Library Issues

Editor: Richard M. Dougherty, University of Michigan.
Contributing Editors: Steve Marquardt, South Dakota State University; William Miller, Florida Atlantic University; Sarah Pritchard, Smith College; Mark Sandler, University of Michigan

Vol. 16, No. 6 July 1996

The “Productivity Paradox” in the Academic Library

by Steve Marquardt

Writing in his monthly "Technology" column of Governing magazine, Jerry Mechling recently lauded the productivity of technology by observing that

Computing today is roughly 32 times (that's 3,200 percent) more productive than it was just 10 years ago, and gives every indication of growing another 32-fold more productive over the next decade as well. That big a change in managing information is beginning to generate big changes in organizing and managing work.1

Higher education has invested heavily in computing, especially in library applications. Given this investment and the increasing productivity of computers, one would expect that the library and its personnel would be substantially more productive than ten or twenty years ago, and by multiple orders of magnitude. And yet it seems that few productivity gains have been dramatic enough to allow significant reductions in total library personnel. Why is this the case?

The answers to this question are relevant not only to library services and budgets, but also to the current discussions of how to exploit information or educational technologies in order to improve the "academic productivity"2 of faculty and the "learning productivity" of students.3

Automation of Library Functions: Phase One

Productivity gains sufficient to allow the reallocation of personnel did occur in the early phases of library automation, when the most routine and simple operations were computerized.

Circulation. The greatest and most obvious gains in library productivity occurred when inventory control was automated. When computers replaced manual filing of check-out cards, calculation of fines, and production of overdue notices, hours of labor were released for other purposes. This was the least expensive labor in the library, however, as most of this work was done by student workers.

Cataloging. In the 1970s, manual card production, using information taken from paper sources, was accelerated by outsourcing to electronic bibliographic utilities. In the 1980s, online catalogs replaced catalog cards and cabinets in most academic libraries. These changes reduced the number of personnel required to catalog the materials added to library collections.

Acquisitions. Automated systems tended to require additional time and effort in book and serials purchasing, receipt, and payment. These manual systems, unlike cataloging systems, were not based upon national and professional standard rules and data structures, and thus were relatively simple, albeit somewhat labor intensive. In contrast, automated systems provided by commercial vendors tended to require new data elements that could be used to produce the budget, subject, and efficiency data desired by library managers but laborious to derive from manual purchasing systems. Here there were at best very small savings in personnel or productivity.

Automation of Library Functions: Phase Two

More recent innovations have expanded the power and range of information retrieval. In this richer, more varied and complex environment, productivity has yielded a greater number of retrieved products through more sophisticated search techniques. Automation of the complex, however, has not presented the number of opportunities for personnel reallocation as did automation of the routine.

Reference. Online public access catalogs (OPACs) have created additional work for public service librarians. The card catalog, however limited in its author, title, and subject access points, was a simple passive system requiring the user to exercise only the fingers. The online catalog offers searches by call number and publisher, plus Boolean logic and the limitation of searches by language or format. The electronic catalog's precise and varied commands require user training and point-of-use documentation.

The desire to make library use as convenient and easy as possible for users inspires further refinements by library staff. Librarians add value to an increasing array of information sources and possibilities. Users benefit by having greater power and speed in deriving information. But rather than save staff time, OPACs, indexes, and databases require that more time and effort be spent selecting, obtaining, installing, training, and instructing in their use.

Advent of New Services and Systems. Increased demand for interlibrary loans have resulted from the arrival of rich and powerful CD-ROM bibliographic databases. Now many of these same databases are accessible through site licenses over the Internet. Time and effort are required to analyze the pros and cons of user access, and to assess the purchase, license, and maintenance costs for each of four access alternatives: local mainframes, stand-alone CD-ROMs, local area networks, or the Internet. Library staff must be trained and must train users in the operation of these new resources. The ability to query other library catalogs over the Internet, through gophers or the World Wide Web, further increases both the need for staff and user training, and the demand for interlibrary loans. Automated file transfer and the explosion of Web sites further add to the workloads at public service points.

"More recent innovations have expanded the power
and range of information retrieval . . . Automation
of the complex, however, has not presented
the number of opportunities for personnel reallocation
as did automation of the routine."

Library Systems Architecture. The early turnkey versions of integrated online systems could be compared to the home entertainment system all contained within one cabinet. These early systems were relatively self-contained packages of circulation, catalog, and acquisitions functions. The newer, more fully featured systems have rapidly evolved into configurations of varied components. They contain hardware from one source, software from another, and tapes of data loaded into the catalog database and accessed through a screen that often links to other nearby libraries, to a local free-net, to the World Wide Web, and to the local campus information system. This transition from automating the simple to automating the complex has frustrated hopes for increasing productivity.

Technology versus Productivity?

