Information
Technology Investments in Research Libraries
by Susan Rosenblatt
Research
libraries have been investing in information technology (IT) for more
than twenty-five years, beginning with the automation of repetitive
processes such as cataloging and circulation. The initial goals of automating
these functions were to reduce the rate of cost inflation, if not the
absolute costs, of these operations and to improve the timeliness and
accuracy of information. Over time, automation also enhanced direct
services to users. The online catalog, with its twenty-four-hour-per-day
availability, sophisticated search-and-retrieval capabilities, and ability
to provide information not included in card catalogs, is the prime example
of the corollary benefits that library users experienced through the
1980s. Today, library automation is mature, and the rapidly growing
availability of digital information -- including indexing and abstracting
services, scholarly journals, popular publications, quantitative data,
and digital copies of materials originally issued in print -- promises
to transform library services and the experience of library users.
Libraries are being transformed
not only by external pressures and opportunities but also by internal
developments and demands, specifically the creation of digital information
and services. During the 1990s, libraries began actively creating new
forms of online publications, facilitating scholarly communication,
establishing digital library standards, and assuming an increasingly
active role in instructional technology. As the twenty-first century
dawns, campuses face limitless opportunities to invest in IT but have
finite budgets. Moreover, competition for institutional resources is
intense. Thus, the choices to be made are crucial. How can the campus
ensure that scarce resources are invested wisely? How can it demonstrate
the value that library IT investments add to teaching and research?
How might planning for future IT investments in libraries build on the
assessment of past successes and failures, and how can rigorous outcome
measurements for future investments be developed? Although our understanding
of the transformational potential of the new technologies may not be
comprehensive, we know enough to discern that certain kinds of investments
may be more likely than others to ensure that library IT meets academic
needs.
The Introduction
of Information Technology to Research Libraries
In retrospect, libraries'
individual and collective decisions to initiate their automation efforts
with cataloging operations, thereby creating a foundation for online
access to information, appear brilliantly strategic. But in reality,
many of the significant benefits of these early efforts were unintended
and unforeseen. The information explosion and postwar influx of students
to higher education in the 1960s led to increased library acquisition
budgets and rising collection growth rates. These changes heightened
concern about cataloging costs and performance. The traditional, almost
handicraft means of cataloging production could not keep pace with the
new workloads; manual processes of card production and filing delayed
providing users with information about new acquisitions; and redundant
cataloging among libraries maintained high costs within libraries as
a system.
Systemic improvements to
cost and performance could be achieved only by reducing these redundant
efforts among libraries. Under the leadership of the Library of Congress,
the library community established standards for machine-readable cataloging
(MARC) so that records could be shared over campus networks. The MARC
standard enabled the creation of national bibliographic networks supporting
online shared cataloging and interlibrary lending. By the late 1970s,
virtually all research libraries were using one of the national bibliographic
utilities -- the Online Computer Library Center (OCLC) or the Research
Libraries Information Network (RLIN) -- for cataloging and interlibrary
borrowing.
During the 1980s, libraries
automated other local processing operations and established online catalogs.
Local automation moderated the costs of labor-intensive processes such
as acquisition, circulation, serials check-in, catalog maintenance,
and detailed inventory control. The online catalog improved services
by providing bibliographic access at all hours of the day or night wherever
there was access to the campus network. The retrospective conversion
of card catalogs -- the core legacy file of the research library --
gradually improved the comprehensiveness of both local online catalogs
and the bibliographic utilities. Through the 1980s, cost reductions
and service improvements resulting from automation grew modestly yet
steadily. Users began to rely on the online catalog, and libraries used
the bibliographic network to initiate shared collecting programs and
online support of interlibrary lending.
How did libraries' experience
with IT investments compare with that of other businesses and sectors?
