at the time oriented towards data and analytics. And they had to go fight for that money from the people who were in charge of the media spend. That's wild. I think you're speaking to change that probably resonates with a lot of people in the consumer insights industry, where there is just a shift in the way people consume data, who's consuming data, how they're consuming data. And it's affecting us all.
Rand - 00:29:07:
The other big one, which I'm sure Consumer Insights World is steep into, is because of the media hype cycle around AI, we made like one of the top graphs in SparkToro, the which search and AI tools is your audience using, in what order and how much more or less than average user, the average American, or we cover UK and Canada as well, but not any other country. So seeing a chart where it's like, oh, okay, ChatGPT is the seventh most used search and AI tool behind like Amazon, Reddit, YouTube, Google, Bing, like that really gets me to a place where despite the media hype cycle, this is not the number one or two most important place. If I'm putting no effort into my Reddit presence or my YouTube presence, maybe I'm being misled into putting all my eggs in the ChatGPT basket right now. That graph, probably more than any other, has been the one that people copy and paste and tell us makes a big difference for them.
Stephanie - 00:30:07:
That's so interesting because it really challenges what the hype cycle tells us. This relates to a question that we have that we wanted to ask you. In your blog, you mentioned the power of listening to raw and filtered audience voices on platforms like Reddit, for instance. Get that, love it, do it too. However, we know that social media can sometimes provide biased or skewed data. How do you ensure that brands are accurately interpreting these kind of informal conversations without misrepresenting or oversimplifying the needs of their full audiences versus what I kind of think of as the vocal minority?
Rand - 00:30:44:
This has certainly been a problem in the political sphere, right? Where the parties have been like torn towards their edges rather than the centers.
Stephanie - 00:30:52:
And then you find out that like 80% of people are not aware of an issue at all. Crazy.
Rand - 00:30:57:
Right. I'll tell you quite honestly, Snackbar Studio, we see this all the time in video games where a vocal minority is complaining about a particular feature or something in a game. You look at the overarching data, right? Like the full Steam reviews, the sales numbers. This is noise, right? This is not signal. I think that two things are important to consider here. One is there's almost always a source of truth. It's hard to get to sometimes, right? But like a statistical source of truth. I really like Clickstream data for this. Clickstream panel data tends to be because you're getting millions of devices. Sure, it has bias. Absolutely. And you can try and control for that. That's what the Clickstream panel data company, you see a SimilarWeb, Datos, which is the one that we use. There's a bunch of other, Comscore, right? Whoever. That's what they try and control for. But compared to, hey, here's what all these Reddit comments say. It's often much more revealing to look at Clickstream data and say, okay, this is popularity. These are trends. This is what people are actually reading. Sure, this has 700 upvotes on Reddit, but that is not indicative of the broader market. And also, if you're a marketer, you should be very thoughtful about potentially being okay with overweighting that feedback. And the reason why is because those noisy few people who make the most complaints on Twitter, Reddit, Facebook, LinkedIn. If you're in B2B LinkedIn, this matters a lot. Those are the same people who, if they weren't saying negative things or if they were saying positive things, would probably help your product reach way more of your addressable audience. So it's a delicate balance, right? Looking at the data and saying, hey, 85% of our customers are really happy. Ignore these Reddit Threads. I understand that from a product perspective. But from a marketing perspective, those angry 15% or 1%, I don't care, they can make a huge difference in whether you're getting a flywheel effect in your marketing or whether that is a source of friction. One of the great ways to look at this is how visible is this negative stuff? How engaged with is it?
Stephanie - 00:33:13:
That's what I was about to say, because I may not ever comment on Reddit, but boy, I read a lot of Reddit.
Rand - 00:33:18:
We see it. You know, if you search Google for a brand name or a comparison or something, and then you see a Reddit thread where users who have even a few upvotes are complaining about a thing, you know it's going to be tough. This is the same thing, you know, that I talk to small business owners about in their Google Maps reviews. A small number of vocal critics can drive away a tremendous amount of business.
Stephanie - 00:33:43:
Watch out. Watch out. I like it.
Matt - 00:33:46:
You kind of nibbled at this a little bit. So I'll just ask, why do you hate AI? You're on LinkedIn. You're picking fights with these people, all these guys that they're just hustling, man. They just want to sell you their ChatGPT prompt list for $14.99. Come on, Rand. Why can't you get behind that?
