We have a III Millennium system.  We manufactured looseleaf volumes filled with pages barcodes (one per bibliographic title).  We had three volumes for our open stacks, one for our microfilm collection, and one for our current periodicals display. The actual data entry did not take that long.  Four people worked on it fitfully over the course of one summer. Our open stacks and microfilm collection have been placed in an automated warehouse on site (and the record of calls therefrom now provide more accurate statistics), so four of these volumes have since been retired, but the fifth one covering our current periodicals display is still in use.

The presupposition of this sort of data collection is that the propensity of users to place something back on the shelf does not vary by title.  I would wager it does vary, but that the variation is due to the physical properties of the item in question, not to the editorial content.  A title with a magazine format (e.g. "Physics Today") is likely put back less often than one published in hardcover (e.g. "Transactions of the Royal Historical Society"), simply because of the bother of finding the right place and jamming it in.  The degree to which the distribution of observed use reflects the distribution of actual use is likely higher with regard to a pool of high-use titles (with regard to which there are many opportunites for observation) than would be the case for low-use titles (with regard to which there are few opportunities).  For all that, the use statistics are your single best guess of the ratio of use between the publications in your collection.

If  the purchase of  electronic editions of the titles in question (through some package deal) is off the table, you could rank-order your titles according to cost-per-observed use and cut from the bottom up.  For purposes of optimizing the use of your budget, the statistics and the prices pretty much exhaust what you need.  If you are of the view that core lists are valuble or that you have to make nice with professor so-and-so, you have a different problem.

IW

I.  Woodward
Serials Office
Colgate University Libraries
Case Library and Geyer Center for Information Technology
13 Oak Drive
Hamilton, N.Y. 13346
Ph.:   315-228-7306
Fax:   315-228-7029

-----Original Message-----
From: SERIALST: Serials in Libraries Discussion Forum [mailto:SERIALST@list.uvm.edu] On Behalf Of Martha Coleman
Sent: Monday, June 23, 2008 3:14 PM
To: SERIALST@LIST.UVM.EDU
Subject: Re: [SERIALST] Print Serials in-house statistics

We also use Innovative Interfaces to gather inhouse use statistics.
Rather than a barcode on the shelf, we created a rolodex file of barcodes.  We don't know which issue was used but we at least know the title is getting some use.  The stats have been very helpful in decisions to cancel.  I know the count may not be perfect but it helps to have a rationale for cancelling a title.

Anil Deshpande wrote:

>We are considering gathering usage statistics for print serials by
using a bar
>code Scanner and the in-house circulation module of our software. Has
any
>one had experience with this? Did the data you gathered help you make
>decision about purchases? What benefits did you realize from a project
like
>this? Any suggestions are welcomed.
>

---------------------------------
Martha Coleman
Reference/Serials Librarian
U A Fort Smith - Boreham Library
5210 Grand Avenue   PO Box 3649
Fort Smith,  AR  72913
479-788-7208     FAX 479-788-7209
mcoleman@uafortsmith.edu