The most recent report from Ithaka S+R, on Library Acquisition Patterns, examines book purchasing data gathered from academic libraries between fiscal years 2014 through 2017 in an effort to discern current acquisition trends. Takeaways include a declining university press presence, continued strong representation in humanities titles, and Amazon not leading the vendor pack by a wide margin.

The data was gathered from libraries that use either Ex Libris’s Alma integrated library system (ILS) or OCLC’s WorldShare Management Services (WMS), and generated two strands of information: an analysis of 124 academic institutions’ library book acquisitions for FY17; and a trend line of print and ebook acquisitions, university press presence, and book vendors at 51 United States academic institutions between FY14­–17. Samples were limited to print and electronic books acquired on a title-by-title basis, as opposed to books purchased by subscription or as part of a package.

FIRST, AN AMAZON QUESTION

The impetus behind the report, coauthored by Ithaka analyst Katherine Daniel and director of the Libraries, Scholarly Communication, and Museums Program Roger C. Schonfeld, with Joseph J. Esposito, senior partner at Clarke and Esposito LLC, “was an interest in the extent to which Amazon has grown to be a real force in the distribution of books, especially university press publications, to libraries—especially academic libraries,” Schonfeld told LJ.

Esposito, who specializes in strategies in digital media, publishing, and education technology, had approached Ithaka in 2014 with a query about how much university press sales were actually declining, and what percentage of Amazon’s sales were to libraries. “Given Amazon’s purportedly prominent role in the book-selling business,” the report’s introduction poses, “is it possible that libraries might not actually be obtaining fewer books published by university presses, so much as their acquisitions are not being tracked because they are being made through Amazon?”

With funding from the Andrew W. Mellon Foundation, Ithaka began investigating that question. Cloud-based library platforms were beginning to see wide use, said Schonfeld, and the investigators saw an opportunity to use ILSes to gather information from their acquisitions files, rather than consolidating vendor sales data. The result, they hoped, would be a database of academic library purchases of wider scope than had been previously possible.

That methodology, Schonfeld noted, has “enabled us not just to answer the Amazon question...but to ask, and indeed start to answer, all sorts of other kinds of related questions about acquisition patterns and sales channels and publisher roles and disciplines."

Schonfeld and Ithaka senior analyst Liam Sweeney launched an initial investigation in 2016, reaching out to several institutions that used Ex Libris Alma. Through that pilot, Sweeney ascertained that an ILS could serve as a viable source of data for library acquisitions. In 2017, Ithaka S+R expanded the Library Acquisitions Project (LAP) to incorporate data from WMS as well; the two systems have standardized data fields, rather than allowing customized fields, which allowed for consistent cross-institutional data gathering and analysis. Publisher information was standardized through each record’s ISBN information, and disciplines were captured using both the Dewey Decimal system and Library of Congress classifications.

PRINT DROPS, E INCREASES

The report’s findings highlighted several trends in academic book purchases, with the caveat that the participating institutions—all of which grant a bachelor’s degree or higher—varied widely. “We're not trying to make a claim here that the findings are perfectly representative of all higher education,” Schonfeld told LJ. “But at the same time we have a really broad sample, a truly diverse set, of higher education institutions."

Over the entire FY14–17 period at the institutions surveyed, although library expenditures as a whole increased, one-time, title-by-title spending on print books decreased year to year across all fields, with humanities seeing the smallest drop and STEM (science, technology, engineering, mathematics) the largest. One-time ebook purchases experienced a net increase, but not sharply enough to offset the drop in print book expenditures. Ebook spending on humanities and social sciences increased.

Spending on university press print books fell by 17.7 percent during those years, but the proportion of university press acquisitions—approximately 20 percent of one-time print book purchases across all four years— remained essentially stable compared to those from commercial presses. Ebook acquisitions fluctuated between 2014 and 2017, but averaged out to remain relatively flat even as the average cost of an ebook rose by 35 percent. The cost of print books remained stable.

On average, the libraries included in the FY17 study spent a total of $3.61 million on books, adding an average of 4,750 print and 345 ebooks.

