Working Paper

Stanford University Libraries & Academic Information Resources

Business models, not economic models for research libraries in the transition to more digitized information resources. *

A talk presented at the December 1996 meeting of the National Digital Library Federation with the NSF/NASA/ARPA Digital Library Initiative projects.

Michael A. Keller
University Librarian;
Director of Academic Information Resources;
Publisher, HighWire Press
Stanford University

The agenda suggests that these remarks will concern themselves with economic models from the perspectives of the members of the NDLF. However, two conditions make such an assignment perilous. First, I am not an economist nor have I ever studied economics. Second, as these remarks are the next thing to off-the-cuff, they have not been discussed nor has consensus been obtained by my colleague university librarians of the NDLF. Having thus leveled my ability to speak with any degree if credibility on this topic, I can now suggest a topic and a set of conditions which might have some utility to this gathering.

I believe that most university librarians in the NDLF would agree that one of their primary responsibilities is the operation of their programs in a business-like fashion. In the context of our not-for-profit operating environments, we are charged with providing as centrally funded academic entities inside of our universities resources and services supportive of teaching, learning, and research, the vast majority of them without extra charge to members of the university community, principally students and faculty. We are supposed to do this without overspending our budgets and most of us do this, most of the time.

So, on this pretext, that of running an academic support organization on a business-like basis, that is, without overspending one's budget even while satisfying a myriad of expectations and demands for academic information resources, I propose to discuss some business models from a couple of perspectives rather than economic models for all of NDLF. I suspect that these remarks and observations are ones which my NDLF colleagues will find common to their experiences well, but if they do not, some are here to emend and extend what I have to say.

What are the perspectives?

First, a few remarks on what university research libraries are and have been as well as what they might become, then…

Second, a few observations on business models to which we have contributed to through our Internet publishing support bureau HighWire Press, and…

Finally, an observation on some other business models we have seen only as consumers or intended consumers.

The 99 university libraries in the Association for Research Libraries expended about $517 million on books, serials, mss & archives, music, microfilm, databases, and so forth in 1995. Of that $517 million, about 61% goes for serial publication. We do not have good data on the percentage of that sum spent on information in digital formats, but based on Stanford's experience, it is about 10%, rising from a few tenths of a percent about a decade ago.

The rapid rise in the percentage of our library materials budgets spent on digital information, or more precisely for much of it, the rapid rise in the percentage of our library materials budgets spent on leasing access to digital information has caused some of us to worry about unforeseen and ill-considered changes in an underlying principle of research librarianship. And from changes in the underlying principle flow many changes in roles and methods of research librarianship. From changes in roles and methods might come serious consequences in the mid- to long-term for research and higher education, some of them quite negative.

The underlying principle is a simple one. University libraries are the result of a long evolution. One of their principle activities is to make capital investments in information of use to current faculty and students as well as for future generations of faculty and students. In making capital investments in the purchase, organization, and preservation of that information, research libraries were and are cultural repositories, highly selective it must be admitted, but in aggregate we in the old and new world have amassed a considerable portion what the intellectual classes, broadly defined, have considered to be the significant records of mankind and the thinking of mankind.

As the percentage of our library materials budget spent on access to information increases, much of that information networked and digital, much of it subject to contractual limitations on its use, manipulation, and capture for long-term storage, most of us can now see the implications for the functions just mentioned.

In the most sinister interpretation of the information-as-commodity model, we will tend to spend our library materials budgets on information as a commodity, not as an investment. Such expenditures are or will be for current faculty and students, with little or no concern for future generations of our own local communities, and no concern whatsoever for our role as cultural custodians. If we relinquish our role as the university's agent for spending on information resources and permit wide-spread "by-the-drink" or "pay-as-you-go" or "just-in-time" acquisition of information for individuals and very small subsets of the university community, we believe that duplicate spending will result. Such a distributed system of information gathering might also lead, some of us believe, to a needless recapitulation of the expensive and inefficient system of departmental libraries of the 19th-century, albeit in digital form. Such a concern is reflected as well in the for-profit publishers for the research community in their drive to perfect encryption, document tracking, and digital watermarks.

