Saturday, October 13, 2007

Search

There is a lot of information out there. There is also a lot of crap. Google has done wonderfully well in providing an easy way to generate an ordered list of crap based upon simple queries. This is not a failing of Google, it's more a statement on the web, I guess. Now, let me paint you a simple scenario, where the problem the user is trying to solve is "Why does my Windows machine crash when I try and watch a DVD":

  • User fires up Google and types in a search term; "windows crash" + DVD.
  • User gets a massive list (30k results at the time of writing)
  • User filters through the results, occasionally refining the search and either gets lucky or frustrated
  • Chances are, someone, somewhere has had the same problem and has fixed it
This is a pretty common scenario, where a user is trying to solve a problem leveraging the knowledge scattered throughout the web. Some other common scenarios include:
  • Searching for information on a person (esp. a famous one)
  • Searching for product information and/or reviews
  • Searching for cheap prices on a desired item
So, the challenges here are firstly to better understand the goal of the user, and secondly, to better understand the content of the indexed information.

TBC.

1 comment:

Mateusz Mrozewski said...

My friend use to say "if I can't find something on google it means my query is wrong".

If think that the key to success is the word "semantic". There is a lot of results for "windows crash", and a lot of for "DVD", but for "windows crash {in context of} DVD" it might be better.