Will Search Engines Learn from Our Fumbling?

      No Comments on Will Search Engines Learn from Our Fumbling?

After quickly reading through Search Beyond Google, I wonder whether any search engine looks at subsequent queries within the same web browsing session to see whether you’re trying to refine the same core search because your first attempts failed. This could be done for instance to give more weight to new keywords added, or or combinations of words turned into phrases.
Let’s say my first query is trial by error query and my second query is “trial by error” query. The search engine would observe my behavior and order results differently than if I had started directly with query number two. (I realize that those two queries are not a good example because of the low number of results, but I hope my idea is clear.)


I find myself rephrasing and refining queries sometimes three or four times in a row to get the information I’m looking for, and I’m sure the sum of those variations is valuable by itself, to triangulate what I’m really looking for. When we explain something to another human being, the sequence of arguments keeps building context until hopefully there’s a breakthrough in understanding. It helps to know what something is not to eventually understand what it is.
I am not talking about long-term personalization here (which I doubt can be very useful to search, but I could be wrong). Neither am I terribly excited by Mooter’s clustering (explained in the article previously mentioned), which asks users to think as a taxonomist (which is OK for me, but foreign if not repulsive to most people). Rather, I’d like search engines to have a kind of conversational understanding within the string of queries you’re likely to go through until you give up or succeed. I find it very crude that every new Google query starts from scratch.
03/01/04 update: Mercury News: New Web tools aim to customize searches.

Leave a Reply

Your email address will not be published. Required fields are marked *