Employee Satisfaction in Online Reviews

Employee satisfaction impacts the efficiency of businesses as well as the lives of employees spending substantial amounts of their time at work. As such, employee satisfaction attracts a lot of attention from researchers. In particular, a lot of effort has been previously devoted to the question of how to positively influence employee satisfaction, for example, through granting benefits. In this paper, we start by empirically exploring a novel dataset comprising two million online employer reviews. Notably, we focus on the analysis of the influencing factors for employee satisfaction. In addition, we leverage our empirical insights to predict employee satisfaction and to assess the predictive strengths of individual factors. We train multiple prediction models and achieve accurate prediction performance (ROC AUC of best model = 0.89 ). We find that the number of benefits received and employment status of reviewers are most predictive, while employee position has less predictive strengths for employee satisfaction. Our work complements existing studies and sheds light on the influencing factors for employee satisfaction expressed in online employer reviews. Employers may use these insights, for example, to correct for biases when assessing their reviews.