Abstract
Rapid growth of market competition has made e-commerce enterprises attach increasing importance to customer relationship management. The call center with a modern CRM system have become a vital carrier for high quality customer service. This study examines the influence that interaction justice and emotion have on call center employees' performance. A total of 267 call audios were collected from an e-commerce enterprise. The theoretical model is tested based on two-stage least square regression. Results show that call center employees who are treated unfairly by customers during calls yield worse performance than those who are treated fairly. The above relationship is partially mediated by employees' negative emotion.
Introduction
In the high-stakes world of e-commerce, Customer Relationship Management (CRM) systems are often viewed through a purely technical lens: better software, faster databases, and smarter algorithms. However, the human element—the call center agent—remains the primary interface for these systems.
While existing research heavily focuses on how to improve customer satisfaction, there is a significant gap in understanding how customer behavior affects the employees themselves. This study addresses the "customer is always right" paradigm by asking: What happens to service quality when customers are unfair, disrespectful, or vague? Specifically, it examines how "interaction justice" (or the lack thereof) influences an agent's emotional state and their subsequent ability to resolve issues or close sales.
Theoretical Framework
The researchers break down "justice" into two critical types. Interpersonal justice refers to politeness and respect; if a customer is abusive, the agent's performance drops. Informational justice refers to the quality of data provided; if a customer is vague, the agent cannot effectively use the CRM to solve the problem.
As illustrated in the conceptual model above, customer rudeness doesn't just distract an agent; it actively depletes their psychological resources. The agent must expend "emotional labor" to regulate their feelings, leaving less cognitive energy for the actual task.
Methodology
The study moved away from traditional self-reported surveys, which are often biased by "impression management." Instead, researchers analyzed a proprietary dataset of 267 valid call audios (approximately 240 minutes) from a Chinese e-commerce enterprise.
- Objective Scoring: Researchers listened to the calls and scored them on a 5-point Likert scale for interaction justice, employee emotion (based on tone), and call results (customer willingness to proceed).
- Statistical Modeling: The study utilized a Two-Stage Least Square (2SLS) Regression Model and a three-step mediation analysis to determine if emotion acts as the "bridge" between customer behavior and work performance.
Key Results
The statistical analysis supported all primary hypotheses. Both interpersonal and informational injustice significantly lowered instant call performance ($p < 0.001$). Furthermore, unfair customer behavior was a direct predictor of negative employee emotions.
As shown in the mediation analysis table, emotion was proven to be a partial mediator. When emotion was added to the model (M4 and M5), the direct effect of customer behavior on results decreased. For instance, the coefficient for interpersonal justice dropped from $0.398$ to $0.323$, proving that a large portion of the performance drop is due to the agent's emotional reaction to the interaction.
Conclusion and Future Work
The findings suggest that e-commerce enterprises should look beyond technical CRM features to support the emotional well-being of their staff. Future CRM systems might include algorithms to identify and filter "low-intent" customers who are more likely to exhibit abusive behavior. Additionally, the paper suggests exploring real-time emotional "nudges" on an agent's dashboard to help them regulate their emotions mid-call, thereby protecting both employee productivity and mental health.
Summary updated