Emotion Detection in Email Customer Care

Proceedings of the NAACL HLT 2010 Workshop on Computational Approaches to Analysis and Generation of Emotion in Text, pages 10–16, Los Angeles, California, June 2010. c⃝2010 Association for Computational Linguistics Emotion Detection in Email Customer Care Narendra Gupta, Mazin Gilbert, and Giuseppe Di Fabbrizio AT&T Labs - Research, Inc. Florham Park, NJ 07932 - USA {ngupta,mazin,pino}@research.att.com Abstract Prompt and knowledgeable responses to cus- tomers’ emails are critical in maximizing cus- tomer satisfaction. Such emails often con- tain complaints about unfair treatment due to negligence, incompetence, rigid protocols, un- friendly systems, and unresponsive personnel. In this paper, we refer to these emails as emo- tional emails. They provide valuable feedback to improve contact center processes and cus- tomer care, as well as, to enhance customer re- tention. This paper describes a method for ex- tracting salient features and identifying emo- tional emails in customer care. Salient fea- tures reflect customer frustration, dissatisfac- tion with the business, and threats to either leave, take legal action and/or report to au- thorities. Compared to a baseline system us- ing word ngrams, our proposed approach with salient features resulted in a 20% absolute F- measure improvement. 1 Introduction Emails are becoming the preferred communication channel for customer service. For customers, it is a way to avoid long hold times on call centers phone calls and to keep a record of the information ex- changes with the business. For businesses, it of- fers an opportunity to best utilize customer service representatives by evenly distributing the work load over time, and for representatives, it allows time to research the issue and respond to the customers in a manner consistent with business policies. Busi- nesses can further exploit the offline nature of this channel by automatically routing the emails involv- ing critical issues to specialized representatives. Be- sides concerns related to products and services, busi- nesses ensure that emails complaining about unfair treatment due to negligence, incompetence, rigid protocols and unfriendly systems, are always han- dled with care. Such emails, referred to as emotional emails, are critical to reduce the churn i.e., retain- ing customers who otherwise would have taken their business elsewhere, and, at the same time, they are a valuable source of information for improving busi- ness processes. In recurring service oriented businesses, a large number of customer emails may contain routine complaints. While such complaints are important and are addressed by customer service represen- tatives, our purpose here is to identify emotional emails where severity of the complaints and cus- tomer dissatisfaction are relatively high. Emotional emails may contain abusive and probably emotion- ally charged language, but we are mainly interested in identifying messages where, in addition to the flames, the customer includes a concrete descrip- tion of the problem experienced with the company providing the service. In the context of customer service, customers express their concerns in many ways. Sometimes they convey a negative emotional component articulated by phrases like disgusted and you suck. In other cases, there is a minimum emotional involvement by …

Content truncated to 500 words