Article Preview
TopIntroduction
Calls centers are key organizational structures in a wide variety of industries, including the insurance industry (Callaghan & Thompson, 2001). Call centers were developed in part based on work done by Agner Erlang, the originator of traffic engineering and queuing theory (Angus, 2001). Erlang C is a mathematical formula that can be used to predict the most probable distribution of incoming calls based on historical data. By using the incoming call distribution, the appropriate number of phone lines and staff can be determined based on the trade-offs between costs and service quality (Townsend, 2007).
However, there is disagreement on the role and purpose of call centers, and there are two views of the cost vs. service trade-off. One view is that call centers are used by organizations as a way to reduce costs with customer service delivery a secondary consideration. The other view is that call centers can increase profits by maximizing customer service (Robinson & Morley, 2006; Li, Tan, & Xie, 2003). From either perspective a key concern of companies is “stickiness” or lock-in. Customers are more likely to leave or switch to another company if they have a bad experience or receive low service quality, and this can be true more-so in comparison with several other types of business interactions (Keiningham, Aksoy, Andreassen, Cooil, & Wahren, 2006). From either perspective, the effectiveness and service level provided by a call center is vital to the competitiveness of an organization (Lam & Lau, 2004).
The metric used for service levels in this study is a commonly used metric by call centers (Koole, 2003), used in industries and call centers well beyond the insurance call center in this study. Prior work shows it is related to customers’ satisfaction with their call center experiences (Cronin, Brady, & Hult, 2000); in this study we examine the degree to which other variables impact service levels to help decision-makers more fully understand the metric they use, and to help them reallocate resources to maximize their service levels.
Determining which factors most significantly service levels is not straight forward, and approaches such as considering the number of calls answered does not necessarily equate to service. Traditionally call center service levels are based on capacity optimization for a given volume of calls. Our analysis shows that call volume is not related to service levels, implying that capacity is sufficient enough not have an impact on service levels. However, since neither capacity nor call volume explains service levels, managers are left wondering which variables do have effects.
Companies collect large amounts of data from their daily operations to help manage call centers. Computer-based decision support models have the advantage of sharpening information-processing skills (Curry & Moutinho, 1994), and the importance of implementing business intelligence tools to analyze and use this data is increasingly realized by many organizations. This paper demonstrates how business intelligence (BI) can be used to identify and analyze different factors affecting call center service levels in the insurance industry, which can lead to improved customer service while at the same time possibly maintaining or reducing costs.