This paper was refereed by the Journal of Electronic Publishing’s peer reviewers.

Abstract

This study illuminates the differentials in book pricing by format, which can appear arbitrary and unpredictable. Using the paperback list price as the base, the research uses a sample of 500 titles selected for purchase by a small, private liberal arts college library during a defined time period. I have recorded the prices of paperback, hardcover, single and multiple-user e-books from Ebrary, EBSCO, and Amazon Kindle. The mean pricing differential is calculated for each version across all titles in the sample, as well as for the top five publishers represented in the sample, in order to identify patterns in pricing decisions.

Introduction

The prices of books, like those of other commodities, are affected by supply and demand. Other factors affecting book prices have included the Net Book Agreement in the UK (in effect 1900–1997), and price fixing of e-books by Apple, Inc. and five major publishers, as a US District Court decided in United States of America v. Apple Inc., et al., in 2013. A Wall Street Journal article explains the factors affecting e-book pricing by Amazon and other direct-to-consumer e-book vendors (Trachtenberg 2015). One recent research project examines the history of library book pricing in print and electronic formats, including the history of net book pricing, price fixing, and responses from libraries (Zhang and Eschenfelder 2016). The researchers find that the pricing variables that determine the prices of print books sold to libraries include the length of time since initial publication (newer books tend to have higher demand and higher prices than backlist titles), genre, version, binding, condition, and paper quality; e-books for libraries are priced by taking into account various factors including DRM, number of concurrent users, perpetual access, age, price of the print version, loan cap, and embargo.

There are also a variety of factors that affect the cost of publishing a monograph, and these may ultimately affect the sale price for print and electronic versions. In an Ithaka S+R report from 2015, Nancy Maron, et al., studied 382 titles recently published by university presses and found that the cost of publishing each title has varied “ from a low of $15,140 to a high of $129,909 . . . [and] the overall average cost per title was $28,747 (basic) and $39,892 (full cost)” (2016, 17). The basic cost “includes the cost of staff time and other non-staff direct costs in Acquisitions, Manuscript Editorial, Design, Production, and Marketing,” while the full cost adds press-level overhead costs to those factors (4). These costs are for monographs from scholarly presses, which are a subset of the titles in the current study. The Ithaka project illuminates various factors that shape the costs of publishing scholarly books and that ultimately affect the sale prices of those books.

Previous research on the e-book market has mainly focused either on sales to individuals for e-reader devices or on acquisition models for libraries (perpetual purchases, short-term loans, and subscriptions). While there is little published research on the comparison of book prices in print and electronic formats, the average purchase prices for electronic and print books in the library market have been compared. Lippincott, et al., found in 2012 that in YBP’s GOBI ordering database, single-user e-books (that is, one user at a time) from Ebrary were typically priced the same as hardcover books, and multi-user (allowing an unlimited number of simultaneous users) e-book prices increased by 50% from the hardcover price. A 2013 study (Bailey, et al., 2015) compared average costs of print and electronic books for libraries and found that e-books tended to be more expensive. Both of these studies have considered prices only in the aggregate. The present study is more granular and seeks to identify more detailed pricing patterns by comparing costs of paperbacks with those of hardcovers, single and multiple-user e-books from two major library e-book aggregators, and Amazon e-books where available.

I will analyze the wide variance in pricing among paperback, hardcover, Ebrary, EBSCO, and Amazon Kindle books. Prior to this study, I had begun to observe that e-books for the library market often have either the paperback or the hardcover cost as the base price for a single-user e-book, with an unlimited-user version available at a higher price. However, I often noticed that certain single-user e-books were priced either much higher or much lower than the print version. Anecdotal evidence suggests that Kindle is often the cheapest option, and that Ebrary and EBSCO are comparably, but not always equivalently, priced. This study will test my informal observations by systematically analyzing pricing data from books selected for purchase by my library. An analysis of this data will illuminate the relationships among the prices for each version, by calculating the differences and seeking patterns in the multipliers used to determine the prices.

Methodology

I analyze books selected for firm orders by the Olin Library at Rollins College, a small, private liberal arts college in Winter Park, Florida. I began with books selected on a specified date (Feb. 13, 2014), moving forward in time until a sample of 500 was collected (Nov. 11, 2014). I compare the prices of the paperback, hardcover, Ebrary, EBSCO, and Kindle books. I use price data for the first four as reported in YBP’s GOBI ordering database and Kindle prices as listed on Amazon.com. To be included in the study, a book must be available at least in paperback, as well as either through Ebrary or EBSCO, and listed with pricing in GOBI (regardless of whether it’s in print or in stock). The sample is not a randomized set; rather it is the complete list of books selected during the specified time frame that fit the criteria above. If a 3-user option or a concurrent access option is offered, that cost is disregarded, and only the single-user and/or unlimited-user prices are recorded. Rental or lease prices are not recorded. I use paperback pricing as the base and analyze the other prices in relation to that. I compare books among broad subject areas, based on LC ranges as found in GOBI, to determine whether the subject affects the price differential.

