Sponsored By

Media Molecule finds success and failure alike in Little Big Planet's sharing systems

At GDC Taipei, Media Molecule community manager James Spafford discusses the company's failings in its initial attempt to share user-created levels, how they improved them, and how their target still hasn't been reached.

Brandon Sheffield, Contributor

June 27, 2012

5 Min Read

The initial motto of Little Big Planet, even when the game was pitched to Sony, was "play, create, and share." "When I joined, though, 'share' was like a smaller sibling to play and create," says community manager James Spafford at a talk at GDC Taipei. "When we talk about share these days, we're more likely to mean [social media] kind of share," he says, though you could also call it the uploading and downloading of content to a service like YouTube. But more than that, "Content discovery is a massive part of the experience," he says. "Without discovery, content is painful, and you wind up watching the wrong thing." Because of how important it is, LBP developer Media Molecule informally added "discover" to the game's motto. Content ownership is another key element. "If you make something, if you created it, it's yours, no matter what the EULA says," he says. "Technically PlayStation owns it, but you made it." In LBP "Share is the reward glue between play and create," he says. Players are rewarded by other players playing their creations. Unfortunately, the first iteration was somewhat problematic. To discover content in the first game, players would rotate a globe to find. There was a thing called "cool levels" which was decided by a robot. "The problem was the content wasn't relevant to anyone at all," he said. "And once people figured out the criterion for 'cool levels' they started to target that, and it became useless." The team also had trouble with ratings. First they tried allowing people to grant hearts to levels, but they made a mistake by adding an achievement for getting 30 hearts, so people just spammed each others levels asking for hearts. Likewise, they had star ratings, which didn't work at all as intended. "You only really rate something if you hate it, or you love it," says Spafford. "You either rate five stars, or nothing. That wouldn't have been a problem, but we forced people to rate levels." People just hammered the X button to get through the ratings, so everyone rated the game its default rating. "Like the star-rating, we forced people to tag levels [with content descriptions]," he said. "And when we did, we only displayed five of the 40-something tags. Again, people just hit the tag button, and every level had the exact same tags." In short, it was a closed world that didn't function properly, with no access to content outside of the game, which prevented ownership, discovery, and the like. LBP was also a victim of its own success, since 30% of their players uploaded levels, whereas with many games that number is much lower. Within 1 month, there were 1 million levels, which broke their systems. With a million levels, there were 30,000 pages of content you had to sift through to try to find something fun to play.. With LBP 2, "Luckily we were just able to remake everything," he said. They allowed people to customize their own levels planet, and also put user-generated levels in a list. "Never underestimate the power of a list," says Spafford. "A list means you can put the most relevant at the top, and then it can scroll as long as you want it to." They also got rid of tags, and added sensible ones that described content of the game. Creators could choose first tags on their own, and others could tag if they wished to. As for the "cool levels" problem, created by a robot, "We left it there, which I really really wish we hadn't," he says. "We thought we'd be able to tweak the cool levels system a lot easier. We didn't and we weren't able to." They also tried adding Media Molecule picks, which is basically the opposite of cool levels. "This one is chosen by a human: me," he says. "Unfortunately that's a bottleneck, because it forces me to play a lot of levels, and put them into a list. And that takes a lot of time." Alongside various other changes, perhaps more importantly the company made a far-reaching web site. "We put every single level -- and every single person that played the game had a unique url on this web site, so they could link to it wherever," he said. The site allows you to "play now," which will download the level on your PS3 for you, queue things up, or put them in a list. "I can send you my level in an IM, and you can queue it for later," he says. "Before, I'd have to tell you the name of the level, you'd have to write it down, remember it and input it later." The LBP franchise now has 6.7 million levels. Around 5,000 levels are published every day to the servers, and the games are now across many platforms, "So the share design systems not only need to work well, they need to work across all the games," he says. The team learned that it needed to design for evolution. People acting in groups are complex and will evolve, so the games should too. "Marketing people will tell you that your website is the most important thing on the internet. But we believe that your website is a tiny tiny speck in the massive world of the internet," he says. While the site was an important step, he believes it's more important that that content can be shared out across the web, concluding: "We don't think it's important that people come to the web site -- we think it's important that the information on the site is enabled to go out to the web where people already are."

Daily news, dev blogs, and stories from Game Developer straight to your inbox

You May Also Like