Maybe it's time to kill your game and move on: Supercell on cutting its losses

There's the Silicon Valley mantra of "fail fast, fail often." It essentially means you ought to get out and make something, expect to fail (and fail) to reach your initial goal, learn from your failure, then upon your next iteration apply what you learned from that failure, for a better result.

There's the Silicon Valley mantra of "fail fast, fail often." It essentially means you ought to get out and make something, expect to fail (and fail) to reach your initial goal, learn from your failure, then upon your next iteration apply what you learned from that failure, for a better result. Then you fail again, then repeat. In theory, you're supposed to get to a great result in a shorter period of time by learning from a concentrated succession of failures.

Following that mantra, Supercell is super-good at failing at making successful games. This might come as a surprise to some -- the company is one of the few mobile game studios able to command positions at the top of the top-grossing charts on iOS and Android. Games like Hay Day , Clash of Clans , and most recently Clash Royale have pushed Supercell's valuation to over $5 billion, reflecting inarguable financial success.

But that success is built on the graves of Supercell games murdered by the very teams that made them. At GDC in March, Jonathan Dower, a game lead and artist (and a self-purported serial killer of games) at Supercell outlinedhow out of the studio's last 10 games, seven were killed in prototype, two were killed at soft launch, and one -- Clash Royale -- actually launched globally.

Clash Royale is currently topping the mobile charts. That's not bad for the one out of 10 games that made it out alive.

In a follow-up interview following his GDC talk, Dower explained when a developer should kill off an in-progress game. "I think it's different with every team and every game," he said, "but once you get that feeling about the 'other game,' or about the game you 'should' be making, in your mind, you've probably already killed your game. I think that's a good indicator."

"Once you get that feeling about the 'other game,' or about the game you 'should' be making, in your mind, you've probably already killed your game."

Note that this is Supercell's way, and not applicable or appropriate to all game types. There are arguments for sticking with a sub-par game and developing it over a long period of time, instead of outright killing it. One recent example is Rocket League , a game seven years in the making.

There's also the argument that Supercell is a company that can afford to completely axe a game (or games) before release, and many game developers aren't in a financial position (i.e. have a $5 billion valuation) where they can throw out so much time and effort.

"Supercell has had some really good luck with their games, but they were that studio [that didn't have a financial safety net]," Dower argued. "They were that company that was killing games before releasing anything successful."

Even with the arguments against the idea of killing games often, there is something to be said about analyzing when a project is dragging a developer down, keeping creators from moving onto better things. "We get emotionally connected [these games], but at some point you have to think [whether or not] you can fix it," Dower said in his GDC talk.

One might think that a company like Supercell runs comprehensive analytics when deciding whether or not to kill a game -- and of course metrics do come into the conversation -- but for Dower, the decision ultimately boils down to the gut feeling of a small team of experienced game makers. The execs at Supercell don't issue a death sentence or grant a pardon, Dower said. The life or death decision is up to the people who've had their hands on the game.

That raises the question of what happens when people on a team disagree with the fate of a game of questionable viability. Dower, whose team jointly decided to kill the soft-launched game Smash Land in a sauna over beers, said it's best to just keep the discussion going with your coworkers.

"We'd probably just keep talking about it, keep hashing it out," if there was a disagreement over a game's fate, he said. "If someone [on the team] really, really believes that they can keep this going, then there might be a discussion...[I'd ask them to] 'Prove me wrong.'"

Supercell has 180 employees worldwide, 70 of which are developers. Dower said developer-driven game-killing wouldn't be possible without the trust of management, an important key if you're to adopt the philosophy of Supercell.

"That's one of the base qualities, the trust," said Dower. "We're hiring people who can make those big decisions. [The CEO's] role is to trust them. Otherwise, the whole system breaks down."