Rand - 00:34:04:
I do kind of hate AI. I hate it because it is poorly regulated when it should be well regulated. I think my wife, Geraldine, asked me maybe a couple of weeks ago, she was like, hey, what's an ethical way to use AI? And my response was, there's no ethical way to use it. I use it all the time, but you can't use it ethically, right? Because you know the source of how it was built. And how it was built was essentially companies with hundreds of millions and billions of dollars of investment took content that they did not have a right or permission to take from all of us, right? People who created content. They didn't ask for permission. We didn't know they were collecting this data. We might have said, yes, Google, you can crawl this, but we never said, hey, Sam Altman and OpenAI, like, feel free to take all our stuff and then use that to train upon for free with no copyright laws or rules around this. It certainly does not, in my opinion, fit the dictionary definition of or the legal definition of free use. That is, I don't know how a legal scholar gets there, but because Sam and his crew are both charismatic and extremely powerful and have billions of dollars and fund a lot of Senate campaigns, they were able to go to the legal frameworks, the powers that be in the United States and basically say, hey, either you give us unfettered permission to do what we're doing or... China will win at AI and the United States will fall behind it. And I don't totally, I don't believe that. I don't buy it. I think that there are plenty of ways to get around the copyright. I do think it is more ethical to use it for text and programming, absolutely, than it is for image generation and video.
Stephanie - 00:35:44:
Or creative uses.
Rand - 00:35:46:
Yeah, right? Stealing from artists and compensating them not at all, that is bullshit. Like, there's just no way around it. So this is why I hate AI. But that does not mean that I don't recognize the value, the use cases, that I don't use it myself sometimes, that we have used it inside SparkToro to do some clever and interesting things. But I don't say we're AI-powered or AI-first or any of that junk, because I don't need to sell VCs on my business. I think it's deeply effing weird to say we're AI-powered, we're AI-first, our AI software. It's like saying, you know, in 2001, our website was built with code.
Matt - 00:36:26:
I mean- Right.
Stephanie - 00:36:28:
Right, right, right.
Rand - 00:36:31:
Yes. Like we understand that technology that exists and is quite popular is almost certainly behind your side. Why in 2012 would you be like, oh, we have a React powered front end UI? Cool, man. Like, why is that of interest at all? It doesn't make any sense to me. I can't for the life of me figure out why people market themselves as AI first. You want to use it on the back end in interesting and useful ways that solve real problems that your customers really tell you that they have? Great. You know what? Go for it. I get the whole like everybody else is doing it. I kind of have no choice. Yes, we're in a capitalist, ethically compromised government not doing its job scenario. I'm out here with the rest of y'all. I'm not going to be a hypocrite and tell you not to do it. Also, anytime there's a hype cycle, whether it was, you know, crypto or Web3 or any of the recent ones, NFTs, virtual reality, like all that stuff. Anytime there's a hype cycle, which there is right now. Always get a lot of hucksters and grifters. Yeah. People with bad intentions and bad behavior, taking advantage of people who don't know what they're doing.
Matt - 00:37:41:
If you put a wanted poster up of all of the tech grifters from the last 10, 15 years, you know who would be there, Theranos, you know there is a spot for the AI face of that side of this whole motion just waiting to be filled.
Rand - 00:37:59:
I think you can trace a direct line from... You know, the Sam Bankman-Fried's to the Bernie Madoff before him and AI grifters. If you do a Google search Sam Altman history and you it is just awful, absolutely awful. I mean, the credible claims of all sorts of terrible behavior, you know, in personal life and professional life. It's just disgusting. And it really surprises me that the board of directors a few years ago when they sort of learned of all this behavior. Remember, there was that brief time where they sort of ousted him from the company.
Stephanie - 00:38:31:
Yes.
Rand - 00:38:32:
And then thanks to nefarious network of power and money, he managed to take power back. It will not fill you with joy and hope for the future of AI. But hey, sometimes words that come after other words are really useful. And AI can do that for you.
Stephanie - 00:38:48:
AI does do that really well. I think that's a fair comment.
Rand - 00:38:51:
Spicy autocomplete. It is great.
Matt - 00:38:55:
Well, you heard it, folks. Rand Fishkin wants us to lose the AI race to China. He said it word for word right there. That's the soundbite. We'll ship that soundbite over.
Rand - 00:39:06:
It feels like a deeply racist argument too, doesn't it?
Matt - 00:39:09:
It's absolutely. Oh, and it's the oldest one in the book. Automotive, energy. It's been around forever.
Rand - 00:39:15:
That's real weird. I will say this. The other thing that I hate about AI is people who use it to do things that it's not good at instead of to do things that it is good at. Yeah, drawing a distinction between those two, being aware of, I think if you're using it, you should understand how it does what it does. There are several really good articles out there. The Nielsen Norman Group put out a really nice, incredibly readable visual article that shows you the simple structure behind what neural networks, which sound all fancy and whatever, are doing. There's another great one from the founder of Wolfram Alpha. There's a great article from Stephen Wolfram on his blog about how LLMs work and how ChatGPT does what it does. After you read that, it ceases to become magic and it starts to become statistical probability and very accessible. And then you really understand why, for example, there's these viral stories. I'm sure you guys have seen these, right? Like the viral stories of when, you know, when someone threatened to turn off Claude, it responded that it would kill the operator first. This is not Terminator. Like, no one gave Claude the ability to fire lasers, you know, at people behind desks. And also, it's not threatening anyone. It cannot threaten.
