If You Build it, Will They Come?: Natural Experiments with Springshare's Proactive Chat Reference

Abstract

This presentation explores the current state of chat reference and specifically addresses using Springshare’s proactive LibChat to increase chat traffic in a natural experiment that took place at two four-year institutions over the academic year 2017-2018, in comparison with two four-year institutions that did not implement this method.
As part of a recent transition of the California State University system
to Ex Libris’ Alma/Primo, several universities embedded Springshare’s proactive chat widget on Primo search results pages (SRPs), which allowed users to begin reference transactions natively within their research sessions. In this presentation, we take advantage of a natural experiment that arose when two libraries deployed chat to SRPs, and two other libraries with similar enrollment profiles in similar geographic areas did not. By analyzing chat transactions collected from the four libraries over the course of academic years 2016-2017 (pre-deployment) and 2017-2018 (post-deployment), we calculate the effects on chat transactions when introducing proactive chat onto Primo SRPs. This quasi-experimental panel study, with two libraries in the treatment group paired with pseudo-control libraries, found that introducing proactive chat onto SRPs significantly increases the transaction count – even up to a quadrupling of the baseline in certain months.
Presenters will also explain how libraries can set up their own natural experiments as an alternative to A/B testing of Primo. We close with an overview of pros and cons regarding chat on search results pages, offering workflow suggestions for libraries considering this implementation and the corresponding increase in reference traffic that it will bring.