Customer Sentiment in Web-Based Service Interactions: Automated Analyses and New Insights

@article{YomTov2018CustomerSI,
  title={Customer Sentiment in Web-Based Service Interactions: Automated Analyses and New Insights},
  author={Galit Bracha Yom-Tov and Shelly Ashtar and Dan Altman and Michael Natapov and Neta Barkay and Monika Westphal and Anat Rafaeli},
  journal={Companion Proceedings of the The Web Conference 2018},
  year={2018}
}
We adjust sentiment analysis techniques to automatically detect customer emotion in on-line service interactions of multiple business domains. Then we use the adjusted sentiment analysis tool to report insights about the dynamics of emotion in on-line service chats, using a large data set of Telecommunication customer service interactions. Our analyses show customer emotions starting out negative and evolving into positive as the interaction ends. Also, we identify a close relationship between… 

Figures and Tables from this paper

Do Customer Emotions Affect Agent Speed? An Empirical Study of Emotional Load in Online Customer Contact Centers
TLDR
The concept of emotional load is introduced—the emotional demands that customer behaviors impose on service agents—to analyze how customer emotions affect service worker’s behavior and suggests that the emotional content in customer communications should be an important factor to consider when assigning workload to agents in a service system.
Integrating emotional load into service operations
Emotions are clearly important. They drive our behavior and how we behave drives our relationships, with family, friends and coworkers. Emotions are influential both in the social world and in the
A Layered Approach for Summarization and Context Learning from Microblogging Data
TLDR
This paper presents a layered graph-based approach using both content and structural data to analyze and summarize tweets at different levels of granularity, which is evaluated over four different real-world Twitter datasets using standard information retrieval metrics.
The Co-Production of Service: Modeling Service Times in Contact Centers Using Hawkes Processes
TLDR
This paper proposes, develops, and compares new stochastic models for the co-production of service in a contact center, and finds that the bivariate Hawkes processes that incorporate the amount of information provided by each party or the sentiment expressed by the customer give the most accurate predictions.
Computational modelling of speech data integration to assess interactions in B2B sales calls
TLDR
The study demonstrates that using delta metrics to assess the interactions leads to new insights and modeling the dynamics between sales representatives and customers in business-to-business sales calls, by relying solely on the acoustic signal.
Psychology of Business Intelligence Tools: Needs-Affordances-Features Perspective
TLDR
The findings suggest that psychological needs motivate the use of BI tools that provide 13 affordances to fulfill five psychological needs, namely autonomy, competence, relatedness, having a place and selfrealization.
Digital Traces: New Data, Resources, and Tools for Psychological-Science Research
New technologies create and archive digital traces—records of people’s behavior—that can supplement and enrich psychological research. Digital traces offer psychological-science researchers novel,
Tüketici Yorumları Üzerine Bir Metin Madenciliği ve Veri Boyutu İndirgeme Yaklaşımı
Veri boyutunun artmasiyla orantili olarak degiskenler arasi baglantilarin analizi daha karmasik bir hale gelmistir. Yapisal olmayan veri kumelerinde, yapisal forma donusturme on islemleriyle
Cognitive and Emotional Load Influence Response Time of Service Agents: A Large Scale Analysis of Chat Service Conversations
TLDR
Use of operational data of 145,995 chat service conversations shows that cognitive load and emotional load increase agent response time both between and within service conversations.
User profiling and satisfaction inference in public information access services
TLDR
A number of models are built using a range of supervised machine learning methods and it is shown that these outperform textual and non-textual baseline alternatives alike, making a case in favour of NLP methods for these tasks, and paves the way for further applications in the public information access domain.
...
...

References

SHOWING 1-10 OF 40 REFERENCES
Predicting Customer Satisfaction in Customer Support Conversations in Social Media Using Affective Features
TLDR
It is shown that affective features such as customer's and agent's personality traits and emotion expression improve prediction of customer satisfaction when added to more typical text based features.
Unveiling What Is Written in the Stars: Analyzing Explicit, Implicit, and Discourse Patterns of Sentiment in Social Media
TLDR
This study provides a fine-grained analysis of the implicit and explicit language used by consumers to express sentiment in text and demonstrates the differential impacts of activation levels, implicit sentiment expressions, and discourse patterns on overall consumer sentiment.
An Unsupervised Aspect-Sentiment Model for Online Reviews
TLDR
An unsuper-vised system for extracting aspects and determining sentiment in review text is presented, which is simple and flexible with regard to domain and language, and takes into account the influence of aspect on sentiment polarity.
Exploring Customer Loyalty Following Service Recovery
Existing research shows that loyalty is a function of customer perceptions of trust following service recovery. The authors propose a cognitive appraisal model that portrays trust and emotions as key
Does Sustained Participation in an Online Health Community Affect Sentiment?
TLDR
There is a significant increase in sentiment of posts through time, with different patterns of sentiment trends for initial posts in threads and reply posts.
From bad to worse
Poor employee–customer interactions influence customer satisfaction and employee well-being. In studying these negative exchanges, researchers tend to take either the perspective of the customer
Lexicon-Based Methods for Sentiment Analysis
TLDR
The Semantic Orientation CALculator (SO-CAL) uses dictionaries of words annotated with their semantic orientation (polarity and strength), and incorporates intensification and negation, and is applied to the polarity classification task.
Sentiment Analysis of Twitter Data
We examine sentiment analysis on Twitter data. The contributions of this paper are: (1) We introduce POS-specific prior polarity features. (2) We explore the use of a tree kernel to obviate the need
Sentiment strength detection in short informal text
A huge number of informal messages are posted every day in social network sites, blogs, and discussion forums. Emotions seem to be frequently important in these texts for expressing friendship, sho...
...
...