Predicting e-Mail Response Time in Corporate Customer Support

Anton Borg, Jim Ahlstrand, Martin Boldt

2020

Abstract

Maintaining high degree of customer satisfaction is important for any corporation, which involves the customer support process. One important factor in this work is to keep customers’ wait time for a reply at levels that are acceptable to them. In this study we investigate to what extent models trained by the Random Forest learning algorithm can be used to predict e-mail time-to-respond time for both customer support agents as well as customers. The data set includes 51,682 customer support e-mails of various topics from a large telecom operator. The results indicate that it is possible to predict the time-to-respond for both customer support agents (AUC of 0.90) as well as for customers (AUC of 0.85). These results indicate that the approach can be used to improve communication efficiency, e.g. by anticipating the staff needs in customer support, but also indicating when a response is expected to take a longer time than usual.

Paper Citation

in Harvard Style

Borg A., Ahlstrand J. and Boldt M. (2020). Predicting e-Mail Response Time in Corporate Customer Support.In Proceedings of the 22nd International Conference on Enterprise Information Systems - Volume 1: ICEIS, ISBN 978-989-758-423-7, pages 305-314. DOI: 10.5220/0009347303050314

in Bibtex Style

@conference{iceis20,
author={Anton Borg and Jim Ahlstrand and Martin Boldt},
title={Predicting e-Mail Response Time in Corporate Customer Support},
booktitle={Proceedings of the 22nd International Conference on Enterprise Information Systems - Volume 1: ICEIS,},
year={2020},
pages={305-314},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0009347303050314},
isbn={978-989-758-423-7},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 22nd International Conference on Enterprise Information Systems - Volume 1: ICEIS,
TI - Predicting e-Mail Response Time in Corporate Customer Support
SN - 978-989-758-423-7
AU - Borg A.
AU - Ahlstrand J.
AU - Boldt M.
PY - 2020
SP - 305
EP - 314
DO - 10.5220/0009347303050314