Search engine logs

From InfoCamp

Jump to: navigation, search

Contents

mining search engine logs

what to put in logs

  • IP
  • Timestamp
  • Cookies
  • Query
  • Query post-transform
  • Referer
  • User Agent
  • number of matches
  • Iist of IDs of items displayed
  • click type
  • if result item, doc ID of item

Type of click

  • result item link
  • search navigation (page next)
  • search help
  • advanced search
  • page navigation
  • Back button

what is search log noise?

  • empty search
  • search field spam (urls)
  • web navigation mistakes (domain names)
  • random characters
  • more than 10 repeats of one query in less than one minute...
  • web crawlers (googlebot, etc)

(note: site navigation searches are not so bad. Sometimes you need to fix the UI, sometimes it's just how users work)

Personal tools