From Unknown Title to Viral Game: 12 Growth Lessons From Spellbreak

This session was part of “Working with the Gaming Influencer Ecosystem,” an event hosted by a16z that explored the platforms, audience behaviors, creator dynamics, and best practices for partnering and collaborating with influencers. In this presentation, Seth Sivak, the CEO of game publisher Proletariat, reveals how he and his team grew Spellbreak from an unknown title into a meme-spawning phenomenon.

This transcript has been edited for clarity.

12 Growth Lessons From Spellbreak

1. Focus on high upside channels
2. Post authenticity, not advertisements
3. Share quality content consistently
4. It’s all about the GIF game
5. Enable your evangelists
6. Focus on a community nexus
7. Control the flow of information
8. Create a cool kid club
9. Force social connections
10. Develop a community content generation engine
11. Establish a content sharing flywheel
12. Use memes as marketing

Today I’ll be talking about how we used Imgur, Reddit, YouTube, and Discord to gain 600,000 pre-alpha signups for a new IP from an unknown studio. I think a lot of the folks are probably in a similar situation to Proletariat’s—we wanted to build a new IP, but we weren’t an existing developer that had lots of previous traction and a large existing customer base. This was how we worked through the summer of 2018 to build our pre-alpha sign up. It’s still relevant today.

First, a couple of details on Proletariat. Proletariat was founded in 2012. Our first game was a mobile title, then we did another PC title before Spellbreak. We’re mostly located in Boston, although right now everyone is remote. We’ve raised about 30 million in total venture capital over a couple of rounds of funding. We have about 85 people on the team currently, but we’re growing to around 150 to support Spellbreak for the rest of this year.

For anyone who hasn’t seen it, Spellbreak is an action spellcasting Battle Royale. It’s going to be free to play. We started developing it back in Q4 of 2016, which feels like forever ago, but we are launching the cross-platform version later this year [ed: September 3, 2020]. We’re going to be releasing it on Xbox One, Switch PS4, and the Epic Game Store with full cross-play and full cross-progression.

But this presentation is focused on the pre-alpha that we ran over the summer of 2018, and how we gained traction for the PC alpha launch in February 2019.

If there’s one thing you take away from this talk, it should be this: No one cares about your game.

I cannot recommend enough how humble you should be about this. There are so many games coming out, and there’s so much noise all time, and you’re competing with everyone. You’re competing with Netflix, you’re competing with Facebook and TikTok. No one cares about your game. It’s your job to make people care.

And it’s not too early. If you think it’s too early to start, I can tell you that we made the decision to go out in a pre-alpha because we were not really getting anywhere with potential partners and publishers. We knew we had to show traction, and it was way before any of us were comfortable having players in the game. I look back at some of those videos now of what the gameplay actually looked like at that time, and it was very early in terms of its course of development. But I think it’s really important to find a way to make players care about your game.

Before I jump into all of the lessons learned, a quick disclaimer: Every game is different. Every market is different. Every team is different. I’m going to give you a bunch of advice here, so I think it’s important to take all of it with a grain of salt. On top of that, I think we made a bunch of really smart decisions, but we also got really lucky. I think it’s important to acknowledge that. And the really unsatisfying thing is that the best way to make players care about your game is to make a game that players care about. And that’s incredibly difficult. 

1. Focus on high upside channels

To give you a sense of where we started, what we wanted to do was capture interest from excited fans, get their email addresses for them to sign up for our pre-alpha, and then get them engaged with the game. The number one thing I think everyone should be doing right now is focusing on their high upside channels. That’s not things like Twitter or Facebook because, at the time, our Twitter had very few followers, a couple hundred. That’s not how you go viral. So what we did instead is we took our development blog and we put it on Imgur. By doing that, we were able to generate 200,000 views, and that would drive tens of thousands of email signups for our email list. And no one had even heard of the game; we hadn’t even announced the name for the game yet.

A big reason why I think we were able to grow is channels like Imgur, like 9gag, like Reddit, where you can actually still go viral without needing a base to build from. There’s a lot of these out here, and they change all the time, but focus on areas where you can still get on a top viral page. We just had another set of Imgur gifs go viral just a couple of weeks ago, and it still drives hundreds of thousands of views. I think it’s really valuable as a channel to focus on the ones that can get you to these viral levels.

