Today, I’m talking to Jason Citron, the co-founder and CEO of Discord, the gaming-focused voice and chat app. You might think Discord is just Slack for gamers, but over time, it has become much more important than that, and for a growing mix of mostly young, very online users steeped in gaming culture, fandom, and other niche communities, Discord is fast becoming the hub to their entire online lives. A lot of what we think of as internet culture is happening on Discord.
In many ways, Discord represents a significant shift from what we now consider traditional social platforms. It’s not a public-facing network like facebook or instagram, and it’s not really a broadcast medium for creators quite like YouTube or TikTok. But it’s also not a forum in the way, say, Reddit is, where you participate in big public threads curated by moderated communities. Instead, as you’ll hear Jason describe it, Discord is a place where you talk and hang out with your friends over shared common interests, whether that’s video games, the ai bot Midjourney, or maybe your favorite anime series. It is a very different kind of interface for the internet.
Jason and I dug into the nuances of how he sees Discord in the landscape of other platforms and how he’s made conscious choices about what he sees as the future of online communication. For Discord, that future is smaller, more intimate, and far from the public eye. We also discussed the inherent tension between the version of Discord that acts as a tool for voice chat and the version of Discord that’s become a social destination mixing public and private in increasingly complex and, at times, legally fraught ways. We also touched on the word “servers” and how it’s played an important role in the kind of IRC-style culture the company was born in and still cultivates.
You’ll hear Jason talk about Discord’s evolving business model — unlike Slack, it never went into enterprise software. Instead, it has a consumer subscription service called Nitro and a growing number of other ways it’s exploring making money, including the platform’s very first ads. Jason also revealed why he ultimately decided not to sell his company to Microsoft for a reported $10 billion and also how the post-pandemic slowdown forced the company into two rounds of layoffs and a major refocusing effort about what Jason thinks the Discord community wants and needs. The short answer: a bigger gaming focus and more outside developers building apps, bots, and games that live exclusively inside of Discord.
Of course, because Discord’s users are so young, it faces some particularly unique content moderation challenges. You’ll hear Jason reflect on his testimony in front of Congress earlier this year around child safety and also why the company has made some pretty major tradeoffs around features like encryption that other platforms have been unwilling to make — because Jason’s perspective is that they have to make the app safe for teens.
This was a fascinating conversation, and Jason’s perspective — that online life will only continue to move toward private group chats built around the ways we spend our time with friends — looks more convincing by the day.
Okay, Discord CEO Jason Citron. Here we go.
This transcript has been lightly edited for length and clarity.
Jason Citron, you are the founder and CEO of Discord. Welcome to Decoder.
Thanks for having me, Nilay.
I am really excited to talk to you. Discord is a seemingly very simple application. It’s also very complicated. It exists in a complicated ecosystem of things. There’s a lot to talk about with Discord all the time. It started as a voice chat for gamers. It grew into a place where people hang out to talk to each other. What do you think Discord is now?
Now, I think Discord is a place where people talk and hang out with their friends online.
A long time ago, I was talking to Stewart Butterfield — this is before he sold Slack to Salesforce — and I said, “Do you think Discord is a competitor to Slack?” He said, “No. Absolutely not. Slack is enterprise software. We do all these enterprise logins. We have to deal with all this stuff at your company. Discord is over there and that’s different.” I always thought that was interesting.
I think that’s true. But the idea that one is very enterprise and one is very consumer, that’s gotten a lot blurrier since I’ve had that conversation. I know entire businesses that not only run in Discord but also talk to their communities, to their customers, in Discord. Has that gotten blurrier for you as well?
We have always focused on Discord as a service and tool for consumers to come together and talk and hang out. And it is used in many different ways, to your point. A lot of companies use Discord almost as a new way to communicate with their superfans online. We love that and we support that use case. But most people who use Discord are in smaller invite-only groups — we call them servers — with people that they know or friends of friends. It’s a place that feels like your dorm common room or your living room, where people are hanging out with people they know.
That server terminology is always really fascinating to me. I’m somebody who came up on IRC. The first tech website I worked for, Engadget, actually ran itself on IRC. We had to teach everybody how to use IRC when they got a job with us, which is wild. Everyone’s moved on from that now. But you’ve kept that terminology, that ethos, alive. You’re starting a server. You’re in charge. Talk to me about that. Why keep that old-school terminology in play?
When we started Discord, our focus was very much on building a text and voice chat app for people who played video games. Back in 2015, the alternative products that people used were, literally, they would host voice servers for their friends. The reason we picked the word server was because, at the time, that was how our customers thought about what the tool was doing for them. It was a server where they could go and bring their friends together.
So we said, “Well, on Discord, you get a free server; whereas, on those other apps, you’re paying for a server.” Now, our service runs in the cloud and doesn’t literally have a server the way that they used to rent computers and get IP addresses. So it just stuck. People understood the concept of a server as a place where you come together, and from 2015 to 2020 or so, Discord’s primary focus was on gaming. Even today, gaming is a huge part of what people do on Discord. So people just get it.
I asked that question very specifically because the word server, to me, implies a bunch of control. As you’re saying, you would go out in the world, you would start a server, and it would be your own. But Discord is a platform. There’s an app store built into it. You’re doing a bunch of developer outreach. It is the user interface for some very cutting-edge products like Midjourney. That control goes back and forth. There’s what Discord the platform wants, there’s what users are doing with it, and there’s what users might do with it that you haven’t even thought of. There’s what they might want to do with it that you don’t want them to do. How do you think about that tension?
