טכניון מכון טכנולוגי לישראל
הטכניון מכון טכנולוגי לישראל - בית הספר ללימודי מוסמכים  
M.Sc Thesis
M.Sc StudentShelly Ashtar
SubjectThe Effect of Customer Emotion and Work Demands on
Employee Unscheduled Breaks: An Investigation of
Chat-Based Customer Service
DepartmentDepartment of Industrial Engineering and Management
Supervisors Full Professor Rafaeli Anat
Dr. Yom-Tov Galit


Abstract

The current study examines the influence of customer emotions and work demands on service employees’ withdrawal behaviors. Available research on such effects relies mainly on self-report measures of relatively small samples, and looks mostly on more severe withdrawal (absenteeism and quitting jobs). We examine chat-based service, which is unique in its great potential for automated analyses of objective measures from large samples. We study service chats conducted through the LivePerson Inc. platform (https://www.liveperson.com), and examine the influence of work demands and customer emotions (identified using a home-grown sentiment analysis tool, adapted for chat data). The dependent variable of the study is employee minor withdrawal behaviors, which are subtle withdrawal behaviors exhibited during the work shift, and we define as spontaneous, unscheduled employee breaks. With a sample of 3,084 time intervals and 835 breaks, we find that: (a) Work demands increase the likelihood and duration of employee withdrawal; (b) Customer positive emotions increase the likelihood of employee withdrawal; (c) Customer negative and positive emotions moderate the effect of work demands on duration of withdrawal behaviors; when customers express high negative emotions, higher work demands lead to longer breaks. In contrast, when customers express high positive emotions, the effect reverses, and higher work demands lead to shorter breaks. Our findings offer strong empirical support for the Job Demands-Resources model, with real-life non-obtrusive measures. The findings highlight the importance of attention to work demands and to customer emotions and open new directions for implementing sentiment analysis in designing chat platforms (e.g., determining staffing of employees and routing of customers).