2. Post authenticity, not advertisements

The next piece is, for anything you’re going to be posting, you want to make sure that it’s actually authentic. So because all of our posts were set up to be dev blogs, we never said, “Hey, download this game right now or sign up.” Instead, we’d wait for the first comment to basically say, like, “Hey, how do I get the game?” And then we’d respond. Or someone in the community would respond with the link. And that gave the sense that this was genuinely authentic interest and that fed on itself.

So when you’re thinking about the posts you’re putting out there, consider the tone you’re structuring and think about how you’re weaving players that come into it. Players will immediately detect if it feels like an advertisement. How can you make it authentic in terms of: you’re sharing this because you’re gamers, too, and you want people to see your game? That sort of stuff goes a really long way.

3. Share quality content consistently

Once we started to get the ball rolling here, it was important to do it consistently. There are two reasons for this. The big one is that players start to expect to see content on a regular basis. So, at this point, the game was under NDA. So we had full control over when we would share content, but we set up this regular cadence for when our players could expect to see something new. The second piece is, I think it’s important to build that muscle into your own team—to be posting content on a regular basis, but making sure it’s good. The worst thing you can do is continue to keep up consistent content if it’s just not good. If it’s not working, you need to reevaluate if it’s the channel, if it’s the content, whatever it is. Don’t just keep putting it out there, because I think you’re wasting your time.

4. It’s all about the GIF game

For games right now, animated gifs and short-form video is incredibly important. If you haven’t already had a conversation on how your game is going to show in an animated GIF, you need to have it. It doesn’t have to be super-fast action skill shots, the way we had in Spellbreak, it could be, you know, poignant moments or touching moments or cute things or funny things.

That stuff is incredibly necessary right now because the water cooler moments that people used to share in the office where they would say, like, “Hey, you know, I was playing this cool game, or I fought this cool boss” no longer exist. They’re all getting shared as animated gifs now, whether it be through TikTok or Instagram or whatever it is. So you really want to be thinking about: how does our game show as an animated gif? And if it doesn’t show well, you should really be thinking about why.

5. Enable your evangelists

Once we had gotten a number of posts to go viral on a few of these channels, players were signing up for our pre-alpha. We used MailChimp for this. And once we had that interest, we tried to immediately tap into it further. So we basically made people wait about a week or two for a key. But we would say: if you come in through a referral, we’ll give you a key immediately. And that drove a lot of people to say, “Hey, of my friends, you guys can get a key right now, you can skip the line.” And that was really meaningful to drive a lot of our growth. At one point during the summer, like 60 percent to 70 percent of our signups that were all referrals.

And some people were like super-referrers because they would be like, “Hey, I’m posting this in my 500-person Discord server,” or “I’m posting this to Reddit” or whatever it is, “be sure to use my name as a referer.” And we didn’t give the referers any actual value. We didn’t provide them with anything. I think that was a miss. I think we could have done something where we said, “Hey, if you refer X number of people, we’ll give you a sweet outfit at launch.” We didn’t do that because we just didn’t have the resources at the time, but even without any reason to do it, it was enough for people. People were so excited to be a referer and to share that they still did it. And it drove a massive number of signups for us.

6. Focus on a community nexus

We knew that email wasn’t frequent enough to actually drive the community. So we focused the community into Discord. And the way we did that is we said, “Hey, once you sign up for a key, it’ll take a week or two to get there. If you came in as a referral, we’ll give it to you right away because we figured you were already tapped into the community because you had a referral.” But, if not, we would say, “If you want a key really fast, join our Discord ” And so what would happen is people would see a viral post on Imgur, and they would sign up for the email.

In the email we would say, “Hey, it’s going to be about a week or two, depending on demand, for you to get a key. But if you join our Discord, you’ll probably get one tomorrow.” That started to drive a ton of people into our Discord server. We knew that Discord was going to be really useful for us because we could engage directly with them and tap into that audience in a much better way than we could through any other social channel currently out there or through email. 

