The Changing Landscape of Academic Publishing
Two Librarians’ Perspectives
When it comes to library collections, it is critical that we not confuse means with ends. The purpose of the library is not to have a great collection; the purpose of the library is to meet the teaching, learning, and research needs of those it was established to serve. The collection is a tool that we use in pursuit of that goal, but it is not the only tool, and the collection itself is certainly not the goal. One problem with building collections (as traditionally understood) is that it involves guesswork, trying to guess ahead of time which exact books, journals, and databases our users will need. But inevitably, we guess wrong. We are very good at selecting relevant and high-quality books but not very good at selecting the exact titles that our patrons will need. So in many libraries, we have been moving aggressively in the direction of what is called patron-driven or demand-driven acquisition: We put records in the catalog for electronic books that we have not yet purchased, and purchase is driven by the actual use of those books. Even better, the process of purchase is completely transparent to the patron, so the only thing the patron experiences is access to a much larger collection of titles than we could possibly offer by preemptive purchase. The result is not likely to be a better collection (in terms of coherence and completeness), but if the result is that our patrons are better able to do their teaching, learning, and research, then many librarians will consider that to be an acceptable tradeoff.
The model that built the great research collections was a reflection of the limitations of the print world. When print runs for scholarly monographs were modest and moving individual books from one institution to another was time consuming and costly, buying monographs just in case they were needed made sense. The principles behind building a balanced collection were well understood. For the disciplines in which an institution had doctoral programs, that institution attempted to collect everything that was published in that field. Masters programs could be a little more selective, and for bachelors programs, general works could usually suffice. A consequence was that many items were never used, but if that meant that a scholar rarely had to resort to interlibrary loan, this was simply the cost of doing business. We did not have the kind of usage data that would have enabled us to do things differently.
Although the move toward demand-driven acquisition has been revolutionary in large research libraries, it has had less impact in health sciences libraries, where collections are much more skewed toward journals and databases. The typical biomedical library spends only 7% of its collection budget on monographs.1 This makes the selection process even more challenging because adding a new journal or database resource is potentially a multiyear commitment. When we buy a book, there is usually no point in paying much attention to usage data. The book is (usually) a 1-time purchase, so there is no way to “cancel” it if it turns out to have been the wrong choice. Journal and database subscriptions, however, represent ongoing commitments of funding, and we can bail out of them later if they turn out to be returning low value in terms of ongoing, real-world usage.
So there is a great deal of interest in ways to extend the demand-driven model to article acquisitions, although it has been challenging. In the past, we relied on interlibrary loan to obtain articles from journals for which we did not have subscriptions. Now we find that even though we can fill most requests in <24 hours, even a few hours may be unacceptably long for some of our primary clientele (particularly in academic medical centers). Standard pay-per-view charges are often too high to be a practical substitute, but services like the Copyright Clearance Center’s Get It Now can provide an appealing option. Participating publishers discount the price they would normally charge, and libraries can split the cost with the requester, cover it entirely, or pass the entire cost on, making it manageable depending on local circumstances. DeepDyve, which pioneered article rental, offers another promising avenue. We hope to see much more experimentation in this area.
Services like these still operate very much at the margins, however. With open-access articles still providing only a fraction of available content, librarians are very focused on trying to make the best decisions they can about what resources to license.
Librarians have been slow to change our models to accommodate the radical changes in the environment. For electronic books and journals, a limited print run is meaningless, and moving an item from one place to another is virtually without cost (leaving aside, for the moment, the substantial investments in infrastructure that make this possible). We can get a level of usage data unimaginable to our colleagues from decades past. Now we have to figure out how to use that data wisely.
The Role of Usage in Making Collection Decisions
Librarians, perhaps unwittingly, overemphasize usage data and give publishers the impression that they play a larger, more stand-alone role in our decision making than is often the case. Provosts would love to know the magic number, the specific cost per use above which we would always cancel something. But it is not that simple. We also need to take into account who is using the resource and how it fits into the university’s research and education priorities. That is much harder to quantify.
During the print era, collection building by prediction was the only option available to us, and this system resulted in some truly great library collections. But the problem is that it does not matter much to the users we serve how objectively great our collections are; what matters is whether our collections actually contain the exact documents needed. The universal existence of interlibrary loan departments in both medical and research libraries (and the fact that those services tend to be very busy) suggests strongly that even the greatest collections fall significantly short of meeting the real-life needs of those they serve. And it is also worth pointing out that a vanishingly small percentage of library users have access to great collections.
Cost per use is only 1 data point we must use in making cancellation decisions. That said, it is a very important one. If a journal is costing us $75 per download and we could have gotten those articles piecemeal at a unit cost of $30, then there is a pretty strong argument to be made, not that we should not provide access to the content but that a subscription is the wrong way to provide it. However, other factors, some of them political, also have to be taken into account. A more accurate way to look at it is that a high cost-per-use figure does not trigger an automatic decision but rather a conversation. When that figure gets above the cost per article via interlibrary loan or document delivery, then we start looking critically at the subscription. We do not necessarily cancel, but we start talking about the possibility of canceling.
One drawback of comparing cost per use to interlibrary loan costs, however, is how unappealing those loans can be to some segments of our clientele, particularly in the biomedical arena. This will vary, of course, by whether the library charges for interlibrary loan and the expectations of the clientele. Several years ago, we held a series of focus groups at University of Alabama at Birmingham to investigate what faculty and graduate students did when they were not able to get an article they wanted immediately. Interlibrary loan was not high on the list. Many of the participants were only vaguely aware of the option, and even if they were, they felt that filling out the form, waiting for the article, and paying the cost were not worth it, even though we make it as easy, fast, and cheap as interlibrary loans can be. Depending on the need, they were much more likely to try to find a different article, rely on the abstract, do without, or, as many of our participants sheepishly admitted, contact a colleague at another university who did have access and would send them a copy. Given the nature of the focus groups, we cannot quantify that type of usage, but clearly, the black market in journal articles is thriving.
