[Sigia-l] Emetrics Summit 2005?

[Casey Jones]

I work in a digital library, and we're also facing problems in
determining website success. Web analytics software is designed for
commercial enterprise. Even the pricing model - the more visits you
get, the more you pay for the software - is designed specifically with
e-commerce in mind. As an educational non-profit, we do not get more
"revenue" (funding) just because more people click through our site. I
have had little luck discovering information on non-commerce
statistics.

Indra's list is a great start, and I hope that more people will
comprehensively study non-commerce implications of web stats. While I
would hardly term this an original approach as Antonella requested,
these are some metrics my group considers...

* Successful registration
* Return visits (some sites are not allowed to set persistent cookies,
so this is impossible to determine) of registered and non-registered
users
* Multiple simultaneous visits for same login ID (for sites that
require login to access materials, this demonstrates multiple
concurrent use in a classroom, which can be transformed into a success
stat for desire of materials - to some degree)
* Download of objects
* View of actual resources vs metadata
* Usage by geographical area (important for funding)
* Paths to object (specifically, direct links from outside sites)


Other issues we're considering...
* Develop models of successful visits (or failed visits) and determine
number & percentage that conform to models
* Site classification (level of importance of page - eg, nav vs
content pages) and determine number and percentage of visits to each
* Successful/ failed searches
* Trending site link ranks in popular search engines (are our ranks
improving over time)


A note on ARL - ARL's main focus has been on using surveys to analyze
the "success" of libraries. While this is a good effort, it falls
short in defining successful metrics for web-based interactions,
especially for sites that aren't libraries (and a digital library is
not a library).

  ..casey jones

On 5/12/05, Indra Chandon <indra at semantia.com.au> wrote:
> Pages bookmarked - which page(s) and how many people have bookmarked the
> page
> Pages as Home Page - which page(s) have been set as the default Home
> Page when opening a new Web browser window
> Download speed - how long (time) it takes to download a page
> Number of return visitors - how frequently they return and style of
> interaction (path of page views)