I noted titles that were advertised in special categories that might result in pricing differences. Of the total list, 5 are reference books, and 66 are marketed as textbooks, including 1 that overlaps both categories. Seven are juvenile books. I chose not to break these categories out of the tabulation, because they all represent actual selections for our library. Thus the aggregate figures would be skewed if books in those categories were removed. As the primary inquiry explores the difference in pricing among versions of a title, the commonly low prices of juvenile books or high prices of textbooks and reference books, would not adversely affect the sample. Additionally, when librarians define the parameters of a demand-driven or approval plan, we do not necessarily differentiate the various genres of books from each publisher.

Expected outcome and significance

This study will illuminate the differentials in book pricing that can appear arbitrary and unpredictable. The data will inform publishers, aggregators, distributors, and librarians as all parties make decisions for pricing or purchasing print and electronic books. In order to identify patterns, I will report mean differentials in pricing by format, simultaneous users, aggregator, publisher, and broad subject areas. These are significant factors for librarians who purchase books through approval plans and/or demand-driven acquisition, as they may specify criteria for the preferred version to be delivered, and these criteria may be delineated by publisher or subject area, for example. Patterns in pricing may inform these decisions which, on a large scale, could have significant effects on library budgets and book sales revenue. For example, if a particular publisher appears to be consistently pricing e-books far above the print price, a librarian might opt to exclude that publisher from a demand-driven or electronic-preferred approval plan.

Analysis of results

The top five publishers represented in the study are as follows: Routledge, 43 titles; Rowman & Littlefield, 39; New York University Press, 27; McFarland, 19; and Oxford University Press, 18. These sets of data are large enough to be generalizable for those publishers, as opposed to the many publishers represented in the study by single titles. The mean prices by format for these top five publishers, along with the pricing differentials as factors of the base (paperback) price, are as follows:

For the entire sample set of 500 titles, the mean prices and differentials across formats and aggregators are as follows:

In comparing the aggregators Ebrary and EBSCO, overall Ebrary appears to be slightly more expensive for single-user e-books (although the difference of 3.94% is not statistically significant), while EBSCO is more expensive for unlimited-user books (this difference of 14.71% is statistically significant). Both are offering single-user e-books at significantly below the hardcover cost but over twice the paperback cost. Their unlimited-user books are above the hardcover cost and over thrice the paperback cost.

As noted in the above set of data, I discovered other patterns I was not seeking. With the exception of Oxford, the publishers with the greatest numbers of titles in the sample did not have certain versions available. Of course, the scope of the sample requires a paperback and an Ebrary and/or EBSCO version for a title to be included in the study. So it is possible that, for example, McFarland does publish hardcover books; but there are none in the sample.

In terms of subject areas, the top 5 broad Library of Congress classification ranges represented are H (Social Sciences), 122 titles; P (Language & Literature), 75; E–F (History of the Americas), 54; B (excluding BF; Philosophy & Religion), 53; and D (World History), 35. Following are the mean prices and factors of difference for these ranges, in LC order:

Considering only these top subject ranges, the mean paperback prices are comparable, with a slight uptick in the D range. The hardcovers are also fairly evenly priced, with B and H a bit higher. The Ebrary single-user prices are at least about twice the paperback, ranging from a factor of 1.99 in E–F to 2.6 in P. For unlimited-user Ebrary books, prices range from a factor of 2.54 in B to 3.59 in P. EBSCO single-user titles range from 1.92 in E–F to 2.41 in B, and for unlimited users, from 2.98 in D to 4.11 in H. Kindle books begin at .57 in three broad subject areas (B, E–F, and P) and rise to .69 in D. Apart from these factors, we see no consistent pattern in prices by subject area, except to note that Ebrary books are priced highest when in the P range; the lowest prices are for Kindle books; and the highest pricing factor is for EBSCO unlimited-user titles in H.

Conclusion

As we see in the Ithaka study by Maron, et al., there are a variety of costs involved in publishing a book. Publishers’ attempts to recover these costs ultimately affect the sale price of each title. We find in the current study that the sale price for a title is not assigned at a single level but rather in tiers by format, and in the case of e-books, by aggregator. With the exception of Kindle, all hardcover and e-book versions in the study are priced at more than twice the paperback cost. Subject areas do not significantly affect the prices. Single-user e-books from Ebrary and EBSCO (priced similarly by both aggregators) are significantly cheaper than hardcover versions; and unlimited-user e-books from both aggregators are significantly more expensive than hardcover versions (with EBSCO priced significantly higher than Ebrary at this level). Meanwhile, Kindle e-books are significantly cheaper than all other versions, including paperback.

