With community based tools and games spreading all over the net lately, I find myself thinking more and more about communities and ways to make them scale. Obviously thanks to the Internet we can create communities of scales never seen before. Physical limits no longer apply rather we are limited by human capacity to filter information and by the technology that helps us makes sense of this information.
I spent some time as a very active member of Consumating.com. During this time I learned to love the overheard feature. Overhears show on top of the screen and quote some random post from a random conversation. It’s a great way to discover conversations and people you didn’t know about so far but how does it scale? What happens when consumating grows to 100,000 users? A million? 650 million?
The simple solution is some static division of users: by geography, age, industry, etc. This solution seems to work pretty well in many cases especially if the division is appropriate to the context. Where this approach lacks, however, is when certain categories of the arbitrary division aren’t popular enough and users end up in what seems like an empty site or when some category becomes too popular and crowded. In such cases it might be better to use a dynamic, user-driven approach.
On tribe.net (and I’m told on Ning as well but I never bothered looking) any user can create a new group (or tribe) based on whatever topic (or lack thereof) she finds interesting. Some tribes are location based, some are interest based and some are community based. One of the pros of this system is that it is self regulating. Small tribes with little interest or activity tend to die off while tribes that get too big can fork off into several related tribes. Any user can start an offshoot of the parent tribe to concentrate on more specific topics or subsets of the community. Of course, this eventually brings us to a similar problem, what happens when you have too many tribes?
This problem becomes even worse on mobile. While on the web we can (somewhat) easily navigate lists or even hierarchical structures of groups, this is not so easy on a mobile phone. Navigating long lists is painful, and hierarchies will likely take too many page loads and end up in confusion. Perhaps a search based solution or a matching/recommendation engine would do better in this case. I’d love to hear about any ideas, experiments or even better, working solutions to this problem.