Analyzing Call Center Performance: A Data Mining Approach

Society is becoming more accustomed to toll-free numbers as an efficient way to request and receive services in all aspects of their lives. While a move can be observed to eliminate humans as handlers of most rudimentary customer requests, responding to telephone calls remains a top priority in customer service. Call centers are either managed in-house or contracted out and provide a variety of services. The performance of the call center depends on the performance of its customer service representatives and the call handling regulations. The research aims to apply some well-known data mining techniques such as neural networks, classification and regression trees, support vector machines and a hybrid decision tree – neural network approach to the problem of predicting the quality of service in call centers; based on the performance data actually collected in call centers of a large insurance company. We also applied the apriori association rule mining algorithm to find interesting features among the variables. We first compared the performance of models built using the above-mentioned techniques and then we analyzed the characteristics of the input sensitivity in order to better understand the relationship between the performance evaluation process and the actual performance to help improve management and performance of call centers.