Many writers have described the "productivity paradox," in which IT
investments do not appear to reduce costs. Michael Hammer suggests that
investments in IT and reengineering efforts often do not yield the intended
results because technology is used primarily to speed up traditional
processes, including their inefficiencies.1
In Hammer's view, significant productivity improvements cannot be completely
planned in advance, nor can they be implemented incrementally. Instead,
they result from discontinuous thinking: a paradigm shift in which the
business processes are reengineered from a cross-functional perspective.
Francois Bar and Michael Borrus postulate that the emergence and consolidation
of the network-based enterprise will lead to greater productivity, whereas
traditional automation processes did not.2
Like Hammer, these authors argue that initial IT investments have been
disappointing because they have largely automated existing tasks and
inefficiencies. Bar and Borrus believe that productivity can be meaningfully
improved only through the complete reorganization of operations to maximize
the advantages of a coherent network infrastructure connecting multifarious
computer-based processes. Similarly, in The Rise of the Network Society,
Manual Castells suggests that productivity improvements in the service
sector cannot markedly improve until technical innovation has permeated
the whole economy, including business, culture, and institutions.3
Libraries' strategies for
implementing IT from the 1960s through the 1980s appear similar to private-sector
investment strategies in industrial and business process automation,
with one important difference. Library investments were made not by
individual libraries in a spirit of competition but through a combination
of local and collaborative interinstitutional efforts. Thus, investments
in cataloging standards, bibliographic networks, and retrospective conversion
had "industrywide" benefits; each research library and, indirectly,
its clientele benefited from the investments of the others. Similarly,
libraries' IT investments built on government and higher education network
investments, without which the national synergies of library automation
could not have occurred.
Beginning in the 1990s,
evidence emerged of the possible larger-scale effects of diffusion of
IT among other participants in the library services and product chains.
Vendors of library materials collaborated with bibliographic networks
to deliver shelf-ready books, together with online catalog records,
eliminating many local operations and improving the timeliness of information
available to scholars and students. Document-delivery firms began to
take orders through the library network. Services linking journal vendors,
online banking, and library systems resulted in a seamless flow of information
-- and money -- among the various business entities and libraries. These
new kinds of integrated services allowed for a more rapid reengineering
of internal library operations and for the restructuring of workto take
advantage of opportunities for disintermediation (for example, avoiding
cataloging and book-processing operations altogether), reducing costs
significantly while improving quality and responsiveness.
Perhaps more significant
for the future is the emerging integration of various types of digital
information -- including catalogs, abstracting and indexing services,
online journals, numerical data files, digital archives, and instructional
tools -- into a seamless environment over the World Wide Web. No longer
is the catalog merely an inventory-control device, listing objects in
a library's collection; along with the Web, it is now a direct gateway
to the information and services that faculty and students need. No longer
is the library the sole repository of scholarly information; that responsibility
rests in a shared, distributed environment including government, private
firms, scholarly societies, various campus departments, university IT
services, and libraries. The late 1990s have provided a glimpse into
the transformational possibilities that the Internet holds for libraries,
learning, and scholarship, once technology has permeated all of the
major facets of the university's mission. Might this constitute evidence
of the discontinuous progress described by Hammer, Bar and Borrus, and
Castells? If so, does it promise both reductions in costs and improvements
in services?4
Assessing Outcomes
of Information Technology Investments
Castells described two other
factors (in addition to the diffusion of technological innovation, noted
above) that inhibit the development of productivity measures in the
service sector. These factors appear relevant to libraries. First, service
industries traditionally find it difficult to calculate productivity
statistically; thus the absence of observable productivity enhancements
may in part be a symptom of the inability to define relevant performance
measures. Second, productivity changes may not be easily measured because
of the broad scope of transformation under the impact of IT and related
organizational change.
Libraries, by contrast,
have gathered statistical data on orders placed, items cataloged, and
circulation transactions in both the manual and the online environments;
as a result, productivity improvements as measured by workload transactions
have been easy to assess. Workload outputs such as these are significant
measures for libraries and thus contradict Castells's general observation
about the service sector's inability to measure outputs. Measuring outcomes
in library services supporting instruction or research -- such as timeliness,
accuracy, or relevance of information retrieval, support for learning,
or contributions to scholars' efficiency -- is more difficult. Yet such
measurements are essential if libraries are to succeed both in garnering
funds to maintain information resources and services and in documenting
the strategic value of new IT investments.
Ultimately, libraries' difficulty
in defining and measuring service outcomes -- congruent with Castells's
observation of the service sector's difficulties more generally -- may
be more important than the ability to count workload outputs accurately.
Qualitative measures of library effectiveness must correlate with the
mission of the university itself, an institution that also has difficulty
defining and measuring outcomes. Individuals on many campuses are beginning
to question the need for a library in the networked environment. Can
the search engine replace the librarian? Why do we need large warehouses
of books if everything can be digitized? Ultimately, the only effective
response to these questions is a measurable connection between the librarian
and the library (both digital and print) on the one hand and the university's
missions of teaching, learning, creation of new knowledge, and public
service on the other.
Cost Benefits
Many campus administrators
find both IT and the library to be budgetary "black holes." They may
not easily discern exactly how investments in either benefit the institution
as awhole. Expenses for libraries and IT are local, whereas many of
the benefits -- both operations costs savings and service improvements
-- are deferred and often accrue outside of the local institution. Local
IT investments may leverage benefits from the investments of others,
but these benefits may take time to materialize. In this world of local
costs and shared benefits, individual institutions may be tempted to
try to minimize local costs yet capture systemic benefits resulting
from the investments of other institutions. However, if the majority
of institutions employ such a strategy, both local and shared benefits
evaporate. The challenge is to sustain a system of networked information
across institutions, measuring local and systemic outcomes more precisely
to see that investments are effective.
For example, local investments
in automating card catalogs reduced the costs of later interlibrary
lending, a benefit primarily to libraries other than the library that
had automated its catalog. Likewise, today's investments in digitizing
unique special collections may benefit scholars worldwide well into
the future yet may never result in measurable reduced operations costs
locally.
The JSTOR project offers
another example.5 Through
initial investments in digital imaging of core journals, the project
hopes to reduce eventual capital costs for library buildings. Improved
access to journals is a secondary -- but not unimportant -- goal, as
are potential operations costs reductions. After an initial foundation
investment in digitization, capital costs for the project are borne
through one-time institutional investments in the project; small annual
subscription fees sustain regular operations. Thus, individual libraries
make rather modest capital investments in digital scholarly content
in the expectation that many universities will avoid significant capital
costs later, that libraries will eventually sustain modest operations
costs reductions (for example, in circulation, stack maintenance, interlibrary
lending, and preservation), and that scholars will benefit from improved
access to scholarly journals.
Service Benefits
Estimating the cost savings
resulting from IT investments is often possible, but measuring service
improvements attributable to these investments has been much more difficult.
Not only are measurements correlating with the academic mission difficult
to devise, but the nature of the networked environment creates qualitative
changes in services and in the expectations of users. Thus, what we
know how to measure -- for example, circulation or reference transactions
-- may not be relevant output measures for the digital library. Traditional
outputs may decline as digital information and services allow users
to become less reliant on print collections and services delivered in
person within a physical library. IT investments may thereby have the
paradoxical effect of apparently reducing productivity. This transformational
potential of the digital library and related networked services is simultaneously
exciting and vexing. How do we measure the positive effects of IT and
balance them with the apparent decline in productivity in traditional
systems?
Futurists and creative thinkers
articulate exciting visions of the fully networked library's benefits
to teaching, scholarship, and the public-service mission of universities.
Even though many of these ideas will never be fully realized, others,
completely unimagined, will come to fruition. As this transformed scholarly
information environment develops, those who depend on the traditional
library of print and place may feel threatened by the prospect that
the university might allocate scarce resources to technological wizardry
rather than to core library resources and services supporting research
and teaching. The library may be caught in the middle of an acrimonious
debate: accused by some of Ludditry, unresponsive to new needs and unwilling
to invest in technology; chastised by others for diverting resources
from collections and traditional reference services.The challenge is
formidable -- to make wise IT investments so that the networked library
meets the needs of scholars and students who use it and to maintain
the print collection and existing services so that the traditional library
meets the needs of its users.
New Service
Models
No one can accurately predict
how networked library resources and services will function in the future;
nevertheless, today's embryonic digital library provides insights into
the digital future. Many traditional library services -- such as selection
and organization of information resources, access to collections, reference
help, and instructional services -- continue in the digital library,
but in new ways. At the same time, new demands from students and scholars,
as well as new opportunities offered by technology, require new kinds
of services. For example, the Web has begun to change librarians' roles
and to create additional demands on their time. As librarians become
more active in developing Web-based resources and services, how might
the impact of these new services be evaluated? To some extent, they
assume the character of traditional library functions (selection and
organization of information, advisory services, and instruction); yet
they also constitute new forms of publication and new tools for teaching
and research. If time spent creating and maintaining library Web sites
results in a reduction in time spent fulfilling traditional responsibilities,
is the librarian less productive or more productive? Which investment
of time provides the greater benefit to the university community? Is
assistance to today's scholars sacrificed in order to create better
services for future scholars?
The productivity paradox
thus emerges in the library. The Web serves as a useful perspective
on a new paradigm because it challenges librarians and staff to begin
to rethink their roles in the networked environment. Will selection
and organization of digital resources have the same importance as they
did for print materials? Must selection and organization activities
remain local, as they generally are for print collections? Are there
new models for the selection and organization of digital resources,
for example collaborative endeavors on a national or even international
basis, organized by the disciplines themselves? Are the computer scientists
right in their assertion that the search engine will replace the librarian,
so that the organization of information will cease to be a professional
function?
Bonnie A. Nardi and Vicki
O'Day ask, "Is it actually possible that software will replace the librarian?"
They approach this question from an analysis of the practices of librarians
(the expert domain) rather than from the perspective of the user's needs.
They find that effective intelligent agents can be designed based on
what reference librarians do but that software cannot facilitate the
subtle progress in researchers' thinking that results from personal
interaction with a reference librarian. The librarian's help is interactive
and dynamic as well as personal. Yet even though the intelligent agent
may not be able to provide as much help as interaction with a reference
librarian might, the digital library user may choose not to interact
with a human mediator. Nardi and O'Day conclude that software agents
must therefore be designed to work effectively at the routine processes
that they do well and must also encourage human interaction to enrich
the information-gathering process: "Reference librarians and software
agents collaborate to perform different but related tasks to serve clients."
They describe librarians' roles as helping users learn how to find and
use technology-based information resources and tools, following through
when agents fail, and creating the vital human-to-human interactions
that cannot be replicated by software. The authors predict, "More human
agents will be needed to help an ever-growing population of computer
users deploy expanding computational resources effectively."6
Bernie Sloan finds a striking
lack of reference to service or librarians in the literature on digital
libraries.7 Instead, according
to Sloan, the emphasis is on universal access to distributed information
resources and the enabling linking technology. Drawing on the literature
describing interactions among librarians and information-seekers in
the traditional library, Sloan finds collaboration to be the most significant
factor in successfully finding information. If Nardi, O'Day, and Sloan
are correct, there is rich potential for librarians to provide services
within the digital library. The question then is: What is the best balance
of digital library investments in information resources, people, and
technology?
Library Users'
Changing Expectations
Users' expectations of the
library are rapidly changing, although users themselves may not realize
it. Researchers expect books and print journals to be available as quickly
as are online catalog records and online journals. Studies of scholars'
information-seeking behavior demonstrate that the typical search for
information begins with those materials that are most easily accessible,
for example in the home or office, or through a conversation with a
colleague. Only after readily available, informal alternatives have
been exploited will researchers make a trip to the library for information.
As IT puts increasing percentages of relevant information resources
on the scholar's desk rather than on library shelves, reliance on print
collections and services is increasingly less compelling, even for faculty.
Thus, demand for high-quality print materials housed in the library
may decline once online information is adequate for most research needs.
Moreover, the experience
of the digital library user will influence expectations for other library
services. Many will prefer self-service over mediated services. For
example, instead of standing in line at a circulation desk, borrowers
expect to use a circulation kiosk and renew library books online; they
prefer to submit document-delivery requests online directly to suppliers
rather than visit the interlibrary lending service. Similarly, when
a critical mass of reference tools is available online, digital library
users will expect to ask questions over the Internet rather than go
to the reference desk, even during hours that the library is open.
Users' changing expectations
for information services have profound implications for the future of
librarians and the print library. The quality of traditional library
services must be maintained while new digital library services are developed;
yet budgets are not likely to be increased to allow dual systems to
operate. The reallocation of financial and personnel resources from
print-based services to digital information services is inevitable,
but how much should be reallocated and at what pace? Developing disintermediated
services for the traditional library is one strategy for reallocating
staff over the long term, since these innovations may both improve the
services and reduce the costs of many routine operations.
Shifting Boundaries
New digital library service
paradigms will span the boundaries of the roles that libraries and other
participants in the scholarly information network now play.8
They offer the potential for the redesigned, or transformed, services
envisioned by Hammer, Bar and Borrus, and Castells. Instructional technology
programs, digital library resources, bibliographic instruction, university
presses, and the bookstore may be integrated into new services. Distance
education will require a full spectrum of technology-based information
resources and services. New services will require that traditional organizational
structures form new collaborations or be rethought, reorganized, or
bridged.
Libraries are becoming actively
involved in facilitating scholarly communication -- collaborating with
publishers, scholarly societies, and other agents of scholarly publishing.
For example, Stanford University Library founded Highwire Press to publish
high-quality online journals in collaboration with scholarly and scientific
societies. The Mellon Foundation has funded a series of library/university
press projects to explore how collaboration might change modes of scholarly
communication and reduce the costs to presses and libraries alike. The
Association of Research Libraries' (ARL) Scholarly Publishing and Academic
Resources Coalition (SPARC) represents a consortial effort to work with
scholarly societies to publish low-cost, high-quality online journals.
And many other libraries are bringing scholarly projects into library
spaces, where they can be functionally, if not organizationally, merged
with information services.
The digital library comprises
more kinds of information resources than does the traditional library.
It can easily include scholarly materials, popular Web sites, museum
objects, multimedia presentations, quantitative data, and working papers.
It incorporates tools that enable digital information to be more easily
found, understood, and used. It can include extraction and analysis
tools for quantitative data, visualization tools, and authoring and
editing tools. It needs tutorials, guides, self-paced instruction, and
collaborative spaces.
To create and integrate
the wide variety of complex information resources, services, and tools
that digital library users expect, library staff must develop new expertise
and build strategic alliances and collaborative relationships with complementary
partners inside and outside the university.
Investments
for the Future
The rapid pace of technological
change and the difficulty of measuring outcomes create acute uncertainty
about how to make wise IT investments. Administrators, scholars, and
librarians have a common interest in ensuring that investments result
in better access to information and improved services for users. Reviewing
the successes of the past suggests that certain types of investments
may be more likely than others to reduce costs and improve services.
Experimental investments often are not robust enough to support dependable
services, but other kinds of investments, although not paradigm-shattering,
can successfully improve services in the near term and form a foundation
for subsequent changes.
Investments in infrastructure
improve users' present experience and provide capacity for future service
improvements. In addition to the obvious need to maintain the campus
network, at least three other infrastructure investments can help to
ensure a sustainable digital library. First, libraries have expertise
in and understanding of the need for standards and best practices.
Just as the MARC standard enables sharing of bibliographic records,
other standards and best practices can ensure the exchange of digital
objects and tools. For example, the Encoded Archival Description, maintained
by the Society of American Archivists, guides the creation of digital
versions of archival and special collections; the Text Encoding Initiative
guides the creation of digital scholarly documents; and the Making of
America II project defines certain kinds of digital objects. A distributed
digital library requires community agreement on best practices so that
access to distributed repositories appears seamless to the user and
so that digital resources can be archived. Second, the development of
interfaces among campus systems and library systems supports
an exchange of administrative information, eliminating duplicate data
entry and paper handling. And third, certain categories of digital
information constitute infrastructure because scholars will accept
the digital library only after it contains a critical mass of reputable
information. Just as the first step toward the digital library was the
online catalog, indexes to print and digital scholarly information,
encyclopedias, dictionaries, and other reference tools are required
for users to find the information they need. Scholarly journals, reports
and working papers from scholarly institutes, and high-quality documents
that would not be published in print because of small sales help to
create the depth that scholars expect from a research library. Incorporating
high-quality digital copies of heavily used retrospective r esources
and core journals, such as is being done by the JSTOR project, will
create the rich base of historical information required to support research
and may eventually reduce operations costs related to managing print
collections.
Investments in the disintermediation
of routine services enable library users to be more independent;
such investments also improve service quality and reduce staff time
spent providing routine assistance. Examples include self-service circulation
kiosks and online systems that enable patrons to renew books and perform
other library transactions remotely. Electronic course reserves can
eliminate the costly processing of print materials, reduce lines at
circulation desks, and allow more simultaneous uses than do paper reserves;
they can easily be created by faculty or teaching assistants rather
than by library staff.
Investing staff time in
focused collaboration between libraries and strategic partners
to reduce duplication of effort, create a broader knowledge base from
which to provide digital services, and develop best practices can have
long-term payoffs. Investments in collaboration on such issues as digital
archiving, improved systems for document delivery, standards for instructional
resources, interinstitutional licensing agreements, scholarly communication,
conversion of heavily used retrospective resources, and interoperable
distributed digital collections and services are all essential. Including
vendors, technology firms, scholarly societies, publishers, instructional
technologists, and many other providers, experts, and stakeholders in
the design of the digital library will strengthen it.
The digital library will
not materialize without experimentation. Institutional and extramural
investments in pilot projects and research and demonstration enrich
our knowledge about users' requirements, the capabilities of emerging
technologies, the effectiveness of online tools, and the changing nature
of scholarly communication in a networked environment. Experimentation
provides valuable new knowledge, but institutions cannot expect that
experimental technologies and methods will necessarily evolve into production
services. Some experiments will fail, many will take much longer to
reach fruition than expected, and a few may achieve rapid success. Thus,
experimentation must be seen as a learning experience and must not be
expected to result in production services.
People provide the values,
energy, commitment, and knowledge to ensure that the digital library
meets research and teaching needs. Investments in hardware, software,
network, and digital content have limited effectiveness if not accompanied
by strategic investments in staff. By redirecting staff focus
from routine operations to complex problem solving and a deeper understanding
of users' needs, the library gains the equivalent of a recurring capital
investment fund. The digital library is not just information -- it is
the services and tools that support learning and research, and it requires
the expertise of specialists across many organizational units in the
university to create these services and tools. Thus, the digital library
requires greater integration and involvement of librarians and information
technologists in academic life, particularly teaching and learning.
The new information professional must be a partner in creating new opportunities
for students to learn, extending university services to fulfill the
public-service mission, supporting research, and facilitating scholarly
communication. Through their interactions with colleagues, partners,
faculty, and students, information professionals will help to maximize
the digital library's benefits for the university as a whole.
The digital library is a
work in progress, not a finished project; thus its potential capabilities
and impact remain infinite, as do the possible investments that might
be made to create it. Nevertheless, in the face of technological change
and unlimited opportunities to expend fiscal resources, there are certain
investments whose benefits to the educational enterprise can be relatively
safely predicted. The best investment strategy is one based primarily
on scholars' and students' needs yet also focused on the potential to
moderate recurring operations costs. Investments that can create systemic
benefits and savings among libraries will likely result in greater cumulative
effects than will a series of purely local IT investments.
