A webometric network analysis of electronic word of mouth (Ewom) characteristics and machine learning approach to consumer comments during a crisis

Abstract

This study explores the effectiveness of crisis response strategies for public response and perception in the context of social media by examining a case about the Samsung Galaxy Note 7 product recall crisis. First, the study investigated the response strategies Samsung used on Facebook through the lens of situational crisis communication theory (SCCT). Next, we applied a webometric network analysis and exponential random graph models (ERGM) to analyze the relationship between the crisis response strategies and the dynamics of electronic word of mouth (eWOM) behaviors. Then, we performed topic modeling and semantic network analysis to examine the public perceptions of and responses to Sam-sung’s crisis communication strategies based on public comments. Samsung used silence, information, and rectification strategies. More participants and comments were generated and stronger ties were found in the eWOM networks for matched responses than for silence. Public responses and perceptions toward the silence and the late adoption of an information strategy were primarily negative and resulted in complaints about poor customer service, whereas positive responses –expressing brand royalty and forgiveness– increased via the rectification strategy. Methodological triangu-lation in this study offers evidence-based lessons on how to systemically monitor stakeholders’ reactions and manage consumer complaints in order to repair a corporation’s damaged reputation after a crisis. Keywords.