Fandom runs some of the biggest communities on the internet — can CEO Perkins Miller keep them happy?

Today, I’m talking with Perkins Miller. He’s the CEO of Fandom, which runs thousands of wikis for everything from Disney and Taylor Swift to The Matrix and Grand Theft Auto as well as several publications like TV Guide, Metacritic, and GameSpot. It’s a big, complicated media company in a big, complicated time for media companies.

Fandom started as Wikia — the commercialized sister company to Wikipedia.

Wikipedia is the noble community-supported encyclopedia, and Fandom is the explicitly profit-driven entertainment platform, and that means a lot of things that often collide with how people think about wikis. For example, Perkins talks about Fandom like you would any other large ad-supported media company with tons of user-generated content: millions of people contribute millions of pieces of content to the platform, and Fandom surrounds all that content with ads and then uses all that data to generate insights about how fans think about their favorite games, TV shows, and movies.

And while you might enjoy the content and communities on Fandom’s thousands of sites — looking up Baldur’s Gate quest walkthroughs or ridiculous Star Wars character names — the commercialization of it all means a lot of people have complaints about the user experience, particularly, the sheer number of ads. So I asked Perkins about all those ads covering every Fandom wiki: How much is too much? Is it different on mobile, where screen real estate is at a premium?

We also talked about what it means to host user-generated content in 2023. If you’re a Decoder listener, you know that one of our big ideas is that content moderation is really the product for a social network. What’s Fandom’s role and responsibility when it comes to managing toxicity and inclusion, as some fandoms, like Harry Potter, outgrow their creators?

I’m also gearing up to host the Code Conference in September, and I’ve been thinking a lot about AI, search, and the web. Every user platform is getting flooded with AI content, and search is changing as Google keeps more and more of that traffic for itself by just answering questions with AI. What does that mean for a wiki platform — especially a commercial wiki platform that’s full of guides about how to beat Zelda shrines? Perkins has a pretty optimistic view here, which I found refreshing, but I definitely pushed him on it.

We also talked about the general state of media, especially games media, which is pretty rocky right now. Like so many media companies, Fandom recently had layoffs, and I wanted to know if Perkins sees a rebound coming. And then, he has even weirder problems than a traditional media company: entire communities and wikis, like The Legend of Zelda community, have left Fandom recently. That’s a very different kind of labor issue, and I was curious to know how Perkins thinks about managing it.

There’s a lot in this one — Fandom is part of the internet’s fabric in a lot of ways.

Okay, Perkins Miller. Here we go.

This transcript has been lightly edited for length and clarity.

Perkins Miller, you are the CEO of Fandom. Welcome to Decoder!

Great to be here!

I am really excited to talk to you! I think there’s something really big happening on the internet with our platforms, with communities. Fandom is right at the heart of it. It’s a company that’s been around for a long time, and it seems like it’s going through a lot of changes of its own. We have a lot to talk about.

I just wanted to start at the very beginning for people who are coming into this. They’ve probably encountered a Fandom site in the past, but maybe not Fandom itself. The company’s been around since 2004. It started as the sister to Wikipedia in a more profit-driven way. Explain to people what Fandom is.

Fandom started, at its core, with Jimmy Wales, who is the founder of Wikipedia, having this moment and thinking, “You know what? There’s probably more that I can do,” believe it or not, and creating the world’s greatest encyclopedia to do more for people to celebrate their nerdom and fandom, frankly. And so, he spun out on the same Wikimedia platform first this idea of WikiCities. And he thought that communities would come together and coalesce on a geographic basis. But what happened over time and over the years is that people started to merge and celebrate these imagined worlds.

You had these people coming around Star Trek and Lord of the Rings and World of Warcraft and every anime series you can imagine and started to document and celebrate the lore. And it became the home of the canon of these imagined fantasy worlds across gaming, movies, and TV.

We renamed it Fandom probably about seven years ago. I joined about 4.5 years ago, and we’ve been focused on this idea that these wikis, or these blogs — 200,000-plus of them, and they reach over 300 million people a month — are something that we can build on in order to try to celebrate people’s passions for these imagined worlds.

And so, we have been growing Fandom, these core wikis, but we’ve also been acquiring businesses. We acquired a business called Fanatical, which is the leading online commerce platform if you want to buy an online video game. We bought a series of businesses that are for people who are trying to navigate the world of entertainment, so we own Metacritic and TV Guide and GameSpot and Giant Bomb. These are platforms that allow people to navigate their way into what to watch, where to watch, how to play.

We reach, now, probably 350 million visitors a month across all these platforms around the world, and the mission is pretty straightforward. I mean, we believe that fans have this sense of identity. Personally, I’m a Star Trek nerd, and I read silly fantasy books when I’m not reading all my silly business books. There’s a part of my identity that’s grounded in sci-fi. And so, I love a place where I can go down the rabbit hole and discover what I want to watch next and what game I want to play next. That’s been, over the last almost 20 years, the evolution of Fandom as this platform, and people do find their way into this, again, this matrix, as they’re trying to figure out what to watch or what to play.

I think about that a lot. You watch something — let’s use Game of Thrones as an example. I think this was peak wiki as you’re watching television, in American culture at least. You’re watching something, it’s complicated, there’s a lot of lore. You need to figure out what’s going on. You’re going to fall down the rabbit hole of some wiki or another that’s going to explain everything to you.

At its core, that is a user-generated content platform. You have people who are contributing for free, you have some moderators, you have some editors perhaps, and you have an underlying platform that is a technology product to enable all of this to happen.

If you squint, it looks more like a YouTube or Instagram or anything. What are the differences and what are the similarities as you think about, “Okay, I run a giant user-generated content platform”? It’s not Wikipedia, right? Wikipedia is constantly asking people to donate money. It’s free — it has a volunteer ethos.

You’re running something at scale for profit, you’ve got a big private equity investor, and you’ve got the same challenges as the other big for-profit user-generated platforms. What are the similarities and differences there?

It’s probably best to start on the differences. Fandom, because we’re dedicated to the pursuit of these imagined worlds, we lose some of the baggage that you get on more traditional social platforms where there’s a fair amount of toxicity that can get generated, where people are out for their own self-interest. You see a lot of this idea that there are big influencers trying to nudge their way to the top.

