Quantifying the Dynamic Effects of Service Recovery on Customer Satisfaction
利用中国手机市场的真实数据，通过多变量时间序列模型和贝叶斯估计，研究了服务补救对顾客满意度的动态影响，发现道歉效果最差、质量改进效果最好，为管理者提供了不同补救策略的量化效果分解。
This study examines two issues which have challenged prior experimental or survey research: (1) whether the time-varying effects of service recovery on customer satisfaction may follow a long decay or short decay and (2) why and what service recovery efforts have a higher and quicker buildup, with respect to the significance and timing of recovering customer satisfaction losses due to service failures. The authors do so with a real-world data set from China’s mobile phone markets. The authors developed multivariate time-series model to simulate the dynamic service recovery process and implemented Bayesian estimation to resolve overparameterization problem. The empirical results surprisingly reveal that apology-based service recovery efforts are the least effective in salvaging customer satisfaction, with the shortest decay and lowest buildup intensity. In contrast, quality improvement is the most effective, with the highest buildup and longest decay but slowest buildup toward the peak impact point. Compensation has moderate and stable impact overtime. Communications' impact on customer satisfaction builds up the quickest, though with mild endurance and magnitude. Also, the decomposition models enable managers to monitor how many percentages of customer satisfaction gains are originated from which types of service rescue efforts.