Slides
By 2007 Google’s preeminence as the primary influencer of search behavior trends had been established. Libraries and vendors began to modify the way goals were described, inserting “googlelike experience” into strategic plans and product pitches. In the decade since then, we’ve seen a proliferation of content types and resources, along with tools to discover them, as libraries and vendors have worked to support discovery beyond the OPAC. At the same time web scale discovery systems (WSDS) have been evolving to provide patrons with a coherent search experience of diverse content types and to add value in the form subject guides, knowledge cards, and connections beyond resource metadata themselves made possible by linked data. We’ve seen significant improvements in patron experience. But challenges remain, particularly in academic environments, which are motivating continued experimentation.
This panel representing two large academic libraries and the largest aggregator of scholarly resources will reflect on these trends and to examine recent innovations in discovery. Panelists will describe how the approach to discovery has evolved to serve diverging needs of both libraries and vendors who outsource their article experience to aggregators.
The audience will learn about two innovations driven by academic libraries, bento box discovery and coordinated discovery, and the resultsof a vendor’s multiyear experiment with contextual search boxes.
Innovations in Discovery Systems: The Promise of the Bento Approach
Library discovery services have evolved through: expanded OPACS, federated search systems employing broadcast searching; webscale discovery systems (WSDS) that aggregate metadata and fulltext content into a single integrated index; and now hybrid bento style systems that use federated techniques over WSDS, OPACs, and local information content and partition search results into separate zoned screen displays with content grouped by format type or local service information. Recent studies on webscale discovery systems have identified a number of user access issues centering on problems with blended result displays, problematical relevancy rankings of search results, fulltext search problems, and the inability of WSDS to adequately provide access to local library services and resources. In particular studies have shown that users have difficulty distinguishing between material formats and interpreting the blended result displays. The concept of “full library discovery”, a phrase first coined by Lorcan Dempsey, has been introduced to refer to discovery approaches that move beyond the retrieval of collection materials to also include local information services and local content and links.
The bentobased systems are an attempt to address the identified problems with WSDS and also provide discovery services that address user needs, in particular knownitem search and streamlined fulltext access. The bento zones typically present meaningful groupings of content extracted from vendor services APIs, WSDS search results (often limited by format), and local information services and content such as library websites, pathfinder information, dataset repositories, subject specialist links, and course management system content. This presentation will provide an analysis of the 35 libraries presently employing the bento approach and will look identified user needs and search behaviors, as revealed in detailed search and clickthrough transaction log analyses. There is a clear need for an evidencebased analysis of user search behaviors in retrieval environments characterized by access to distributed information resources.
Innovations in Discovery Systems: Coordinated Discovery Building on the premise of bento box discovery that discovery is best partitioned into categories by type of resource and service, coordinated discovery aims to optimize the discovery of each category while providing effective navigation between discovery categories and suggestions from other categories when sufficiently relevant. Suggestions are produced through the propagation of a likely known item search query to other search categories via their search APIs (as in bento box discovery). Results from each discovery category are then evaluated for closeness of fit, and those satisfying the match threshold are presented to the patron in a sidebar “suggestion box” alongside the main search results. The UX leverages patron’s familiarity with major search engine behaviors.