Fandom communities are really about celebrating what’s going on with Game of Thrones. What are the icons that are involved in that story, and what is the story arc, and what is the dynamic? How do dragons play the role? And they want to debate the conditions and canon.

The content creators for our platform — who are awesome, they really are the experts in these spaces — come together as a community, and we have a whole team dedicated just to support these communities. We want to make sure the tools work, we want to make sure the platform’s stable, we make sure that they have the best tools in hand to do the work they want to do. And so, I think that’s a little bit distinct and different from a traditional user-generated content platform, where it’s really about celebrating the IP and creating the tools to do it. We run advertising on the platform to fund that. We are not apologetic about it. We do our best with data to try to make sure it’s as relevant as possible, and we try to get it as right as we can.

But at the core, the fact that we have creators creating content in celebration of these IPs is what’s really distinct and unique for Fandom. We’re similar, I guess, in the sense that we provide a platform for voice for people, which I think is really important. In this new media landscape, everybody can be a creator, and everybody can have a voice. And I think, just like every other platform, Fandom provides the tools to do that.

It just so happens that our community is really focused around galvanizing around the IP and around the story and less interested as much in personal aggrandizement or personal performance. So it’s much more about: What do I know about Star Trek? What do I know about Fortnite? Those are the things that really are celebrated on our platform.

So first Decoder question, how is Fandom structured? How does your org chart work?

We are very functionally oriented, and we try to be as flat as we can. And so, we have a chief product officer, chief technology officer, chief marketing officer, chief business community officer. And then we have a chief revenue officer. And then we have a strong G&A group with a CFO and a CHRO and a general counsel. So that’s broadly how we’re organized in terms of functional, and that’s product and tech, community, business strategy, marketing, sales and commerce, and then the G&A structural support. That’s how we’re organized.

That cuts across all the brands, so the TV Guide app is in the same product organization as the main wiki for Game of Thrones?

It does. It does.

Do you see overlap there? Does that work for you? Do you move as fast as you want to?

We do. I’m very sensitive to the idea of focus, and I think the test for us is, we think about fan identity, how broad a reach is it between “what do I want to watch?” and “what is it I’m watching?” Or “what game do I want to play?” and “how do I play the game?”

And those are actually pretty connected issues for me, even to the point of “what game do I buy?” That’s one of the reasons why we’ve been pretty disciplined about our operating structure that we’re like, there’s not a lot of daylight between “what game do I buy, what show should I watch, and how do I play that game?” and “what does this show mean and can I buy that game?” As a consequence, because the connective tissue is so tight, having a structure that’s functionally oriented gives us actually a lot of leverage.

An interesting part of the puzzle here is you’ve made a bunch of acquisitions since you started. We’ve talked about TV Guide, Metacritic, GameSpot, Giant Bomb. You also made big cuts at those, right? You acquired these properties, they had staffs, they had identities, maybe they were in service to your mission or maybe orthogonal to it, but you bought them and immediately cut a bunch of editorial employees.

Was that in service of integration? Were you trying to say, “Okay, these are just brands, they’re the face of an operation that helps you figure out what to watch and then watch it”? Or, “Actually, this editorial property doesn’t make any sense at all.”

As we talk about restructuring and acquiring businesses, there’s always going to be overlap. I mean, there’s just going to be some redundancy that you just have to work through, and that’s what a restructuring is. And then, there are some decisions that are business-driven around how we set ourselves up for success with the right cost and revenue basis.

I’m an operator that believes that we really want to be oriented around profitable growth and being able to at least see down the road far enough to decide which way to turn the wheel. As a consequence, when you acquire a business, especially one that from a business standpoint was struggling, in order to give me enough and give us, as operators, enough perspective to see down the road, you’ve got to pump the brakes and be able to see down the road far enough to know, “Okay, we need to go right with this business or left with this business.” And that, if we’re able to do that, get enough visibility, then we can push behind investment.

I mean, you look at Fandom as a whole, and that’s been our practice. We’ve added a lot of staff every year, and we’ve hired a lot of people, but we’ve done so because we’ve always been able to see down the road far enough that we know which way to turn the wheel and where to make the investments.

Again, it’s a tricky business, and I really understand the impacts, of course, on individuals. I take it incredibly seriously, but I’m also a CEO of a company that is tasked with driving profitable growth, so that’s the balance you have to strike, and that’s the approach I took when we were looking at how do we build this set of brands to the next level of growth where we just need to get some perspective and put them in the right orientation so we know which way to turn and which way to place investment.

I understand that. I think my question is: when you look at the acquisition of a Giant Bomb or a GameSpot, these are two of the largest editorial outlets in games, and you say, “Okay, we’ve acquired them, I’ve looked at them, and now I’ve got to cut most of the editorial staff,” what did you think you were buying? How did it help you achieve the larger mission for Fandom?

I want to be clear, we didn’t cut most of the editorial staff. We left a very significant editorial staff in place, and they’re very good. We also have editorial staff that was in place at Fandom before. We had an incredible team and Screen Junkies and Fandom Productions, so it is always a difficult decision to think about, “Okay, well, how do we get two plus two to equal five as we grow these companies?”

But it wasn’t one where we said, “Look, we’re going to leave these businesses in a position where it’s going to be difficult to invest and drive growth” because we believe in the brands. I 100 percent believe in the team at GameSpot and the team at Metacritic and the folks at Giant Bomb and a couple of people who work in GameFAQs and TV Guide. I mean, these are really good brands. The question is, again, we’ve got to be smart about how we place them in the right position to drive that growth. And that meant to prioritize the work before us.

The strategy for buying these businesses, I think, was pretty straightforward. It was, as I talked about, that connective tissue or those connective dots. Its fans have questions about what to watch, where to watch it, what game to play, is it good. And we have this huge bulk of information to help people enjoy the movie better or play the game better. So, we just really wanted to have this point of discovery and this point of engagement, which we didn’t have. And I think those brands, the GameSpot, the Metacritic, TV Guide, Giant Bomb, are really representative of the best in class for getting people engaged at that point of discovery.

