2. Contributors:
• Jenny Walker, independent consultant
• Teddy Hein, analytics coordinator, JSTOR
• Ross Houseright, senior data analyst, JSTOR
3. Origin of Content Accesses in JSTOR
0%
Serials
Solutions
5%
Self Referrer
15%
Direct to JSTOR
33%
Linking Partners, etc.
13%
Institution
11%
Google
13%
Google Scholar
11%
4. Origin of Content Accesses in JSTOR
0%
Serials
Solutions
5%
Self Referrer
15%
Direct to JSTOR
33%
Linking Partners, etc.
13%
Institution
11%
Google
13%
Google Scholar
11%
5. Origin of Content Accesses in JSTOR
0%
Serials
Solutions
5%
Self Referrer
15%
Direct to JSTOR
33%
Linking Partners, etc.
13%
Institution
11%
Google
13%
Google Scholar
11%
6. Origin of Content Accesses in JSTOR
0%
Serials
Solutions
5%
Self Referrer
15%
Direct to JSTOR
33%
Linking Partners, etc.
13%
Institution
11%
Google
13%
Google Scholar
11%
7. 2011 Usage – JSTOR Small
30,000
25,000
20,000
15,000
10,000
2011 JR1a
5,000
0
COUNTER Journal Report 1a (JR1a) counts the number of successful full-text article requests by month and
journal from an archive. The metrics that drive into this are Article Views and PDF downloads, excluding
Article Views and PDF Downloads of the same item in the same session if occurring within 30 minutes of a
previous View of the same item, or 30 seconds of a previous Download of the same item.
8. 2012 Usage – JSTOR Small
30,000
% Change from
2011 to 2012
25,000
-24.72%
20,000
15,000
2011 JR1a
10,000
2012 JR1a
5,000
0
COUNTER Journal Report 1a (JR1a) counts the number of successful full-text article requests by month and
journal from an archive. The metrics that drive into this are Article Views and PDF downloads, excluding
Article Views and PDF Downloads of the same item in the same session if occurring within 30 minutes of a
previous View of the same item, or 30 seconds of a previous Download of the same item.
9. 2013 Usage (YTD) – JSTOR Small
30,000
25,000
20,000
15,000
2011 JR1a
10,000
2012 JR1a
2013 JR1a
5,000
0
COUNTER Journal Report 1a (JR1a) counts the number of successful full-text article requests by month and
journal from an archive. The metrics that drive into this are Article Views and PDF downloads, excluding
Article Views and PDF Downloads of the same item in the same session if occurring within 30 minutes of a
previous View of the same item, or 30 seconds of a previous Download of the same item.
10. Getting Good Data … Is Hard
• Survey of JSTOR participating institutions (May 2013)
o 422 responses
o No consistent implementation dates (< 100)
o Too few responses across institutional archetypes to be
statistically relevant
Discovery Service
# responses
AquaBrowser (Serials Solutions)
1
EDS (EBSCO)
154
Encore (Innovative Interfaces)
10
Primo (Ex Libris)
69
WorldCat Local (OCLC)
36
Other¹
48
Summon (Serials Solutions)
102
Vufind (Villanova University)
2
11. Getting Good Data … Is Expensive
• Supplemented with data from lib-web-cats database
(Marshall Breeding)
o Increased # of institutions with data to 1,480
o Again, no consistent implementation dates
Discovery Service
Number
AquaBrowser (Serials Solutions)
45
EDS (EBSCO)
379
Encore (Innovative Interfaces)
121
Primo (Ex Libris)
410
WorldCat Local (OCLC)
117
Other¹
73
Summon (Serials Solutions)
259
Vufind (Villanova University)
41
Backlight (University of Virginia)
5
Enterprise (SirsiDynix)
20
Locally-developed
10
12. Getting Good Data … Requires Negotiation
• Worked directly with EBSCO, Ex Libris, OCLC, and ProQuest to
get customer list and implementation dates (July - Sept. 2013)
• Supportive of effort
• Confidentiality required
Discovery Service
Institutions
provided
Matched
in CRM
JSTOR
participants
Higher
Ed
% JSTOR
participants
% higher
education
A
4,992
3,149
1,781
925
36%
19%
B
760
645
576
417
76%
55%
C
63
57
53
48
84%
76%
D
623
540
397
308
64%
49%
13. Caveat Emptor
Paraphrasing the late Prof. Aaron Levenstein (Baruch)
Statistics are like bathing suits …
What they reveal is interesting ….
But what they conceal is essential.