Our focus is very much on creating tools that give people the capability to design their own space. That was part of the intention from day one. That’s part of the server idea like you’re talking about. We give you these tools to make a server, and then you can choose: what are the text channels, what are the voice channels, what do you want to name it? How do you want to decorate the different people and have them stand out? We have permissions and role capabilities where you can say, “Well, these are admins and these are newbies and they show up differently in chat and they have different powers in chat about who could kick people off and invite people on”
We’ve always had this ethos of leaning toward user control of their spaces. Frankly, what’s so cool about that is that it has extended to customizing Discord with our API and bots platform because we knew that people were going to want to customize the service and connect it to other services outside in the world. That user control and open ethos is what enabled things like Midjourney to flourish. There are over half a million apps that people have built on Discord that are used across our user base, so it’s a really intentional posture that creates conditions for exciting things to happen.
How much of the evolution of the product itself is guided by what people are doing as they build applications and bots, and how much is guided by what you want from it?
We take a mixed approach where we spend a lot of time talking with our customers of all different kinds trying to understand what people are doing with Discord, what they want from it, and what their challenges are with it. Then, we mix that with what excites us as product creators and builders and how we think the world is changing and where it’s going and what we want to create for people. We try to put that all in a pot and shake it up, and then stuff comes out, so in some places, things that our bots community and apps community have built have driven our roadmap. I’ll give you one example. In the early days, we had a hunch that Discord would be used for public communities like some of these we’ve been talking about, but we didn’t actually design that in a first-class way into the product.
When we initially launched, I think the user cap was 30 or 50 people in a server, for example. As people started using it for more public spaces, we kept raising the cap, which was infrastructure work to make the product work better. We had basic moderation tools, but it became clear that when you have thousands of people on a server, you need different kinds of moderation tools. A lot of bots sprung up, and that made us realize we needed to invest in this, so we created AutoMod, which is now built into the platform that allows these communities to moderate in much more advanced ways. Now, we have a whole trust and safety team. That whole effort was really a response to what people were doing with our product that we thought might happen but weren’t really sure, and it wasn’t the original focus.
Let me put that into contrast with, say, Reddit, which is another huge user-generated platform that is really driven by its community. There’s a tension there. There’s what the community wants, the tools it builds for itself, and there’s what Reddit wants. Those things come to a head. They wax and they wane. Have you had those moments where you know that the platform needs to do something that will make the community mad but you need to do it anyway? Or have you been able to integrate what the community is building in a more healthy or more stable way?
I think that one thing that’s fundamentally different about Discord from Reddit is that we are much more a group chat app for friends than this public space with moderators and (user-generated content). We don’t think about Discord as a UGC platform, for example.
I think about it as a communications app. It’s a group chat app. If you look at where people are spending their time and what they’re doing most of the time, most people are texting in invite-only group chats with their friends or on voice chat, playing games, talking about their day, cooking dinner separately, falling asleep together. That’s what people do. It’s a place where people talk and hang out with their friends, primarily. And then they do go explore these other spaces in their interests and participate in these big communities. Some people really love that part of the service. But Discord is a communications tool. It’s not a UGC platform in the sense that I think you’re describing.
I want to stick with that for one second because I think the difference is pretty finely shaded. It’s a communications platform, but it’s not one to one. It’s one to many by default. You log in to a channel, you’re talking to one person, but you could be talking to lots of people. It’s not encrypted, which I want to come back to. That’s a choice you’ve made to make sure you can monitor what’s going on. And because you’re unencrypted and you can moderate, you do moderate it. You do have a trust and safety function. What is the actual distinction between a communications platform and a user-generated platform like twitter or Reddit?
Well, all communications are user-generated, so maybe that’s what you’re getting at. But I think what I was reacting to is more what you were describing trying to make this comparison between Discord and Reddit, which is a great product. People post content on Reddit and the content that people post for other people, usually strangers, is the primary thing that I think people get from Reddit. Discord’s very different. Discord is more like a group chat app where you’re sending messages, frankly, often to one person. Direct messages are very popular. Our servers are also popular. But in that case, it’s like three to 10 friends. If you’re playing a video game, it could be your guildmates or the people you regularly play with or a club that you’re a part of.
It’s not a broadcast medium in the way that a lot of these other more social media-type services are. We do moderate it because we know that there are a lot of teens on the platform. When we do have those public spaces, we treat that more like the public UGC stuff. But most of the time, people are hanging out with their friends in their virtual living room.
That’s fascinating, and I want to come back to it because Discord is so many things. Like I said at the start, you can look at it through so many different lenses. The idea of it just being a direct small group chat app, I put that right next to the fact that it’s the user interface for Midjourney, which is one of the hottest ai tools out there. And I say, “Well, most ai tools are text-based, they’re prompt-based, and the chat interface is the way we think about using most generative ai tools.” Discord has become that interface for at least one of them. Is that a future that’s in conflict with being a small group chat app, or is that the next extension? Or is that even something that needs to go off in another direction on its own?