Do you think that that is compatible with how games journalists might have wanted to see themselves in the past? There’s a function there that you’re talking about, which is basically getting you to buy something, right? All the way down at the end of that road is, you’re going to spend some money and play a game or watch a movie, and part of your advertising partnerships are obviously with studios.

Fundamentally, every entertainment property has this inherent tension. I don’t think it’s unusual or new, but there’s something there in games in particular where games journalism gets pushed and constrained, and you see it across the entire industry of games journalism, to being like, “Okay, this is a marketing function for games.”

Is that something that you feel now, or is that something that hasn’t really come up?

I don’t feel that. We have a lot of gamers on our team here, and I’m probably the weakest among them, to be clear. But I find, and again, what we find from our research, too, is that people are so curious about games and the community around game launches and changes to games and that role that the journalists provide in terms of giving people those signposts: your favorite franchise, GTA, where’s it going? And I think that role of essentially being able to put the markers out on the road to get people to say, “Here’s where we need to go follow.” I think that role is really important, and I think the team does a really good job, and I think there’s relevancy there. It’s not just all transactional. And I think what we think about is, “Well, where are the places our voices need to be broadcast?” And I think those platforms change.

So we have to do things on TikTok and have to be on Snap and have to do things better on YouTube. And we think about, “Well, where does that loop back on a wiki? How do you show that content so people get a sense of what’s there?” And some of it is transactional, so we have Fanatical, and if you’re reading a review on GameSpot, can you go by that game? Sure. Do we want to make that as easy as possible? Absolutely.

But I do think that the role of games journalism to provide, again, those signposts for these franchises for people who are passionate about them is incredibly important. And I don’t see that going away because we know fans have these passion points, and they want to consume everything they can about the things they love.

Last Decoder question: How do you make decisions? What’s your framework?

Well, I am very much a calculated risk person, in my personal life and as a leader in business. I’m pretty data-driven. I consume a lot of data because I find it to give me really important context. And so, the way we make decisions here is, let’s focus on a couple of things. No. 1 is: does it serve the mission and vision for the business? No. 2: what does the data tell us? Is it something that is strong and growing, or is it weak and failing? Just so we can get some directionally correct information. No. 3: can we see a comparable in the market around us? I very much believe that there are many lessons to be learned by just looking around. What did Reddit do? What did Pinterest do? What’s been happening over at IGN?

So these are things that we can look at to give us directional guidance on what works and doesn’t work. Then we again pretty quickly frame up the opportunity. So is it a breadbox or a T-shirt exercise? How big is it? And then, we go from there and make the decision as quickly as we can, and we know we do so with imperfect information.

I think Bezos’ note about 70 percent good is good enough is exactly right. So I try to be as data-driven as we can, and again, use those scenarios and the idea of: can we look around us and see what’s working to give us a little bit of directionally correct confidence? And then make the decision as quickly as we can with imperfect information.

You’ve mentioned IP several times now. I’m an old IP lawyer, the listeners know this is bait for me. The properties you’re mentioning, Star Trek or Game of Thrones or Zelda or whatever it might be — you don’t have a formal relationship with those, right? Those are other people’s IP that you’re building communities around. Is there a connection? Do you have a team that goes to Nintendo or one of the streamers or Netflix or whoever and says, “Hey, we’re building communities around your IP, we should work together in some way”?

We do. I mean, we have hundreds of official wikis on our platform, and we hear all the time about how producers and writers actually use our platform when they’re creating content. Because a lot of these major franchises have long story arcs, really complex narratives, and they’re like, “Wait, what happened in that episode?” Or, “When was that? Okay, that was 20 years ago. Wait a minute, we want to get this right so we can connect the dots correctly.”

So we do, we have these official relationships because I think we provide a real service because the communities are so dedicated to this IP. They’re really, in a lot of ways, the source of truth. And as you think about the evolution of IP holders, I mean, people are going to transition. Somebody who’s writing for Star Trek 20 years ago is different from somebody who’s writing for Star Trek today. And to have one place where it’s a source of record, I think is incredibly valuable. We work very closely with the IP holders around it, and we’re a good source of information. We do surveys and polls to help everybody get perspective.

So we actually think of ourselves as this key partner to these folks. I mean, we’re not creating movies, we’re not making video games, so we’re in the business of just celebrating the content that they create.

There’s a real tension embedded in that. I think that’s fascinating to push on. The official community for a given property has some reputation associated generally on the internet and then an unofficial community has a different reputation. Do you see a difference in usage, engagement, and loyalty between your official communities that you’re partnered with studios on perhaps and your unofficial communities?

I don’t see a lot of it. I mean, it’s very hard to paint this with one brush because those communities are all different, and there are different degrees. I mean, you can think about the Harry Potter communities and Pottermore, and that’s a very large community of people that actually has got a lot of official tailwind behind it. We still have those communities on our platform, and they’re very engaged.

And one of the things that I find is, because there’s sometimes a change in objective that a company may have who may own the IP, they may have a change of heart about what to release or a change in creative vision or whatever it may be or a delay in the game. And the distinction I see is that, because we’re community-driven, it’s really just about what fans want to talk about and look at and play. And so, there’s no agenda other than to celebrate your nerdom.

So actually, 85 percent of the traffic to our platforms is on IP that’s already released. And this is all just people going back to look at The Original Series or TNG [The Next Generation] on the Star Trek side, or to go back to World of Warcraft Classic and they’re trying to level up because they’re rediscovering the game, or whatever it may be, that they’re an anime series that is just replaying on Crunchyroll.

So there’s all this consumption of content and entertainment that happens all around the world that’s not necessarily tied to the sharp end of the spear of a release date. And so, I think that’s a really key distinction, and we try to work with our partners in this and explain: you’ve got legacy and heritage behind a lot of these franchises, and the great thing about Fandom, as a category and as a business, is that we actually dedicate ourselves to supporting these franchises over the duration of their existence. So we’re about the last 50 years. We’re not necessarily just about what may be launching in six months, so that’s a really key distinction.

You opened the door to this question by bringing up Pottermore: That’s a really complicated fandom. There’s a studio, there’s J.K. Rowling, and there are a bunch of fans whose relationship with J.K. Rowling and that studio have dramatically changed over time as some of her views have come out. You sit in the middle of it, so if that community is mad at J.K. Rowling, what is Fandom’s approach to that if they want to go somewhere else? You want to keep those users, but her relationship with them might be driving them away. How do you manage that?

We have a couple things that maybe you can tease out of that. So one, just to remember, we’re communities in celebration of these virtual worlds, and we think that the approach to them is really about: let’s talk about Harry Potter, let’s talk about Hogwarts, and let’s talk about what the magic is that’s around that franchise. There’s of course going to be discussion along the way if things get politicized or there are different agendas. We’re really, again, focused on the imagined world that people want to celebrate because that’s the IP that is meaningful.

And we also, though, on the other hand, as it relates to inclusion and diversity, we have a very strong policy of inclusion. If there’s toxic behavior on the platform, we don’t allow for that. The admins on our platform are incredibly supportive of that because it’s really important that we allow people’s voices to be heard across the world. And again, it’s an imperfect task. As I remind everybody every day, this is not something where there’s a finish line, and there’s going to be nuance, and there’s going to be debate to occur.

But really, if we sort of hold two things true for ourselves, which is: we’re dedicated to these imagined worlds and all they mean to people for the imagination, and we just make sure that the platform and the people who are on it to celebrate it behave as best they can and are inclusive of one another. That’s the role we play. And we let the other business side that may be churning around a given IP or some other political questions that churn around an IP, we leave those to the side and just stay focused on that community in the imagined world.

Well, there is a tension there, though — Harry Potter, I think, it’s all in there. It’s a huge franchise. It’s a massive business for Universal. J.K. Rowling is a famous person. There’s a huge community. You have a policy of inclusion. My thesis about every platform is that the product is fundamentally content moderation. That’s the thing you’re making that makes everybody participate.

Are you at the point where you’re saying, “Look, Harry Potter fans on the official Harry Potter forum, we are not going to talk about J.K. Rowling’s views toward trans people. We’re just going to stay focused on the imagined world”? Is that the content moderation decision that enables the business? Because that really is the heart of all this, right? You’ve got to make that kind of decision at scale and get more of them right than wrong.

And I think if there were anti-trans commentary — and again, this idea of these ad hominem remarks that are directed at individuals, of course, we don’t stand for that. I consider myself, as best I can, an ally to the LGBTQ+ community, especially as the father of a trans child. It’s very personally meaningful to me to make sure that I’m modeling the best behavior I can for myself personally as well as for our company. But again, it’s not necessarily about my point of view. It’s all about what is the right focus for the company.

And the right focus for the company is around these imagined worlds and making sure that, as long as you’re not going into that ad hominem so you’re not attacking individuals. If you’re really talking about, here’s how to celebrate my love of Harry Potter, that’s the lane that we want to make sure we stay clearly focused in and drive straight down the middle of.

So I think, on the one hand, you can make it really complicated, and there’s always going to be nuance, and there’s going to be bumps in the road, and you’re not going to hit it right all the time. Like I said, there’s no finish line to this, and there’s no perfection. It’s all about setting the right intention. So I think, generally speaking, we do get it right, but you’re 100 percent correct, which is, if you stay really focused on the mission of, we’re about fan identity, we’re about imagined worlds. That keeps us in that lane pretty cleanly.

And it’s one of those things, again, if you stay really focused on the mission of, we’re about fan identity, we’re about imagined worlds. That keeps us in that lane pretty cleanly.

I think content moderation questions for platform CEOs are infinitely complex, but I wanted to just ask that set of questions because it gives you a sense of what Fandom is, right? Making a user-content platform this commercial carries with it a different set of challenges than I think the more wide open platforms you might otherwise compare it to, the Reddits and Instagrams of the world.

Here, you have a focus, you have partnerships, and I think the shape of that is meaningfully different, and I wanted to make sure we got a hold of that because my next question is, okay, how does that make money? Is it just display advertising? Is it integrations with the brands? Where does the revenue come from?

The revenue comes in two big buckets. I mean, one, we do sell games. So we’re a reseller of games on Fanatical, which is a smaller part of our business, but it’s growing really meaningfully because we have a great team of folks that create great bundles and packages of games for people to buy and play, which is awesome. We also have a small subscription business on Giant Bomb. So we have some of the greatest voices in video gaming that people subscribe to on Giant Bomb. It’s also awesome.

And then, we have advertising, and the advertising that you’re always trying to do is get the right balance between engagement, so we want to make sure people come to the platform and get satisfied with the answers they get and also the creators can come create the content they want. On the other hand, it does cost us money to support and develop and try to build these experiences, which you think are really meaningful, and give people this discovery.

So the advertisers come in, and we run brand campaigns, we do live events, we do the largest party at San Diego Comic-Con, for example, where we had a Korean band play, and it’s a huge event. And we do the same thing across other comic-cons around the world. So we have a really good events activation team. And then, we do work across our social platforms to make sure people’s voices get out there.

And then, we do things that essentially cross over to different platforms. So if you’re thinking about a movie release or a game release, that’s going to be better architected around: what can we learn from Metacritic and driving into the wiki? Or what can we learn from GameSpot, and how do we marry that up in a way that maintains our editorial independence but then gives those brands a way to connect to the fans? So, it’s a mix of those things — the mix of display and brand advertising, events and social activation — that we bundle together in a way we try our best to be elegant about.

It sounds like your big clients there are game studios, entertainment companies.

We do about 50-50. So we have a split between what we consider our endemic partners — that’s the game studios and the movie studios and the TV streaming companies. And then the non-endemic, so everybody from the FDA to Comcast. Those are platforms and businesses that are trying to reach audiences, and we happen to be a very significant collection of young folks who are interested in gaming and movies and TV. And so, we will work with them to help them reach those folks.

The advertising market right now is really weird. Full disclosure: The Verge is an advertising-supported business, and I’m confident that Decoder listeners will hear an ad soon in this episode. It’s a tough market right now, right? Meta and Google dominate online advertising. The ad agencies seem to be doing well, but that isn’t trickling down. I don’t know what’s going to happen with Twitter’s advertising business. Instagram Reels seems to be making more money than TikTok, which is not what we’d expect from their cultural relevance.