The Association of Research Libraries has probably exacerbated the trend by adopting as a community the phrase "access instead of ownership," which a generous interpretation might regard as a call for improved collaboration among research libraries. The NDLF has, I believe, a more complicated and layered view of what might be differing roles and responsibilities in the continuum from the access model to the ownership model.

The payment-by-the-drink software and systems we see under development, some of them in the DLI projects worry us as we see them as part of a larger problem, one of the conversion to commodity spending on information rather than investments for the long term in information. There does not seem to be much concern for the long term among our academic councils and at the national level, if the Clinton administration's and the European Community's proposals for WIPO are exemplars, there is such a pro-information-as-commodity bias that dire consequences are likely. Among the dire consequences is the increased rate of spending upon fewer sources of information. With the rapid acceleration of prices for serials in science, technology, and medicine over the past 15 years, we have actually seen a decrease in the amount of information, as measured by the number of titles and numbers of pages ultimately made available at individual universities. This new trend in spending on leased or highly controlled and limited access to digital information products could exacerbate the situation.

We in the research library community are of course very much involved in developing the means and the policies for creating and maintaining digital archives. Don Waters will address this next.

So what business model would affect this situation for university libraries? It is too easy to say that a single model will make the difference. Essentially we have to find ways to continue as best we can to make investments, long term and enduring investments in information, whatever the communications format. In order to do this, I think we have to alter some of our patterns of behavior AS A COMMUNITY of research libraries. I am tempted to say because this is an NDLF/DLI gathering AS ONE COMMUNITY OF LEADING RESEARCH LIBRARIES. NDLF has an opportunity to provide some models of behavior.

The behaviors I think worth considering are these:

  1. We need to understand ourselves as investors and act as a community of investors;

  2. Hard-nosed business decisions need to inform our individual selection decisions. For instance, we need to ask why we would chose to make available over our networks serials of low impact, even if the "deal" seems attractive.

  3. Correspondingly, we need to make common cause with the suppliers of academic information whose principles are consistent with ours, those who are more interested in scholarly communication first and in funding the scholarly communication processes essentially at cost. We need to reject payment and access schemes which disadvantage the university-centric view of information development and exchange as well as the university-centric view of long term investments in libraries - collections and services.

  4. We should encourage and even initiate marketplace corrections. Clearly our scholarly societies and university presses share our goals and principles. Yet we are all aware of the lessening of their production for all of us expressed as a percentage of all academic publications. We need to consider how to help them enhance their productivity and then proceed to implement such enhancements.

Distributed computing, larger and cheaper magnetic and optical memories, consistently good networks with constantly improving bandwidth, and the continued development of software tools and systems (and here I think of the browsers, systems such as the Infobus, digital publishing and navigation tools, as well as discovery and retrieval engines) each can and do contribute to the ability of NDLF libraries and others like them to preserve our core values, principles, and functions. The truth is that the same list of factors is as usable by those who have and would treat us as cash cows, on both the production and consumption sides of the business equations.

Let me turn now to some sanitized descriptions of business models and drivers for them we have experienced and contributed to through the operations of HighWire Press. HighWire as most of you know is an Internet publishing service bureau; we like to call ourselves co-publishers on the basis of our contributions to the Internet editions of the scientific journals we bring to the net. A list of the journals we have assisted with is available at Many of them are high-impact and high output journals. HighWire is an enterprise inside the Stanford University Libraries and is already self-sustaining as an enterprise. I should not reveal the identities of the scholarly societies whose business models I allude to next, but I believe that I can present some of the considerations without violating any confidences.

