skillRank

Who owns a specific knowledge in your Enterprise? How easy is it to find knowledge holders? Can you do it with a few clicks? And how do you award valuable knowledge holders?

The sad truth is that big enterprises suffer from the massiveness of their size and their burocracy. For talented, or just willing to contribute, is difficult to emerge. Managers often hire a consultant, whom skills are well known and “advertised”, without knowing they have the right people in their company, ready to shoot!

Web 2.0 teach us something: the larger the information set the easier it is to find useful information - allowing favouriting, commenting, tagging and rating important information emerges from a flat schema.

As pageRank (google trademark) brings order to the Web, skillRank can be a solution to bring order to the Enterprise.

SkillRank is an analysis algorithm that assigns a numerical weighting to each tag associated to user’s contribution, be it a document, a picture, an article, a video, … basically everything that can be tagged, rated, favourited or commented.

The idea behind skillRank algorithm is that when a user contributes to an enterprise environment she’s actually contributing for what her skills and interests are. Other users can appreciate the quality of the contribution performing certain actions on the contribute. This propagates the user herself. The algorithm is tag-centric, i.e. the user inherits the tag associated to the contribution.

The algorithm works more or less like this:

• CONTENT gets tagged when published (editor tag has more value, let’s say 10 votes). The tag vote propagates to the USER (just 1 vote)
• CONTENT can be tagged after other users read it. This count 1 vote (for each user tagging) for the CONTENT. This vote propagates to the USER
• CONTENT can be tagged when other users save it to their “favourites”. This vote (one for each user favouriting) propagates to the USER
• CONTENT gets rated: the vote propagates to the USER (-2 -1 0 +1 +2 based on the number of stars)

This algorithm is moving its first steps. Any idea or comment will be very appreciated.