Where do you fit into that ad market? Are you competing for dollars against the Metas and Googles, or do you have a different set of competitors there?

I mean, it’s hard to lump ourselves into a Meta or a Google conversation, just the scale is so–

Well, it’s just like if you look at the pie chart. It’s like, that’s where the share would come from.

Exactly. That’s right. I think, what we do differently… you’d probably look at the business and see the same thing as I do, which is we’ve got a lot of signals as folks come in. If you’ve got 350 million people every month coming in and looking at entertainment content, you’ve got some really interesting insights in terms of what people want to watch or what games they want to play.

What we’ve built is this platform called FanDNA. And what it does is basically say, “Okay, hey look, we’re not taking any personally identifiable information, but we’re able to see patterns of behavior.” And what we’ve been able to build through FanDNA is a set of insights, which we work with these partners to help them just make a better match between the ad message they have and the customers we are trying to reach.

And I think we do a really good job with that. And because we also know what’s relevant to fans because we are at Comic-Con and I put on my Star Trek cosplay, we know the language.

Wait, are there actually photos of you in Star Trek cosplay?

There are actually photos of me. Because we are fans, we know what fans like and love, and we have these insights, that’s where we’re distinguishing ourselves. You can go try to buy just basic demographics anywhere, you can buy reach anywhere, but it’s really hard to do reach and targeting to the degree we can do. Especially if you’re in the market of reaching fans, there’s really no better platform to do it. And so, that’s where we’ve been winning share.

There are a lot of very nerdy people on the Verge staff, as you might imagine. We told these folks, “Hey, the CEO of Fandom is coming on the show. What do you want to know?” By and large, the main question we got was, “These pages are totally loaded with display ads. The user experience of these pages is not what I like. Can they tone it down?” That has to be feedback that comes to you, right? This is the money. We have video ads on the pages. We have huge amounts of display advertising. That’s obviously the revenue. Is there a balance there that you think you have to strike? Do you think you have the right balance now?

You’re asking the age-old question, which is: how much advertising is the right amount of advertising? And I think we’re doing our best to make it as elegant as we can, and yet, we need to drive the economics. I would like there to be better matches overall because people, when they see an ad they like, don’t complain about it. They see ads that aren’t relevant to them, that’s when it tends to set them off that this is irrelevant and therefore unnecessary.

So we’re working on this idea that, hey, we’ve got these insights, can we make things more relevant for folks than it is today? And I think that’s, again, it’s one of those things that doesn’t really have a finish line. It just got to do a better job. And I think the other part of it is, we’re consumed 60 percent on mobile, and that’s just a smaller amount of surface area. And yet, we have a massively dense amount of content that people want to navigate through. And I’m very sensitive to it. I spend all my time evaluating and figuring it out with our product and tech teams.

So we do think there’s a path to make that an even more elegant experience, but it’s harder just because the surface area is more difficult to work with. So I’m never going to claim that we have it right. I can tell you, we work on it a lot, and actually, we study the data all the time because we don’t want people to have a bad experience and we do need to show ads, though. So again, it’s a bit of a contradiction that we have to manage through.

You look at what happened to Meta and Google when Apple rolled out app tracking transparency. Their revenues fell down because their ability to target changed. Meta appears to have figured it out in their latest earnings. Did that impact you the same way? You run a huge advertising business at scale on mobile phones. You have apps, you have websites. Did you feel that the same way? Were you able to say, “Okay, Meta’s got problems. We’re right next to the IP that people care about. Shift your spend to us”?

We’re much more an MWeb — mobile web — and DWeb — desktop web —  platform. We have a native app, and it’s good, but we’re much more focused on the MWeb and DWeb experiences because we have such a complex platform. We have 45 million pages of content, and we’re really trying to get people to the places they want to be in order to get the experience they want and the answers that they’re seeking. And so, we haven’t been able to find a way to elegantly do that in the native mobile experience yet.

So we weren’t impacted as much by that change that occurred that affected more native mobile apps because we’re so DWeb- and MWeb-oriented, which is fine. I do think that there’s definitely opportunity. I mean, for example, our TV Guide app is a great app, and it’s actually doing really well. And we find that’s a great platform to begin to solve a singular utility.

I think our approach and, again, I’m getting a little bit to the side of your question, which is

we weren’t as impacted by the bigger changes that Apple will deploy simply because our app strategy is not as material on the core fan wikis. But where we’d be very specific on utility, because we’re also really narrowcasting, we’re just looking at what to watch, we also were okay and have been able to weather that storm pretty well.

I’m asking these questions because I just think the shape of the internet is about to go through a little bit of a reset. And figuring out where you think the money might come from or where it might head out to, that seems very important for all of the platform companies to sort out right now. And the notion that we can just produce millions and millions of pages of user-generated content and put programmatic or cheap advertising on it, it just feels like maybe that world is over, especially as some of the distribution that we’re used to on the internet changes.

In particular, what I mean is, I don’t know what’s going to happen to Twitter, and I really don’t know what’s going to happen to search. And that, to me, just feels like the change that’s underlying the internet right now, so I’m asking these very specific questions, but I’m wondering if you feel that bigger change as well.

Maybe the underlying question is, how does Gen AI and these large language models and the indexing of the internet change over time? And I think our thesis, and I’m, as you can tell, kind of nerdy, I also do a lot of thinking in scenario structures. I try to say, “Okay, faced with an uncertain future, how do we narrow that uncertainty?”

The way I tend to approach it is to say, “Well, what outcomes could we see happening?” You could see an outcome where search is decimated by Gen AI. You could see an outcome where it’s really lumpy, and it’s not clear that people know exactly how Gen AI will impact discovery. And then there’s a path where you say, “Well, verticalization will be more important than ever because domain expertise and understanding will always have its place.”

And so, in those three scenarios, I tend to think that the third one, this idea that deep verticalization and communities will be more important than ever. Because there’s just so much information all the time, it’s all about context and relevancy, and it’s what we actually specialize in.

