Academic books are an important part of scholarship and have traditionally been integral to academic libraries as they develop collections to support the research needs of students and faculty members. However, as library budgets shrink and students and scholars turn toward away from the liberal arts, university presses and other associated organizations have begun to express concern that book sales are in decline. But another phenomenon started happening simultaneously in this industry: Amazon began selling academic books, competing for customers with the likes of GOBI Library Solutions and ProQuest, among other established wholesale vendors whose metrics are often used to measure university press sales. This led Joseph Esposito to question whether Amazon’s role in the academic book market was causing sales to appear artificially depressed because they simply weren’t being counted toward university press sales.

With support from The Andrew W. Mellon Foundation, Ithaka S+R undertook a national study to examine the extent to which academic libraries were purchasing books through Amazon. Along the way, the study grew to encompass a number of other questions, such as how print and e-book expenditures have changed over time, how libraries spend to obtain books in different disciplines, and which are the largest vendors and university presses active today. The data used for this study came from academic libraries that use the integrated library systems Alma and WorldShare Management Services (WMS) to manage their acquisitions data.

The final analysis is divided into two sections. The first examined library acquisitions at 124 academic institutions in fiscal year 2017, and the second examined trends in acquisitions at 51 academic institutions that were able to provide continuous data from fiscal years 2014 to 2017. Our analysis was limited to print and electronic books acquired on a one-time, title-by-title basis, where the former were capped at $215 and the latter at $350. This was done to mitigate the inclusion of items that were miscategorized as books and to remove very expensive books and books obtained as part of a package deal, whose individual contents are not recorded.

Our findings in FY 2017 suggest that ongoing resources account for 75 percent of materials expenditures, and one-time resources, including books, account for only 20 percent. The libraries in this sample put 24.5 percent of their expenditures toward books, and of this, approximately seven percent of expenditures are for those books included in our main analysis. Libraries spent 42.6 percent of their print book budgets on humanities titles—by volume, humanities books made up almost half of the books added to libraries’ collections within our subset. Social sciences titles comprised the largest share of e-books at 32 percent of expenditures. Libraries put the highest share of their expenditures with university presses toward acquiring books published by Oxford and Cambridge, and overall, university presses held 23.6 percent of the print book market and 18.5 percent of the e-book market. And to answer the main question, Amazon is the second largest vendor of print books, but trails GOBI Library Solutions by a wide margin and has no discernible presence in the e-book market.

From FY 2014 to 2017, library material expenditures increased in real dollar terms, but print book expenditures saw a year-on-year decrease while e-book expenditures saw a slight net increase. Additionally, the average cost of an e-book in our sample rose by 35 percent, while the average cost of a print book remained relatively stable. Print book expenditures by disciplinary field also dropped, although expenditures for humanities titles saw the lowest drop; for e-books, humanities expenditures, along with those for social science titles, actually saw an increase in spending of 2.2 percent and 7 percent respectively, where other fields decreased. Expenditures for university press print books fell by 17.7 percent during the four-year period, but the proportion of expenditures for UP titles relative to commercial press titles held steady at around 20 percent. University press e-books effectively saw no net change in expenditures, and GOBI and Amazon were again the leading vendors, although libraries continue to also acquire print books from specialized vendors.

Although the findings in this report provide a compelling look into the acquisitions at academic libraries over the past few years, they should not be viewed as representative of all U.S. academic libraries, nor should they be viewed as conclusive. Throughout the course of this study, we grew to appreciate that integrated library systems are best suited for print resources in an increasingly digital environment, and that they are not optimized to act as a data source for cross-institutional analyses. These gaps in our data generated more questions. For instance, to what extent do packages offset the declines we saw in print book expenditures? As libraries move to acquire more electronic resources, how will publishers and vendors adjust to meet these needs while meeting their bottom line, and how will this affect libraries’ acquisitions?

This study only serves to dip our toes in the water, so to speak. As the community continues to explore our findings and add to ours with their own experiences, greater sharing of data and improved tools to better manage data and electronic resources will be important to continuing the conversation.