The business models for these journals is based on tradition and on each organization's publishing strategy. Here are the range of possibilities:

  1. Some boards want to move to an Internet-only publishing scheme. Their literature is growing fast. New modes of scholarly communication made possible by network-based publishing are proving to be important additional tools for research. They are making Internet-only subscriptions available at rates below the paper-based subscriptions, but not bundled with them. They depend upon institutional subscriptions for first copy costs and can see tremendous benefits to their subscribers in rates, but librarians are not leaping to the Internet editions and discarding duplicate paper subscriptions much less sole paper copies as fast as the boards would like. In the light of ongoing experiences with Internet publishing, the boards are continuing their long-term strategy, but altering their tactics and spreading out those tactics over time.

  2. Some boards need to continue the paper-based subscriptions to individuals even while improving and expanding the Internet editions of their journals. They will levy surcharges on readers who wish access to the Internet editions on top of the paper-based subscriptions. On the other hand, regional and national site licenses under some conditions are being developed in order to gain access to new marketplaces and as investments in the future of those marketplaces.

  3. Some boards do not imagine that their journals are of much interest beyond the societies' members, so will offer member-only access on the basis of generally increased dues and subscriptions for personal members. Some of these boards are experimenting with advertising in the Internet editions involving not just advertising panels, but also links from citations in the articles of commercial products to the Internet sites of the commercial providers.

  4. Some boards can afford to run their Internet editions as loss-leaders for a number of years and are apparently willing to do so as experiments to test how their members would like best to gain access to the disciplinary literature. They intend to develop business models and strategies later.

  5. Not one of the societies with which HighWire is working knew the demographics of their subscribers, whether personal or institutional, as well as they would have liked. We have invested quite heavily in research to give the boards better bases for thinking and acting. We think that this characteristic is quite common as well in the for-profit sector and we think also that such experiments as Tulip and Red Sage in the past as well as the current "deals" to provide access to whole lists of journals were and are intended in part to get better information on readers and their behaviors.

  6. All of our boards have members who believe as well as some who do not believe that Internet editions can pay for themselves. All see the benefits of these new modes of scholarly communication and some see new business opportunities among the new modes of communication. As a result of both the conservative and avant-garde thinking, we see new partnerships forming to develop new information products and services. How all of this will ultimately affect the marketplace AND the roles, principles, and functions of university libraries remains to be seen. We are all taking risks.

This recital reflects the experiences, some of them, of one NDLF member. Perhaps the same considerations will arise as the NDLF gets more involved in the production of collections of digital collections. At the moment we are considering what might be termed "the new retrospective conversion," the conversion of paper-based editions of texts and images to digitally based ones. We operate as subsidized units of universities, as have university presses. We expect and want to continue providing "free" information goods and services on the basis of subsidies. Yet we know that some business models will need evolve to account for the possibilities of very general, global access to the new digital collections. Some of us can conceive of business models which provide new capital for new conversions. Some of us can see return on such subsidized investments as to allow us to expand our digital collections more rapidly. Most of us have started conversion of paper-based information resources with subsidies from our operational budgets and with some outside funding. Few of us, I think, prefer to imagine settings at our universities or by our investment decisions on information in which large profits are amassed, stockholders and corporate officers get rich, and information monopolies result. Yet, it is precisely such settings which we observe and which cause us to worry.

Obviously I believe and Stanford believes that there are ways to continue making investments in information for teaching, learning, and research. I believe that my NDLF colleagues share to greater or lesser degrees this belief. Yet we are also self-conscious players in a great transition. While the earth is moving, we are unsure where our hundredth and thousandth steps will land. We would like to understand more about the economic and business models inherent in the DLI Projects to see whether there is common cause and shared values.

Thanks for your attention.

* Copyright 1996 by Michael A. Keller. This paper may be copied and redistributed on the Internet freely provided that the entire paper with this copyright notice is transmitted. Other rights reserved to the author.

Those wishing to contact M. A. Keller can do so via e-mail at [an error occurred while processing this directive]