‘Let’s just fail quickly this time’: Semiwork took a ‘big risk’ on REPO after its first game took 6 years to make and didn’t sell very well

The co-op horror game REPO is, like Schedule I, a surprise indie hit. Released in February, it’s got an “overwhelmingly positive” rating across more than 91,000 user reviews on Steam and remains near the top of its most-played chart: As I write this, nearly 170,000 people are playing.

In a new developer video diving into REPO’s creation, semiwork’s Pontus Sundström explains how it all came about, and the circumstances surrounding its creation were not particular fortuitous: Semiwork had spent six years working on its previous game, Voidigo, and while it also has an “overwhelmingly positive” rating on Steam, it wasn’t a big seller. That left the studio in a spot where its future wasn’t assured, and so its approach to REPO was a little bit different.

“Okay, we can’t take this next step,” semiwork co-founder Walter Tischkewitz says in the video. “We did it anyway. ‘Let’s just fail quickly this time.’ And we took all the money from Voidigo and we put it all in this game, REPO.”

While REPO is similar in some ways to Lethal Company, another big indie success, semiwork has previously said it wasn’t actually inspired by that game but instead emerged from an idea for a singleplayer horror game about cleaning a haunted mansion while hiding from a creature called the Headman. Today’s video dives deeper into that evolution, revealing that the decision to ditch cleaning came after it was decided to make REPO a multiplayer game: “Co-op cleaning in real life, it’s not that fun.”

“Yeah, think about cleaning in real life,” Tischkewitz says. “You never go like, ‘Hey friend, let’s help each other clean here, let’s grab the broom together and clean co-op and have a fun time.’ It’s always like, ‘You take bathroom one, I take bathroom two.’ So there wasn’t really anything in the gameplay loop that kept the players together, and that’s when we realized, okay, maybe we’re not doing a cleaning game.”

The idea for what REPO would become grew “organically” from there, while the team took sort of a learn-as-you-go approach to actually making the game because it was their first project made with Unity. That’s also why the first content update, promised in March, is taking a little longer to arrive.

“We are still learning things as we go, and we want to make sure that we learn stuff the right way,” Sundström says. “Which means that future updates after the first update will probably be released a little bit faster because we have a little more knowledge on things and how to do it.”

Plans for that update, whenever it arrives, include various quality of life features, a new map, and a “duck bucket,” a bucket you can put over the duck. Trust me, it’s important.