I think that the fact that the chat input box has become the primary way to interact with a lot of these generative ai tools and that we have a really popular and extensible chat input box is great. Midjourney is a really cool product and people love using it. They have one of the largest servers on Discord, if not the largest server on Discord. But a lot of people actually take the Midjourney bot into their invite-only server with their friends, and they’re using it there in a more creative space that is not in the public view or that is a server that you can just go join. I love Midjourney and the things that other generative ai apps are on Discord, but for us, when we think about the service that we’re offering to users, it’s a group communications tool.
One of the things that people do when they’re hanging out with their friends is they play around with these generative ai products. They share their creations with their friends, and they act as conversation pieces, as a shared experience, to do together. From that lens, we love it, and that’s why we encourage and support it, but we really come back to what most people are doing on Discord most of the time, and that’s chilling with their friends and hanging out. Video games continue to be a huge part of what people do every month on Discord.
You add this all up and you get this strikingly different view of what being on the internet should look like. It’s not whatever TikTok is turning into, whatever Home Shopping Network instagram is turning into. It’s text. You’re typing a lot; you’re looking at photos that are being generated; you’re interacting with other computer systems through text prompts. Discord is a window into that. You’re maybe writing some applications that are inherently text-based inside of Discord. But it’s almost like a command line vision of connecting on the internet. It’s old school. Do you see Discord as being that big, as in this is a different way of thinking about connecting and computing, or are we very focused on it as a chat app?
The way that we think about it and our vision for where we think the future goes in regards to Discord, it really comes back to how people spend their time with their friends. When I started the company back in 2012, the bet that I made was that video games would continue to become a bigger and bigger form of entertainment for people. They would become more and more social. They would be across more and more devices and that there wasn’t a great communications product that was started with your friends around gaming. So that was the original thesis. Even today, when we picked our heads up after covid to reevaluate what was going on in the world and what our customers cared about — we spent a lot of time last year with folks — I have more conviction today that (gaming) will continue to become the future.
If you go back over the last 12 years, it’s really played out, and gaming has very much gone mainstream now. I think 93 percent of Gen Z plays video games. When I was a kid, I was weird playing multiplayer games by myself or with my friends, but it was a niche thing. Today, it’s quite normal. Our vision for the future is a world where people have really rich shared experiences, and they can spend quality time with their friends no matter where they are in the world. A lot of that is going to be video games that exist just on platforms. Some of those will be video games that we will serve directly through our platform.
The bots platform, we’re evolving it to include embedded experiences because that’s the part we think will matter. But it’s really about this idea of how I think the internet’s going to evolve. There’s a need for more cozy, intimate spaces where people can spend quality time with their friends away from the broadcast performative stuff that we see a lot of. And we’re very focused on creating those cozy spaces for people to talk and hang out with their friends and deepen their friendships.
This is a theme I see everywhere right now that the United States is heading into an election year: what is our social media going to do to us and what is it doing to teenagers? It’s all colliding, and a bunch of these social networks are not ready to take the weight, or they don’t want to. In the case of Meta, I don’t think they want to. You’re not positioned in that fight at all. You’re saying, “Look, the internet should go back to being smaller, more fragmented, more among people you know and less about these giant culture-defining social media platforms.”
How comfortable are you in that bet? I mean that in the big way, not the little way — you’re the CEO, you have to say you’re comfortable. But is the internet actually moving in that way? Because I know a lot of people who want it to, but I’m not sure that it actually is.
I don’t think it’s an either-or. Over the last 15 years, as we went from the internet being new to Web 2.0 and now the rise of mobile, I think that we saw a lot of this aspirational promise of these broadcast social media services and what they could do for us as people. I actually think they do a lot of really great things. I think, 50 years from now, we’re still going to have something like this. It’s undeniable that all of these services create value for people. I think there are questions we’re working through as a society around some of the negative externalities of those things and how we want to manage through that. But I think we’re going to have public photo sharing and public video sharing apps for the long run.
What we’re seeing on Discord — and I think this has been a trend for the last five, six, seven years toward group chat messengers in general — is that people understand that those public spaces are interesting, but there’s something else that they want in their lives, too. That’s more intimate, cozy spaces where you can spend time in a relaxed way with people you know and spend quality time with friends even though you can’t maybe be in the same physical space. I think that that is going to continue to grow, and I think that the social media stuff will grow. I think both of these things will exist in big ways in the world if you go 20 years in the future.
When you look at Discord right now, what part of it is growing faster? Is it the small cozy spaces part, or is it the place where a bunch of crypto startups talk to their customers? That was a big growth moment for Discord, but that’s the big public broadcast version of it. So is it the smaller cozy part that’s growing faster, or is it the more public part?
It’s the smaller cozy part. In fact, that’s the part that’s always been growing the fastest. The thing that we find that’s interesting is, because it’s not publicly out there, people don’t really know about it so much. The crypto thing was big and now the ai thing is big, and those things did bring a lot of people to our service. But at the end of the day, the people who come to our service and love it the most are the ones who come with their friends or find their friends and it ends up becoming this place where they hang out online and keep in touch with the people that they care about.
That brings me to the Decoder questions. You’ve described a lot of things Discord could be, a sense that you need to focus, and the main thing that it always has been and should be in the future. I want to ask you how the company’s structured, but there’s a little bit of context here. You did recently have some layoffs. You cited the economic slowdown and your headcount was growing too fast, so you cut 170 people, which is 17 percent of the company. You cut 4 percent previous to that. How is Discord structured now? Did these cuts change that structure?
