Operations-driven Analytics for Multifunctional Call Centres’ Outbound Services: Automated Workforce Performance Evaluation
DescriptionOutbound call centres play a critical role in the industry of opinion survey and market research.This study focuses on the operations driven data analytics for outbound call centres. The mainfocus is to automate workforce monitoring with CATI data, including audio files, log-book data,etc. Some of these data have typical count structure with heterogeneity and sparsity.The proportion of successful interviews plays the most role to determine the operations efficiency,and cost, of an outbound call centre. It is affected by the performance of the agents such as theirwork attitude and ability to complete the calls. The response rate of outbound calls has decreaseddramatically in recent years due to reasons like tele-marking and fraud, resulted in larger turnoverrate of the call agents. It is important to develop better strategies regarding agent selection andincentive schemes to ensure sustainability of the call centres through reducing the project costswhile maintaining the talented call agents.We have accumulated intensive experiences in utilizing CATI systems during the past 20 years.We shall summarize the data structure generated by the CATI systems to build up a useful databasecapturing the information like length of the conversation, and other log-book information. Thisdatabase will be used to develop a set of indicators for measuring the performance of outboundcall centre workforce. By recognizing the heterogeneity among call agents, fieldwork supervisorsand the respondents, we shall develop regularized non-normal mixture regression models, withcomponent selection and shrinkage features. After validation, these models can identify factorssuch as conversational dynamics, workers’ attitude and characteristics that help to explain thechance of making successful interviews. The existence of inflation points in the data will also beconsidered. Because the workforce heterogeneity arises from one-time to repeated measures of anindividual agent, and from frontline agents to the fieldwork supervisors to the project manager, themodels will be developed suitable to handle cross-sectional, longitudinal, and hierarchical datastructure. The result of this project will provide a foundation for studying workforce performanceof outbound call centres based on data analytic models. Furthermore, the results can helpimproving call centre operations through identifying the characteristics and best practices of goodagents, provide better performance measures and incentive schemes to stimulate workforce’spersistence, and identify any potential misbehaviours.
|Effective start/end date||1/01/18 → 22/12/21|