By Aaron Mann

Social search engine Aardvark just published a very interesting white paper: "Then Anatomy of a Large-Scale Social Search Engine".  

I found this a fascinating read.  It expresses search as a social "village" model as opposed to the keyword driven "Library" model. We have all been trained to think in keywords for search, but there are many questions that just don't fit in a few keywords. The combination of natural language nuance plus contextually relevant people to answer your question is a powerful proposition.   

Of course this type of social search happens all the time.  We often join or create sub-networks with a particular interest area so that we can scope our natural language queries to the right audience.  If you are a parent in Berkeley you might join Berkeley Parents Network; if you are a cyclist, Road Bike Review; a foodie, Nourish Network and so on.  We then adjust our question to the relevant network, group, or sub-set of friends.  Or you just blast it out to all your followers,  friends and connections.  The promise of social search is to make this process much more efficient.  Does it work?  According to Aardvark:

• 87.7% of questions sent to Aardvark got answered (very high answer rate!)
• 75.0% of users who asked Aardvark a question also answered a question for someone else (very high participation rate!)
• 70.4% of answer feedback had a rating of ‘good’ as opposed to ‘ok’ or ‘bad’ (high quality!)

What kind of questions do people ask:

The concept of modeling users as content generators really struck a cord.  Socialarc uses a lot of database modeling and parsing techniques to accurately identify what targets are most likely to be receptive to an outreach message.  Making sure the "offer" is contextually relevant (and value-add) to the audience is a key to a successful social media campaign.  Aardvark, and other social search engines, take that much further - to the level of the individual question and user.

You can read a summary on their blog and download the paper white paper here.  And we added a poll question on the side - tell us how you would find the latest cool local restaurant!