The way that we’re structured and those cuts come back to, “What are we trying to accomplish and what are we trying to build for the world and for people?” Over the last year in particular, I mentioned we spent a bunch of time going back into the market and talking to customers in a way we hadn’t over the last few years through covid to really refresh our mental model of what people are doing (on Discord) and what they want from us. What we really clicked in on was this insight that while we were focused on being more mass market through covid, gaming actually went mass market at the same time as people grew up.
Because our service is so good for playing games with your friends, while people do lots of things on Discord, what we realized was that gaming actually is still one of the main ways that people spend a lot of time with their friends on Discord. I think it was like 95 percent of our users play video games. Last month, 1.5 billion hours were spent playing games across Discord on 60,000 titles, so people spend a lot of time playing video games.
What that made me realize was that we should really focus on gaming because it’s a huge thing that people do on our service, and there are a few billion people that play video games in the world. It’s the largest form of entertainment that’s growing the fastest. I think we have a very unique role to play there, and we love it. Me, my co-founder (Stanislav Vishnevskiy), and a lot of the people at the company grew up playing games, and games are such a core part of our social lives and our best relationships. With that insight of saying, “Okay, we’re going to focus in on gaming,” we realized that we had too many people at the company, and we were not focused on the right stuff. We went through a reevaluation of what the next chapter of Discord was going to be, and through that, we realized we needed to shrink the company a little bit and shift our focus.
Going forward, we are very focused on gaming as our core use case — group chat around gaming. But we’re continuing to enable other things because people who play games do lots of other stuff, and that’s how we got here in the first place. As a result of that, we are organized as a functional company, meaning we have engineering, product management, marketing, finance, and talent. We organize functionally because Discord is one product. We essentially have one product. So we need to organize and coordinate in a way that it comes out as a coherent experience for people. I talk about it like a symphony. Our people hear it as one song. Even though we have a hundred people playing instruments, we have to coordinate it effectively.
A functional organization allows us to do that, and then we break it into three different types of work that we organize against. We call them our foundational initiatives. This is the stuff that is the bread and butter of what we need to deliver on for our users, like performance, trust and safety, and core messaging and communication features. We have cross-functional teams dedicated to each of those with a design leader, a product leader, and an engine leader, and they have a roadmap.
Then, we have what we call our core priorities. These are the step function things that we’re betting on. We have a few of these, and notably one of them is our new Quests feature, which we announced. We’re getting into how we help game developers bring their games to life and build their businesses. Another one is our embedded activities platform that I mentioned, and then there are a couple of other things in there.
Then, we have a much smaller… I call it a venture initiatives team. It’s a team of six people that report directly to me, and this is the crazy innovation lab where we’re trying stuff that may never come to life, but it’s the bigger swings to see what we could create and innovate on for folks. So that’s how we organize the company, and we run it in an interesting way, too, with Loom videos. I’m happy to get into that if you want, but I’ll pause.
Well, two things. I definitely want you to get into Loom videos. That’s the first time anyone’s ever said that on the show, which is wild. How do you run the company with something called a Loom video?
Loom is actually a specific product that makes it really easy to do video, like screencap and camera recording. But basically what we do is we run the company hybrid. We use Discord to run Discord. So we’re all over the country in America. We’ll get groups of people together to look at marketing plans or review creative assets or product strategies or product demos. I found there was a lot of coordination overhead around how to get a meeting together with all the people that want to be in the conversation. It was this chaotic thing. And we had this idea, “Well, what if they recorded the presentation ahead of time and sent it to me, and then I could just watch it in my free time? And then I could respond with a video as well, so they could see my excitement around the feature and I could have their presentation up and be clicking through it.”
That gets the initial presentation and reaction out of the way, and it turned out that we started doing this in just a couple of spots, and now it’s becoming this thing across the whole company where people record video of them doing a presentation. Usually it’s five, 10 minutes long, and then folks can watch it. I often watch these things at 2x, sometimes when I’m brushing my teeth. I always have free time, but I don’t always have an hour to have a meeting, and then sometimes I’ll watch it two, three times, let it marinate in my head, and then I will always have 10 or 15 minutes to just respond or react. I essentially make a response video and send it to them. Then a hundred people can watch it. They don’t have to be in the meeting, and most of the time then, we don’t need a meeting. The communication happened and they can go.
Sometimes, if we do need a meeting, we can get folks together. But we’ve already basically had a full brain dump back and forth, and it’s just really accelerated our product development and creativity. I think it’s also created a more human connection between our employees because now we’re seeing each other in video in these ways that is much more natural and casual that sometimes is missing when you’re hybrid. So it’s been really cool.
Alright. You can’t tell me about your crazy six-person renegade crazy idea squad without telling me a crazy idea. What are some crazy ideas that you’ve tried?
I put the chip on the table so that y’all know it’s there, but the challenge with talking about this stuff is that most of it is not going to see the light of day. Our users will be listening to this and they’re going to hold me to it.
Oh, yeah. If I say anything, they’re going to hold me to it.
I think my idea on my team is to have the worst ideas so that everyone else can have slightly better ideas. What’s one of your worst ideas that you would just never do that’s obviously a bad idea?
I don’t know. I’m sure I have a lot of bad ideas. I just don’t know which ones are bad before I try them.
