I had a great breakfast today with Ben Foster, my old PM counterpart at Ebay.  He is now the director of product management at Adchemy, a leader in computational performance marketing.  We were talking about a concept I’ve been thinking about for a while – a Semantic Browser. 

I was inspired to try to solidify my thinking on what exactly a semantic browser could do and what problems it would solve.  Here’s my list so far:

1.  Know who I am wherever I am on the web.  It’s silly that I have profiles in different places all over the web.  Any advertiser should know my demographic info on any site I visit, regardless of whether that data resides on the site or not.

2. Know my social graph wherever I am on the web. Make my social network portable, and use that data to do a better job at providing me recommendations, content, etc. Bijan mentioned some applications of this in an earlier post

3. Know my click stream. Or, simply know my history.  Wouldn’t it be valuable to know how deep I went into specific sites to determine if a lead or a search was indeed a good one?  

4. Give me a “best of” button.  Analyze everything above, and give me the chance to say “show me the best sites, media, and products out there based on what you think I am interested in at this moment.” 

This is a huge technical undertaking, and it’s unclear exactly how the data should be processed.  My opinion is that it’s time to create a competing browser product that collects all this information and allows advertisers to access the data in a useful way.  Then, just get a small % of any ad revenue that results from the use of this information.  We are investors in a couple companies that are going after pieces of this opportunity (most notably, Me.dium).  But it’s a huge one, and there’s a lot that has not been done. 

Would love to hear people’s thoughts on this one.  I wonder if Google is already working on this (they should).