We’ll tell you what matters because we have 45 million pages of content around these imagined worlds, and we’ve got tens of thousands of community members who are constantly weighing in about nuance and talking about how things connect to one another and laying that out. And again, there may be a moment in the future where these language models and the way Gen AI works will be able to solve for the questions you have in a curious way and with the right context and relevancy.

But right now, and I think for the foreseeable immediate future, I think the sense of vertical deep communities being really important to people who are trying to get some esoteric information is really, really valuable. I think generic stuff like “What’s the temperature in Phoenix today?” I mean, sure, that’s a very broad, general piece of information.

Google killed my Phoenix weather business. Very upset about it.

I think that’s a perfect setup to the AI conversation, because I don’t think anybody knows what’s going to happen as AI takes over the web. It seems like that’s happening on a number of fronts. There are the changes coming to search. I’m looking at some older interviews you’ve done. You said a huge amount of your traffic comes from organic search. If Bing takes over for Google, and Bing Chat starts answering all the questions about Glee or whatever other fandom, is that an outcome you’re prepared for? Or do you think, “Okay, people are going to come to us directly because we are where the communities are”?

I spend a lot of time inside these services to evaluate and test. We do consider one outcome, which is Gen AI disrupts all discovery and the chat partner is able to quickly interpret your intention and get you exactly the right answer. This is this idea of the perfect storm scenario, that the AI evolves to the point where, of the 350 million folks we have every month, every query that that person has is exactly interpreted correctly by the AI to get that person the exact answer they were seeking.

I tend to score that probability fairly low right now in terms of any immediate action. And then, I tend to then think about: well, what are the other likely outcomes? Have people built up a set of behaviors around discovery where they’re putting random keywords into a search engine and seeing what shows up because they’ve got three things they’re wanting to put together, and Google’s done a really good job of saying, “Oh, you give me three things. I’m going to take the indexing capabilities I have and give you a series of results that I think may be in the landing zone of what you’re looking for.”

I feel like that’s a pretty good way for people to process, which is like, I don’t know exactly what I’m asking. I have a few things. I want to bang them into a box and see what shows up. And if I throw them all in there, and I put in Kirk and Enterprise and Worf, that’s when I’m like, “When did he show up on the Enterprise?” Yeah, I’m going to get Fandom. I’m like, “Oh, there’s the Memory Alpha platform on Fandom, which is our Star Trek platform, and I’ll dive into the rabbit hole, and sure enough, I’m going to find out when Worf showed up on the Enterprise. I think that feels to me like a scenario that’s likely pretty durable, just the way I see behavior working today. It doesn’t mean I don’t think that perfect storm will happen where there’s an AI–

But let’s say there’s, even to go with Bezos, 70 percent of the perfect storm. I Google “How do I beat some shrine in Zelda Tears of the Kingdom?” And Google has scraped enough of Fandom and enough of IGN and enough of whatever to deliver the answer to that question in the Search Generative Experience. And maybe they can’t for all the other stuff, right? You’re asking your question about Star Trek, and maybe you do land on Fandom for that one, but there’s a huge chunk that Google could take today because the answer is fixed, right? There’s a way to beat the shrine, and people have typed it into text boxes on the internet, and Google can read it, and they can spit it right back out at you. How do you account for that scenario?

I think that scenario, again, if there’s a query that is precise, that you really know there’s one answer to, yeah, I think that there’s absolutely the case where Gen AI and that result will work. I think you’ve seen Google–

So a huge part of games publishing right now is Zelda guides. Zelda guides targeted to search is some massive amount of games publishing right now.

Every game site publishing right now, including the ones here at Vox Media / Polygon — totally devoted to Zelda guides. Eventually, Google’s just going to be able to read all of that because part of the deal with Google right now is that we allow them to index our sites such that we might get the search traffic, and that relationship might change as they roll out more and more of their AI experience.

That’s not the perfect storm; that’s just a linear progression from here to there with Google. How are you accounting for that? Have you gone to Google and said, “Look, you need to start paying us for this training data”? Have you said, “Look, we’re going to block you from robots.txt or whatever brute force mechanism there is”? There are some big tradeoffs in there, and they’re not all “the perfect storm.”

Trust me, I have a model that calibrates for some loss of organic traffic. That’s because the tooling that powers Bard gets to the point where it’s effectively, and again, you’re probably on the prototype as I’m on the prototype to see how the Gen AI results. And today, I’m still going through to see the series of links because the answer typically is incomplete or not quite correct. So I think there’s absolutely a future — and I don’t know what the time horizon is — where a very nuanced AI who can get intentionality correct and can consistently give you the answers to the questions you’re seeking results in a very serious challenge for anybody who has content anywhere.

What we’re focused on right now is: what are the communities doing together in support of these IPs they love? And if we stay dedicated to that, I think that that remains incredibly powerful because I think you’re going to slice off some of those transactional queries, and that’s okay. Because if you have the communities of interest who are galvanized together around the IP, that’s going to be durable regardless.

Let me ask you the same question I actually asked Meredith Kopit Levien from The New York Times. The Times obviously has the same sort of relationship with Google as every other publisher, and I said, “What are you going to do if there’s Google zero? In the perfect storm, if Google goes to zero, what does your business look like?”

You can go listen to that one — I encourage the listeners to go listen to it. What I’ll tell you is, there’s the answer she gave me, and then, a couple of weeks later, The New York Times signed a huge distribution agreement with Google to use their tools because obviously they’re thinking about this, right? So I’ll ask you the same question, and tell me if you’re on the cusp of signing a huge deal with Google. What happens to Fandom if Google goes to zero, if Google zero occurs?

I think, if Google is search, if you’re saying that Google zero means every search query is consumed in its ecosystem, I think there are a couple of things that I think are interesting to consider. One is: how does new content get created? And I think that, right now, there are a lot of communities creating content everywhere, and if you lose publishing, this is publishing broadly, if you lose publishing broadly because there’s no economic basis for people to go to a website and watch or read content and see an ad or subscribe to a platform, it’s recursive. It’s going to be a problem for the AI — what will it be sourcing?

And you’re probably left with social media as your primary source of content, which, as we know, tends to unfortunately be polarized. The content at the extremes tends to be more and more prominent. And so, you can possibly assume that, in fact, the AIs will start to train on more polarized information that’s probably incorrect.