Specifically, the figures are analyzed as overall mean prices in relation to the base price for each title. For the purpose of comparison, we consider the paperback price as the base figure for this study. The overall average sale prices of the 500 books in the study indicate that, on average, hardcovers are about 2.5 times the base price of paperbacks. EBSCO and Ebrary single-user e-books are between 2.2–2.3 times the base price. Ebrary sells unlimited-user e-books at 3.2 times the base, while EBSCO prices them significantly higher, at about 3.7 times the base. Kindle tends to be the cheapest version, at about 62% of the base price.

We conclude that book prices are not, on the whole, as arbitrary as they might seem in some cases. Although some titles are priced at wildly varying figures for different versions, there is still a discernible overall pattern to the average pricing of a given publisher’s books at each tier, in relation to the base price. Although there are outliers, it seems we are seeing that there is a “rhyme or reason” in book pricing. In other words, in the case of each publisher represented by significant numbers of titles in this study, there is a logical pattern to the pricing of their books by format and version.

The aim of this study is not to editorialize about the ways in which pricing decisions are made. In the case of e-books offered by aggregators, those prices depend on negotiations between publishers and aggregators. Whether or not either party might be using a mathematical formula as a starting point, there is a set of de facto patterns in the pricing of books by specific publishers that emerges. As librarians create and refine parameters and preferences for approval plans and/or demand-driven acquisitions, and as they update collection development policies, they may use these patterns to inform decisions and to support cost-effective unmediated purchasing.

Suggestions for further research

This study has analyzed book prices from a sample based on actual purchases by a small college library. Similar studies could be performed at different types of libraries. There is other data that has yet to be studied, including prices from aggregators beyond those discussed here. EBL and Ebrary are both owned by ProQuest, and their pricing could change depending on organizational structure and strategic direction, as the two platforms are currently being combined as ProQuest Ebook Central. YBP is now owned by EBSCO; this change occurred after the data for this study was gathered. Other aggregators such as JSTOR, Project Muse, and Credo Reference could be studied as well.

Pricing for textbooks could be studied in contrast to other books. Textbooks are so designated by the publisher, and that designation is reflected in marketing and retail pricing. However, a list of books adopted for courses at a particular institution could be analyzed, perhaps also in relation to the costs of e-textbook versions marketed directly to individuals, beyond library aggregators.

I have not sought to identify price changes for specific titles. The prices documented were gathered only once, at the time of selection. Changing prices could be analyzed in a future study. The speed of publication of print and electronic books could also be studied and compared across e-book aggregators, as some e-books on library platforms take years to become accessible after being listed for sale but “not yet published.”

Jonathan H. Harwell is Head of Collections & Systems at Rollins College’s Olin Library. He joined Rollins in 2012, and was previously a librarian at Georgia Southern University, the University of Alabama at Birmingham, and Berry College. He holds an MLIS from The University of Alabama, an MA in Social Science (emphasis in anthropology) from Georgia Southern University, and a BA in English (minor in Spanish) from the University of Southern Mississippi. He worked with a colleague, Geoffrey Timms, to develop the OARS system as an open-source mechanism to manage serial reviews. In his former life, he was a teacher in Albania for two years. His passions include collection development in academic libraries, and researching the cultural history of Quakers in the South.

Bibliography

• Bailey, Timothy P., Amanda L. Scott, and Rickey D. Best. “Cost Differentials between E-Books and Print in Academic Libraries.” College & Research Libraries 76, no. 1 (2015): 6–18.
• Lippincott, Sarah Kalikman, Stephen Brooks, Aisha Harvey, Judy Ruttenberg, Luke Swindler, and John Vickery. “Librarian, Publisher, and Vendor Perspectives on Consortial E-Book Purchasing: The Experience of the TRLN Beyond Print Summit.” Serials Review 38, no. 1 (2012): 3–11.
• Maron, Nancy, Christine Mulhern, Daniel Rossman, and Kimberly Schmelzinger. The Costs of Publishing Monographs: Toward a Transparent Methodology. New York: Ithaka S+R, 2016. http://www.sr.ithaka.org/publications/the-costs-of-publishing-monographs/.
• Trachtenberg, Jeffrey A. “E-Book Prices Rise, and Sales Suffer.” Wall Street Journal, Sept. 7, 2015.
• United States of America v. Apple, Inc., et al. US District Court, 2013. https://www.justice.gov/atr/case/us-v-apple-inc-et-al
• Zhang, Mei, and Kristin R. Eschenfelder. “Changing Perceptions of a Good Book Price: Comparing the Net Price System of the 1900s to Contemporary E-book Pricing Debates.” Poster, iConference, 2016.