7. Control the flow of information

So the next thing we did here is we had people under NDA. But here’s a little secret. It wasn’t a real, legally binding NDA—even though NDAs are kind of questionably legally binding. This was just a checkmark on the actual signup form. And there was just a checkmark that said, you know, I’m going to participate and I agree to not stream or share any game details outside of the official Discord server. And that was it. We basically just threatened people that we would ban them if they did this. And almost no one did. A couple of people would do it occasionally, by accident. And the community would actually self-regulate. They would come in and go to that streamer and say, “Hey, this is under NDA. You can’t share it.” And it worked, even though we were up to, at one point at the very end of the beta, almost 200,000 players. At that point, we actually had to ban a couple of people because they were intentionally trying to share stuff. But it went nine months without us ever needing to actually do that.

I think controlling this is really important because it allows you to build a lot of interest around anything that you’re going to put out. And there’s a real scarcity of content that you keep control over, and that becomes really valuable for every piece of content you then put out.

8. Create a cool kid club

Next, inside of the Discord we started to really build a community by having a number of reasons you wanted to be there. So if you were an actual tester, we had this Discord bot set up where you would send the Discord bot your key—the actual string from Steam, or we went through a couple of different platforms at that time. But you take your Steam key, and you would send it to the bot. The bot would check that to see if it actually was a real key and then give you a Discord role. That role would come in and unlock a series of different rooms for you. They were all private. They were all for testers only.

And so it created this cool kids club for people that actually had that key and allowed them to engage with other testers and feel like they were getting secret info. Because at any given time only about a third of the people in the Discord were actually testers. There was a big group of people that were not. So those people were constantly asking, “What’s going on? Is the game fun? Is it worth waiting for?” And all these people that were testers were like, “Yes, the game is super fun, definitely be excited. Your key is going to be coming soon.” And it created this really interesting dynamic. 

On top of that, we also set up a bunch of reasons to be in Discord. Discord became the first place we posted everything. In our work in progress channel, we had developers on the team just drop-in a video of whatever they were working on that week, whenever they wanted to. At that time, we were doing weekly updates to the game.

And it drove a bunch of interest because then the community knew that if they were on the inside paying attention to the Discord, they would see the newest stuff before anyone else. There was a meme that was shared on 9gag and a number of different sites about folks that didn’t have a key yet. I think it really represents the build-up and value people get from actually being involved and engaged. 

9. Force social connections

Once we had all these folks in Discord, we basically forced them to become connected to each other. Spellbreak is a battle royale game. At the time, we didn’t have enough developers on the Proletariat team to fill a full lobby ourselves. So we needed players in order to even test the game. We set up regularly scheduled playtests where we basically forced everyone that was playing to get into voice chat in Discord. What that meant was all of these people were starting to make friends with each other because they would get into Discord, they would start to chat, and they’d be like, “Oh man, good play. You guys, you beat me last time. I got you this time.” And we started to develop these social connections across this entire group of a few hundred players that became really strong and was really the undercurrent and basis for the community. Without us really pushing these playtests, I don’t think it would have happened. Because I think that that sort of loose social connection of just being in text chat or whatever is not the same as forcing you to play with the same group of players, over and over again, for weeks.

That was really valuable to create the undercurrent of the initial community for Spellbreak. It also gave the developers an opportunity to really engage with the players. I was often in these, frequently playing with players. And so there was tons of great feedback to be had out of that. But the real value, I think, was in establishing the social connections between the community and making them all comfortable playing with each other.

10. Develop a community content generation engine

Having the game under NDA meant that no one could share content, but they wanted to make it. We built a highlights channel in our Discord, and we said, “Hey, you guys can use Streamable,” which doesn’t have any externally facing links, “to create your highlights and share them with each other.” We really tried to push them to give feedback to each other on what made a good highlight. And we eventually instituted a thing where we said, “We’re going to take the best montage highlight reels, and we’re going to post those to the official Spellbreak YouTube page.”

That meant that people were training themselves to make great Spellbreak content by working as a community to give each other feedback. And they were incentivized because, since the game was under NDA, they could be one of the few people who got their sweet plays shown on the official Spellbreak YouTube. For us, it was awesome because it was just content we didn’t have to create. These people did a better job than we could have at the time.

