Bottom up, Top Down, and Sideways Categorization

I’ve been having fun with the new My Yahoo! Feature, “Recommendations for You”, that recommends RSS feeds based on the current feeds you are monitoring – sort of a Tivo for RSS. You can accept the recommendation and add the feed to your page, reject it, or just go on to the next recommendation. The system seems to have a learning feature, because the new recommendations appear to be progressively more on target, although not without occasional surprises.

Consumer-generated media generally lacks “top down” categorization schemes like those painstakingly applied news and trade content in traditional online databases like ABI/Inform or PROMT. So, we rely on keyword matching, contextual search and text mining to help keep us aware of which blogs or boards are talking about the subjects we’re interested in. For consumers, such personalization features have come to be expected. In a B2B setting, the ability to “discover” new relevant content is emerging as a key requirement - critical for managing reputation, staying on top of new technologies and business opportunities, and competing in a global economy. Since it’s based on the most granular data of all, let’s call this the “bottom up” approach.

User-generated tags, as in Technorati, Digg.com, and del.icio.us might be thought of as the “sideways” approach to categorization, since it relies on peer networks sharing common taxonomies. While one can speculate on the broader implications of “social tagging” and “folksonomies”, one advantage to this approach is its sheer scalability. Purists may cringe at the sacrifice of standards, but if the success of other collaborative internet projects, like Wikipedia, are any indication, this form of categorization is here to stay. Expect to see new search and syndication technologies and new practices in professional media analysis and measurement that will help businesses to harness the power of these new directions in categorization.