Adina Levin, VP of Products at Socialtext, has been writing some great posts of late (Social context for distributed social networks, Tags for ActivityStrea.ms, and others) on the need for creating social context in order to focus attention. This is becoming increasingly important as the volume of content (blogs, microblogs, activity streams...) continues to grow.
Social context as defined by Levin is "the way that people think about what’s relevant to share to whom". In order to capture and maintain attention in distributed social networks we need to be able to share profile data and content at a much more granular level than we can do today. As an avid road cyclist new to Vancouver this would allow me to share aspects of my profile of La Bicicletta Pro Shop Cycling Club on Facebook -- my background in triathlon and cycling, a blog where I write about cycling and so on. Other members of the community would be more likely to be interested in this information as opposed to professional information or other interests or hobbies. Similarly, when I'm sharing information through my blog or twitter, I can more finely target the intended audience using tags.
Last week Ross Mayfield, President of Socialtext, responded to an article in the WSJ (Who Knows What?) that makes the argument that blog, wikis and tagging should be actively used in locating expertise in large organizations. Mayfield's response was supportive but he noted that microblogging and activity streams should be included in locating expertise. Including activities and interactions from blogs and activity streams will provide more valuable insight into credentials than static profile data.
These posts are really addressing the same issues from different perspectives -- Levin addresses the problem of disclosure and Mayfield addresses discovery. The same capabilities and tools are needed to address both. I need to be able to share profile information (that may be distributed) and content (including microblogs and activity streams) based on the different personas I have -- professional, academic, personal. Tagging can provide context to content including blogs and activity streams and profile data. Levin argues that distributed profile data can allow individuals to create different personas.
Having worked in a large organization I can attest to the fact that using profile data exclusively can be difficult and time consuming even when profiles have social network capabilities. It's also true that leveraging social media in expertise location is useful -- because it shows both expertise and interest. Someone blogging on specific issues demonstrates knowledge that a credential alone can't. And it also shows an interest and passion for the subject matter which is equally important. Monitoring activity streams confirms interest and engagement that is also useful in the credentialing process.
But credentials also contain information that is important in the credentialing process -- professional and academic experience and designations, and even interests and hobbies. As the WSJ points out adding tags to profile data can aid in searches.
As Levin points out the practice of tagging content isn't very explicit or well defined -- although she also point to the ActivityStrea.ms proposal as a possible solution to sharing content. To really make tags useful in either syndication or aggregation of content some measure of standardization should also be considered. Having spent the past decade in the master data management world it seems that some of the core capabilities such as name and address standardization could be used to make tagging more effective. Name standardization for example uses a thesaurus's in order to improve duplicate suspect processing (so for example I can recognize Smith, Smythe, Smithe etc.).
Here's and example that I ran into recently with twitter lists. I was looking for web site designers lists. What I found was that many of the same names showed up in design, webdesign, webdevelopment and similar concepts. Providing a thesaurus capability that would identify and group similar concepts is essential for success in tagging. Search analytic capabilities -- clustering and categorization could play an important role in this process.
Making profile data more granular by allowing individuals to create personas would greatly improve its usefulness in expertise location. The basic capabilities are already available. JanRain's RPX solution (based on OpenID/Oauth) supports aggregation or extended profile data from multiple sites to build a rich profile. Extending the RPX data model to support personas can be informed by how MDM solutions allow for similar capabilities. Personas can then manage how rich profile data is shared
What would really make this compelling and really useful in locating expertise though would be some additional level of automation in the credentialing process. For this why not look to the credit industry for help? FICO type scores that analyzed both rich profile data and social media and generated "credential scores". Scores could be generated and stored for each persona and added to profile data to support and simplify the process of locating expertise.
Expanding the use and capabilities of both rich profile data and tagging to will aid in both the discovery and disclosure -- whether in the support of credentialing and expertise location or community building.