This meant that we had a whole bunch of people making this content. Then when we actually went fully live into the closed alpha, when we lifted the NDA, we saw that so many of these people had this whole library of videos they dropped on day one. So right when we went live, there were a whole bunch of YouTube videos where people had 5, 6, 8, and they were already in their backlog and they were all awesome. They just dropped them all. And that was a huge benefit to us. It created interest and this positive attitude around content creation inside of the community really early, that I think was incredibly valuable to generate that kind of content.

11. Establish a content sharing flywheel

Next, because we told the Discord community they were going to see everything first, we would share any of the public stuff that we put up there initially. The best example I can give of this is the pre-alpha teaser video that launched our YouTube channel in October of 2018. We had no followers in there. So initially we said, “Hey guys, can everyone follow the Spellbreak YouTube channel because we want to get enough subscribers to have the custom URL”. Once we had that, we posted 90 seconds of gameplay. And what ended up happening is so many people engaged with it coming out of the Discord, because we said, “Hey, you guys are going to see this an hour before we flip the switch to be public.” That made the video go viral on YouTube.

This is one of the things that vaulted us in terms of size: it went viral on YouTube, and it got picked up organically by a number of different gaming websites and a number of different gaming influencers. Our Discord server doubled in size in four days. We went from 30,000 to 60,000 people in the course of 4 days. And that was all driven by this one piece of content. We basically said, “Everyone come in here. Share, like, update, like, upvote this.” And, you know, we started to do that with everything. So we’d say, “Hey, you guys get to see this early, go check it out, like, comment, whatever.” And that would really help us seed the engagement with all of our content when it started.

12. Use memes as marketing

We also had a Discord meme channel set up, in the same way that we had a highlights channel. Lots of people would hang out there. We made the meme channel accessible to everyone, not just testers. I think that was really important because it meant that tons of people who didn’t have a key yet could still be making memes. And the people that did have a key, the testers, could also be doing it as well. They built some of our best marketing campaigns ever. These memes would get shared out through their own channels, on their own Discord servers, on their own Twitter.

They were giving us the best possible marketing about how valuable it was to have a Spellbreak key and how great the game was. Having that meme channel was incredibly useful. I kind of thought it was silly when we first added it. I was like, “Why did we do this?” Then I saw the impact of these things getting posted all the time and I was like, “We really need to double down.” I think we’re considering even more, how do you take this one step further? Because I think one of the big things to consider right now with an ongoing live-operated game like Spellbreak is the opportunity for it to become the memes that people create in the game. It almost becomes the medium of communication between friends.

That’s something that we’re starting to see more and more in the way people are sharing and communicating. The more you can enable and train your community to do that, the more they’re going to use it as a tool. And it doesn’t only benefit you, because mostly this is all inside baseball, right? It’s all people in this public community sharing memes themselves. But occasionally, some of these memes break out, and they go outside of your community. And then you end up in a situation where a bunch of people in the Speller community are like, “Haha, that’s so funny,” but there’s a few people that are like, “What game is this? What are you guys talking about?” You’re able to attract free engagements, free new players because you’ve built this community around sharing these memes.

[Audience questions]

Andrew Green (a16z market development partner): Thank you, Seth. That was awesome. So we’ve got a bunch of questions here. The first one is: Have you tracked pre-alpha player retention versus launch or toward launch? 

Seth Sivak: Yeah. Some subset of those people are insanely engaged. They’re some of our most engaged, they play every day. Some of them have thousands of hours in the game, which is just crazy. We did a calculation on the top 1 percent of Spellbreak players, and they basically play 40 hours a week. And they have since the summer of 2018. That’s insane.

One of the challenges in going through the process from pre-alpha to alpha to beta to where we are now, though, is that we changed the game a lot. The thing I would be conscious of is that while those users can be really great, you’re basically torching them. So a bunch of players came in, and they really loved some of the original movement mechanics and the really, really high-skill cap the game had initially, and we changed that over time. And it’s just, like, it’s not the game they love anymore, and we’ve lost them forever. That’s fine because they were a really small subset. And I think at the time if we’d kept on that same path, the game would be too niche. But I think you should expect to lose a lot of those players.