The creative process is an interesting thing. We find this a lot at the company when we build features. If you’re trying to innovate, a lot of the time, things that end up working seem like bad ideas upfront, and oftentimes, they seem like bad ideas to people who are even really good at innovating. I tell this story that, when we were originally building Discord back in 2015, we were a team of maybe 12 people at the time. Half of the people working on it thought Discord was a bad idea in a 12-person startup. It turned out to be a good idea.
It’s really hard to tell beforehand whether innovation is going to work. It’s really important to have space to try things, react to it, and innovate and iterate to see where it takes you. Sometimes magic comes out the other end, and a lot of times, you just get duds. But we don’t ship those. We try not to ship them.
We’re 870 employees at the company. We have a little over 200 million monthly active users all around the world.
One thing I think about all the time is when The Verge was small, we were able to try things and get rid of them really fast. We had a small team, everyone knew we were trying something, the vibes were shared, and we had a small audience, so we could get rid of things and only three people would ever know. Now, we’re big and we have a big team. Some people are really committed to some ideas; some people aren’t. Some people can’t tell from a Zoom call that I’m on that I just want to see what happens. And then we have a big audience that’s paying a lot of attention to us, and it gets much harder to take risks and shut things down. Discord is in that spot. You have a big team. You have a huge audience of people who care a lot. You won’t even mention a bad idea because they will hold you to it. How do you think about taking risks?
It’s a unique situation. Part of the reason I didn’t say anything is because we take an incremental approach toward exposing users to risk or to ideas based on how confident we are in them. Practically speaking, what we tend to do is launch features to very small segments of our user base and see how they respond. But that even comes after what we call a closed beta or closed alpha, where we just recruit 50 people who sign an NDA to try something. Before that, we use our employees. A lot of them are customers, so that’s a free couple hundred users to test something. And before that, the team has to be confident in it. So there are these gates that things go through.
It’s not a highly structured process because it depends on the thing. But we frequently will eventually get to a point where we might, let’s say, launch something to 50,000 people in our customer base and then let them try it, send them a survey, see how they interact with it, and then, based on that, decide which way to go. While it can be frustrating for those folks if we, let’s say, remove something from the product, we usually only remove something if it’s not actually that popular or useful. The funny thing about it is when we remove something, some people care, but most people don’t, and that’s why we removed it. So it works. We don’t remove something that a lot of people love because, if a lot of people loved it, we wouldn’t remove it.
Do you ever foresee Discord having the Microsoft Excel problem where someone has built an entire business around one button in the toolbar and you can never take it out?
We do have a lot of developers who build apps on Discord, and so we do think about this around companies that have built these products and depend on Discord to deliver their service and as a user interface for their service. We really value that. To some extent, we already have this dynamic where people rely on us. Users rely on us, too, and this is one of the places where we get a lot of friction with our users when sometimes we realize that a certain segment of people really like the way something has been designed. Maybe it’s tens of millions of people who love it. Then, we realize that there are 50 million people or hundreds of millions of people who want something different.
Managing that tension can be quite challenging because that’s a case where maybe we ship something that people loved, and now, it’s years later and our customer base has grown. The dynamics have changed. Managing all of those competing interests is not easy, and sometimes we get it wrong.
This brings me to the Decoder question. You have a lot of decisions to make spinning stuff up, spinning it down, growing, focusing. How do you make decisions? What’s your framework?
We always just focus on our customers. We try to prioritize our customers — “what can we do to better serve people today than we did yesterday?” Then we try to mix that with “what do we want to do” and “what are we excited about?” Because great things only come if the people who are making it are excited and passionate about it. We actually put a lot of stock in that, and then it’s like, “How do we build a great business and make money and do all that stuff?” Obviously, we’re a company, so that’s part of it. But I really do believe that over the long term, the best way to build a great business is to serve your customers really well. We just keep coming back to our customers and try to figure out how we best serve the most of them. That really guides everything.
Is that the tiebreaker in every scenario, or do you sometimes say, “Look, I’m the CEO. I’m just making this decision”?
Sometimes the way actual decisions get made is, “Well, we think 30 percent of our customers want this and 40 percent of our customers want that and 15 percent want this, and how do you figure out which ones to listen to? How do you even know if that’s the right breakdown of what customers want?” Sometimes we look at data and we’re like, “Well, people don’t interact with this, but they tell us that they love it.” So there’s a lot of judgment that goes into making these decisions. We’ve evolved this over time, but there are different models that people have around how to do this.
We essentially try to pick a single person who is the owner of making a particular decision, and it’s up to them to farm for dissent in the company and collect the insights that they have from customers. Depending on the decision, sometimes if I’m not happy with it, I’ll pull the veto card as CEO. But I try not to do that because great people want to have autonomy and make great decisions and want to collaborate with other great people. So I like to participate in that collaborative process. But from time to time, it’s like, “Okay, Jason, what do you think we should do?” I apply my judgment and make a call and we see what happens.
One big call in 2021: you decided not to sell the company to Microsoft. Can you walk us through that decision?
Every great company along its life will get acquisition offers. It’s just a fact of life because, if a company is great, someone’s going to want to buy it. That was not the first offer we’ve gotten, and it wasn’t the last one we’ve gotten. It was just the one that became most public. Whenever we get into any of these situations, I try to ask myself a question around where we are as a business. What’s the best thing for our customers? What’s the best thing for our shareholders? What do I want? What does my team want? And how do we work through that?
