Tagging and Enterprise Knowledge Management

There was an interesting series of workshops in Edinburgh, which focused on collaborative tagging. It had some interesting presentations from various social tagging industry types. I’d like to bring to light three themes that I found interesting in terms of inside the enterprise. Tagging brings to an enterprise a valid bottom-up mechanism of harnessing intelligence in their human network. If you can somehow combine the bottom up, with a top-down ontological view, you would have a pretty confluence of knowledge management approaches.

Finding expertise in the organisation

There were two papers that dealt with finding expertise in the enterprise. This is a difficult problem for larger, geographically dispersed organisations. Fringe is a system developed in IBM, that allows IBM’ers to tag people. I could tag you with “agile” to serve as a marker for the fact that you’re a fellow agilist. Or I could tag you as “websphere+expert” to denote that you know you are a websphere expert. A long time ago, 1999/2000 I was inspired by a KM system that mined email to discover who knew what. I can’t seem to find such an organisation anymore, but this Fringe tool seems to be able to take it's place.

A group at Avaya tried to induce expertise based on a persons tagging behaviour. They built a tool that would help them find experts quickly called Hermes. They use some pretty fancy maths to figure out the “ExpertRank” based on this behaviour.

When you consider these two methods as two dimensions that can be part of an individual’s profile, you can see how you could quickly find people interested in becoming experts, and perhaps even the experts.

Tag Quality

Tag quality is hard. If you open the ontology to uncontrolled labels, you create an opportunity for mess. Everyone has a different understanding of what tagging is for, what tags to use etc. However frustrating, these differences are exactly what you want. A couple of papers talk about fixing this; A paper on suggestions talks about using some fairly clever algorithms for suggesting tags based on a variety of factors, content, semantic normalisation, having a canonical model, etc. They’ve implemented some of this in My Web 2.0 (which is some yahoo tagging thing).

A second article discusses trying to figure out an ontology based on flickr tags. Again, they’ve got some pretty fancy algorithms for comparing tag clouds and understanding structure.

Findability

It was actually, the site, findability where I found a link to this conference. There are two articles on findability. The first discusses an alternative interface to wikipedia, which is designed to use bittorrent to serve large media files. (wow! How web2.0 compliant is that?). They mine wikipedia and generate tags based on text then present the tags via their own p2p interface.

The final article I found interesting was about tag clustering. This paper discussed trying to find tags that were related, based on common tags, and trying to create some kind of alternate navigation mechanism. In fact, this is seemingly what the rawsugar guys are trying to do.

With all this data, there's definitely space for clever time based visualisations of tag clouds.. if only I had the skill and expertise..

tags:tagging, ontology, knowledge+management

 

Comments

[Dennis D. McDonald]

I just discovered your blog. Great stuff and directly relevant to the topic "expertise management" that I am currently researching under the assumption that many folks are already applying "web 2.0" and "social networking" techniques to knowledge and expertise management.

Dennis McDonald
http://www.ddmcd.com/expertise_management_index.html

[Yasin]

Great Article. I would suggest socialmarc for its simplicity.

[salt]

Sir/Madam(sorry!, couldn't find out about u)

I find the content is pretty good and exhaustive and met my requirements but still half a min of reading and i wished u had made an effort to create a better contrasting effect for better readability ....
just a thot....
thanks