One of the real reasons why we limited the number of keys we were giving out is we just wanted to have enough players in the game to get reasonable retention and metrics without torching those players. Because we figured a player will come in and we’ll lose them, whereas if they’re still waiting for a key and they haven’t had a chance to play it yet, we know we’re going to continue to increase retention. So let’s use the minimum number of players possible to continue to get that information without using all of them. That’s one of the reasons why we never did an open beta, because we figured that would be like a freeway to get rid of a lot of really great users knowing we’re going to continue to invest in the game and increase their retention.

Andrew: Do you foresee an issue in scalability when using Discord as your central community hub?

Seth: Yeah, so we’ve bumped into the upper limit a few times. The soft cap is about 300,000. So we basically call anyone who hasn’t been active in a month. And that’s like posting on a channel. I think you’re seeing some of the bigger games now. There are some games that are bigger than us that are approaching the 500,000-person limit. I think there is some scalability concern. What we see now is a lot of sub-Discord spun-up by region. There’s a number of competitive Spellbreak Discords. There’s a Discord for Brazil and Germany and France and Japan. So there’s a bunch of them, but we use Discord as the central hub for our main community stuff, mainly because we felt like it’s the lowest barrier for us to possibly engage. But yeah, it’s definitely a concern. We would not be able to actually moderate a community that size without an army of volunteer moderators.

We have several people on the community team, but they spend a lot of time working with volunteer moderators. We have something like 15 or 20 of them. Some of them speak multiple languages, and we’re very lucky to have that because otherwise, it would be completely unmanageable. But I think it’s important these days to decide on a single focal point where you want to send people to be like, “Hey, go ask your question in this server or in this Discord server or ask this question on Reddit” or whatever it is. More and more, I think you’re better off choosing an existing platform than running your own forums. You know, our original game, we had our own forums on our own website, and I just don’t think that’s valuable anymore. I think you want to pick a single place and drive people there.

I think it’s important to do that early, because if you don’t, the players will choose one and it might not be one you want. So, for example, if you don’t have control over your own Reddit. We have our official Reddit because we were the creators of it, but that’s not the case for a lot of other games. But then if you’re not the owner and the moderator of that, that can be really scary. Or, you know, if all of your community is just going to the Steam forums, for example. You want to be thoughtful on where you want to send your community. Pick a place early and then stick to, “Hey, this is where everyone should go. You have a question, go here.”

Andrew: Merging off that, how do you leverage Reddit?

Seth: So, Reddit is a much smaller audience for us than Discord. We use Reddit for longer-form stuff that needs to be less ephemeral. Things in Discord move really quickly, as with any kind of chat sort of thing. So our official patch notes, for example, are anchored in Reddit because we think it does the best job to give people a chance to communicate with us. We do patch notes in there. We do other forms of content in there. That’s where our official roadmap sits, rather than having kind of our own separate blog—which we do have and use occasionally. We’d rather do it straight in the Reddit.

I don’t know if Reddit’s the right thing long-term because, again, I think we only have 19,000 or 20,000 people in our Reddit or something like that. It’s definitely much smaller than our Discord, which has like 220,000 or something. So it’s something that we don’t use as much. Right now, though, the vast majority of the things on Reddit are memes. And that’s what we’ve seen from all of our channels, that they slowly get taken over by everyone just sharing memes. And we’re kind of okay with that. But that’s what we use Reddit for.

Andrew: How do you think TikTok changes things today? Does it at all?

Seth: Yeah, I think it will. I think it’s still very much the Wild West. I don’t think anyone really knows what to do about it. Spellbreak has a lot of views. But at the end of the day, the vast majority of the stuff that’s getting shared for Spellbreak is still gameplay clips, whether it’s TikTok or Instagram or any of these other places. Now, we might get to a point where we have some sort of insane cultural relevance, like Fortnite does, where people are sharing the Spellbreak dances or whatever. But I think right now I see it as another avenue for sharing your sweet gameplay clips or your memes.