Alternatives to Interlibrary Loan
Given the limitations of interlibrary loan, there is growing interest in alternative means for acquiring individual articles. Purchasing directly from publishers is one avenue, although often it is too expensive to be practical at scale. Services like the aforementioned Copyright Clearance Center’s Get It Now service provide a mediated process for acquiring pay-per-view articles in a more cost-effective way.
ReadCube is an example of another new service providing a variety of options: One is temporary read-only access, which costs the library only a few dollars; the second is permanent read-only access, which costs more; and the third and most expensive is a downloadable, DRM (digital rights management)–free personal copy. Libraries that have experimented with such models report considerable success so far in terms of both patron satisfaction and library savings. With these models, the cost-per-use cutoff becomes much more effective.
The Rise of Institutional Repositories
As we move in this direction, some questions are raised about the roles that librarians (and the library as an organization) play in supporting the diverse needs of their communities. As libraries reorganize to adapt workflows to the demands of managing digital content, we see shifts such as renaming the collection development function “content management,” a reflection of the fact that we are not building collections anymore as much as we are developing services to help people effectively access and make good use of content in a variety of forms. This shift has been more pronounced in medical libraries because of the heavy emphasis on licensing electronic content. For many of the disciplines that a research library serves, the print monograph collection will continue to be a major investment for quite some time to come.
This shift is also reflected in the investments that many libraries are making in institutional repositories (IRs). Although much of the focus on IRs in the library community has been on using IRs to provide open access to versions of published articles, libraries are also using them to provide access to local materials that fall outside the traditional publishing streams. At the University of Alabama at Birmingham, for example, the digital collections encompass short films from the ethnographic studies course, oral histories covering a variety of Alabama topics, a variety of the University of Alabama at Birmingham publications, and masters and PhD theses and dissertations. At the University of Utah, students and faculty are invited to deposit (or to let us deposit) versions of their published articles in our IR, and we too use the IR as the depository for University of Utah theses and dissertations. We are also increasingly looking for new ways to make the IR program useful on campus. In other institutions, we are seeing partnerships between libraries and university presses expand as libraries start to move into a variety of publishing support services.
Activities like these may seem to move the library far afield from its traditional role of developing a collection of published materials for the use of the members of its communities, but they can also be seen as merely an extension of the library’s traditional mission of connecting people to the knowledge they need to do their work.
Although open access remains a hot topic for librarians and publishers, it appears to have had little impact so far on the decisions that librarians make about what to acquire or license. Librarians are overwhelmingly supportive of the general concept of open access because it reduces barriers to connecting people to content, and most libraries will sponsor the occasional workshop on author’s rights and do a fair amount of one-on-one interaction as faculty have questions. Many research libraries are developing their own publishing services programs to help develop open-access content. Theoretically, if we could get the open-access versions of articles as easily as the publisher versions, we would be likely to consider canceling licenses, but the current state of things has open-access versions scattered across multiple repositories, and there are no easy ways to determine whether all of the content of a journal is available in some free and easy-to-access version. Even the National Institutes of Health’s public-access policy has not affected this much. It is great that people can now find freely available versions of articles from journals we do not license, but we do not see it leading us to significant cancellations anytime soon.
We can certainly expect to see an expanded emphasis on data curation: librarians working with the Offices of Research and Sponsored Programs, along with information technology and other university entities, to develop appropriate infrastructure, policies, and services to support the research enterprise in complying with regulations and requirements from funders at all levels and, increasingly, from journals. The Transparency and Openness Promotion (TOP) Guidelines that were recently announced (the American Heart Association is a signatory) and similar efforts will require institutions to develop services to support their investigators, and librarians have considerable unique expertise to add to these efforts.
Data curation is a huge and complex topic, and it is not yet entirely clear how the IR will work as a data host. Most of the conversation over the past couple of years has been about “big data,” and it seems clear that library-based IRs, which were set up to handle images and conventional text documents, are not likely to be the home to really big data sets, but a lot of “small- to medium-sized data” are being produced on our campuses, and the IR has a vital role to play there at least. Librarians are engaging in ongoing conversations with the Office of Research, campus information technology, and other units about how we can best help the campus solve its data-curation problems, both current and future.
Libraries and Institutional Priorities
Our bottom line, regardless of the type or size of library, is ensuring that what we acquire and the manner in which we acquire it are tightly aligned with mission priorities. We can anticipate that this focus on local/institutional mission will become more controversial as pressure grows on libraries to act as agents of change in the larger scholarly communication system, an orientation that inevitably pulls attention away from meeting local, immediate needs even as it, we hope, creates an environment that will be more amenable to those needs in the future. Striking the right balance while remaining mission-critical to our sponsoring institutions2 is our constant challenge.
About the Authors
T. Scott Plutchak (Director of Digital Data Curation Strategies, University of Alabama at Birmingham) and Rick Anderson (Associate Dean for Collections and Scholarly Communication in the J. Willard Marriott Library, University of Utah) are longtime friends and colleagues.
- © 2016 American Heart Association, Inc.
- 1.↵Annual Statistics of Medical School Libraries in the United States and Canada. 37th ed. Seattle, WA: Association of Academic Health Sciences Libraries; 2015.
- Anderson R