GroupSwim is a provider of a hosted software solution for creating intelligent communities. The software provides a place for people to author content in community-centric discussion forums, and is a reaction to document-centric collaboration software.

Multiple communities can be created within a given GroupSwim account. That means that when a business signs up for the GroupSwim service, they can then create different areas within the software for the different groups within the business. There can be an area for marketing, another for research and development, and another for post-sales customer support. These different areas are called communities, and the information in one is walled off from the information in the others. Of course people can be a member of multiple communities and the service provides quick access to all of the communities that any one person is involved in. This is one valid way of setting up GroupSwim within a business; another is to create a single community for the entire business and then create a number of interest-aligned groups where everything is in one place but is segregated in a soft-way (group-interest) rather than a hard-way (community-alignment).

If this was all that GroupSwim offered (essentially a Pillar 1 product with a focus on discussion forums), then there wouldn’t be a lot of reason for writing about them—there are many, many tools that provide such capabilities. What makes GroupSwim interesting and worthy of some thought and attention, is the secret sauce that is spread out after the content has been created. The two key capabilities that stand out to me are its semantic engine and its ability to infer interest based on user behavior.

Two Key Capabilities in GroupSwim
The first key to the power of GroupSwim is its semantic engine, whereby content that is entered into a community-centric discussion forum is analyzed for key phases or terms that will be of interest to the wider community members. The semantic analysis and suggestioning (new word!) actually happens when the author is still typing the original material; it’s a bit like having a reference librarian looking over your shoulder and suggesting where you should be filling a new book. It’s essentially knowledge taxonomy without having to be explicitly taxonomical. Users definitely get the ability to type in their own tags, and GroupSwim offers tag type-ahead to assist with normalization of tags across the community members.

The second key to the power of GroupSwim is its ability to infer interest in content based on normal, everyday user behavior. When someone reads a post in a given community, interest in that topic and the related semantic tags is noted. When someone initiates a watch on a particular post or tag, interest is inferred. When someone emails a link to a post to some of their colleagues (“hey, check this out!”), that infers interest. The main point is that interest is implicitly inferred from user behavior, rather than having to be explicitly constructed because the user rated the content as being a “4 out of 5”, for example. Within GroupSwim you can give a thumbs-up or thumbs-down to certain content, but the act of doing so is not the only measure of content interest, nor by extension, expertise of the author.

Other Interesting Things
One of the design decisions within GroupSwim is that a user can not describe their own expertise. Expertise is inferred based on the content a user actually creates and the reaction of the community to that content. Thus if you want to be known as an expert in a certain area, you can’t self-describe yourself in those terms, you actually have to act in that capacity and have other people within the wider community recognize that expertise. It’s a bit like the knowledge management axiom that “knowledge is socially constructed”. You become an expert in GroupSwim because others recognize you as such.

When a new person joins an established community, the service highlights the existing users that are the top contributors, and the areas in which they are seen as having expertise (or “authority”, in GroupSwim speak). Thus a new user is immediately informed about the key people within the community, and doesn’t have to trawl through the entire collection of posts to form their own opinion before deciding whether to trust someone’s viewpoint or not.

The GroupSwim system currently applies its secret sauce to posts in GroupSwim discussions, but with the April 2008 release of the service, the secret sauce will apply more broadly to uploaded documents and files. Exactly how those capabilities will work has not been disclosed, but it is directionally good. Apparently wikis, as the new way of authoring documents, will be the next capability to be added in a future edition of GroupSwim.

Summary
So in summary, there are great capabilities within GroupSwim to share knowledge and expertise around a business within targeted communities. With only supporting discussion forums in its current iteration, the service lacks well-rounded maturity, but good things are apparently on the way. Once the ability to support documents and files has been added, and assuming that it is done right, then GroupSwim will become of broader interest and relevance for organizational collaboration initiatives.

For another perspective on GroupSwim, see sazbean.