Insights from the Data

The University of Virginia Library's support for sustainable scholarship is informed by a variety of relevant data sources. Below are some of the key data sources and insights that inform our sustainable scholarship decisions.

Local Budget Data

UVA Library keeps track of long-running trends in what we spend on resources, both in broad categories and for specific vendors. This helps us predict what our collections will look like in the future if past patterns continue to play out. Insights from budget data have helped us to conclude we must change our approach to Big Deals:

  • The proportion of our collections budget consumed by four Big Deal vendors (Elsevier, Wiley, Sage, and Springer) has doubled in less than a decade, from 21% in 2009 to 43% in 2018.  
  • Electronic subscriptions now consume an amount equal to 100% of our collections funding from the University. Purchase of physical materials and other resources is only possible thanks to library endowments. Neither source of funding is growing as fast as the cost of Big Deals. 
  • Locked-in annual inflation rates have led the prices of some of our biggest deals to double over the last decade.
  • The COVID-19 pandemic has had a significant impact on our budget, and on the budgets of campus units who help support our purchase of resources. The choice we face is not whether to stop subscribing to some resources; it is which subscriptions to cut.

As part of our review of our spending on these biggest deals, we joined several of our colleagues around Virginia in publishing Spending Disclosures, which included annual cost for these deals as well as some basic usage data and projections into the future. Additionally, we started conversations around campus about these costs, including a presentation to the faculty senate laying out the challenge we face in controlling information costs.

Local Usage Data

UVA Library has a rich dataset reflecting the usage of academic journals across several axes (downloads, citations, publication) over the preceding decade. We learned that most of the journal titles in our big bundles are rarely or never used. For most bundles, 80% of usage was concentrated in the top 20% or so of journal titles, and 95% of the usage was from the top 30-40% of the journal titles. Looking closely at this usage data, we saw that we could design a custom package of journals that is tailored to the interests of UVA researchers, rather than buying access to hundreds of unused journals.

UVA-affiliated users with Netbadge credentials will soon be able to explore this data in a visualization tool created by our Research Data Services team.

National Big Deal Data

The non-profit group SPARC maintains two interesting datasets relevant to Big Deals:

  • The Big Deal Cancellation Tracking site shows the international trend of breaking up big journal bundles, from Iowa State to Finland. Where available, the tracker also shows how much money institutions save by ending their Big Deals. 
  • The Big Deal Knowledge Base shows prices paid and other terms for Big Deals at hundreds of institutions. This data seems to confirm what past research has shown: that journal package prices have as much to do with the profit motive of the vendor and the negotiating savvy of the institution as with any objective measure of institutional size or use. 

Alternative Access Data

To help determine whether a subscription is worthwhile, we use data provided by Unsub to compare the cost of a subscription to the cost of alternative forms of access, including open access, interlibrary loan, and on-demand article purchase. As alternatives to subscriptions grow more seamless and robust, we can look to invest in the most efficient pathways to access, creating a strategy tailored to UVA's unique research profile.

Qualitative Data

In addition to quantitative measures of use and value, we consider input from library liaisons who are experts in their field, as well as feedback from faculty and departments about the resources they value, including how often they expect to use a resource and how quickly they need access in most use cases. We strive to support a broad, deep, and diverse array of resources, and we realize numbers don't always tell the full story of a resource's value to researchers.