Sociologist Paul Attewell has identified several factors that contribute to this productivity paradox, most of which can be applied to the academic library.4 Attewell contends that word processing has created a shift to slower channels of communication. Although word processing has accelerated the production of words at a keyboard, it has also shifted communication away from the more effective modalities of speech, gesture, drawing, and demonstration. Under the influence of word processing, human communication has become more formal and wordy than face-to-face conversation. E-mail is preferred to conversations because workers prefer not to intrude upon and interrupt their colleagues.

Workers fall prey to the temptation to improve the quality of their messages and their words, formatting and embellishing the most mundane of communications, catalog records, spreadsheets, and presentation graphics. Attewell cites a study in which "writers using word processors made nearly five times as many modifications and corrections compared to persons carrying out the same task by handwriting."5 The same tendency may well be true of catalogers, or of reference librarians who perform just one or two more searches in just one or two more powerful and databases.

The Power of Software. The increasing power of the software itself undermines productivity. Improvements in product quality undermine quantity or output, and thus reduce productivity. One suspects that Jerry Mechling's 32-fold increase in productivity is measured in the software's features, i.e., in its capacity to store and process data and information. The emphasis on more features, however, creates "new learning demands [that] are repeatedly thrust onto employees whose major responsibility is to do work, not to learn about information technology."6

Because academic librarians need to learn about information technology, the rapidly evolving software creates a double burden in their information work and also in the office, library, and Internet automation that supports that work. As their work tools become more numerous and fully featured, librarians experience the truth that computerization pays off in productivity only if accompanied by a drastic simplification of work processes and procedures, an impossible simplification in the increasingly varied world of the academic library.7

"We need to keep in mind
the increased productivity of those who use libraries,
even while the number of workers
. . . is not being dramatically reduced in
. . . proportion to the productivity of
their machines."

Computers are highly productive generators of data and information. But the people who must analyze the data, read the information, and respond to demands for more data or further revisions of prose do not necessarily become more productive. Although computers have generated more output, much of it is unproductive administrative paperwork. Computers create more work, but workers do not necessarily become more productive. In libraries, extra paperwork poses a danger of detracting from the amount of face-to-face service and research support performed by librarians and staff.

Lower level jobs have been replaced by automation. These include the tasks of computing overdue fines; generating fine notices, statistical reports, and purchase orders; and loading more data into more user friendly databases. However, the gains realized in automating these tasks either have resulted in increased numbers of library services, or have been reinvested in additional technical support personnel who operate the automated systems, thus adding to the administrative overhead.

More Data for Administrators. Administrative hunger for the data produced by computers also undermines productivity. The very fact that computers can generate tables of numbers and graphs of trends leads managers to request increased numbers and frequencies of such reports. The tables and graphs are often more relevant to managerial concerns than to operational issues, more relevant to control than to process. Operational decisions can be delayed until more management data is obtained. W. Edwards Deming described "the fallacy of management by numbers" as follows:

To manage one must lead. To lead [the manager] must understand the work that he and his people are responsible for . . . It is easier for an incoming manager to short-circuit his need for learning and his responsibilities, and instead to focus on the far end, to manage the outcome get reports on quality, on failures, proportion defective, inventory, sales, people. Focus on outcome is not an effective way to improve a process or an activity . . . management by numerical goal is an attempt to manage without knowledge of what to do, and in fact is usually management by fear.8

These criticisms speak to internal processes, but what about the services provided by libraries?

Effects on Services. Computers have resulted in a dramatic increase in the number of services offered by libraries. Faster retrieval of reference data and index citations, knowledge of cataloged holdings of distant libraries, and full-text online retrieval of publications are a few of the new services that benefit users who are able to use library resources and services more efficiently and productively. Simply put, users now retrieve more information in less time.

Library staff, however, must cope with the need for more technical support, with steeper learning curves, and with the added administrative overhead. A further drag on productivity is the need to maintain older technologies. For example, librarians are frequently reluctant to cancel paper indexes and other traditional materials, and are often as reluctant to dispense with traditional interlibrary loan and document delivery processes in favor of exclusive reliance upon automated alternatives.

The library's dilemma is analogous to Attewell's example of the electronic teller machine that has resulted in customers making more frequent withdrawals of smaller sums, but has not eliminated the need for tellers to service the walk-in traffic. Libraries can mount their indexes and other reference resources on a campus network, and even extend access beyond the campus and become involved in support for distance education, but librarians must continue to serve students who enter the library. And librarians are understandably reluctant to abandon the buy-once-read-many-times paper or CD-ROM resources in favor of "pay-per-view" fees for resources under the exclusive and permanent control of publishers. Therefore, in many cases, both the traditional and electronic paradigms are supported, enhancing services but eroding productivity.

What is To Be Done?

There is no turning back to the manual age. Rather, the decisions to be made now include:

  • which processes or services to automate,
  • what form of automation to implement,
  • how the implementation should be managed, and
  • what should be expected in terms of library productivity as well as user services.