14. Initial Usage Results: U.S. higher education
• Culled customer lists to JSTOR participants in U.S. higher
education for EDS (EBSCO), Primo (Ex Libris), WorldCat
Local (OCLC), and Summon (Serials Solutions/ProQuest)
o Looked at average content access per month for each JSTOR
Class for 12 months prior/post implementation date
o JSTOR average usage change for all U.S. higher education
(August 2009 – September 2012): -3.2%
Discovery Service
Usage Change Post-Implementation
A (218)
-8.7%
B (100)
-0.4%
C (13)
-13.3%
D (117)
-4.4%
15. Initial Usage Results: Worldwide higher education
• Culled customer lists to JSTOR participants worldwide for
EDS (EBSCO), Primo (Ex Libris), WorldCat Local (OCLC),
and Summon (Serials Solutions/ProQuest)
o Looked at average content access per month for each JSTOR
Class for 12 months prior/post implementation date
o JSTOR average usage change for all higher education
(August 2009 – September 2012): -0.7%
Discovery Service
Usage Change Post-Implementation
A (541)
-4.6%
B (340)
-1.3%
C (18)
7.1%
D (238)
-1.3%
16. Initial Usage Results: U.S. higher education
• There are significant differences by JSTOR Class within
each discovery service. Why?
Discovery
Service
Very
Large
Large
Medium
Small
Very
Small
Any
A (218)
-3.1%
-11.1%
-9.9%
-11.1%
-12.1%
-8.7%
(7)
(16)
(70)
(54)
(65)
1.1%
7.1%
-11.6%
-7.4%
-3.9%
(21)
(15)
(32)
(11)
(20)
NA
NA
-8.6%
-19.2%
5.7%
(0)
(0)
(3)
(5)
(5)
-10.0%
2.1%
5.6%
-5.7%
9.3%
(21)
(13)
(46)
(20)
(17)
-2.9%
-3.2%
-2.7%
-3.2%
-5.9%
(86)
(87)
(390)
(363)
(638)
B (100)
C (13)
D (117)
JSTOR
(1,599)
-0.4%
-13.3%
-4.4%
-3.2%
17. Initial Usage Results: Worldwide higher education
• There are significant differences by JSTOR Class within
each discovery service. Why?
Discovery
Service
Very
Large
Large
Medium
Small
Very
Small
Any
A (541)
2.1%
-4.9%
-4.5%
-9.7%
-4.9%
-4.6%
(11)
(51)
(134)
(109)
(220)
-1.2%
-0.3%
-2.8%
-4.0%
5.2%
(26)
(80)
(114)
(53)
(62)
NA
15.3%
-10.8%
-19.2%
30.6%
(0)
(1)
(4)
(5)
(8)
-7.3%
2.5%
4.6%
-3.4%
-2.6%
(24)
(30)
(89)
(36)
(58)
-2.5%
0.2%
-0.9%
-2.0%
0.9%
(111)
(359)
(917)
(864)
(2,637)
B (340)
C (18)
D (238)
JSTOR
(5,258)
-1.3%
7.1%
-1.3%
-0.7%
18. Deep-dive into discovery partners
• Subject metadata matters … a lot
o Relevancy ranking is driven by subject metadata
o Subject metadata is a higher priority for JSTOR than full-text
• Libraries don’t spend enough time configuring their system
for implementation
o A vast majority of systems are left at “default” state
o Link resolver configuration is critical
o Libraries need guidance
• Publishers/content providers don’t spend enough time on
their data syndication, including how that data is received
and used
o Archive collections vs. CSP vs. Books
o Importance of good KBART files for link resolvers
Hinweis der Redaktion
is an allegory used by the GreekphilosopherPlato in his work The Republic to illustrate "our nature in its education and want of education“. It is written as a fictional dialogue between Plato's teacher Socrates and Plato's brother Glaucon.Plato lets Socrates describe a group of people who have lived chained to the wall of a cave all of their lives, facing a blank wall. The people watch shadows projected on the wall by things passing in front of a fire behind them, and begin to ascribe forms to these shadows. According to Plato's Socrates, the shadows are as close as the prisoners get to viewing reality. He then explains how the philosopher is like a prisoner who is freed from the cave and comes to understand that the shadows on the wall do not make up reality at all, as he can perceive the true form of reality rather than the mere shadows seen by the prisoners.Applying this analogy to the discovery space, I think many resource providers – and probably more than a few librarians – feel like those prisoners chained to the cave looking at the shadows on the wall (the claims of the wonderfulness of web-scale discovery) and viewing that as reality. Is it reality? I’m not entirely sure … so, we need to begin shedding some light on what is ACTUALLY happening (from a usage perspective) so that we can begin to make better determinations about what is real and what might be imagined.