It’s easy enough for an online service to start clustering people into like groups based on physical location, topics of interest, family status, or any number of other factors that a site may take into account within its profile system. However, it gets much more difficult when the site allows you to manually cluster connections instead of doing it automatically. The fact is that the majority of people will not participate in this type of tedious exercise just to gain some simple communications benefit for interacting with a specific group. In addition, users may cluster only some of their connections, leaving a large portion in a group “void.” The real task here is finding a method to automate this more meaningful type of user clustering.