The study used a time-sampling method to test aspects of A. Grandey's (2000) emotion regulation model of emotional labor. Eighteen customer service employees from a call center recorded data on pocket computers every 2 hr at work for 2 weeks. Participants completed ratings of emotion regulation, events, expressed and felt emotions, well-being, and performance on 537 occasions and completed questionnaires containing individual and organizational measures. Multilevel analyses supported many aspects of the model but indicated that it has to be implemented precisely in terms of regulating emotion for organizational goals. Results also showed that deep and surface acting had different consequences for employees. Overall, the study found that emotion regulation is a viable platform for understanding emotional labor.