So far, every time, it’s fallen out on the “Let’s stay independent. Let’s keep building. Let’s keep growing. We’re having a blast. There’s a ton of opportunity in front of us.” So that was all it was. It came back to our customers and what we think we could build and what we wanted to do, and then we made a call.
I just talked to Dropbox CEO Drew Houston, who told me every company tried to buy Dropbox at the beginning. They would say, “Look, this is a feature. It’s not a whole product.” But Dropbox has managed to survive because it is cross-platform, because it works with everything, because they prioritize working with everything well, because you can use it without being sucked into some other ecosystem and locked in. Discord is like that. You just use it and then you’re playing a game over here, wherever you’re playing a game.
Do you think about that — that if you’d sold to Microsoft, they would’ve pulled more people into that ecosystem and that would’ve constrained you in some way? I’m sure Sony has talked to you at some point — or Nintendo or whoever. Is it that actually being cross-platform is the thing that lets you be successful?
I think that the thing that lets us be successful is it is cross-platform, but it’s also our focus as a business on building communications tools. There are a lot of other gaming companies that have communication services as part of their offerings. Most major game publishers have something. But it’s not their priority. Whenever they’re forced to make a tradeoff around, “Where do we put our best engineers and our best creative people?” they focus on the gameplay, as they should, but we focus on communications.
That business structure combined with cross-platform, combined with the macro trends around gaming becoming more cross-platform and more device-agnostic, I think that is really a big factor in what’s helped us get here. But at this point, a lot of people are on Discord, and we’ve become known as the place to talk and hang out with your friends, so there’s a lot of momentum around that that I think we appreciate and seek to cultivate. I think that’s what really makes us successful.
Do you think about how many fronts of competition you have with the big game companies? Microsoft has Xbox Live. They also obviously make video games. You are slowly starting to make some actual games as part of the Discord platform. That’s another front of competition. Is that something you think about — how many ways you’re competing with those companies and the many ways you might stand apart?
Our focus is very much on communications tools. We have a very small games team. They’re mostly to help us figure out how our platform should work so we can open it up to other folks, which we did just before the Game Developers Conference a couple of weeks ago. So, right now, anyone can come to Discord and opt in to the developer preview and build an HTML5 game embedded right in Discord. That exists because we’ve built some stuff ourselves to figure out how it should work. When I think about competition, I don’t think we’re competing with the gaming companies. In our earlier days — like the 2016 to 2017 era — all of them tried to launch competing Discord services, and I don’t think any of us know too much about them anymore. I think we won that battle.
At this point, most game developers use Discord as some part of their go-to-market and development process. This is one of the other things that we learned over the last year as we were talking to customers and game developers, too, was how essential Discord is today and our community capability as part of development. We heard from a lot of folks that they bring in early playtesters into smaller private Discords and hop on with the developers, and they’ll do nightly playtests to get feedback to guide development of the game.
A lot of that informed our roadmap for this year and really trying to elevate game developers and make them our customers and really collaborate with them. We got a lot of positive excitement from folks around Quests and our HTML5 platform and some other things we’re doing. I very much think that we are helping game developers with the services that we provide, not really competing with them.
In terms of revenue and where you’re growing, it seems like Nitro, the subscription service, that’s where the focus is. How’s that going? Are you getting a lot of consumers actually paying for yet another subscription service?
Nitro is doing very well. People really enjoy the features that they get: higher game streaming; animated custom emoji; being able to give their friends benefits through server boosts. I haven’t publicly shared how much revenue it’s making, although you could Google it and figure it out, but it’s doing very well.
We recently launched another consumer revenue line, which we call our avatar decorations, in our shop. Just a couple of weeks ago, we launched a partnership with Valorant, which is one of the big FPS titles, and you can buy Valorant decorations for your profile, and players love that. That’s going incredibly well. I’m really excited about what could come next for us through our sponsored Quests format where we’re helping game developers reach their audience on Discord to help them build better businesses. Of course, we’ll make some money. And then players will get free rewards in their games that they love. And then there’s our platform, where we can help support games and other things like Midjourney build businesses and create new shared experiences for people.
Those are some pretty direct revenue models. You have a subscription service. You buy some cosmetic items for an avatar. Those are some microtransactions. Some developers come on the platform. You monetize those developers directly. You’re also poking at ads. You’ve launched some gamified ads inside of Quests. Is that the bigger revenue opportunity, or is that something you’re trying out?
We’re just trying it out right now. It’s hard to say how big it will be. What we have heard from game developers and game publishers is that they know that their players are on Discord, and they really want to be able to reach them. When we talk to players, and we’ve run some of these Quests experiments, players really love getting free rewards in games they like.
People love free stuff. That’s a universal truth.
Yeah, I’m like, “Please give me free Magic: The Gathering packs. Just send them right there.” I think it’s possible that at some point every game will be running Quests on Discord, and if you like to play games, there’ll be free stuff you can get in every game that you care about. I’m excited for that. I think it could be a big business, but we just started it like a week ago.
One of the things I think about when I think about a platform like Discord, ads, and games is that a huge market for ads around games is app install ads. Download this game — that was a huge market for facebook and instagram before Apple introduced App Tracking Transparency. It seems pretty clear Apple wants a piece of that. But you’ve got a whole community of gamers who like video games, and you could show them ads to download video games. That could be a big business. Is that something you would do?
