I've run the Global Game Jam in San Francisco four years in a row, along with a variety of other events like the SFMOMA GameLab-A-Thon and SF Indie Game Jam. Here's a few things I learned:
Don't make it a competition
The fastest way to disenchant people from being creative is turning it into a competition. The fastest way to make beginners and newcomers feel unworthy is turning it into a competition. The fastest way to make the majority feel bad about their creations is turning it into a competition.
Just making a game in 2 or 24 or 48 hours is a momentous accomplishment worth recognition. Don't make it feel like wasted time because people didn't "win". If you want to build a strong community, keep everything collaborative. Celebrate accomplishments, recognize standout talent, but don't rank. Creating winners also creates losers.
At the beginning and end of the jam we emphasize that it isn't a competition. At the end, we let the audience vote and we recognize outstanding excellence in different categories: technical excellence, excellent use of theme, etc. We do them quickly, humorously, and there are no prizes
. We remind people that everyone completed the challenge. Everyone who made a game won.
Have an inspiring theme
I was delighted when GGJ this year returned to their phrase-style themes. Phrases, especially adages or common metaphors, give players an abstract dynamic with plenty of space for personal adaptation. Go too literal, and most of the games will be similar. Go too abstract, and people don't have jumping off points.
One suggestion (courtesy of Ian Schreiber) which particularly tickled me and was successful for our SF Indie Game Jam, was to go to a random wiki page and let that be the theme, or go to several until majority accepted the theme.
Let people hear each others' ideas
Common creative advice is to throw out your first [X] ideas because they're the default, the obvious, the sputters before the gears really start turning. In jams, those are unfortunately also the ideas that get consensus - because they ARE the most normal, or obvious. To combat this, we started having each team pitch their game to everyone after about an hour of concepting. This lets people see which ideas are common, and we've wound up with much more diverse games since we started doing this.
Give specific opportunities for break ups
I didn't start doing this until recently, but we now have a set amount of time where groups can "break up" with each other. We call it out explicitly and always have food and drinks available, so people can excuse themselves gracefully from a group and try out other options. We got a hugely positive response from doing this. 48 hours is a long time to commit to a team you don't like, and we've seen our retention increase since we gave people explicit opportunities to search out better teams.
Healthy food energizes best
Food is energy, and at jams people need plenty of it. Most of the time we default to what's cheap: soda, chips, peanuts, etc. When we started getting the budget for healthy snacks (hummus, vegetables, fresh fruit, dried fruit, bread, cheese), we noticed people weren't crashing, and the energy was more sustainable over the course of the jam.
There IS a "just right" for team formation
Team formation is always a sticking point at jams. We've done it differently every year, but I think we finally found the right mix.
- Jammers need autonomy: One year we tried assigning groups prior to the jam based on surveys jammers filled out. The majority of jammers were not happy, perhaps because skill compatibility doesn't equal social compatibility, but perhaps because people felt forced into something they hadn't chosen. Some pre-formed teams also showed up and were upset about getting split, or caused us to break our already fragile arrangement. Now we form teams on site.
- Make the invisible visible: From the first jam, we've utilized crowns to help visualize people's skills and experience. Jammers create strange paper hats, the color(s) of which corresponds to areas of expertise: 2D art, 3D art, sound, design, production, and various programming languages and environments. This works for two reasons: it gives people something to talk about (admittedly this isn't the most inherently social crowd), and it lets people seek out others.
- Suggest, don't dictate: We encourage group sizes of 2-4. We encourage people to come without pre-formed teams. But we have no rules. Ultimately, they just get in the way and make people unhappy. Team formation rules are useful to the extent that they enable people to have a fun, successful jam. Past that? Ditch 'em.
- Emphasize team flexibility: Remind people that these aren't their final teams, that they'll have opportunities to switch later. They'll be far less stressed and much more open if they're not worried about making The Big Decision. (And with this method, they usually stay with their first team anyway!)
- Facilitate ratios: We split people up by expertise and then make a game out of finding missing team members. "Each pair of programmers needs to find one artist and one designer or producer" is usually all we need to say. Thanks to the visibility, they handle the rest on their own.
- Have a late team: Some people inevitably show up late. While sometimes they can sneak into existing teams, we found it beneficial to have a place where people who were still looking for a group could hang out. More often than not, the stragglers wound up forming their own coalition.
Two projectors are better than one
This one is small, but at the end of 48 hours when people are exhausted but still want the payout of showing off their work, having two projectors is a godsend. Have one team setting up on one projector while another team is presenting on the first. It basically doubles the speed at which teams can present, letting everyone go home and sleep all the faster.