[Web4lib] recommendations for web and catalog "visits" statistics solutions

Thu Apr 3 11:33:33 EDT 2008

Sarah:

        I have been using two statistical programs to create statistical
reports about our Library website usage.  The easiest is Google Analytics.
Once it is set up it provides many reports of interest with minimal effort
on your part.  It defaults to reports for the most recent 30 days, but you
can change that easily and combine months, or zero in a a specific day.  If
you want to look at activity in a specific sub-directory, there is a search
function that will limit the lines displayed to the characters you type in.
It creates lovely charts and graphs, and has multiple export options; pdf,
csv, sml, tsv.  It will tell you the country and city from which specific
numbers of  request originate and provide you will accompanying lovely
maps.  And the price is right! (free).  Hopefully Google is as benevolent
as it appears, but I figure they aren't learning anything new from this.
They already know everything!

        We have been focused on gaining an understanding of how our patron
base has moved out of the building.  It is important to understand just how
few of our current library users actually come to the building, or phone or
email us.  I use a program called Alterwind Log Analyzer to analyze that.
On our campus, the ip ranges are neatly assigned.  I can tell, by range, if
a request originates in our building, or from our campus.  Alterwind has
wonderful filtering features, so you can take large logs and "filter in"
only the relevant html pages, and ignore all the little gif files in your
analysis.  I did this for a full semester 2 years ago, and I will probably
do it again at the end of this semester.  It is also possible to pick a few
days -- some during the middle of the semester, some during breaks, and see
how usage changes.  It was quite eye opening to see such usage.  We were
able to use the results to get more computers put in a space used by
commuter students.  AlterWind cost something under $100.00, and is well
worth it.  Running a semester analysis took A LOT of time.  We think it is
worth doing every couple of years because the changing demographics of our
users is otherwise almost invisible.

         I hope this helps.

Katy

Kathryn K. Silberger
Automation Resources Librarian
James A. Cannavino Library
Marist College
3399 North Road
Poughkeepsie, NY  12601
Kathryn.Silberger at marist.edu
(845) 575-3000 x.2419

             "Houghton-Jan,                                                
             Sarah"                                                        
             <sarah.houghton-j                                          To 
             an at sjlibrary.org>         "'web4lib at webjunction.org'"         
             Sent by:                  <web4lib at webjunction.org>           
             web4lib-bounces at w                                          cc 
             ebjunction.org                                                
                                                                   Subject 
                                       [Web4lib] recommendations for web   
             04/02/2008 08:12          and catalog "visits" statistics     
             PM                        solutions                           

Hi all,

I am asking for recommendations on two things:
1) software you are using or would recommend to capture accurate user visit
numbers for your library's website
2) any method or vendor-provided built-in tool that you are using to
capture
accurate user visit numbers for your library's catalog.

I am doing a webcast for Infopeople, a California project for training
libraries, on April 24th on recommended tools and best practices for
capturing accurate numbers for the two categories above specifically for
the
purpose of having accurate numbers to provide for the Annual Public Library
Survey.  They are adding a new category for statistics this year that
libraries will be asked to submit: virtual visits.  The definition provided
in the survey is as follows:

Virtual visits include a user's request of the library web site or catalog
from outside the library building regardless of the number of pages or
elements viewed.  This statistic is the equivalent of a session for a
library's website. Exclude virtual visits from within the library, from
robot or spider crawls and from page reloads.

So, near as I can figure it, the tricky parts to the above definition are
that they want all those exclusions (the "within the library" is
particularly difficult for all but the most robust stats systems) and that
they're looking for catalog "visits" (not # of searches, # of holds, etc.
which is what the vendor provides us with).  There are a number of kludgey
ways that one can try to guess the number of visits if the vendor doesn't
provide that functionality.  Tell me what you've kludged!

But the biggest problem, and the wording of the definition here is key, is
that they want "a user's request of the library web site or
catalog...regardless of the number of pages or elements viewed."  Does this
mean that if a user starts in the website, moves into the catalog, then
leaves, that it should count as one visit?  I don't think there's a way to
do that since they are on two different servers, two different systems, two
different databases (correct me if I'm wrong).

Any insight or recommendations you can offer are much appreciated.

Thanks!
Sarah Houghton-Jan
Digital Futures Manager, San Jose Public Library
Author of LibrarianInBlack.net

_______________________________________________
Web4lib mailing list
Web4lib at webjunction.org
http://lists.webjunction.org/web4lib/