Jason Price is the Head of Collections & Acquisitions and John McDonald
is the Director of Bibliographic & Information Management and Faculty
Relations at the Claremont University Consortium, a non- profit service
company serving the five elite liberal arts colleges and two graduate
schools of the Claremont Colleges.
This project was a follow up to this request…
Overall, our strategy was to set aside usability and discipline- related ebook issues,
which have been covered in other presentations over the years, and instead to
analyze the collections available to assess how they matched our own collecting
patterns in print.
But we wouldn’t feel right about focusing on availability without at least
mentioning the bigger picture.
• With ebooks, Digital Rights/ Usability issues are at least as important as collection
breadth and depth
• The only aggregator our library currently uses is NetLibrary, which has made
ebooks unpopular with a number of our students because of simultaneous use and
page- at- a- time only issues
• The massive continuum of use rights represented here should be a MAJOR factor
in each libraries ebook aggregator decisions
• Jason will present data addressing collection profiles of each of the 4 major
• John will address print book purchasing profiles from 5 libraries and availability of
their print purchases as electronic books
• Aggregator Collection data was updated Oct. 2008
•‘ Unique identifiers’ for books more numerous than for journals, isbn’s much harder
to work with
• Deletion justified because focused on titles that that would match print collections
• Requested data from all 4 major aggregators
• Included Ebrary Subscribable collection ( i. e. Academic complete) as a separate
collection, but NB: the Ebrary collection includes all books available for purchase
from Ebrary ( including some that are included in academic complete)
• Data received from NetLibrary didn’t include call number
• Ebrary has a deeper collection in terms of age with four times as many 80’ s- 90’ s
books as any other aggregator
• Very few from 1970s or before
• MyIlibrary has twice as many books as any other aggregator with ‘ pub year’ 2006
( turns out 23,615 were from ICON Group and were republished books or economic
reports) – These books were removed from the rest of the analysis since their
dates were not available and likely put them out of the range of current collection
•( Click) Next slide focuses on the ’ 00’ s’
•( Blue) Net Library consistently has more books per year, even in the most recent
•( Orange) Ebrary Subscribed collection lacks books from 2007 & 2008
•( overall) Pattern of number of books available per year extremely consistent across
•( overall) Are there signs that the market place hit a growth peak in 2005?
• There are about 225,000 unique ebooks currently available from this set of
• This data can be understood from the book’s perspective: if I’m a particular
ebook, what are my chances of being available from one or more aggregators?
• In the overall marketplace, more than half of ebooks are unique to a single
aggregator ( 56%), and only 1 in 25 is available from all 4
• 40% of the books are unique to NetLibrary, 7- 10x as many as any other aggregator
• Now we can look at availability from the libraries’ perspective [ Proportion of
books available by aggregator ]
* note: pie chart excludes Ebrary Subscription ( 3958 unique titles) and MyILibrary
ICON published titles, q. v.)
Ebook aggregator Marketplace = Available from at least one of the big four
Green pies show about 1/ 5 of marketplace available from MyI, 1/ 3 from Ebrary &
EBL, and ¾ from NetL
Looking just at books published 2005 to 2007, EBL and MyI have about 3 in 10,
Ebrary 4 in 10, and NetL 8 out of 10.
We asked four other libraries to send us data extracted from the ILS for all
purchases from 2006 to 2007.
We cleaned the data and standardized it to the greatest extent possible.
• ( Pale Blue) St Mary’s College bought 6,091 books
• ( Blue) Claremont Colleges bought 17,218 books
• ( Red) Azusa Pacific bought 9,614 books
• ( Green) Denver University bought 42,423 books
• ( Purple) Loyola Marymount bought 32,968 books
• This chart shows the percentage of books purchased by publication year. Using this ratio
allows us to normalize across the range of large & small schools in this dataset.
• 75 to 95% of purchases were books published 2004- 2007
• Libraries A & S purchased a much greater proportion of older books. And while this isn’t a
bad thing, it could impact how ebooks will relate to print collection patterns
• Using Call numbers to define LC classes, the greatest ratio of print books purchased were
in the Humanities and Social Sciences, which is as expected from this group of liberal arts
• The relative distribution of purchases across disciplines was very similar for all 5 libraries,
with only Claremont’s Arts collecting being above normal for the set of schools.
•< 5% of books purchased by any given library were purchased by all five libraries
• 40 to 60% of books bought by any given library were not bought by any of the other four
• This indicates, that although the schools are similar in size and demographic and academic
mission, they still collect print books in relatively unique patterns. Obviously with a larger
set of schools included, the uniqueness factor would diminish by some amount, but it is
clear that schools are relatively individualistic in their collection development.
The most important part of the analysis is to compare the library print purchase
patterns to the ebook vendor supplied lists.
* 7 of 10 print books these libraries bought are not available from any ebook aggregator
* These numbers have been confirmed since the presentation using OCLCs xISBN workID
ensure that isbn differences in purchased book format ( e. g. hardback vs paperback) did not
affect ebook availability values
* Numbers presented during the conference were 3- 5% higher, so the effect of format was
Despite significant variation in the sets of print books libraries bought, there is little
variation across libraries in the percentage available as ebooks from any one vendor
In every case the minimum to maximum match rate ( within a column) varied by 2- 8%,
( Standard error less than .02), suggesting that match rates for each aggregator collection did
not vary significantly across libraries
( Click) So overall the average single aggregator march rate varied from 5 to 23%
Turning this analysis around, we compared the collective set of print book purchases from
the 5 libraries and matched them to the vendor lists.
Each bar indicates the ratio of books that matched one through all five of the libraries print
Th i t ti d t i t t The interesting data point to note is the long dark blue bar of NetLibraryfor the “ One”
category – indicating that this vendor most often matched the unique material from each of
the libraries, well above any of the other libraries.
In the future, we could use this ratio to assign each vendor a “ uniqueness factor” associated
with their match to the local collection.
We offer an initial observation related to publisher and subject
• On the left are the 12 publishers with the most ebooks that did not match the aggregated
and on the right are the 12 publishers with the most print purchases that were not available
• Publishers on both lists appear in bold, showing that it is common for ebook aggregators to
have SOME, but not all of the books from a particular publisher
• Only 4 of 12 publishers appear on both lists, bringing into question why some of their
books are available as ebooks, but not others which conceivably have some demand in e-format
from the libraries
• 7 of the publishers appearing on the print purchases not available as ebooks list are
University Press publishers ( initalics), indicating that this group may be behind in providing
to their content as aggregator- available ebooks
• This table shows the top 12 LC classes purchased by Claremont
• As expected about ¾ of the books were not available from any aggregator
• There was some variation across these locally popular subject areas:
• With more than average e- available in Economic history, and
• Less than average e- available in the arts and romance literatures
• Further subject analysis would certainly be of interest
• Overall, we found that aggregator title lists are still largely unique
• Library print purchase patterns are still largely unique as well
• Only 30% of print titles are available in the aggregated book marketplace
• Some publisher- type differences may account for the non- match rates, as
evidenced by our large % of non- matches in our university press purchases
“ Quiet challenge” to NetLibrary to expand their use rights to make their extensive
Conclusions– No possibility of duplicating our current print monograph print
profiles in the ebook world
Any questions, thoughts, or reasons to call us crazy?
© Jason Priceand John McDonald
Contact us: Jason. price@ libraries. claremont. edu and
John. mcdonald@ libraries. claremont. edu
What about the Google Books Settlement announcement?
Click tabs to swap between content that is broken into logical sections.