Web Statistics

Many publishers request statistics for their websites. Statistics generation is incredibly process intensive, and provide results of limited use. This page outlines the current provision and limitations of web statistics for publishers in the University.

What are Web Statistics?

Web statistics can take a number of forms - from a basic "hit counter", which simply records the number of times a page is loaded, right through to advanced statistical analysis of user locations, browser types, visit times and so on.

What Statistics Tell You (and, what they don't...)

Many publishers requests web statistics in order to measure the "performance" or "popularity" of their site. However, this is not as straight forward as it first appears...

Whilst statistics will give an indication of relative "popularity" - say, in identifying trends from one month to the next, they are unable to provide definitive information on a wide range of other factors:

  • How much of the traffic came from humans (rather than web robots - such as Google)
  • The precise geographical location of the user
  • Other information about the user: age, gender, education/employment status, primary language
  • How long a given page was viewed for
  • Whether the page was actually relevant to the visitor (perhaps they reached it by mistake, or maybe the content was not exactly what they were looking for)
  • How useful the content was to the user - whether they "liked it" or not
  • How much of the traffic was "development testing" by the site publisher

Current Provision

TBC