Supercell's killed games: Celebrating lessons, not failures

To celebrate the launch of Brawl Stars, Supercell paid for flight and lodging to bring GamesIndustry.biz to its headquarters in Helsinki, Finland for two days of interviews and discussions about the studio, its games, and its culture. This interview is the second in a multi-part series covering those discussions. You can read the first part on the studio's team-led culture and structure here or continue to part three on its relationship with its game communities.

For many members of the Brawl Stars team at Supercell, the game's launch marked their first full release with the company. But for those who had joined the team after working on a series of killed titles within the studio, it was even more momentous.

Since its founding nearly 10 years ago, Supercell has fully launched five mobile titles including Brawl Stars -- a surprisingly small number for such a successful company. I was told that, not including the studio's live titles, Supercell typically has between four and six games in development at any given time, and sometimes as many as eight. With only five total full launches, that's a lot of killed games.

"Since its founding nearly 10 years ago, Supercell has fully launched five mobile titles... That's a lot of killed games"

Even Brawl Stars had to face the prospect of its own death at one point. The game soft-launched on the Canadian App Store in June 2017, more than a year before its full release. Game lead Frank Keienburg told me that the unusually long soft-launch was a result of Brawl Stars being outside the norm for a Supercell title in a number of ways. It was a real-time action game on mobile, a genre for which Supercell has no internal benchmarks, and few external examples from other studios. This unusual situation required a shift in strategy before the team could call the game viable for full launch.

"Before summer, we needed to make a decision," he said. "But we really felt we needed to do one more thing, which was launch it on Android, in Asia specifically. Asia to test if Brawl Stars had a chance of success in Asia, and Android because generally Brawl Stars is a game which is played by younger players.

"If you look around the world in this age bracket, it's not like everyone is on iPhones. So you get the game, you're really excited about it, you go to your friends and say, 'Let's play this together!' because the game is made to be played together. And they're like, 'Is it on Android?' And then it stops.

"When we took this barrier away and had both platforms together, that really opened up a viral effect and we saw with very minimum marketing efforts from our side, without actually creating a lot of content throughout the summer, we had a very engaged audience. After the summer, that gave us the confidence to say, 'Yes, this is it, we are on a good track. If we can do more content, if we can ramp up our internal pipeline for content, this can really be a great game and will be played by the community for a long time.'"

What Keienburg and others at Supercell told me throughout the Brawl Stars launch event was that decisions such as these - putting a game in soft launch, keeping it there while trying different tactics, and eventually releasing or killing it - are all made by the game teams themselves, and are never decisions made on high by executives or outsiders. That's exactly what happened to another Supercell title, Smash Land, which was soft launched in April 2015 and killed three months later.

Jonathan Dower, one of the developers who worked on Smash Land, was hired to join two other Supercell employees who had the idea for the project. He told me that while games across Supercell might be killed for any number of reasons, in Smash Land's particular case it was a combination of the game changing dramatically from its original vision over time, and an understanding that the title might not be sustainable in the long-term.

"The original idea was to make a 'build-without-the-battle' game; more of a Settlers vibe where you just short of chop wood and have a lovely life," Dower said. "When we got here, I was here for a week and helped out with some other teams doing concepts for some characters, but then we started day one and we saw that game Monster Strike from Japan and we loved the mechanics, so Day 1 we changed it to this kind of ping-pongy game.

"In soft launch, we started changing the game to try to get players to play longer, better retention. It was okay, but it wasn't awesome. And the more stuff we did to change the game, it started to feel messy. It wasn't the core that we were originally building. It almost felt like it would be better to start from scratch again. And we just looked at ourselves and thought, 'Could we work on this for two years?' And none of us felt like we had that original passion."

The official decision to kill the game wasn't easy, Dower said, but it did come from an understanding that the game's entire team (by soft launch, around ten people) had together. It was on them to take the decision to the rest of Supercell at the company's regular Friday meeting, and as is tradition for the studio when a team kills a game, they prepared to send Smash Land off with a champagne toast.

"We wanted to have a break from the team room, just to relax and have a sauna together," he said. "The decision came out of there. We kind of knew we were going to do it before anyway, but it was a team decision. [At our Friday meeting,] I got up at the microphone and said, 'Sorry guys.' I was almost in tears. I had the champagne up there. I was going to toast everyone and say thanks for working on the game because so many people were involved, but as soon as I picked up the microphone, I knew I couldn't drink, so I had to put it down. It sucks. You get very attached to what you're working on."

Despite the high number of killed games at Supercell, Dower and others told me that it had not affected staff retention. Though multiple employees told me there was no specific process for members of a killed game's team to move to other projects, those same employees also said that after a game was cancelled, everyone just seemed to "find their place." Dower specifically mentioned that there wasn't anyone he could think of in charge of letting people go when projects ended.

CEO Ilkka Paananen was able to outline the pattern of teams being reabsorbed elsewhere in the company with a bit more specificity.

"If a team decides to kill their game, sometimes the team decides, 'Actually, we work really well together as a team, so let's give it another shot and work on something else.' And maybe they have another idea and they can do that. In other cases, they think, 'Yeah, this was okay, but let's all go to different teams,' and someone's interested in this game and someone's interested in another game, and then the team gets separated."

"Failure sucks. We spent ten months on [Smash Land], and we had to throw it away"

Jonathan Dower

No one I spoke to at Supercell downplayed the difficulty of killing projects, especially when individuals had been on multiple killed projects in a row. Dower said that while there isn't any formal support system for employees struggling with these decisions, the studio's culture is such that this discouragement is frequently and openly discussed.

"Failure sucks," he said. "We spent ten months on [Smash Land], and we had to throw it away. Failing definitely sucks, but the learnings we get from that are the sort of thing that we're trying to celebrate. We do it for many things; even the admin team might have champagne to celebrate some failing they've done. It's not like we're just chugging champagne, I don't want to give that impression. Failing sucks, and we take it super seriously, but celebrating the learnings is the point of it.

"The most important thing about killing [a project] is that people can actually have impact in the company elsewhere. A couple of the [Smash Land] guys went to Clash Royale to help get it over the line. They were quite crucial. I helped out with some character development on Clash Royale as well, then I jumped on another new game team and we killed that, and then another one, and we killed that as well, and then I ended up as game lead on Clash Royale for two years.

"That's the whole point of killing stuff. If we kept that going, we'd have to grow the team to maybe 20 people to get it global, and then if we just kept going global for every game we make, this company would be a thousand people, easy. You have to weigh up. It's really hard to do. I think we made a good call at that time, because we all did stuff that was better."

Keienburg echoed Dower's statements on the importance of taking lessons to other projects, saying that this was part of the entire company's culture, and added that failure didn't always mean that a team had done something wrong in how they made the game.

"There's also the realization that you can fail without making mistakes," he said. "Failing and making a mistake is not the same. Sometimes it's just not within your control. Sometimes you have a great game, you have everything together, and it still doesn't work out because the market doesn't want it. That's maybe more disheartening."