Wednesday, January 13, 2010

Which pieces loosely formed make a difference

My favourite definition of Information is "a difference that makes a difference."

I'm trying to make sense of the use of our library and of our online resources and there's so much information that is not just context-dependent but multi-context dependent.

Let me try to explain what I mean.

Our library's SFX resolver allows can generate a list of our most "popular" serial titles and according to it, the most popular title on campus is "Dissertations Abstracts International". The second most popular title is "Dissertations Abstracts B". Then comes, "Science" and then, "Nature" and the other usual and unusual suspects.

And I *think* I know why.

At MPOW, most of our indexes are hosted by Scholars Portal Search (using CSA's Illumina interface) and the default that we have set for these indexes is that all linking to full-text articles happens dynamically through our SFX "Get It!" buttons.

But there are a small handful of exceptions to this policy because the links occur automatically from Illumina including...

digdissert

But is that a difference that makes the difference?

MPOW's Web Stats Take Two

I made a second presentation of some statistics surrounding my library's web site and online resources and it was a tougher crowd this time around as they asked lots of good questions about the points I was making.

After I was done with my talk, I returned the data behind one of my key slides to reassure myself that it was sound and realized that it wasn't.

I noticed from our Google Analytics dashboard that our traffic to the library's front page appeared to slowly ebbing downward.


To illustrate this drop, I chose a random day in October 2007 and a random day in October 2009 and there was about a 30% difference between the two. But, on closer inspection, I realized that there was a drop in traffic in October 2009 during the University of Windsor's first Fall Break (called UWin week) which probably exaggerated the results.

So, in these updated slides, I compared the traffic of the entire site (that Google Analytics can understand) for the entire month of November. And I found out that November 2009 had 88% of the traffic of November 2007.