Evaluate call center performance using Big Data Analytics

Abstract

An assessment of the quality of call centers (CCS) can be described as the process of listening to recorded conversations between an operator or technical support service and a customer to assess the effectiveness of the operator and its performance. The main problem with quality control is that managers or supervisors do not have time to listen to all records, and therefore only a few of the total number of saved conversation records are randomly selected. This leads to inaccurate measurements of performance, since most of the records of calls are not tapped. This article presents a distributed call monitoring system to evaluate all recorded calls using multiple quality criteria. In the proposed system, we analyze a large number of call records using the popular Hadoop MapReduce platform, and using text algorithms such as cosine transformation and N-gram. Lists of slang words were also integrated into the monitoring system. Empirical call records are used to demonstrate the performance of the proposed call monitoring system.