I've found that JR2 data can be quite a good indicator of subscription demand, but as others have mentioned, there are nuances.  For example, I find that the more general search traffic the site receives, the better the JR2 data is, because users are searching directly on the site.  If they are looking for articles cited in databases, our link resolver will redirect them to ILL before they hit the site and generate a turnaway.  For that reason, I've found that ScienceDirect has excellent JR2 data *for our campus* because we license a huge package, while Taylor & Francis is iffy because we only subscribe to a few of their journals, but this will absolutely vary by institution.  And of course, as others have also said, it is essential to look at multiple years of data.

I've also found that ILL data can be a good indicator of subscription demand when the numbers are quite high, but they are often so low that it's impossible to draw any solid conclusions.

Interestingly, I was meeting with one of our faculty members yesterday and he mentioned wanting to be able to follow citation trails.  He wanted to be able to view articles in the moment, and repeatedly going to a separate form to submit requests for articles that wouldn't arrive until at least the next day, and often turn out not be relevant, wasn't terribly friendly to his workflow.  He still submits some requests, but far fewer than the turnaways he's generating on journal sites.

I have a personal (unsubstantiated by any organized evidence, so take with a grain of salt) suspicion that ILL numbers are inevitably LESS than subscribed use due to the barriers of time and extra effort, while subscribed use is likely to be MORE than the actual utilization of the content due to factors like the citation-following described above or the phenomenon of revisiting/redownloading the same article from multiple workstations (work vs home, e.g.).  In other words, the real utility of the journal is somewhere in between ILL/turnaways and subscribed COUNTER reports. This would also help to explain why cancelling a title rarely leads to a sudden flood of ILL requests.

I did a project some years ago where we bought ScienceDirect backfiles on the basis of turnaways, with modest success.  I have been slammed with other duties and haven't updated my findings, but I presented at NASIG at the time.

Nikki DeMoville
Coordinator - Electronic Resources, Acquisitions, and Resource Sharing
Robert E. Kennedy Library
California Polytechnic State University
San Luis Obispo, California

Direct 805-756-5780
Fax 805-756-7711

On 10/31/2018 9:31 AM, Bluhm-Stieber, Hella wrote:

I have been using the number of turn-aways to add new journals subscriptions, but I noticed that I need to look for several years of data, because sometimes a certain journal spikes in one year, but is not used much in other years. I also look at ILL requests and our overall journal coverage in the subject area.



Hella Bluhm-Stieber, MLIS, AHIP

Medical Librarian

M.J. Chatton Medical Library

Santa Clara Valley Health & Hospital System

751 S. Bascom Ave.

San Jose, CA 95128

(408) 885-5654

Fax (408) 885-5655



Library opening hours: Monday, Wednesday-Friday, 8 am – 1 pm, Tuesday 8 am – 5 pm, closed on weekends and county holidays




From: Serials in Libraries Discussion Forum [mailto:SERIALST@LISTSERV.NASIG.ORG] On Behalf Of Herraghty, Maureen
Sent: Wednesday, October 31, 2018 9:08 AM
Subject: Re: [SERIALST] ILL data as basis for new subs (or not)


The scientists I deal with (all post-graduate level) tend to assume that ILL will take too long, so if we don’t have electronic full-text access, their first-choice alternative is usually to ask a colleague to retrieve it for them from another institution, or request a copy directly from the corresponding author. Only if those options fail will they request it as an ILL.


I only take time to fully analyze those situations where they seek to reach full-text. My method for cross-checking turnaways for true demand is to do citation analysis on papers recently published at my institution.  If they are citing it, they are reading it, and if they are not getting it from the Library’s collection, then they must be getting it somewhere else.   I’ve been able to pretty accurately identify potential demand with that comparison.





Maureen E. Herraghty

Director of Library Services

Elizabeth Gray Danforth Library

Donald Danforth Plant Science Center

975 North Warson Road

Saint Louis, MO 63132


t  314.587.1081

f  314.587.1181



From: Serials in Libraries Discussion Forum <SERIALST@LISTSERV.NASIG.ORG> On Behalf Of Melissa Belvadi
Sent: Wednesday, October 31, 2018 4:13 AM
Subject: [SERIALST] ILL data as basis for new subs (or not)


Hi, all.


I occasionally see an article in our professional literature about how well ILL article request data does (not) reflect likely demand, and that definitely seems to reflect our own experience from my somewhat informal analysis of our own data.

But I seem to have trouble persuading my colleagues here about this.


For example, I'll present JR2 turnaway data and "abstracts viewed" in EBSCO and Proquest as evidence for adding a subscription, but they'll respond that :"if they didn't bother to ILL it, then they don't really need it".


Has anyone done, either for publication or for internal use that you can share with me, some kind of "systematic review" on this issue?


Or even if you have a clear and concise explanation of why that "they didn't bother" reasoning is not an appropriate conclusion to draw, I would appreciate that too. I haven't found the right way to articulate why I think that's wrong.

Or if you agree with my colleagues, tell me that too!



Melissa Belvadi

Collections Librarian

University of Prince Edward Island

mbelvadi@upei.ca 902-566-0581

Make an appointment via YouCanBookMe




To unsubscribe from the SERIALST list, click the following link:


To unsubscribe from the SERIALST list, click the following link:

NOTICE: This email message and/or its attachments may contain information that is confidential or restricted. It is intended only for the individuals named as recipients in the message. If you are NOT an authorized recipient, you are prohibited from using, delivering, distributing, printing, copying, or disclosing the message or content to others and must delete the message from your computer. If you have received this message in error, please notify the sender by return email.

To unsubscribe from the SERIALST list, click the following link:

To unsubscribe from the SERIALST list, click the following link: