Changing Times

Assessment of Continuing Resources Due to Budget Cuts Necessitated by COVID-19

Jaclyn Lee Parrott

Jaclyn Lee Parrott (jackileeparrott@gmail.com) is an Assistant Professor and Collection Management Librarian in the Collection Services Unit at Eastern Washington University Libraries & Learning Commons in Cheney, WA.

Manuscript submitted May 14, 2021; returned to author for July 1, 2021; revised manuscript submitted August 6, 2021; accepted for publication December 29, 2021.

Due to COVID-19, a purchasing freeze was implemented, and budget cuts mandated by Eastern Washington University. This necessitated a review of all the library’s continuing resources with a short turnaround time for decision-making due to subscription renewal deadlines. Considering quantitative and qualitative factors, a collaborative effort from internal stakeholders ensued. A tiered collection assessment decision making approach was designed and implemented. Cancellations ensued, and 25 percent of the collections budget was cut. This project involved a systematic review of databases, individual journal subscriptions, and print standing orders. A project of this scale could not have been as efficient and effective without the cooperative effort between those in collection services, public services, faculty, and administration.

Academic library budgets at public universities are often cut or remain flat during normal operations, and depend on many factors, including enrollment, endowment contributions, or state government support. Scholarly information costs continue to rise while library expenditures are seemingly under constant scrutiny.1 When the unexpected global COVID-19 pandemic occurred, it placed more pressure on libraries’ funding for resources and services. The pandemic adversely affected public universities’ budgets, including that of Eastern Washington University (EWU). Librarians faced the undesirable decision to cancel continuing resources to save the university funding.

This paper discusses how EWU stakeholders collaborated during the COVID-19 pandemic to reduce their library’s collection budget, which involved a systematic review of databases, individual journal subscriptions, and print standing orders. Librarians designed a tiered collection assessment approach tailored to the university setting, considered quantitative and qualitative factors, and cancelled subscriptions based on the methodology employed. A project of this scale could not have been as efficient and effective without the cooperative effort between several collaborators who determined what continuing resource subscriptions would remain financially sustainable during unprecedented times. This paper addresses the library’s context, how other libraries have evaluated their collections, how EWU assessed theirs with a tiered ranking approach that relied on quantitative and qualitative factors, the results they achieved, what could have been done differently, and what may be done in the future. Other libraries may be interested in adopting a similar approach and model for making sustainable budget cuts to continuing resources.

Background

EWU Libraries and Learning Commons support a regional public university with approximately 12,000 students and 500 faculty members. It offers a plethora of undergraduate and graduate degree programs. The main library provides access to approximately 1.1 million physical items, 515,000 e-books, 150,000 e-journals, and 239 databases (299 before recent cuts). In May 2021, the library employs thirty-one individuals, including one library dean, one library faculty chair, and thirteen tenured or tenure-track faculty librarians. There are ten reference and instruction librarians, one collection management librarian, one metadata librarian, and one discovery services and systems librarian. All librarians serve as subject liaisons.

Subject liaison librarians serve the university’s various programs and communicate regularly with teaching faculty. They perform collection development and provide instruction in the subject areas that they represent. Three librarians and four staff members make up the Collection Services unit, comprised of acquisitions, cataloging, discovery services, electronic resources, and other technical services functions. The librarians in this unit include a collection management librarian, metadata librarian, and discovery services and systems librarian. Until recently, there was also a collection maintenance librarian, but this position was first frozen and then cut permanently. The staff consists of four library and archives paraprofessionals (LAPs) and one library and archives paraprofessional (LAP).

When the COVID-19 pandemic began to spread in the United States, it became apparent that it would be some time before Washington’s medium-sized, regional public university could safely resume on campus classes. Therefore, a campus-wide teleworking plan for employees was implemented. Executive administration proactively initiated purchasing and hiring freezes when classes switched to online. The library returned year-end money that had been reserved for emerging resources to the university. Severe budget cuts were mandated before one fiscal year concluded and the next began. State budget cuts ensued, while enrollment continued to drop. Rather than cutting more personnel and placing an added burden on existing library faculty and staff, the collections budget was targeted.

At the author’s university, changes continue to occur. In July 2021, the university’s seven colleges were restructured into four. EWU Libraries and Learning Commons, which functioned as a stand-alone college with its own library dean and budget, became the School of Libraries organized under the newly formed College of Professional Programs (CPP). The incoming CPP dean had a background in psychology and would oversee six other schools in addition to the library. These include the School of Accounting, School of Business, School of Education, School of Psychology, School of Military Science and School of Social Work. The workforce at the top levels of the university and the library also changed. The provost resigned, and the interim provost became interim president after the previous president resigned due to a vote of no confidence from the teaching and library faculty. In December 2021, the new provost stepped down and the CPP dean became provost. The CPP associate dean became dean. Modifications in the top of library leadership continue to occur. In the past two years, the library has had three deans due to a resignation, death, and one serving in the interim. Now the library has a director instead of a dean and faculty chair who reports to the CPP dean.

Labor shortages in library staff resulted from the same reasons as the transition in library leadership (resignations, retirements, and death). Due to circumstances, the library lost twelve employees. Only one of the staff members was replaced, and no new librarians have been hired since these changes transpired. With permanent reductions of university staffing, these vacancies will most likely not be filled due to the hiring freeze implemented during COVID-19 related budget cuts. Subject liaisons now represent more programs outside their field of expertise because of the hiring freeze and reduction of the library labor force. In the new fiscal year, the library lost three more employees. Personnel changes have impacted library personnel over time (see figure 1). In 2018, staff and faculty thrived, in current times they are surviving, but the future looks bleak.

With already sparse personnel resources, the library’s collections fund budget faced needed reductions. As seen in figure 2, the collections budget consists of state funded operations money (the university is currently 50 percent state supported), endowments, replacements, and distance learning funds. The endowment fund includes various grant foundation monies with stipulations regarding how the donated money can be spent. The replacements fund is used to be replace lost or damaged books. The distance learning fund pays for the library’s streaming media and shared consortial e-books. The main collections fund is used for subscriptions, one-time purchases, and maintenance fees. It also includes service charges, shipping, tax, or bindery charges. This collections fund became the target of the cuts.

The collections fund budget had remained fairly stable in the past, but cuts were necessary before the pandemic. Fiscal years 2015 and 2016 saw slight increases in the collections fund budget, fiscal year 2017 brought a 9 percent decrease, fiscal years 2018 and 2019 remained flat, and fiscal year 2020 realized a cut of 1.4 percent. In the fiscal year 2021, the library budget was cut by 6.8 percent. This resulted in a cut of $89,000 of continuing resources from the collections fund budget and returning $31,000 to the university that would normally have been used for one-time purchases. In fiscal year 2022, the collections fund budget faces a permanent reduction of $300,000, which is another 25 percent cut.

The library previously had a set budget amount allotted for collections. The collection management librarian and collection maintenance librarian managed all subscriptions and monies associated with the collections portion of the budget. They consulted faculty librarians and the library dean as needed. When the collection management librarian resigned in July 2019, the collection maintenance librarian assumed all resource budget management duties and took over as the collection management librarian.

An allotted portion of the collections budget is normally allocated to acquire new print materials pertaining to subject liaison librarians’ areas of expertise, and is based on past average spending. This portion of the budget covers teaching faculty requests or collection development choices. Librarians use Choice book reviews or Global Online Bibliographic Information (GOBI) to assist in their decisions. GOBI is an acquisitions interface used to place orders for print and electronic monographs. Besides offering reviews and various vendor options, it also provides approval plans for librarians based on their subject areas. Each librarian is notified when there are new publications within their individually profiled subject areas to help enable them to make more informed selections.

GreenGlass, OCLC’s collection management web application, generates and analyzes custom holdings data, and was used as an assessment tool in 2017–2018. This helped librarians determine how well the library supported each program’s curriculum. while simultaneously informing deselection decisions. Additionally, interlibrary loan (ILL) requests are tracked, revealing the most frequently used journal titles.

The focus shifted to e-resources when the library physically closed due to COVID-19. Librarians were no longer purposefully building the collection with their own selections. Their requests were restricted to those that came directly from students or teaching faculty. Subsequently, all print standing orders and all print serials subscriptions had to be evaluated. Print serials are normally bound, but the bindery budget was also slashed.

Before the pandemic, librarians met regularly with teaching faculty to determine the addition or cancellation of e-resources. Usage statistics are routinely considered as part of this process. Wish list spreadsheets are maintained, and regular Collections meetings take place to facilitate dialogue between all librarians. Not everyone feels empowered to advocate for their program’s needs. To give everyone an equitable voice, a collections survey was distributed, and each librarian voted on emerging resources they deemed necessary for the areas that they represented, while others could also advocate for these resources (see appendix A).

Previously, the collection maintenance librarian tracked statistics for all e-resources with data available from vendor records in Alma, the library’s Library Services Platform. However, there was no longer a position dedicated to collection assessment and ensuring that these vendor platform and journal statistics were current. COUNTER dashboard data generated from SUSHI in Alma Analytics lacked the detail needed to generate reliable Cost per Use (CPU) metrics for collection decisions on an as needed basis. These statistics needed to be updated before renewal decisions could be made.

A Global Pandemic

In March 2020, the pandemic forced operations to cease in-person. All courses were moved online, and a purchasing freeze was implemented. The library building closed to the public. All print shipments were placed on hold. As a Federal Depository Library, this included government documents. Resource sharing, borrowing, and lending stopped. All collection development was suspended for monographs until February 2021. Only e-books requested directly from faculty were ordered. The monographs budget was reduced by $10,000 so that this money could be used to ship books to and from users with an EWU Libraries account who were unable to access them on-campus during the library closure. Due to the budget freeze, none of the year-end money normally spent on one-time library purchases could be directed towards new collections, nor could remaining funds be set aside for a new subscription. All monies were taken back by the university. Librarians no longer had the autonomy to make selection decisions. Every decision was filtered through the library dean, and items he approved to renew or cancel were submitted to the president’s office for final approval.

This centralized model was problematic when trying to be proactive with cuts, as library subscriptions include different licensing stipulations and renewal deadlines. Some licenses are multi-year renewals, and managed through a consortium, for example. It was necessary to quickly retrieve data for the remaining encumbered fiscal year 2020 renewals and for the upcoming fiscal year 2021 items. Generally, collection librarians do not seek approval to pay for encumbered materials. With enrollment numbers still in flux, hard financial data was not readily available for the new fiscal year budget; therefore, the budget was fluctuating. Collections’ decisions were based on the timing of renewals despite usage being high for these continuing resources. Email threads of feedback between library colleagues proved inefficient and chaotic as subject liaisons were not familiar with every database and journal, but still provided feedback on all resources even though certain ones did not fit into the subject areas they serve.

Librarians realized a more measured approach needed to be prioritized to make informed and balanced decisions relating to all resources due for renewal in the next fiscal year, while seeking approval for those still encumbered in the current fiscal year (see figure 4). A systematic approach was devised in May, and designed in June. Library staff coordinated the update of resource statistics in summer, and by August, most feedback had been collected. This allowed time for the remaining subscription decisions to be proactively submitted for approval before it was necessary to negotiate terms with vendors. At the beginning of this project, the library anticipated a 25 percent cut to the library budget, with the majority of cuts being made to the collections budget. Reviewing how other libraries have assessed their collections and the various methods they used to evaluate their resources was the first step in determining what factors EWU’s library would consider in its resource review and collections budget cuts.

Literature Review

It is evident evaluation of library resources has occurred throughout time. Kennedy et al. point out that this is often necessitated by ever-increasing serial costs.2 Wilde and Level advocated for extensive interdepartmental collaboration when undertaking assessment.3 Kelly determined that only collection development individuals should chart the course.4 Wilde and Level acknowledged that there appears to be a lack of routine, formal assessment taking place in libraries, and that most collection assessment seems to be done on an as-needed basis.5 According to Murphy and Buckley, the cost of serials has increased by 43 percent since 2013.6 Arthur saw the rising costs of serials as an opportunity to negotiate with vendors to reduce continuing resources contract rates.7

Concerns related to a periodical’s perpetuity when there is no guarantee that the electronic version will endure after its print counterpart is cancelled. Financing the electronic version is usually more expensive. Foudy and McManus noted that the price for electronic and print options are frequently offered at a discount when bundled, which further complicates the process for evaluating journal titles.8 Furthermore, journal packages are not always flexible. Vendors expect a certain spending threshold to be met. Title swaps may be allowed, but not cancellations. Sometimes an entire collection is cancelled versus having the opportunity to customize a title list by selecting specific titles.

Quantitative methods help prevent bias in decision making. Wilde and Level explained how statistics help narrow down which titles should be evaluated if usage is low, rather than wasting time assessing heavily used titles.9 Libraries vary in how much they are willing to spend on each use (uses measured vary depending on what metric is utilized, e.g. search, download, click, etc.). CPU is calculated by taking the subscription renewal price and dividing it by a year of usage. Enoch and Harker used seventy-one dollars per use as their threshold.10 Arthur chose $201 as the measure for determining his library resource’s CPU.11 Murphy and Buckley based theirs on how much an ILL transaction would cost; a cost of thirty-fifty dollars or more merited an ILL request since this is often what a library is charged, depending on how many copies have already been requested or what another library may charge others.12

Hoeve stated that involving teaching faculty in the assessment process through qualitative surveys or other communicative means is helpful because they can provide feedback on issues such as program accreditation needs or university mission, and explain which journals directly support their course curriculum or research.13 Departmental response versus relying only on individual faculty responses is important, according to Hardy, Zimmerman, and Hanscom.14 Many librarians focus on what their collections lack, rather than what they own or to which they can provide access. They also seem to undertake evaluation projects with longer timelines periods of time versus the shortened timeline a global pandemic necessitated.

Various methods of e-resource evaluation and collection assessment have been used. Wilde and Level employed analytics such as usage data, collection overlap, and statistics from link resolvers to help inform how well their resources were being used or duplicated.15 Hardy, Zimmerman, and Hanscom tracked only searches/sessions/full-text abstracts data elements for subscriptions that were not part of their consortial arrangements.16 Enoch and Harker focused on evaluating journals with access restrictions. They ensured that all institutional users could access a resource, and considered the length of embargo periods.17 Range of scope for each journal (journals reaching a broader audience versus a narrow range of users) was important to Kennedy et al. at the University of Florida.18 Foudy and McManus factored in rankings, rate of inflation, breadth, uniqueness, cost-effectiveness, and available authentication options.19 IP authentication is a preferred access method, making it easier for students to access resources remotely as identified users. EZproxy is used as an intermediary so authorized users can log in seamlessly, regardless of their location. Often, a journal or database does not support IP authentication or EZproxy.

Besides usage data and faculty feedback, Hardy, Zimmerman, and Hanscom prioritized retention of journals with content on diverse cultures and populations.20 Jensen described how most libraries rely on subject liaisons to build and maintain satisfactory collections, but since her library opted to activate more demand driven plans, such as a pay-per-use model, based on what users sought to access directly, since her library no longer has liaisons. This model allowed for short-term loans, and a title was automatically purchased after the fourth use.21 With this method, only articles directly accessed incurred a cost, rather than supporting an entire journal collection.

Enoch and Harker initiated cuts by not automatically processing their approval plans. They converted anything possible to an e-resource that cost the same or less as print titles. In another mandated round of cuts, they developed a rubric that outlined specific criteria that each of their resource subscriptions needed to meet (e.g. restricted access, title duplication, usage, and ease of use).22 They used Pareto’s Principle to determine a package’s value to their users.23 This principle expects eighty percent of outcomes to result from twenty percent of their causes. For a library, this would mean that only 20 percent of a collection is valuable to its audience.

Sutton focused on comparing citation lists, overlap data, usage, or a journal’s impact factor for resources that required further analysis.24 Source Normalized Impact (SNIP) is a complex metric used by Moisil at the California Digital Library.25 It reflects differences in each field’s citation practice. Carroll and Cummings discussed how their library developed a Serials Decision Database to aid in collections assessment. The database incorporated serials information into a single spreadsheet, and pulled data from their integrated library system, interlibrary and citation databases, journal usage reports, and subscription agents.26

Libraries have checked their holdings against bibliographies, used OCLC’s WorldCat Collection Analysis, followed the Conspectus method (an inventory of a library’s strengths and collection intensities), or used other standards-based perspectives for each subject. Acknowledging that all these methods tend to be one-dimensional, Kelly argued for a more holistic approach, believing that various perspectives and tools should be incorporated into any collection evaluation project.27

Some librarians have methods to track circulation of print titles, but most rely on a “dust test,” Moisil notes.28 Document delivery options are a good alternative when a serials cancellation project is underway. Nash and McElfresh confirmed this when they determined that none of the titles they cut had generated a significant number of ILL requests.29 Jaskowiak and Spires’ cancellations did not significantly increase ILL’s workload.30 Murphy and Buckley shared a new model that integrates a library’s OpenURL link resolver with document delivery to make articles more readily accessible.31 They explained how specific services such as Get It Now, the A–Z Academic Article Collection from Reprints Desk, ReadCube Access, DeepDyve, and IngentaConnect offer access to articles on demand in varied forms. Contracting with one of these platforms provides unsubscribed content directly to end users when they seek full-text for an article.32

Method

To conduct an effective evaluation of materials for their collections’ assessment project, EWU Libraries and Learning Commons’ librarians involved several stakeholders in the project and did not rely solely on Collection Services staff. They also incorporated both quantitative and qualitative measures and methods. Statistics are informative when evaluating materials; however, data is only one aspect of what makes a resource valuable. Each discipline’s journals vary in cost. Science journals often cost more than humanities journals, which is why CPU should not be the only factor when considering disciplinary trends.33 It was also important to solicit librarians and other teaching faculty members’ input, individually and by department since they are more familiar with the journals and databases in their respective areas of expertise.

The primary question addressed by the project was: How could stakeholders collaborate effectively to reduce the collections budget by 25 percent and still support curriculum needs? This question resulted in three objectives: 1) To collaborate with stakeholders so that the approach would be fair and consistent across all subject areas; 2) To design a method that stakeholders could use to identify which continuing resource subscriptions could be canceled; and 3) To cut 25 percent of the collections budget based on selected criterion.

The first objective was to collaborate with stakeholders to be fair and consistent across all subject areas. Communicating via email with librarians or having group discussions at meetings were no longer effective mediums to make decisions. The university had mandated that every purchase be approved through the president’s office, and therefore, this project involved several stakeholders. Internal stakeholders included executive administration, the library dean, business manager, faculty chair, subject liaison faculty librarians, other teaching faculty members, and collection services staff. Collaborators in each category were involved in the collection assessment decision-making approach. Librarians were asked to communicate with teaching faculty in subject areas where they served as liaisons. Once renewal decisions were reached, the collection management librarian corresponded with the library dean, business manager, and faculty chair. The library dean communicated with the appropriate person in the president’s office, requesting final approval to purchase or cancel materials. When approval was received, collection services staff took the appropriate measures to renew or cancel resources.

External stakeholders included students and vendors, as they would be affected by the decisions to renew or cut resources. Teaching faculty members were both internal and external stakeholders since they were part of the decision-making process, and their teaching and research was directly affected by the outcomes of these collective choices. All collaborators and stakeholders who were involved in the project are detailed in figure 5.

Collaboration between stakeholders was conducted via shared documents and Zoom meetings since the library was closed due to COVID-19. The library developed a comprehensive plan to involve everyone listed in figure 5 in a way that was both strategic and effective. Without the well-coordinated collaboration, cooperation, and communication between all parties, such drastic cuts to the collections budget could not have been made as efficiently or effectively.

The second objective was to design a method that stakeholders could use to identify which continuing resource subscriptions could be canceled. The library’s faculty chair and collection management librarian consulted and agreed on a tiered ranking and decision-making approach. This approach allowed librarians to rate each resource based on various factors and not limit decisions to quantitative data or qualitative feedback. A plan was needed that would enable cuts to be made across all departments and subject areas. This necessitated direct feedback from librarians regarding the resources within their subject areas, plus collective library faculty feedback for larger packages covering interdisciplinary areas. Usage data and librarian opinions could no longer be used as the single decision point to inform resource cuts. Considering multiple qualitative and quantitative factors before ranking each resource presented best cases for keeping or cutting subscriptions.

The library faculty chair created four tiered categories (see table 1) to determine priorities to assess the library’s collection. Unique to the methods cited in the literature review, these categories enabled objective data to support any subjective arguments from faculty librarians and teaching faculty to keep a resource. The purpose was to focus on librarians’ professional judgment based upon their liaison expertise. It revealed other areas that would merit further analysis. This approach spotlighted resources that were not used as frequently, forming a baseline for the collection management librarian and library dean to reference when decisions were due. It was emphasized that librarians should not evaluate resources unfamiliar to them to keep their focus on their individual subject liaison areas.

The collection management librarian created a spreadsheet of all renewals that required evaluation by librarians, plus a master tracking spreadsheet that contained all collections budget information and all renewal decisions when finalized. The data was initially saved on a shared drive, and was later switched to Google Sheets, which provided an easier platform for all librarians to simultaneously edit. One tab listed the current college, department and programs offered. If the program included graduate areas of study, that was noted. Since many librarians had assumed new subject areas due to turnover, the next tab listed subject areas and the librarians assigned to them. All e-resources were listed under each subject area with their price first and were then divided into their renewal months by tabs. This “subject” tab allowed each librarian the ability to quickly find the resources that required their feedback. A “month” tab was created for each month of the fiscal year (July 2020–June 2021) and showed when the subscriptions were due for renewal. Librarians could make decisions on materials in advance of their renewal dates by using the “month” tab.

When this project was implemented, July renewals were finalized, and librarians started tackling August and each month thereafter. They inserted their initials under the appropriate tiered category for the resources they represented, based on their subject liaison area. Each tiered category was listed in a separate column. Their feedback was based on their knowledge of each resource and teaching faculty’s feedback from the respective departments represented. Some resources were ranked by a single librarian, and others required multiple librarians to rank them. If an interdisciplinary resource was ranked, all librarians provided their initials under the tiered rating that they felt they could justify. Additional tabs provided separate title lists for interdisciplinary packages, which needed further evaluation

The spreadsheet provided a description of each expense, vendor, and cost in separate columns (see appendix B). A column used to justify each expense was included. The contract end date or perpetual access was specified for any subscriptions that were cancelled. Another column was added to ensure that the LibGuide created to communicate these changes was maintained once a decision was reached. Another column provided database and journal statistics links (see appendix C). These included the last two years of data with CPU outlined for each invoice cost and the twelve-month period most closely aligned with it. Forecasted CPU information was provided for some resources.

Collecting statistics was also a challenge, and staff and librarians collaborated to generate and provide available data for analysis. Not all current costs were available because vendors may inflate their renewals by 4–6 percent on average. This information becomes available closer to the renewal period for each resource. The collection maintenance librarian wrote instructions, met with staff, helped to generate statistics, and provided cost data on file that is regularly tracked for each continuing resource. A library and archives associate and the metadata librarian helped to generate the remaining statistics and CPU metrics.

Some smaller publishers do not provide statistics, or their reports vary (i.e. not all vendors provide SUSHI harvesting), which made it more difficult to navigate the data. Most resources provide COUNTER usage reports, which can be obtained through administrative portals. Although COUNTER converted from using COUNTER 4 metrics to COUNTER 5, some vendors had not converted to the latest version. The conversion to COUNTER 5 made analyzing data difficult, since two different years of reporting varied in formatting and metric types. The metrics librarians primarily used when reviewing COUNTER 4 Database Report 1s were Regular Searches, Result Clicks, and Record Views. For COUNTER 4’s Journal Report 1s, Full-Text Article Requests were counted. COUNTER 5 reports included Database Master, Title Master, and Journal Requests (Excluding OA Gold) with the metrics Searches_Regular and Total_Item_Investigations and Requests (see appendix D).

During the project, librarians relied heavily on the evaluation of continuing resources based on CPU data. They debated over how to determine the average CPU threshold when considering cuts. The collection management librarian flagged items to be considered for cancellation that exceeded ten dollars per use. Some of the librarians believed that an average ILL cost should be the determining factor. As previously noted, Murphy and Buckley shared that the average cost of an ILL is thirty-five dollars, assuming that the first five loans for a journal title are free through the Copyright Clearance Center (CCC), which is the case at EWU.34 If the subscription costs more than ILL, it was cut unless it was needed for an accreditation. The librarians chose to rely more heavily on ILL and the CCC for articles the library could no longer access. As a result, the EWU signed up for RapidILL, Ex Libris’ software to improve and expedite resource sharing.

Before assigning a resource to a tiered category, certain quantitative and qualitative factors were considered (see table 2). CPU, a quantitative factor, was not always the determining factor, although this was certainly consequential. Besides relying on database and journal usage statistics and CPU data, LibGuide statistics were informative. These statistical reports were generated from administrative assets data sets available within the LibGuides. One report showed the number of clicks for a database link was when it was accessed through the LibGuides within customizable timeframes. Peer library and open access holdings also played roles in decisions. If there was duplication or overlap with the other full-text aggregators that the library used, titles were cancelled. Since Science, Technology, Engineering and Mathematics (STEM) journals are often more expensive, this factor was considered before cutting a journal in this field simply due to high CPU. Available formats and access options were reviewed. Less expensive alternative sources were explored. When consensus could not be reached by librarians on a cross-disciplinary resource, Survey Monkey was used to poll librarians to reach a deciding vote.

Qualitative factors involved group discussion at meetings. EWU’s Collection Development Policy was referenced. Deselection guidelines in this policy include factoring in how relevant the resource is to the university’s mission and curriculum. Since the university is not classified as a research institution, broader research needs and range of scope were prioritized over the specialized needs of graduate students and teaching faculty. The university is focused more on student success and retention, and not as heavily on faculty research and support.

While the collection development policy states that subject librarians are responsible for deselection of resources, it seemed prudent to involve teaching faculty as much as possible. Seeking departmental feedback helped weigh a resource’s political capital, in what programs or courses it was used, and uniqueness of content. Titles were retained if teaching faculty justified the need for them in their field or program.

Diversity was another factor that impacted continuing resource retention. This is a priority at the university since EWU’s goal is to be the premier public diversity-serving institution in Washington state. Diversity, equity, and social justice are included as an initiative in EWU’s current strategic plan. EWU is also a recipient of the 2019 Higher Education Excellence in Diversity (HEED) Award. This award honors universities with an exemplary dedication to diversity and inclusion. If a journal or database that fit into this framework was not as highly used as desired, it was still renewed based on this value. The policy also encourages the library to use the buying power of their consortia as much as possible.

If a librarian decided to keep a journal and provided strong justification why the expense was essential for a particular title, approval was sought. If a title was considered core or regional, librarians advocated for it. If a title lacked justification, the collection management librarian contacted the appropriate subject liaison librarian for additional feedback. The library dean made the final decision for any outstanding titles that required a decision before a deadline.

For the library’s individual journal title subscriptions managed by ESBCO, information on duplicate holdings and open access information were included on a separate spreadsheet in addition to the format, title, metric type, cost, usage and CPU for the last two years. Rather than assigning each title to a tier as with most continuing print and e-resources, librarians reviewed each title relevant to their subject area, and added feedback in a separate column. They provided their reasoning for keeping or cancelling a title before submitting a decision to the library dean via the collection management librarian. For certain packages, a journal’s impact factor was considered, or if it was a key journal for a field. If an embargo was a year or less, a title was often slated for cancellation.

Due to title transfers or title name changes, not all titles had the data needed to help inform decision making. These scenarios raised the question of whether some statistics were reliable given any lapses in coverage that may have taken place, or if they had been linked correctly from the start. Feedback was not received for all journal titles due to the sheer number that needed review or based on the lack of knowledge pertaining to them. In these cases, the library dean decided whether to keep or cancel these titles. If an electronic version was available for print titles under review, it was preferred. If online access for journals was not IP authenticated, that option was ruled out. Sometimes electronic access was tied to maintaining serial coverage, or print and online formats were bundled together. Cutting too many titles would increase service charges, and was a consideration.

Streaming videos are in high demand at EWU since instructors prefer online accessibility, and the pandemic accelerated this need when classes moved to online. It is hard to manage their cost with the Patron Driven Acquisition (PDA) model maintained by the library until the end of the fiscal year 2020 when the library’s set funding threshold was reached and no extra funding could be allocated to continue using this model. With this model, four uses of a film triggered a purchase. A PDA play is incurred for a title when an end user accesses the title in a unique session and watches 30 seconds or more of consecutive footage. The absence of a set annual subscription fee meant that costs were unpredictable, and depended on how many times a video was accessed and for how long it was viewed. The library began managing this vendor platform by request only. Not all requests are approved unless they directly support course curriculum. The library opted to add AVON’s streaming video service to supplement Kanopy when it became available through the library’s consortial arrangement since this platform also offers a variety of educational films. It is more cost effective than Kanopy since ProQuest offers an annual subscription rate. With instruction continuing to remain online or hybrid for the foreseeable future, the demand for streaming video will remain high. Due to licensing and copyright restrictions, not all physical DVDs can be duplicated for online use.

Once all the resources were assigned a tier, the collection maintenance librarian created a separate spreadsheet to order each resource by tier according to how many votes received. Each tier had a total cost for the expenditures assigned to it, reflecting how much savings each provided. Figure 6 shows the distribution of resources by percentage for each ranking. All Tier 1 expenses were most likely to receive approval if they included the appropriate justification. Some Tier 1 resources were cut if cheaper alternatives were available. Tier 2 resources were kept when possible, and only cut when absolutely necessary. Tier 3 items were further scrutinized, and many were cut. Tier 4 expenses were automatically cut.

Results

The third objective was to cut 25 percent of the collections budget based on selected criterion. All the decisions for this fiscal year have been made, and approximately $330,000 was cut from the continuing resources budget, which is approximately 27 percent of the collections fund budget. This included several standing orders, databases, packages, and individual journal titles. Seventy percent of print journals were cut, 12 percent of the monographs budget was cut, 100 percent of standing orders were cut, 100 percent of the journal binding budget was cut (although some of this may be restored in the new fiscal year), 100 percent of year-end money was cut, and 40 percent less was spent on streaming media by mediating requests. The library increased spending on new subscriptions by 4 percent. Two percent of funds allocated for new subscriptions will be slashed, resulting in a 25 percent permanent reduction of $300,000 to the collections fund budget index that will take effect in the new fiscal year.

Each resource’s row on the master spreadsheet was highlighted with a different color, signifying the action taken regarding it. Red highlights indicated that a resource had been cancelled. Orange meant questions remained about it. Green showed that the item was renewed. The collection maintenance librarian communicated changes to vendors as necessary. Some vendors tried to negotiate cancellation notices with lower renewal prices, but this did not influence decisions once they had been made. Holdings were updated in Alma when the renewal expired.

In some situations, the decision was made to subscribe to individual journal titles rather than an entire database. In one case, EBSCO’s Academic Search Complete subscription was upgraded to their Academic Search Ultimate version for broader coverage with more journals. In doing this, Science, an expensive journal title with a direct subscription, was cut, and could still be accessed via this upgrade.

Clarivate’s Web of Science was exchanged for Elsevier’s Scopus for a much lower price. Access World News was reactivated due to being partially subsidized through a state library arrangement. ProQuest’s Newsstream was added back through another prior consortial agreement since it was no longer part of the larger package deal through a different consortium. The library’s Junior Library Guild print elementary books standing order was changed to digital eBooks, and AVON was added to help meet streaming video demands.

The collection management librarian, business manager, library faculty chair, and library dean were creative with the available library funds. Because the library was closed in summer due to the pandemic, student staffing was eliminated, and a summer index fund normally spent on staffing for a percentage of collection costs was used. A library staff member helped track and pay invoices partially with this fund. This same team advocated that departments split the cost of a resource with the library or fund it fully if they had their own funds. The Education department funded Education Week fully and split the cost of ProQuest’s Education Database with the library. A special music fund was used to procure IPA Source. A decision was made to subscribe to Linguistics & Language Behavior Abstracts due to a vendor credit. Funds were taken from a special Library fund that is used at the library dean’s discretion to support Kanopy requests. Endowment funds were re-evaluated quarterly to reallocate these monies according to the library’s needs. One subscription supported by these funds was cancelled by faculty librarian vote so that Project Muse could be added. Benzeit Dictionary of Artists was added temporarily to fill a gap resulting from the fact that art students were not able to access all print materials readily during the pandemic.

Since it was necessary to communicate cancellations to teaching faculty, librarians created a LibGuide (see appendix E), “Subscription Renewals, Cancellations, and Additions.” This LibGuide kept library and teaching faculty updated with ongoing changes to library resources. It was provided to all faculty members for a centralized and transparent place to communicate decisions made pertaining to the library collections budget and includes a statement that explains the project. It provides a list of all renewals, cancellations, and additions with their formats, perpetual access coverages, and subscription end dates. It includes a total amount in cancellations and a link to another Research Guide for academic resources made available by vendors during COVID-19.

When the pandemic occurred, publishers and vendors began offering extended trials to e-resources. The library took advantage of this despite the labor involved to activate and deactivate the resources. The collection management librarian developed a COVID-19 LibGuide that highlighted all the resources new to the university and extended coverage of current resources that were being offered, updating it as access ended. The discovery and systems librarian updated the Database A–Z list each time a staff member activated or deactivated various collections. Open Education Resources (OERs) and e-books were also highlighted on this guide.

Recognizing that some faculty would not be happy with all the decisions made, the collection management librarian created a spreadsheet to track feedback received from faculty to support resubscribing to cancelled subscriptions if more funds become available later. Only one database cancellation received negative feedback from more than one department, but since there was strong justification to cancel it, the librarians stuck by their decision. Based on other feedback received, two resources were added back that had initially been part of a larger package. A wish list is being maintained for any emerging resource requests.

Analysis

The objectives were met. Everyone collaborated, a measured method to assess continuing resources was designed and implemented, and the budget was cut to meet expectations. The pandemic provided an opportunity to reduce excess spending and to re-assess what continuing resources programs truly need in order to support curriculum and intellectual inquiries. Besides offering extended or new access to resources for a period of time, most vendors offered or were amenable to negotiating flat or reduced renewal rates to resources, which helped tremendously. The library was still able to add some new resources by upgrading a subscription, finding cheaper alternatives, and through taking advantage of consortial deals and a subsidized trial. Without e-resources and technology available, the library would not have been able to serve faculty and students as effectively. The library never really closed for this same reason, aside from the building being inaccessible for a time. This enabled the library to continuously serve its stakeholders.

There are still quite a few challenges with this project. Libraries using a similar approach may want to be more proactive and have a plan in place to cut resources before a crisis occurs. For example, the author and her colleagues were not able to collect subject liaison ranking feedback until July and August. Since several renewals take place in July, some were cut or renewed in May when it was necessary to notify vendors without further evaluation of resources using the tiered ranking system approach. A checklist could be provided to ensure that subject liaison librarians had considered all quantitative and qualitative factors before submitting their ranking feedback to ensure thoroughness and consistency.

This project was not comprehensive due to timing constraints. Cuts needed to be made before renewal deadlines and in the same fiscal year that they were mandated. In the future, if time permits, it would be helpful to send formalized surveys to teaching faculty to help inform decision-making. Despite librarians regularly engaging with other teaching faculty in subject areas they represent or being familiar with the resources critical to program support based on their own expertise, they could not consistently provide strong justification for why they would rate a resource in a particular tier. In some cases, it was difficult to solicit feedback because many library and teaching faculty members are not under contract during summer or were on leave when decisions were needed. This made it difficult to get the specific counsel needed for certain titles. It was unavoidable, considering the timing of the mandated budget analysis and cuts. This meant that some things were cut that might have been more important to teaching faculty than assessed by librarians. Implementing decisions was a lengthier process than average since approval had to be solicited from executive administration.

When there was not an opportunity to gather qualitative input, librarians ranked resources based on the available quantitative data. Usage statistics and CPU were used the most to attain the mandated 25 percent in cuts to the collections budget. This data heavily influenced the rankings and decision making for cuts. Quantitative data did not weigh as heavily when journals were not canceled for strategic or political reasons. It also was a significant factor for titles duplicated in other databases or for the more costly STEM journals.

Working from Google Sheets enabled everyone to simultaneously work on a document in real time. However, navigating so many spreadsheets resulted in information overload, which was overwhelming or confusing for some. Librarians often asked questions without first referencing the spreadsheet(s) for relevant information. This revealed how much the collection management librarian is relied upon for collection development and assessment despite the attempt to make this project as collaborative as possible

After such a labor-intensive cancellation process, questions remain if the library will re-purchase multiple individual titles in lieu of a larger database package, which may save on cost, but not on the time management required to activate and maintain these titles. Although staff analyzed Elsevier’s Science Direct package of journals for each title’s impact factor, not all journals were evaluated with this level of detail. Journals have not been evaluated based on any of the university’s faculty publishing or citation factors, and is another reason the study was not considered comprehensive although every continuing resource was ranked using the information available to librarians. Some journals were cut when perhaps those in which faculty have published should have been kept, including those that they most frequently cite. Citations could be analyzed to determine if faculty and students prefer a particular publication year range or format. Future title cancellations should perhaps consider the ISI impact factor during the review process.

It would also be useful to develop an attractive visual platform to report usage to stakeholders. This would illustrate on an ongoing basis which resources are most used or underutilized. Attempts have been made to explore using Tableau, data visualization and analysis software libraries use to present statistical data in more automated and user-friendly ways.

ILL could be tracked to see how many article requests resulted from journal cancellations. Continued feedback from faculty will be useful to determine future needs. Regarding a long-term plan, a continuing resources committee should be formed with representation from various areas to ensure that regular evaluation of resources occurs. Once the restructuring of all colleges takes place, a more thorough program and enrollment review should be conducted, and funds reallocated accordingly.

Conclusion

Collaboration was essential to make all necessary cancellations for continuing resources to ensure that program curriculum would continue to be supported, and to meet renewal decision deadlines. Having a tiered ranking system for collection assessment designed to evaluate resources thoughtfully with set quantitative and qualitative factors helped the process flow in an organized and consistent manner. All resources the library dean submitted for approval to executive administration were approved based on justification provided using the tiered ranking approach. Each resource submitted for expense was justified well quantitively or qualitatively. Permanently reducing 25 percent of a collections budget was not a small undertaking, particularly when it had to be accomplished in a short timeframe. All objectives were met, and programs and course curriculum had the library resources to support them.

Fiscal stewardship of the library is taken seriously, although the administration should take note that libraries need to be adequately funded to support meaningful teaching and learning for both university professors and students. This assessment accomplished what it set out to achieve, yet continuously eliminating library resources due to budget cuts is not a sustainable approach to supporting research, education, and student success. No one can predict what the future holds for library budgets and collections or higher education. A project of this magnitude would not have succeeded without the cooperative efforts of all stakeholders involved. This collaboration exhibited the librarians’ ability to reduce the collections budget to be sustainable in challenging, unprecedented, and continuously uncertain pandemic times. Any library facing similar challenges could benefit from taking a similar systematic approach involving multiple stakeholders.

References

1. Maria Savova and Jason S. Price, “Redesigning the Academic Library Materials Budget for the Digital Age: Applying the Power of Faceted Classification to Acquisitions Fund Management,” Library Resources & Technical Services 63, no. 2 (April 2019): 131, https://doi.org/10.5860/lrts.63n2.131.
2. Kathryn Kennedy et al., “Evaluating Continuing Resources: Perspectives and Methods from Science Librarians,” The Serials Librarian 55, no. 3 (2008): 429, https://doi.org/ezproxy.library.ewu.edu/10.1080/03615260802059064.
3. Michelle Wilde and Allison Level, “How to Drink From a Fire Hose Without Drowning: Collection Assessment in a Numbers-Driven Environment,” Collection Management 36, no. 4 (2011): 223, https://doi.org/10.1080/01462679.2011.604771.
4. Madeline Kelly, “Applying the Tiers of Assessment: A Holistic and Systematic Approach to Assessing Library Collections,” Journal of Academic Librarianship 40, no. 6 (2014): 587, https://doi.org.ezproxy.library.ewu.edu/10.1016/j.acalib.2014.10.002.
5. Wilde and Level, 226.
6. Julie A. Murphy and Chad E Buckley, “Document Delivery as a Supplement or Replacement for Serial Subscriptions,” Serials Review 44, no. 3 (2018): 241, https://doi.org.ezproxy.library.ewu.edu/10.1080/00987913.2018.1525238.
7. Michael A. Arthur, “Managing a Comprehensive Cost-per-Use Project in a Large Academic Library,” Serials Review 44, no. 4 (2018): 300, https://doi.org.ezproxy.library.ewu.edu/10.1080/00987913.2018.1558936.
8. Gerri Foudy and Alesia McManus, “Using a Decision Grid Process to Build Consensus in Electronic Resources Cancellation Decisions,” Journal of Academic Librarianship 31, no. 6 (2005): 533, https://doi.org.ezproxy.library.ewu.edu/https://doiorg.ezproxy.library.ewu.edu/10.1016/j.acalib.2005.08.005.
9. Wilde and Level, “How to Drink From a Fire Hose Without Drowning,” 228
10. Todd Enoch and Karen R Harker, “Planning for the Budget-Ocalypse: The Evolution of a Serials/ER Cancellation Methodology,” The Serials Librarian 68, no. 1–4 (2015): 284, https://doi-org.ezproxy.library.ewu.edu/10.1080/0361526X.2015.1025657.
11. Arthur, “Managing a Comprehensive Cost-per-Use Project,” 302
12. Murphy and Buckley, “Document Delivery as a Supplement,” 242
13. Casey Daniel Hoeve, “Resource Management in a Time of Fiscal Scarcity: Combining Qualitative and Quantitative Assessment for Journal Package Cancellations,” Serials Librarian 75, no. 1–4 (2019): 46, https://doi-org.ezproxy.library.ewu.edu/10.1080/0361526x.2019.1576571.
14. Beatriz Betancourt Hardy, Martha C. Zimmerman, and Laura A. Hanscom, “Cutting without Cursing: A Successful Cancellation Project,” Serials Librarian 71, no. 2 (2016): 116.
15. Wilde and Level, “How to Drink from a Fire Hose Without Drowning,” 222.
16. Hardy et al., “Cutting without Cursing,” 113.
17. Enoch and Harker, “Planning for the Budget-Ocalypse,” 284.
18. Kennedy et al., “Evaluating Continuing Resources,” 433.
19. Foudy and McManus, “Using a Decision Grid Process,” 535.
20. Betancourt Hardy, Zimmerman, and Hanscom, “Cutting without Cursing,” 119.
21. Karen Jensen, “No More Liaisons: Collection Management Strategies in Hard Times,” Collection Management 42, no. 1 (2017): 6, https://doi-org.ezproxy.library.ewu.edu/10.1080/01462679.2016.1263812.
22. Enoch and Harker, “Planning for the Budget-Ocalypse,” 287.
23. Enoch and Harker, 286.
24. Sarah Sutton, “A Model for Electronic Resources Value Assessment,” Serials Librarian 64, no. 1–4 (2013): 246, https://doi-org.ezproxy.library.ewu.edu/10.1080/0361526x.2013.760417.
25. Ingrid Moisil, “Renew or Cancel? Applying a Model for Objective Journal Evaluation,” Serials Review 41, no. 3 (2015): 161, https://doi-org.ezproxy.library.ewu.edu/10.1080/00987913.2015.1065466.
26. Diane Carroll and Joel Cummings, “Data Driven Collection Assessment Using a Serial Decision Database,” Serials Review 36, no. 4 (2010): 227, https://doi-org.ezproxy.library.ewu.edu/10.1080/00987913.2010.10765322.
27. Kelly, “Applying the Tiers of Assessment,” 586.
28. Moisil, “Renew or Cancel?,” 161.
29. Jacob L. Nash and Karen R. McElfresh, “A Journal Cancellation Survey and Resulting Impact on Interlibrary Loan,” Journal of the Medical Library Association 104, no. 4 (2016): 297, https://doi-org.ezproxy.library.ewu.edu/10.5195/JMLA.2016.143.
30. Megan Jaskowiak and Todd Spires, “Subscription Alternations: Usage of Canceled Journal Subscriptions via Article Delivery Methods,” Serials Review 44, no. 2 (2018): 99, https://doi-org.ezproxy.library.ewu.edu/10.1080/00987913.2017.1396099.
31. Murphy and Buckley, “Document Delivery as a Supplement,” 242.
32. Murphy and Buckley, 243–44.
33. Wilde and Level, “How to Drink From a Fire Hose Without Drowning,” 229.
34. Murphy and Buckley, “Document Delivery as a Supplement,” 242.

Appendix A: Prioritize New Resources

Appendix B: Library Resources Tiered Librarian Feedback Spreadsheet Categories (tabulated my month each resource renewal due in)

Vendor Payment

Description of Expense

Expense or Range of Expenses

Access Ends/Contract Ends

Added to Collections LibGuide?

Describe Why Expense is Essential

Cost Per Use Spreadsheet (Google Docs URL)

Based on your subject liaison areas, please place your initials under the appropriate Tier you think the resources you represent fall under.

Tier 1: A resource we cannot cut if we intend to keep operating as a university

Tier 2: A resource we could only cut in an absolute worst-case scenario, since it could affect department accreditation or require that a program stop offering certain classes which cannot substitute for this resource

Tier 3: A resource which is highly useful, but which we could bear to cut—doing so might require adjustments to curriculum and student assignments but that’s feasible

Tier 4: A resource which has some value, but which would be easiest to cut right now, since doing so would likely not require faculty to make any curricular changes

Column A: Vendor Payment

Column B: Description of Expense

Column C: Expense or Range of Expenses

Column D: Date Access Ends/Contract Ends

Column E: Added to Collections LibGuide (Y/N)

Column F: Describe Why Expense is Essential

Column G: Cost Per Use Spreadsheet (Google Sheets URL)

Column H: Tier 1: A resource we cannot cut if we intend to keep operating as a university

Column I: Tier 2: A resource we could only cut in an absolute worst-case scenario, since it could affect department accreditation or require that a program stop offering certain classes which cannot substitute for this resource

Column J: Tier 3: A resource which is highly useful, but which we could bear to cut – doing so might require adjustments to curriculum and student assignments but that’s feasible

Column K: Tier 4: A resource which has some value, but which would be easiest to cut right now, since doing so would likely not require faculty to make any curricular changes

Appendix C: Database Statistics Example

Metric_Type	Reporting_Period_Total 2018-2019	FY19 Cost	Cost Per Use 18/19	Reporting_Period_Total 2019-2020	FY20 Cost	Cost Per Use 19/20
Searches_Regular	16189	$3,021.00	$0.19	13745	$3,282.24	$0.24
Total_Item_Investigations	3678	$3,021.00	$0.82	2478	$3,282.24	$1.32
Unique_Item_Investigations	3121	$3,021.00	$0.97	2099	$3,282.24	$1.56
Unique_Title_Investigations	97	$3,021.00	$31.14	76	$3,282.24	$43.19
Total_Item_Requests	7	$3,021.00	$431.57	8	$3,282.24	$410.28
Unique_Item_Requests	5	$3,021.00	$604.20	7	$3,282.24	$468.89

Appendix D: COUNTER 5 Metric Types

For further detail, see: https://www.projectcounter.org/code-of-practice-five-sections/3-0-technical-specifications/

Metric_Types, which represent the nature of activity being counted, can be grouped into the categories of Searches, Investigations, Requests, and Access Denied.

Metric_Type	Description	Host_Types	Reports
Searches_Regular	Number of searches conducted against a user-selected database where results are returned to the user on the host UI. The user is responsible for selecting the databases or set of databases to be searched. This metric only applies to usage tracked at the database level and is not represented at the platform level.	A&I_Database Aggregated_Full_Content Discovery_Service eBook_Collection Full_Content_Database Multimedia_Collection	DR DR_D1
Total_Item_Investigations	Total number of times a content item or information related to a content item was accessed. Double-click filters are applied to these transactions. Examples of content items are articles, book chapters, or multimedia files.	All Host_Types: A&I_Database Aggregated_Full_Content Data_Repository* Discovery_Service eBook eBook_Collection eJournal Full_Content_Database Multimedia Multimedia_Collection Repository* Scholarly_Collaboration_Network	PR, DR, TR, IR DR_D1, TR_B3, TR_J3

Appendix E: Subscription Renewals, Cancellations, and Additions Research Guide