Transaction Log Analysis within a Bento Discovery System

Abstract

Academic libraries have introduced bento-style discovery systems to address identified user issues with web-scale discovery systems. The bento systems partition search results into discrete areas on the screen grouped by content format type and/or local service results and address key user needs such as known-item search and streamlined full-text delivery. This poster reports on a transaction log study on the bento discovery system employed at the University of Illinois at Urbana-Champaign Library. The analysis looks at known-item vs exploratory searches, search argument characteristics, and user clickthrough behaviors.