Books constituted 24.5 percent of materials expenditures, versus the 60 percent spent on serials and other continuing resources such as databases and subscriptions. Of libraries’ book budgets, 42.6 percent went to humanities titles, which accounted for nearly half of the books added to their collections. Social science titles accounted for the greatest proportion of ebook expenditures, at 32 percent.

These purchases were made through a number of vendors, but GOBI Library Solutions led the pack overwhelmingly, with 68.7 percent of total print books and 86.4 percent of ebooks. As the second largest vendor of print books, Amazon only captured 10.9 percent of the academic market. For ebooks, the margin was even wider, with ProQuest/Coutts coming in second at 6.9 percent. University presses represented 23.6 percent of the total print book market and 18.5 percent of ebooks, with Oxford and Cambridge holding the largest share. The majority of university press book acquisitions were made through GOBI and Amazon as well.

However, Daniel noted, while GOBI commanded the largest proportion of the market, “It is important to note that GOBI…also distributes other publishers’ and vendors' products, so that number may be a bit inflated." For example ProQuest, which commanded 2.5 percent of the print book market in FY17 and 6.9 percent of ebooks, also sells through GOBI.

NORMALIZING DATA

Information was gathered using queries developed by Ithaka’s analysts. While the consistency of the Alma and WMS data fields helped simplify the collection process, there was still work to be done on the back end normalizing fields to conform to Ithaka’s database. Certain miscategorizations also needed adjustment, noted Daniel, particularly when it came to the metadata libraries acquired from vendors and publishers, which “can be incredibly messy.” Books had often been recorded as journals, and journals or videos as books.

In addition, data concerning ebooks sold as part of packages needed to be excluded from the analysis. These bundles, often recorded as one invoiced item, can potentially comprise thousands of books, and have become more popular with academic libraries in recent years, and may warrant a closer look in the future.

“Some of the packages that the science publishers, Springer Nature and Elsevier and so forth, have modeled...are very much like journal bundles in a lot of ways,” noted Schonfeld, although these haven’t gotten the same scrutiny from the library community as journal packages.

"I think there's still a lot of changes yet to take place in terms of how libraries are interested in having ebooks bundled together, if at all, under what terms, and how that fits together from the publisher perspective,” he added. “When Springer Nature offers a package of ebooks it's basically ebooks that Springer Nature itself is the primary publisher of, but when...EBSCO or ProQuest offers a package of ebooks…they in turn are licensing from a set of primary publishers, packaging them together in one way or another, and then in turn licensing to libraries."

Some of these models have resulted in less revenue for publishers than they would see from selling books on a title-by-title basis—balanced by potential exposure to a larger market. Other purchasing models such as demand-driven acquisitions are also on the upswing, said Daniel, and will likely influence future data.

LOOKING TOWARD TRANSITIONS AND MORE

Along with the acquisition data itself, the authors hope the report will help shine a light on format shifts. While this has been examined with regard to journal sales, Ithaka’s LAP is an opportunity to dive more deeply into one-time sales of print and ebooks. “The format transition for books is proceeding with more fits and starts than the more linear format transition that we see for journals,” said Schonfeld. “So I think there's quite a bit of data that speaks to some aspect of that and gives libraries a chance to see their planning in that context."

The findings, noted Schonfeld, should appeal to librarians interested in the scholarly communications ecosystem, and how their local purchasing decisions could affect the landscape as a whole. Much of the data collected could also be used to look at acquisitions or licensing of journals, or how libraries handle primary source collections. “We'd love to get input on what other kinds of questions the community would like to see us use this dataset to explore," he said.

Much of the praise, added Schonfeld, goes to Ithaka’s ILS partners. "This was a project that couldn't have happened without some incredibly close partnerships,” he explained. “The cloud-based system providers, Ex Libris and OCLC, went above and beyond the potential in this to make it possible…. I think it's helped us to see the value of the cloud based systems, not only for some of the management and collaboration tasks and priorities that a lot of the libraries…have been taking on, but also at an analytical and reporting and policy-making side. I think these platforms are making it possible to do data gathering and analysis that just wasn't previously feasible.”