And it’s also not training on new data because those sources have essentially been cut out of the market. So I tend to think that users, though especially in our world, want authentic communities to talk about the content they love, and they want to create that content themselves. And so, we’re focused on this idea of my fandoms and creating the way for you to come and be a part of our community. We have several million people registered with Fandom today, and we think that’s going to continue to grow.

And in fact, it may be the case that, if AIs tend to get more isolated and tend to train on more polarized information that may be incorrect, there’s probably going to be an opportunity for platforms that are, again, deeply vertically focused around community that will essentially deflect that trend.

We’ve covered the changes at Reddit very closely — the proposed changes, the user outcry, the protests. A lot of that is driven by Reddit as a company saying, “We need to make more money on this. We are seeing the OpenAIs and the Googles of the world come to Reddit, train on all of our user-generated content, and then make money selling AI elsewhere. We want to get paid for that.” Squint, and Fandom has the exact same problem. Have you gone to Google, OpenAI, Microsoft, Anthropic, and whoever else and said, “Look, if you want to scrape our site, you have to pay us”?

We have not done that. We are very focused on the community creating the content and the community members themselves and creating a better platform. It’s really the thing we can control. We want to make sure that people have access to it and make sure our admins and creators are having the best tools possible, and again, we’re striking that balance of a great ad experience and great engagement, and that’s what we’re really focused on right now.

And we feel that, by having that focus on the community and focusing on the content, it’s probably the lane for us to play in. We’re very much, I think, unlike Reddit has a lot of APIs that go out to a lot of different platforms, I can understand that strategy. But for us, it’s not a right analogy.

Well, Reddit is the face of a problem with Google, right? People are googling questions and putting the word “Reddit” at the end because the best answers are on Reddit, and obviously, Google is scraping that. But you see that with Fandom as well. You see that with wikis in particular. That’s where the right answers are. That’s where the community is. Is there a point where you would go to one of the AI providers and say, “Look, we know you’re reading our site. We definitely want you to pay us”?

Yeah. I don’t know what that business model is, though, for them. I mean, I think that it’s not clear to me right now that I understand the payment that Google would make for accessing information that frankly on our side is public. And so, it’s not clear to me. I think that a Quora or a Reddit, that it’s really those discussions and the kernels of insight, the individuals providing those discussions that people are seeking — and the Redditors who are in those groups curating that. I think they just have a different model in the sense that they’re talking about everything from politics to news and how to fix a car. I mean, we’re very much in this imagined world, and so, it’s not necessarily quite the analogy that I’m as concerned about.

This is what I mean. This is all deeply uncharted territory. I don’t think anybody quite knows these answers, but it’s true that, particularly in fan communities, fanfiction communities, and creative communities, the use of AI and having AI trained on fan work is deeply controversial. Certainly in our reporting, we’ve seen controversies in the fanfiction community, we’ve seen controversies in the creative community. Giving the data to the AI is as controversial as it gets.

Using the AI is even more controversial. I want to come to that in a minute. Have you felt any call from your communities, “Hey, we’d like you to protect our work from the Google AI machine” or “We’d like you to protect our work from the OpenAI machine”? I think that’s the other pressure. There’s the economic pressure, which was going to play out as executives like yourself make different kinds of decisions across the platforms they run. There’s also the user pressure, saying, “I would like to contribute to my community. I would not like to contribute to Google’s bottom line.”

Our communities — and I can’t speak for everyone — we tend to query them and talk to them a bunch. And remember, our strength as a community is in celebrating these imagined worlds. And I think there’s a lot of concern about how Gen AI and these models are used to–

Let me give you the example much more specifically. If I go to ChatGPT today and say, “Write me a story about Kirk and Worf on the Enterprise” it just does it, and part of that story is because it has gone and read the Fandom wiki and has all of that information in it. Is there a value exchange there that feels fair to you as the CEO of the company, and then, I think more importantly, to the members of that community?

Well, that was exactly the creator answer I was getting to, which is that, I think that there’s a question about what is a derivative work, and is AI a tool for creators to do more work or is it a tool that disrupts the creator’s vision for what it is? I think we don’t know that right now, and I think that our view is that we want to make sure that we are dedicated to those creators who are building these imagined worlds. And it may be that those creators in the gaming world, you can give voice to NPCs, and in giving voice to NPCs, is that going to create a more rich game environment? Would you want Gen AI to assign personality and voice to the NPCs and allow you to play your game in a much more rich environment? And it’s much more dynamic.

There’s an example where the positive benefit of being able to apply large language models and learnings and training data to something, that’s great, and I think we all would potentially celebrate that. So I think just as the knife cuts one way, it can also cut the other. And I think, right now, the view is, let’s use the tools as best we can in order to create great content and celebrate it, without necessarily working from the position of fear, which is not yet manifest.

I want to take credit for not immediately taking the bait at the words derivative works. I’m just very proud of myself, moving right on. But everyone should know that you said it, and I avoided it.

There’s another side of this, you said the knife cuts both ways. Like you, I’ve been trying all the AI tools, and my view of the generative AI tools right now, especially when it comes to text, is that this is an absolute cannon of C+ content, right? It’s not great, it’s not wonderful. I think I can do a better job at writing than ChatGPT can do today, but it never gets tired. It never stops. It can write about anything you want for as much as you want. It is a cannon of C+ content.

And a lot of times what we’re finding is any text box on the internet, people will find it, and they will point that cannon at that text box. You run a platform of user-generated content. It has to be that AI-generated content is coming onto your platform, perhaps at scale. Has this become a problem for you yet? Or is this something that you’re actively moderating against? Or are you letting the community decide what to do?

Right now, since we’re community-driven and anyone can create a blog, a wiki, on our platform, we want that to continue to have the same flexibility it always has had. But if we see, just like we would look at spambots back in the day, which may have just been as long ago as last week, if we see a lot of content that’s being created that is clearly artificial or clearly missing the mark — most importantly, that it’s wrong — the community absolutely flags it. It’s something that doesn’t get a lot of traction, and that’s one of the things about being so dedicated to these imagined worlds.

