In the summer of 2010, Drake University’s Cowles Library implemented EBSCO Discovery Service™1 (EDS). During the implementation and throughout the first year of use, the library faced challenging decisions regarding resource selection, how to present resources in a way that maximized their utility, and configuring the public interface to appeal to both first-time users and experienced researchers. This chapter provides libraries considering or having purchased a Web-scale discovery tool a look at the various issues and potential solutions they may face when implementing a discovery service. It will specifically target those aspects of implementation that concern resource selection and configuration of the public interface. Issues and examples discussed refer to EDS™, but many of the issues are common to all discovery solutions.
TopLiterature Review
In early 2010, when Cowles Library began searching for a discovery solution, very little had been written in the library literature about Web-scale discovery products. Perhaps the first mention of this kind of tool was made by Marshall Breeding shortly after Google Scholar™2 was introduced (Breeding, 2005). Breeding made a distinction between the many limitations of on-the-fly federated searching and what he termed “centralized search” (p. 27) approaches; namely, scalable systems that are able to store and search harvested content in a central index. Prior to 2010, the only other significant examination of this category of tools was a discussion of a prototype library discovery system developed by the University of Nevada Las Vegas Libraries (Dolski, 2009).
Since early 2010, Rowe (2010) and Vaughan (2011) have reviewed the major Web-scale discovery products. Rowe examined Serials Solutions'®3 Summon™4, EDS™ and OCLC's WorldCat®5 Local, scoring each on content, user interface/searchability, pricing and contract options. Vaughan took an extensive look at those three products, as well as Primo Central™6 from Ex Libris™7, and pointed out several important realities to keep in mind when evaluating discovery services: 1) Web-scale discovery tools do not provide access to a library's complete collection; 2) libraries will need to maintain some separate data silos and interfaces; 3) existing discovery services cannot anticipate the researcher's exact needs, but progress continues to be made on relevancy ranking and resource recommender features; and 4) Google and Wikipedia will continue to be popular destinations for researchers, and libraries must continue to develop strategies for connecting users with library resources.
Julia Gross and Lutie Sheridan (2011) conducted usability testing on Edith Cowan University Library's implementation of Summon™ and found that, while users were able to quickly and easily find large amounts of information with Summon™, they remained confused about result formats and usefulness of the information. Sarah Williams and Anita Foster (2011) carried out a usability study of EDS™ on a small sample of six participants at Illinois State University's Milner Library. While engaging in five research scenarios, participants in this study had little trouble identifying and applying both pre- and post-search limits, but all were confused about the role of federated search results in EDS™. One of the discoveries made by Williams and Foster is that, while participants generally found it easy to retrieve relevant results using EDS™, they still require (and desire) some kind of instruction to fully utilize the capabilities of a powerful discovery tool (Williams and Foster, 2011).