Funny you should ask!  I'm in the midst of our second HUGE cancellation
process in less than a year.  I'm finding that publisher-supplied
statistics don't match the EMS, either.  However, for one of the
publishers we believe the discrepancy, (EMS is hugely greater than pub
stats) may be due to the pub requiring login instead of using IP
authentication.  We believe that users may be turned away from using the
publication, even though we have a page that supplies login information.
They've registered as a hit in the EMS because they clicked the local
link.  So for the EMS we're counting raw hits and not downloads.  We
expect that it will be somewhat larger.  That's as close as we can get
with our EMS, though.

We have a horrible time collecting statistics from publishers.  They
give us our downloads for all their titles, not just those to which we
subscribe.  It's not fun editing them.

We don't use Shibboleth.

Have a good weekend!!

Mary

-----Original Message-----
From: SERIALST: Serials in Libraries Discussion Forum
[mailto:SERIALST@list.uvm.edu] On Behalf Of Shipley, Tara N.
Sent: Friday, May 01, 2009 3:37 PM
To: SERIALST@LIST.UVM.EDU
Subject: [SERIALST] Electronic Usage Statistics

Hello All,

Does anyone else experience problems collecting accurate usage
statistics for their electronic journals? We find that our statistics
from the publisher's website don't always match those of our e-resource
management system. What are your library's procedures for collecting
statistics (i.e., where do you get your statistics)? Does anyone use
Shibboleth?

Thanks in advance,

Tara Shipley
Manager, Library Services
University of Texas Medical Branch
Moody Medical Library
Serials Department
301 University Blvd.
Galveston, TX 77555-1035
phone: (409)772-2394 ; fax: (409)762-9782
email: tnshiple@utmb.edu<mailto:tnshiple@utmb.edu>