Matt - 00:40:34:
It doesn't have intention. It doesn't understand.
Rand - 00:40:37:
All it is doing is saying words that frequently come after other words, probably from a lot of movies and television scripts.
Stephanie - 00:40:44:
I was going to say, we have too much experience with Skynet. This is what happened.
Rand - 00:40:49:
So that kind of breathless, non-investigatory journalism about AI really bothers me too.
Stephanie - 00:40:56:
I will say just in this vein, though, that it's part of what makes it challenging is that AI doesn't tell you very often when it's bad what it's doing. Right. Although I did have a use case where it was like, hey, no, no, I can't do this, which was I was trying to remember some specific song lyrics. So it's like, hey, ChatGPT, this, this, this. It was like, I absolutely cannot do this thing that you're asking me to do. It was the only time it's ever said that to me. It was like, don't use me.
Rand - 00:41:25:
Is it because of licensing? I don't think
Stephanie - 00:41:27:
so. I think it's because it's probabilistic words. Right. Like you said. And so it can't tell you like, oh, it's this song. It's going to write a new song for you.
Rand - 00:41:39:
Great point. I also find it fascinating when people post in this is one of the OpenAI forums where people give feedback. There was one that had gotten a lot of recent attention where someone had said, hey, I asked ChatGPT to add some numbers together from my spreadsheet. Why can't it do addition? Like surely, it has a calculator function. And then a bunch of people are like, okay, you don't understand what this technology is.
Matt - 00:42:03:
What it does.
Stephanie - 00:42:04:
It's a bum one.
Rand - 00:42:05:
So these are all problems that require people to be carefully thoughtful about what it is and what it does and what it doesn't do. I think that care, unfortunately, is missing because it's a product that, with the exception of, you know, Stephanie's example, will almost never tell you, I can't do this or I'm bad at this. Or, hey, this thing that you're asking for is probably not a great thing to ask me. You should go use another thing.
Matt - 00:42:28:
Totally agree.
Stephanie - 00:42:30:
Hey, Rand, what's one piece of advice as we close out here that you would offer to somebody who's just starting out in like the audience research space, the consumer insight space, or if you want to, just the startup space? I'll leave it to you.
Rand - 00:42:44:
At the early stage, the thing that I generally recommend is rather than go and try and collect a bunch of data and then see where it applies, I would list out the problems you actually want to solve. What are you trying to solve with consumer insights, audience research data, whether it's from SparkToro or surveys or interviews or another piece of software? Rather than going exploratory, let me collect everything and then see how it fits. I think there's use cases for that in later stages. But if you're starting out, list your problems, then let's see which piece of data can answer them.
Stephanie - 00:43:17:
Makes a lot of sense.
Matt - 00:43:18:
And then kind of in that same vein, if you had to list something you are excited about coming our way, be it a technological shift, change in the way audience research might be used by brands, what do you see coming down over the next few years that you're excited about?
Rand - 00:43:30:
So at the broad level, I'm not a super proponent of what all the platforms are doing, where they're taking traffic away. Google is sending far fewer clicks than ever. AI tools, of course, are answering things right in there. Pinterest and YouTube and Reddit and LinkedIn and Facebook, like all the algorithms are biasing against content that contains links. That kind of sucks. But the response to it from marketers is something I'm excited about. I think there will be renewed creativity in how people post, in what they post, in how they drive interest in products through Zero-Click Marketing. I'm really excited about the world of Zero-Click Marketing. It's a little bit like 20th century marketing for the digital age, right? You get to go back to like, how do I make a TV commercial that's compelling, but for LinkedIn? You know, how do I make a billboard for Pinterest? And I think we're going to see some cool stuff there.
Stephanie - 00:44:23:
I love that.
Matt - 00:44:25:
A reduction in quantity, leaving to an increase in quality. I think that's very, very reasonable to see.
Rand - 00:44:30:
Necessity is the mother of an invention.
Stephanie - 00:44:32:
Exactly. I love that. Thank you so much for joining us today. It's been an absolute delight.
Rand - 00:44:38:
Thanks for having me. This was a pleasure and I hope we get to do it again soon.
Stephanie - 00:44:43:
The Curiosity Current is brought to you by AYTM.
Matt - 00:44:47:
To find out how AYTM helps brands connect with consumers and bring insights to life, visit aytm.com.
Stephanie - 00:44:54:
And to make sure you never miss an episode, subscribe to The Curiosity Current in Apple, Spotify, or wherever you get your podcasts.
Matt - 00:45:02:
Thanks for joining us, and we'll see you next time.
.jpeg)