So I think it will be another useful channel. We’re all trying to figure out how to leverage it effectively. It seems like there are not quite as many tools to do that, because it’s not really a social network. It’s not like, you know, everyone that follows Spellbreak will see the Spellbreak content all that often, because of the way the algorithm currently works. But it’s definitely something where you absolutely should be caring about it and thinking about it. With the monster growth in the last six months, it’s become a much, much bigger channel when we do surveys of our community. We ask, “What social networks do you use?” And I think it’s grown from like only 12 percent or 14 percent of our community to like 60 percent of our community is now on TikTok, just over the last six months.

Andrew: Wow. Interesting. Can you give a high-level overview of who managed these efforts for the studio? Was it the dev team or the marketing folks? How did you scale that team, and how much time did you dedicate to that ramp?

Seth: We had a couple of people focus on this. We had, at the time, our community director, who’s now our global head of marketing. He did a lot of the early stuff with Imgur. And then it was really a group of only like three or four people that were working part-time on it, as well as the full-time community manager for the early stuff. That was really built around wrangling the content internally, getting help from the development team to get, like, cool animated gifs made and doing the work to actually post it. We sort of relied on the entire dev team to help with the engagement on Discord. Having them drop clips of what they were working on or join in the playtest was really valuable.

You don’t need a large team to start this, but I do think it’s really important to have one or two people that are focused on your growth side and creating that content and actually doing that work. And the sooner you can start to slice off resources to actually be developing that sort of content or enabling your community to develop that content, the better off you’re going to be because it’s always going to be a challenge to prioritize it against the development team.

Andrew: Did you engage in direct moderation and admin of your communities on Discord, or was it mostly pushed content or moderated by moderators that you onboarded? When did that start to change?

Seth: So when we were small, when it was only like 15,000, 20,000 people, it was all internal. We didn’t have any moderators. It was basically a couple of people on the development team and our director of community at the time. We were all moderating it because it wasn’t… We could wake up in the morning and read all the activity really quickly.

Then as soon as we kind of broke out of that level, it became completely unmanageable really fast. But we have always done a lot of work to train the volunteer moderators. So it was still really hands-on. At the end of the day, there’s only a handful of admins that can actually do stuff. It would still rely on the moderators flagging, “Hey, this person is spamming” or whatever, and then someone would have to go in and manage it.

It wasn’t until we grew to the size we are now that we had a lot of people managing it overall. When we were small, though, that smaller community, you can do it with just a handful of people kind of part-time, because there isn’t a lot of interest and upside and people coming in and kind of ruining the community at that point because there aren’t enough eyeballs. But once you get above, I would say 20,000, 30,000 people, you need to take it much more seriously because then there’s enough people in there for people to find it interesting to spam.

Andrew: Got it. So yeah, we’ll do one more question. How do you plan on carrying the momentum? As you said earlier, you changed the game a bunch, and some of the earliest folks in the community got burned. A lot of them were retained. As it keeps growing, how do you continue to engage, from after the beta ends to the upcoming full release? Especially when you don’t have a formal release date.

Seth: Yeah, things definitely shift. If I could do it all again, I would have kept the game under NDA longer. I would have kept it so that we could have been in control of that information and sharing content on a more regular cadence. The biggest thing that we’re focused on right now, as we get close to release, is we’re up to about 1.7 million or 1.8 million people that have signed up during the course of the pre-alpha, the alpha, and the beta. And so we’re obviously going to be leveraging all of those people on the email side. We’re also doing some straight, traditional marketing, we’re doing some digital spend, and we’re doing a bunch of influencer spend, as well as trying to do as much as we can organically with influencers, too. So that’s a big piece, but what we’re hoping to do is start to activate the super-engaged that are still in our Discord to really help propel that.

So our lead-up to launch includes a 10-day countdown that is going to be a bunch of content that we really want them to start to amplify. And we’re doing something similar here, where effectively we’re going to be opening up several different play sessions for different regions to come in and see the launch version of the game early and capture video, whether they be from the journalist side or the influencer side, so that they have their videos ready to go that are under embargo for launch day. We’re still sticking to a lot of the similar tactics, it’s just at a different scale. We’re trying to back that up with some more traditional marketing spend, now that we’re like on the very edge of launching.

Andrew: Awesome. Thank you so much for the presentation and for the Q&A, Seth. That was super informative.