In the context of Quests, we’ve explored showing a quest for a game that you haven’t played yet but is similar to something that you’ve liked to play before, maybe that your friends are playing, and people respond quite positively to it. At the end of the day, people who play games like to play games, they like to try new games, and then developers want to build games and create these businesses and reach players. I think there’s a really interesting win-win-win type of product experience that we can create to matchmake players and developers, and Quests is our starting point. I’m not exactly sure where it’s going to go, but so far, the response from players and developers has been quite positive. I’m optimistic about it.
One thing that happens to every company that goes from straight direct-to-consumer subscriptions and direct monetization to ads is they realize the companies that buy ads have vastly more money and will just keep spending money. It just happened to Netflix. People like free stuff. People run ads on Netflix. That ad-supported tier is growing really fast, much faster than the pure subscription tier. You can see that they’re monkeying with the prices to get more people onto the ad tier and double-dip on the revenue. Are you worried about that, that you’re going to open Pandora’s box here and just start shoving ads over the platform?
No, I’m not worried about that. Our decision-making always comes back to what the best thing is for our customers, for our users. Over the long term — and the long term is a series of short-terms, right? — prioritizing the end-user experience is what’s going to build a durable business for us. That is our frame in which we think about, if we’re going to have sponsored content, what is the product experience and how does that show up in a way that is tasteful that users enjoy? That is our focus. I don’t think that we’re going to open Pandora’s box. We’re making the calls.
Fair enough. I had to ask — there are a lot of Discord users who really wanted me to ask that question, so I had to do it. You have said you want Discord to finally become profitable this year. Are you on track to do that with all of these new revenue streams?
Yeah. It’s looking good.
Which one of these revenue streams do you think will most help you get to profitability?
We haven’t really broken this out, but I think that we don’t need anything dramatically different to happen to become profitable. Most of our revenue comes from Nitro. I think that that can continue to be the case, and Discord’s going to be in a great spot.
Alright. So I’ve asked you the hard questions about money. Let’s ask the even harder questions about content moderation. Every CEO gets a question on content moderation on Decoder. It’s just the way of the world. You’ve described Discord a few times now as a communications tool, which is really interesting. We think of communications tools as having much less moderation and much less acceptable kinds of moderation. I would not want anyone looking at my iMessage conversations. I would not want anyone looking at my Signal messages. It’s not allowed.
But with social media companies and user-generated content companies, we want a lot of moderation. Discord is obviously somewhere in the middle of those things. You have a lot of moderation. You have a big trust and safety team, but you think of it as a communication tool, even though it’s not encrypted. How do you land on what an acceptable amount of moderation is on Discord?
Our priority when we think about this is keeping teens safe.
That’s the priority when we think about how we approach Discord as a company doing moderation. That’s one. The second pillar of that is giving people tools to moderate their own spaces so they can decide what the rules and norms are for their spaces. This is the fact that every server can have admins who can kick off and ban users and delete messages and enforce rules and norms. That is also something that’s been built into Discord from day zero.
The way that we think about it is to keep teens safe, give people tools to moderate their own spaces, and then really focus on those public spaces and make sure we’re applying moderation there. So if you join the Midjourney server or the Minecraft server, we have expectations of the moderators there, and we have systems in place to make sure that that is a good experience when people join. Because that is a little bit more like a social media experience, even though it’s a chat surface.
Whereas, when you’re in your DMs, the level of the things that we do are actually very standard. I think many other messaging apps do these kinds of things like scan image uploads for CSAM (child sexual abuse material), for example. That’s the stuff that we do there. Now, in the case of teens, we have a product we call Teen Safety Assist. Other companies have stuff like this in messaging services where, when a teen is, let’s say, interacting with someone, maybe sending photos or doing something, they have a sidekick that’s checking out the conversation, seeing what they’re doing, and giving them tips on how to keep themselves safe or report things. That’s how we think about that because it’s really important to us that people feel comfortable and safe using Discord, and adults are more equipped to manage their own stuff. We think teens need some more help. But sometimes adults don’t want to manage their space, and that’s where our teams come in.
Have you ever thought about launching Boomer Safety Assist? Just putting that out there as an idea for you. I think a lot of what Gen Z needs and a lot of what the older folks need turn out to be the same thing.
Anyone can turn on Teen Safety Assist, if you really want to.
Just an idea for you. Just putting it out there. No bad ideas, right? That’s interesting. Earlier, I talked about this idea that this is a different way of thinking about connecting and using the internet. You have these different tools. You have these different expectations. Do you think this is a place where you’re going to increase the level of moderation over time, or are you at a steady state?
I think that as a society, our priority is to get all the bad experiences, the bad crap, off of Discord. If I could wave a magic wand, there’d be none of that stuff there today. But at the scale that we operate, we basically have a city or a country of people on our service. There’s a lot going on, and human nature kicks in. So how do you manage in that kind of an environment? What we found is that, as a society, we’re all still working through the expectations we have of companies in this kind of a world. I think that there will probably be more regulation that comes that will require us to do different things, which may cause us to moderate more. But really, what drives us from a first-principle standpoint is giving individual users control with our moderation capabilities and then making sure teens are safe.
