Paper presentation: 'More than Meets the Eye' – Analyzing the Success of User Queries in Oria
Library catalogs provide quick access to a wide variety of materials available in academic libraries. However, searching for literature is not always straightforward, and users encounter issues in finding the right material for their work and studies. To get a better understanding of these issues, we have analyzed user interactions with the library catalog of the University of Oslo Library, which we will present at the VIRAK conference on the 13th of June.
The slides of our presentation can be viewed here.
Since January 2015, the discovery system Oria (an installation of Primo) has been the main user interface to the catalogs of Norwegian academic libraries part of the Bibsys consortium. Within Oria, various user actions, such as searches and clicks on search filters are recorded, and aggregated statistical data can be accessed via Primo Analytics. This wealth of data provides new opportunities to analyze user interactions with current library catalogs, and to potentially improve digital access to library collections based on lessons learned.
To improve our understanding of the issues at play when users interact with library catalogs, we explore what types of queries users perform, determine how successful user queries are, and look at underlying reasons for unsuccessful queries. Hence, our main research questions are:
|1. Which insights can we gain from classifying user queries within Oria by their popularity, specificity and intended target resources?
2. To what extent are the most popular user queries successful?
3. Which underlying reasons for unsuccessful queries can be determined?
To address our research questions, we have done an analysis of data available from Primo Analytics for the University of Oslo. In particular, we looked at two datasets: the most popular queries (from January 2015 to September 2016), and the “zero-result” queries, i.e. the queries for which searches did not obtain any results (from August 2015 to September 2016).
In addition to a number of basic quantitative analyses, we manually annotated the 50 most popular queries in the dataset, accounting for a large degree of the total number of queries, as well as a random set of 50 zero-result queries. To further validate to what extent the queries were aimed at retrieving popular library materials, we performed cross-comparisons with a set of frequently loaned materials from Alma, the library’s resource management system.
Our conclusion is that the library catalog contains more than meets the eye: even though requested materials may be available in the library, they do not always show up on the first results page. In our analysis we observed a prevalence of queries related to users’ reading lists, but these queries frequently fail to get immediate results. This is caused both by issues in users’ query formulations, and by issues in system support. For instance, frequent user queries consist of pasted citations, which often do not fare well, while in other cases, minor spelling errors impede retrieval of certain items. This suggests the importance of improving system features assisting users in their library catalog searches. Using a variety of findings as a basis, we provide concrete recommendations for improving library catalogs and digital library services at large.
|Details and conference registration: see the VIRAK conference website|