We’ve seen other platforms, the moderators there — Reddit, Stack Overflow, you just go down the line — they said, “Look, the flood of spam, effectively AI-driven spam, is so much that we can’t keep up with it.” The moderation workload here has gone up substantially because of the amount of AI content that’s coming into the system. Some of the platforms have just banned it. They’ve said, particularly coding platforms, “We can’t keep up with this volume, we can’t trust it, we don’t think it’s good. Don’t do it yet. We’ll figure out a path.” Have you written a sitewide policy yet? Have you felt that pressure from the moderation side?

We take the approach that you know it when you see it. But at the same time, you probably have read this as much as I have: even the folks with AI moderation tools can’t figure out what’s AI-driven or not. So I think, if someone says that they’ve got an infallible AI detection tool, I think they’re making that up. I don’t think it exists because I think that it’s very possible to use AI and recognize the nuance that needs to be introduced and use it as a tool to create great content.

How do you just designate between the tool that created great content and just helped accelerate it versus the tool that is spamming a platform with irrelevant information? As I said, we have a pretty good system for detecting the irrelevancy, but at the same time, I also think that someone who says they’ve got a perfect AI moderation tool is not seeing straight.

Yeah, OpenAI released one, and they pulled it. They said it wasn’t effective. So I think that’s the state of that, but you’ve got actual moderators on your platform. Have you heard from them? Are they complaining yet?

We are not seeing that at volume, and it may just not be that, because of the content that we’re focused on, we’re not seeing that as a mechanic. Creating a wiki somewhat takes some effort. You have to really want to do it, which obviously, we want to make it as easy as possible. At the same time, it also means you have to really want to work at it. And how many people want to add to the Zelda wiki? You have to want to do that. And so, I think that’s maybe one of the benefits of having the platform that we have.

Zelda wiki actually brings up a really interesting point, which is: these communities sometimes move. Zelda wiki actually left last year before Tears of the Kingdom came out. I think the Minecraft community is thinking about moving on from Fandom. How do you react to that to say, “Okay, this huge community is unhappy with us. They’re thinking about leaving our service and going somewhere else”? Do you actively get involved in, “Okay, we should try to keep them”? Is there strategic planning? Or is it, “Hey, okay, no harm, no foul, you can move on”?

I’m actively engaged. I take it very seriously. I work very hard every day to try to make sure that we build a good platform that people can use, that’s stable and has great tools. And I think we’re someplace that I think is really community-focused, and I think we’ve got the right values. So I really take it seriously if a community says, “Look, we’re going to go somewhere else.” And at the same time, we also have seen lots of year-over-year editor growth. So we’re actually seeing more people than ever creating content on the platform.

So I think, because these IPs have big, broad fan bases, the great news is there are lots of people who are interested in creating the content, which is great, but I’d also take it very seriously that we have a good experience for our editors to come to the platform and produce the content. So it is a balance, but I do take it very seriously.

Did you have any conversations with the Zelda wiki folks to say, “What would it take for you to stay?”

I didn’t at the time. I think there’s also a role to play as the CEO, which is that a lot of people think the platform is theirs to use, and largely, that’s the way I’d like it to be felt. I’d like to think the folks at the Zelda wiki have no idea who I am because they’re so happy with the tools they’re using, they just don’t care. It’s like, I don’t really care about who runs Gmail because Gmail works fine, and I just want to use it to do the thing I’m passionate about.

On this show, we care a lot about who runs Gmail. I just want to be clear, that’s this show.

I appreciate that, but that’s my orientation. I tend to work with our teams really closely, and we do a huge event called Community Connect every year where myself and Jimmy Wales spend time going one on one with some of our leading community members and admins, so I’m very devoted to the communities. At the same time, there are some communities who see us as a platform and want to be left to do what they want to do. And I also respect that enormously, and let me just do my job to make sure you have a great set of tools.

Did you take any lessons from what’s happened with Reddit? I think Reddit was in the same position, right? They were somewhat neutral. It was hard to even see what they were doing on a day to day, and suddenly, they appeared and they became personalities and they tried to take control of their platform. And maybe that will work out in the end financially, but certainly, in the moment, it has had a series of negative repercussions for them. What lessons did you take from all of that?

It’s the same lesson I think that I’ve had since I’ve been here, which is, we restarted Community Connect when I got here, and I worked very closely with Michael and Brandon, our team of community leaders, to have a forum and listen and make sure we build great tools. One of the things that is very clear to me is that if you don’t make the platform serve the creator, the creator may go somewhere else.

And so, we spend a lot of time. We have a roadmap of creator tools. It goes through the next year. We do a Community Connect every year to coalesce and bring our product and engineering and marketing and content teams all together in one place with our creator communities and listen to them and figure out what to build and when we can launch it. And then we publish our roadmap and say, “This is where we’re going. This is what we’ve heard.”

And we’ve actually, again, over the last year, launched more tools than we ever have. And so, I think that, for me, is where we’re never going to be perfect, and there’s always more we can do, but we tend to spend a lot of time on making sure that we have a good say-do ratio, as they say. We do what we say, with our community, and I think we’ve got it. And if you do that pretty well, then hopefully, on balance, you’re able to retain more folks.

That’s as good a place as any to wrap it up. Tell us what’s next for Fandom. What kinds of things are you looking to roll out that you’re excited about?

We’ve got a lot of new launches coming up on the core platform over the course of the next six to nine months, which is going to be in the form of making discovery easier on the platform, getting our search more optimized, and actually working more closely to provide the quick view of content on Google. We are really focused over the next 18 to 36 months on this idea of having a better, deeper personal experience and this idea of collecting things that you love around your identity at Fandom. That may be in the form of going to our events or collecting goods or bringing your identity more directly into Fandom so we can provide you more relevant content.

That’s a little bit of a teaser for something we’ll be talking about next year, but I do believe that, because we are so close to fans, we can do much more than we do today to give them relevant context and help them celebrate their fandom better than ever.

Amazing. Well, Perkins, you’ve given us so much time. I really appreciate it. You’re going to have to come back next year and talk about some of these new product launches.

Thanks. It’s been great to talk to you. I really appreciate it. You ask excellent questions.