Operations Management of Call Centers – OPERA

The models and quantitative methods to be used in this project come from the discipline of operations research. They are the stochastic models (queueing systems, Markov chains), the mathematical programming (mainly stochastic programming), the discrete-event simulation, the applied probability for the analysis of transient and stationary Markov chains (computation of system state probabilities, computations of first passage times which models waiting times, etc.).

We are finalizing a first study on the routing optimization between calls and emails in call centers where a natural break is included in the middle of a call conversation. The results of this work were presented at the international conference ICMSAO'13 at Hammamet, Tunisia at the end of April 2013. The associated paper will appear in the proceedings of this conference. A paper (a longer and more completed version) was also submitted to the international journal Stochastic Systems.

From now until mid 2014, we will work on 2 research questions related again to real-time operations:

- Adaptative optimization of the agent staffing dedicated to emails. Here we consider classic call conversations without a break. We have just started this project. The objective is to develop threshold routing policies that are adapted to the load variation over a working day. Since the arrival parameters are in reality non stationary, a stationary regime does not exist, which makes the determination of the routing policies a hard task. We will then proceed by approximation. We first assume that the arrival parameters are stationary, and derive the associated routing policies. Then, we adapt these routing policies to the non stationary case. We aim at submitting a related paper to an international journal.

- Optimization of a call center with a call back option. This case treats call centers where a customer has the choice to wait for service, or to give up and gets called back later by an agent in case he estimate that his wait would be long. The objective of this work is to optimize the staffing under service level constraints in terms of direct waiting (for the inbound calls) and indirect waiting (for the customer that agents call back). We again aim at submitting a related paper to an international journal.

From mid 2014 until the end of this project in September 2015, we want to work on two additional research questions:

- Development of flexible call center architectures.

- Analysis of call centers with delay prediction and announcement (impact of delay information on the customer behavior).
The results of these two works will also be submitted for publication (international conferences and journals of operations management/operations research).

Benjamin Legros, Oualid Jouini, Ger Koole. Imbricating Tasks in a Multichannel Contact Center. In Proceedings of ICMSAO'13, Hammamet, Tunisie, April 2013.

Virtually all businesses are interested in providing information and assistance to existing and prospective customers. In recent years, the decreased costs of telecommunications and information technology have made it increasingly economical to consolidate such information delivery functions, which led to the emergence of groups that specialize in handling customer phone calls. For the vast majority of these groups, their primary function is to receive telephone calls that have been initiated by customers. Such operations, known as “inbound” call centers, are the topic of this proposal. Inbound call centers may be physically housed across several different locations, time zones, or countries.

Given the size of the call center industry and the complexity associated with its operations, call centers have emerged as a fertile ground for academic research. Aksin et al. (2007) provide a survey of the academic literature associated with traditional call center problems areas. They point out also the lack of research on call center operations management integrating psychological and behavioral aspects, which is by far more relevant and useful in practice. According to marketing theory for example, customer satisfaction toward the interaction will depend on the capacity of firms in identifying and understanding the related attitudinal process during the call. The objective of this project is to contribute to the literature related to these key emerging phenomena. It is included under a multi-disciplinary perspective on operations management research.

In this project, we will consider various problems related to the three decision levels of the operations management of call centers; long term, mid-term and operational term. The first problem, related to strategic decisions, is to develop and analyze flexible architectures of multi-skill call centers. The second problem, related to mid-term decisions, is to optimize personnel staffing. The last problem, related to operational decisions, is to develop and analyze call center models with delay information. Our approach consists on developing and analyzing queueing models that incorporate various customer and agent behavior features. This new multi-disciplinary perspective would enhance our understanding of such complex systems, and allows to gain useful guidelines for the practitioners.

To the contrary to traditional works where usually a pure queueing analysis is performed, we will account here for the impact of the human element on the call center performance. For the design problems, we will focus on the impact of the human resource management of the agents on performance. We do so by accounting for the agent responsibility and motivation as a function of the design of the call center teams. As for the delay announcement problem, we will study the customer reaction to delay information. We will model how customers would react in terms of balking, abandonment, dissatisfaction because of waiting uncertainty, etc. These novel developed models will be quantitatively analyzed using mainly stochastic processes methods. Various managerial insights will be derived thereafter.