Social filters are a must for communities. I’m aware this may be perceived as a bold statement, but communities rely on the production of small bits of content from a whole variety of people. Users engage the system and the others in it by producing content for them.
The inevitable question then is how to prevent this pile of highly unstructured, user-driven information from being just that? How do we make meaning out of this content? And also prevent it by being highjacked by a few users that don’t represent the intentions of the sponsor? Social filtering is one of the answers to these questions.
Social filtering is a concept that has been around for a while, evolved from the earlier collaborative filtering. One of the well known implementations of collaborative filtering is the iTunes recommendation engine. Social filtering is essential in a community system as when you have more than 30 pieces of content or 30 people it becomes impossible to identify the most relevant content and people.
Traditionally, one way of doing this was to manually survey the data, read it, find what seems the most interesting and appropriate for the target audience, and highlight that. This is a very common practice, and works well for communities that have the resources for an ‘editor’. Another way is to make visible what other people like and show that. This has a dual purpose: firstly, it allows communities to exist with less moderation by people because the members and their preferences are taken by the system and used for moderation; secondly, it taps into a second phenomenon, which is people’s curiosity to know what others are thinking. A great example of this second phenomenon is on YouTube – one of their main ways to emphasize videos is to show what others are watching right now. And to drive the point home even further, Gartner is also starting to see the same thing with their emphasis on analytics.
I’m sure I’ll have more to say about social filters. What are your thoughts? Comment below.
