IL2006: What is the big deal about social search?

Search Engine Report by Chris Sherman

I am only putting excperts into my blog posts as many other bloggers were at the conference and have information up as well. Check out the IL Wiki to see other bloggers and their comments.

What is the big deal about social search?

-wayfinding tools informed by human judgment
-Yahoo is biggest proponet
-all guides to the web
-search engines reflect human bias, because programmers have to make choices. They also watch us while we search to see how good they are.
-personalization to refine search for everyone.
-algorithmic search has plateaued. Google was last big breakthrough.
-social search companies are leveraging the work of volunteers

Types:
*delicious, Furl, Diigo, MyWeb – shared bookmarkes. MyWeb is good for group projects
*tag engines – Technorati, bloglines
*collaborative directories – Yahoo directory, ODP, Prefound, Zimbo, Wikipedia (adding lots of links to articles with the info they already have)

*Personalized verticals: Kosmix, Eurekster, Rollyo. Latter two create a list of sites you find interesting with community based aggregation system.
*Collaborative harvesters: submit articles to a community where the articles get voted on- popurls.com, Digg, Netscape, Reddit
*Social Q&A site: Google Answers, Yahoo Answers, Answerbag – Yahoo has a “stupid filter” so good info rises to the surface.

Problems:
*scale & scope – many non-English sites
*Tagging: language is inherently ambiguous (orange the fruit or orange the color?), lack of controlled vocabulary + chaos, human laziness (nobody adds metadata to word docs, etc), idiots= people doing something to get attention and not to help people, spammers

What will work:
*people efforts are good – need blend of people/algorithmic searching
*trust networks (rating)
*increased personalization & user control over result filtering
*social search will work best with non-text content (photo, music, video) because regular text was missing and other textual aren’t generally associated with the file.

Future trends

-blurring of content & process (blog commenting)
-more personalization
-more dynamic interaction with both search & web
-more vertical/mashup specialization