Seppänen says there's one very important rule he abides by when they start looking at those potential partners. "We only work with outsourcing companies that don't create their own games," he says. "There's a big temptation to have your best people on [your game], and then the juniors on the outsourcing, and that shows in quality." There's another benefit, too, in that a company focused on outsourcing is used to working with other people. "Their processes tend to be very good. ... They're used to dealing with external people, and they can push the time it takes for asset [delivery] down. And they may even have libraries of assets, like a lot of cars that they can pull from really fast."
Once it's narrowed the field, Remedy sends out the package - the documents and sample model - and waits to see what it gets back. Sure, the team is looking for a good "pilot" car as an example of the company's work, but Remedy's also evaluating the company itself, he says. "We look at how they deal with it, what kind of questions they ask. Even though we put a lot of effort into the documents, they're not perfect. Never. There are always questions. So, if someone's not asking any questions, that's a warning sign. [We look at] how they deal with [us], do they keep us informed, do they send us work in progress?" And, naturally, they look at the result, too, how it looks, how it fits in the game, whether it meets the specifications they set out. "And that's like the holistic package of how we can evaluate which company we want to work with."
Working with third parties allows Remedy to take a different approach to development. "We tend to work on all fronts simultaneously, and do it all iteratively," Seppänen says. Rather than waiting to develop the art until they have the game completely defined, they're able to start work on it right away, "so, our graphics start to look pretty from the start, but there's not much of it," because a lot of the game is still undefined. "We try to get to a position where we have something running, and then work from that ... because it's so much easier to reach consensus if we have 30 people, [and they can] sit it on the screen, try it, [and] figure out what needs to be done next. It's kind of an exploratory process, especially when we're working on a new IP."
They take this approach because they want to "create something really, really new, and really push the boundaries on multiple fronts simultaneously." For Alan Wake, they want to build a beautiful world, but they also want it to be very open and free-roaming, while telling a cinematic story. He describes it as "research and development, in a sense," which "requires fast moves. Today, we might be going west, but tomorrow, we might be going north." A larger team can be a hindrance when you're trying to move quickly, he says. A team of 80 has what he calls "implementing staff, which are 3-D modelers or level designers. ... And these people, to be effective, they really need to have something specified. So, you would need a long queue of things for them to do." When you're still exploring what you want to do with the game, "you don't actually know what they'll be working on next week. You actually waste a lot of money if you try and explore with those kinds of teams," he says.
