What Happens When Ebooks Are Free-to-Read?

Image of a Google impact map, depicting content requests by world location.

Between March 20 and August 31, 2020, the University of Michigan Press made all the titles in the Library-hosted ebook collection, UMP EBC, free-to-read. This was in response to the sudden restriction of physical movement and building access due to COVID-19 pandemic “lockdowns” across the world. Because they came with little warning, these restrictions left many faculty, students, and librarians struggling to maintain continuity of research and learning. The removal of access barriers was decided upon with support from Library leadership and the Press’s faculty Executive Committee on March 18, around two days after the Publishing division’s staff vacated their office spaces and moved to 100% remote working -- a situation still in place at the end of the calendar year. The reimposition of access restrictions on UMP EBC at the end of August was a reluctant move. However, as a business that relies on sales revenue to sustain at least 70% of its budget, U-M Press could not afford to maintain free access indefinitely and September starts the sales season for ebook collections sold to libraries. 

Even though the immediate benefit for users of free-to-read content ended in September, we have been eager to learn from the data we gathered in order to shape our future business strategy. With this in mind, the Press did several studies of the impact of free-to-read. (Note that “free-to-read” means that all titles were available on UMP EBC for anyone to read without charge, but only users within the IP ranges of library purchasers of the collection were able to download books.) Three different studies are described below.

Study 1: IP Registry Analysis

With the generous pro bono help of the IP Registry (the global IP address database for libraries and publishers), we analyzed weblogs from UMP EBC for the first period of free-to-read availability, March 20 to the end of June 2020, to try and understand the pattern of “logged” and “anonymous” use. Logged use is traffic that can be recognized as coming from a static IP address and thus connected to an institution or an organization. Anonymous use is traffic that cannot be associated with a static IP address, although the protocols managed by IANA can allow the association of this traffic with, for example, a particular country.

  • The IP registry weblog analysis reveals 85% of the use of UMP EBC came from outside institutional IP ranges. Of these “anonymous” (or fairly — thanks to IANA — anonymous) uses, 43% came from the USA and 57% from other countries. The top ten countries were the USA, France, Singapore, Germany, UK, Finland, Canada, Russian Federation, Turkey, and India.

  • Of the uses recognized as institutional, users from 739 institutions/organizations used UMP EBC. The institutions came from 65 countries, although 80% of use came from the USA. Outside the USA, the top 10 countries or territories that users came from (in order of volume) were the UK, Germany, Canada, Italy, Portugal, Lithuania, Taiwan, Indonesia, Belgium, and Hong Kong.

  • The top 20 institutions by usage were: University of Michigan; University of Oxford; University of California Los Angeles; The Ohio State University; University of Washington; Northwestern University; Free University of Berlin; Columbia University; University of Chicago; University of St Andrews; University of Manchester; the University of North Carolina at Chapel Hill; University of Pennsylvania; Stanford University; New York University; University of Sao Paulo; Indiana University at Bloomington; University of Mississippi; Royal Holloway, University of London; and the University of Oregon.

Study 2: Google Analytics

The proprietary Google Data Studio reporting implemented on the UMP EBC home page (see “Visualize Impact and Engagement”) provides a similar, but less detailed, message to the IP Registry analysis. The attractive Google impact map (designed by Albert Bertram in Library Information Technology) which we implemented on the site provided a daily snapshot of the spread of use.

Please make more materials open-access. It would help world scholarship a lot. I am from the Philippines and much of our research would be better informed if we had access to more materials. Thanks.

Study 3: Qualitative Feedback

During the period of the free-to-read initiative, we conducted several surveys with key stakeholders to get a sense of their reactions to the initiative.

  • Respondents to our survey of authors were overwhelmingly positive with 97% supportive. Quotes from their responses fill out a picture of a group of scholars grateful for access especially on behalf of their students.

  • While we only implemented a pop-up survey of readers on UMP EBC relatively late in the process, we received a large number of responses. What is striking from these is the geographic spread of use and the degree of interest outside the academy. Some of the responses were very moving:

    • “Yes, I vastly prefer print books, but a lot of academic presses have prices so high that even specialized university libraries are not buying books these days. Being an independent scholar, eBooks at universities are not accessible to me. eBooks just make the scholarly caste system worse in this country.”

    • “During the pandemic access to campus and to libraries has been denied. All research and all assigned readings now consist only of materials available online. You should send out regular email announcements to scholarly lists and librarians to advise them of your free digital materials. It would be GREAT if university library systems could link directly to your pdf versions.”

    • “Please make more materials open-access. It would help world scholarship a lot. I am from the Philippines and much of our research would be better informed if we had access to more materials. Thanks.”

    • “I think the initiative to make usually incredibly pricy books easily and freely accessible is great (I was very excited when I found this website) and I will ask if my previous lecturers (I’ve already graduated) know about the website. If not, I will share the link with them and I think they might share it with currents students.”

Next Steps

Despite the limited term of free-to-read ebooks, University of Michigan Press learned a lot from the experience. Many other scholarly publishers who undertook similar initiatives are also studying what they learned and the results are starting to be released informally through conference presentations. Project Muse, for example, provided free access to over 25,000 books from 70 publishers over the summer. The Director of Johns Hopkins University Press, which runs Project Muse, has reported similar results to ours. JSTOR has also resulted similar results from its program.

The bottom line is that if a press is able to make its books freely available in ebook form, untapped global demand for the sorts of scholarship university presses publish emerges that is entirely at odds with the popular perception that monographs “are books that do not get read.” We’ve learned this not only from statistics (that could be clouded by the activities of digital bots and spiders), but also from the feedback collected from real individual users whose gratitude for access is not only moving but inspiring — to us as a publisher and to the authors who have written their books with the expectation that they will be read but rarely get such direct feedback.

The results have suggested several actions that the Press is now taking:

  • The majority of U-M Press sales are currently in North America, but there is clearly untapped demand for ebooks in several other countries. We have now signed a license agreement with Jisc to make UMP EBC easier to buy for libraries in the UK. We are also exploring a relationship with a sales agent for India, a growing economy where we saw a large amount of free-to-read use.

  • Even when the books were free-to-use, the libraries that had purchased the ebook collection and had catalogued it were still the highest institutional users. This has led us to focus on making sure that it is easy to “turn on” UMP EBC collections in the Alma Community Zone and comparable services supplied by EBSCO and OCLC. We have particularly focused on ensuring that UMP EBC open access books are easily available to include in a library catalog even if a library can’t afford to buy the collection. The high quality of the MARC records we can provide libraries thanks to the work of the U-M Library Technical Services department makes this possible.

  • The realization that much of the use was occurring in countries which could not afford full price access has led us, in the short term, to create international pricing discounted based on World Bank classifications of country by income level. Even with this cheaper pricing in place, the comments we received indicate that readers without institutional affiliation might still be unable to get access. In the longer term we are now exploring opportunities to expand open access availability of Press publications with an end goal of making all the specialist monographs published by the Press (at least those for which we have ebook rights) open access.