Administrators must have realistic expectations of automation and its effect on the productivity of the library. Before investing in automation, they should consult the experience of others, read the product reviews, and think through the changes in processes both automated and human with special attention to the implications for workers.

Everyone in the library and the university needs to recognize that automation is a fundamental and ongoing technological revolution that has just begun. In the words of a computing facilities manager at M.I.T., "complaining about the lack of productivity gains at this point is analogous to complaining about your 10-year-old's failure to earn a good living." The productivity benefits and economic and social restructuring driven by movable type, the telephone, the automobile, and television all took more than a decade to realize.9 Online public access catalogs have been in the average academic library for not more than ten years. The presence of computers on the majority of staff desktops in the average library can be no more than ten years old. Local area networks have been in use at best for only five years in most libraries. Telecommunications and network applications (notably gophers, the World Wide Web, and distance education) have exploded over the past five years. All of these varied systems, services, and devices have yet to settle into a strategic synergy that can be said to have increased the productivity of library and information workers, as traditionally measured.

Looking at the Big Picture. When seeking greater productivity, managers must look not merely to the tools, but also to overarching systems, strategies, and purposes and to people. As Peter Drucker recently observed, executives now require an information system linked to strategy rather than individual information tools used primarily to calculate and record past events.10 In the library, this means attention to the mission, strategy, and process. Examples include the outsourcing of some functions and fundamental changes in other processes, such as allowing users to make direct interlibrary loan requests and receive direct shipment of loans from other libraries. Above all, we need to keep in mind the increased productivity of those who use libraries, even while the number of workers within libraries is not being dramatically reduced in direct inverse proportion to the productivity of their machines.

Libraries are making significant contributions to increasing "academic productivity" or "learning productivity" that is sought in the instructional areas. The automation of libraries and information retrieval is one manifestation of the "mass customization" of information technologies that accommodate different learning styles, goals, and schedules.11 The library's direct instructional efforts, incorporating computer technologies in the area of information literacy and critical thinking, can contribute to increasing the "learning productivity" of students. By combining its traditional user-oriented and instructional approaches with computer technology, the library can assist the instructional enterprise to achieve greater productivity.

Total costs in libraries continue to rise, but it must be recognized that automation has reduced at least the rate of that rise in labor intensive areas such as cataloging and purchasing. Library cost pressures are also driven by the requirement that the library continue to support information acquisition and use in paper and microformats, as well as expand into digital and online formats. Despite the rise in costs, there has been a greater rise in the number of library and information services delivered to students and faculty. If total productivity equals total benefits divided by total costs, then the benefits to users faster and more accurate catalogs and indexes, faster full-text and document delivery, a wider range of data to explore are forms of productivity not calculable in the traditional industrial measures of work produced per worker.


1Jerry Mechling, "Don't Hang Back on Technology," Governing 8 (August 1995): 82.

2William F. Massy and Robert Zemsky, "Information Technology and Academic Productivity," Educom Review 31 (January/February 1996): 12-14.

3D. Bruce Johnstone, "Learning Productivity: A New Imperative for American Higher Education," Studies in Public Higher Education 3 (April 1993): pp. 1-32.

4Paul A. Attewell, "The Productivity Paradox," Chronicle of Higher Education (March 15, 1996): A56; Paul A. Attewell, "Information Technology and the Productivity Paradox," unpublished typescript (New York: Graduate Center of the City of New York, 1992); and Douglas H. Harris, ed. Organizational Linkages: Understanding the Productivity Paradox (Washington, D.C.: National Academy Press, 1994).

5Attewell, "Information Technology and the Productivity Paradox," p. 18.

6Ibid., p. 20.

7Ibid., p. 21.

8W. Edwards Deming, Out of the Crisis (Cambridge: Massachusetts Institute of Technology Center for Advanced Engineering Study, 1986), p. 76.

9Robert Bruen, letter to the editor, Chronicle of Higher Education (April 12, 1996): B6.

10Peter F. Drucker, "The Information Executives Truly Need," Harvard Business Review 73 (January/February 1995): 54-62.

11Massy and Zemsky, "Information Technology and Academic Productivity," p. 13.

Mountainside Publishing, Inc.

Library Issues: Briefings for Faculty and Administrators (ISSN 0734-3035) is published bimonthly beginning September 1980 by Mountainside Publishing Co., Inc., 321 S. Main St., #300, Ann Arbor, MI 48104; (313) 662-3925. Library Issues, Vol. 16, no. 6. ©1996 by Mountainside Publishing Co., Inc. Subscriptions: $40/one year; $68/two years. Additional subscriptions to same address $15 each/year. Address all correspondence to Library Issues, P.O. Box 8330, Ann Arbor, MI 48107. (Fax: 313-662-4450; E-mail: Subscribers have permission to photocopy articles free of charge for distribution on their own campus.

Library Issues Home Page

last modified: