Saturday, 21 May 2011

How search engines prioritise results

I've been doing something thinking, and some research, on how search engines prioritise results. Some known data sources:

  • PageRank - the number of links to a particular page (and the authority of each of those links)
  • The searcher's previous search history
  • What previous searchers have clicked on in response to the same or similar search
  • Trend info (perhaps from Twitter)
  • The searcher's location (as indicated by the IP address)
  • Other data on the searcher
A couple of areas that I don't know if search engines use (but would be interested to find out) are:

  • time of year (for example, when searching for "tree" around Christmas, are you more likely to be shown results for Christmas trees, and if so, is it the time of year that is taken into account or just the trending information mentioned above)
  • time of day
  • what previous searchers of the same location have clicked on in response to the same or similar search

No comments: