QED Q's: Quality - and Efficiency-Driven Call Centers Data-Based Service-Research: Science, Engineering, Management
<-- Return to the list
Date: 09-11-2008
Start Time:
1:00pm
End Time: 2:00pm
Speaker: Avishai Mandelbaum, Technion
Location: Warren 208
ABSTRACT
This seminar starts with a bird’s-eye view of Service Engineering and Telephone Call Centers. Then some empirical findings of service systems will be presented, which motivate or are motivated by (or both) interesting research questions. These findings give rise to model-features that are essential to capture in useful service models, for example customers’ (im)patience, time-varying service demand (predictable variability), heterogeneity of customers and servers (skills-based routing), over-dispersion in Poisson arrivals, generally-distributed (as opposed to exponential) service and patience duration, and more. Empirical analysis also enables validation of prevalent models and protocols, either supporting or refuting their relevance and robustness.
Our main data-source is a repository of call centers data, the birth of which was a joint effort of the Wharton School and the Technion’s IE&M faculty. This repository is maintained and developed at the Technion’s SEE Center. The data is unique in that it is transaction-based: it details the individual operational history of all the calls handled by the participating call centers. For example, one source of data is a network of 4 call centers of a U.S. east-coast bank, spanning close to 3 years and covering over 800 agents; there are 218,047,488 telephone calls overall, out of which 41,646,142 were served by agents, while the rest were handled by an answering machine (IVR = Interactive Voice Response).
Data-bases within most call centers are inadequate for operations research. Hence a universal data-structure (schema) had to be designed and implemented, under the heading DataMOCCA = Data Models for Call Centers Analysis.A friendly flexible user-interface is accompanying DataMOCCA, which we call SEEStat. This interface will be demonstrated, in particular its ability for online analysis, at resolutions that span the whole range from single seconds through minutes, hours, weeks and up to months.The SEE respositories and the scope of DataMOCCA are now being extended to accommodate healthcare and internet services. The underlying theme is again the archiving of transaction-based operational history of patients (eg. Within the ED = Emergency Department) or surfers (within the websites, via clickstream data), all interfacing with SEEStat for online analysis, similarly to call centers.
I shall have with me DVD’s with data from the above-mentioned U.S. bank, as well as SEEStat (the DataMOCCA interface) and some relevant tutorials and documentation. You are encouraged to copy this material to your PC and experiment/play with the data yourself. (Note: The DVD has one-year worth of data; a full version of the raw (cleaned) U.S. bank data takes close to 40GB and an abridged version (tables) is about 7GB – please approach me if interested in these comprehensive versions).
BIO
Professor Avi Mandelbaum has a B.Sc. in Mathematics and Computer Science and an M.A. in Statistics, both summa cum laude from Tel-Aviv University. His PhD is in Operations Research, from Cornell University. After graduation, in 1983, he joined the Graduate School of Business at Stanford University. He left the U.S.A. in 1991, to assume a position at the Technion. Professor Mandelbaum has been an associate editor of the leading journals in his field, in particular Math. of OR, Management Science and QUESTA (Queuing Theory). Professor Mandelbaum’s teaching has been acknowledged by Technion prizes, in particular the Technion Excellence in Teaching Award for the course “Service Engineering”. His research has focused on stochastic models (analysis, asymptotics, control) and statistics, with applications to service systems, notably tele-services (eg. telephone call centers) and healthcare (eg. patients’ flow through emergency departments). His research has been acknowledged by prizes, for example the inaugural Markov Lecture of the Applied Probability Society, INFORM 2005. Professor Mandelbaum is the founder and director of the Technion’s SEE Center/Laboratory (SEE = Service Enterprise Engineering), which serves as a focal point for service research at the Technion, as well as a data repository for call centers and hospitals.