Good Abandonment in Legal Research

Posted: 23 Nov 2021

Date Written: September 21, 2021

Abstract

Modern search systems assume a query feedback loop where the user expresses their intent and signals usefulness of results by explicit engagement actions. A key search innovation has been the use of engagement events, such as explicit document access, or dwell time on search results or documents, to improve re-ranking models. Further, query-engagement patterns are taken as signals of customer states including satisfaction or confusion. Search interfaces have evolved from ten blue links to a complex presentation rich with knowledge units, such as answer cards and search heat maps. Users may acquire the needed information silently without explicit engagement signals visible in logs.

Good query abandonment occurs when the user satisfies their information need without explicit engagement or further query refinement. We present search system experiments demonstrating instances of good abandonment, along with a model to recognize good abandonment in session logs. We discuss the implications for search system performance and user search metrics, which are central to new feature testing and search system improvements.

Keywords: Good abandonment, online metrics, multi-component search page

Suggested Citation

Khazaeli, Soha and Cole, Michael, Good Abandonment in Legal Research (September 21, 2021). Proceedings of the 5th Annual RELX Search Summit, Available at SSRN: https://ssrn.com/abstract=3970259

Soha Khazaeli (Contact Author)

LexisNexis ( email )

P. O. Box 933
Dayton, OH 45401
United States

Michael Cole

LexisNexis ( email )

P. O. Box 933
Dayton, OH 45401
United States

Do you have a job opening that you would like to promote on SSRN?

Paper statistics

Abstract Views
140
PlumX Metrics