We will launch more things for Teen Safety Assist to help teens be safe, and we will probably launch more moderation tools for people. A couple of years ago, we launched something called Slow Mode. So if people are spamming, you can turn that off. We do have a whole team working on anti-spam, which is a different version of this where people are just annoyingly sending messages. It’s a continual investment for us. The problem continues to persist, so we continue to have people at our company working to make Discord a safe place for folks.
The idea that teen safety is the first-order bit, and you’re just going to keep focused on that, our Congress seems very interested in this idea. It is the skeleton key that unlocks speech regulations for them in a lot of ways. You just testified in Congress for the first time in January. You talked about protecting children. That’s where you said you tech/live-news/meta-x-discord-tiktok-snap-chiefs-testimony-senate/h_557918e403c4caae01f369d13b560b8a”>don’t encrypt Discord messages because you want to be able to protect children on the platform. That’s a big tradeoff. Most other communications platforms are headed toward encryption, and they’re waging big fights to encrypt messages. Why did you make the other tradeoff?
For the reason we said. Discord is a place where people talk and hang out with their friends and is coming from the lens of creating a casual place where you’re having fun with people you care about. Our priority is making sure that people can relax and have fun. When you focus on that priority, then we made certain decisions in order to make sure we could deliver that kind of an experience for people. So it’s really that simple.
That is very much in opposition to how the other big companies think about their communications systems as opposed to their user-generated content systems. As you get bigger, as more kinds of people use the service, you do have weirder, newer problems, particularly as the community starts to build things and do things that maybe teens aren’t doing. I’ll give you an example. You all recently banned a number of servers related to Nintendo Switch emulation. Nintendo sued a group called Yuzu. They basically disappeared. But there were some forks of that software, and now those forks, those development communities, are gone from Discord. They’ve been banned. What happened there? How do you make those calls? Because it doesn’t seem very clear to a lot of people.
That is an ongoing situation that I really can’t comment on. But generally speaking, we comply with DMCA (Digital Millennium Copyright Act) requests, and we treat copyright law very seriously. When those processes get instigated by companies, we take it very seriously. In that situation, my understanding is it’s something related to that, but I can’t get into all the specifics.
I just wondered because that is happening all over the place in the platform. You look at other companies, they’ve developed massive systems to deal with copyright infringement at scale. YouTube has a notice and takedown regime that is so powerful that I think most people think it is copyright law. They don’t know there’s actually a federal copyright law. There’s just whatever YouTube thinks, and that is a good enough substitute for most YouTubers. Are you in a position where you’re going to have to start building some of those systems as well, like a content ID for Nintendo Switch emulators?
Discord is quite different from a broadcast service like YouTube.
But it is and it isn’t, right? These emulation groups were broadcasting their work.
Well, I guess what I mean is the majority of things that people are doing on Discord are just talking with their friends. If you look at a business like YouTube, their fundamental business is that someone uploads a video and then they show it to you. In that core loop, it’s quite easy to upload something that maybe you shouldn’t in that context. On Discord, we do allow people to upload files and videos, so of course, we do scan every uploaded file for malware and viruses and things like that.
But Discord is not a broadcast video platform where people can upload videos and then other random people can see it. The volume of that thing on our platform is much smaller, I suspect, than something like YouTube. But from time to time, we do get DMCA requests. Or sometimes we get court orders and we have to interact with law enforcement. So we do have a process in place to respond to those things effectively. We do have a process for those things. But the shape of Discord is quite different from the shape of something like YouTube.
I know you said you can’t talk about this specific case, but here, I think these people don’t even know what they did wrong. YouTube has an entire infrastructure. It has a bureaucracy. That’s not actually the only way to describe it. There’s a YouTube bureaucracy that will take your stuff down if you use some song that you’re not supposed to use. Do you have a bureaucracy like that that can effectively communicate in these cases? As the range of uses for Discord expands, they’re going to run into people using it like that, whether or not that’s your intention.
What I can say is that in this situation, we acted in accordance with our policies, and they’re based on a court order injunction. I can’t get into much more than that. But I think we shared that in an article that The Verge wrote about it, so that’s what I can say about that. But again, broadly, when we can, we try to communicate to people the rationale behind why actions are taken if they break our terms of service or something.
We recently released a warning system actually in Discord, where if you break a community guideline or break a terms of service in a way that doesn’t necessarily make it such that we should delete your account, we will give your account a warning and a fractional disable. You can go see exactly what you did, and you can’t send a message for two hours. Part of that is we think sometimes teens just do stupid things, and it’s better to teach them than to just kick them off a platform. But of course, it’s commensurate based on the intensity of the infraction.
in many ways, Discord is a store of knowledge now. It’s replaced wikis and forums for a lot of people, for a lot of things: ai comes to mind, crypto comes to mind. But there are a lot of communities now that are actively updating their Discords with what they know. Do you think about that responsibility to preserve and make searchable all this knowledge that is going into the system?
Yes, we do. And we’ve actually… Okay, this is one of those things where I was about to say something, and if I say it, then our users are going to hold me to it.
No. I’ll just say that I understand that there’s a lot of really important information in public Discord communities that people are worried about being stocked and locked in there. And we understand that. And we intend to try to solve it in a way that makes sense for people. But I don’t have anything specific to share right now.
Alright. That’s very exciting. Jason, I know you got to run. Thank you so much for joining Decoder today.
Decoder with Nilay Patel /
A podcast about big ideas and other problems.
