Opinion Mining of Customers Using Social Media and CRM Social Improvement
PRESENTER:
Vasile-Daniel Pavaloaia
"Alexandru Ioan Cuza" University of Iasi, Romania
ABSTRACT:
Within social networks, the emotional status of customers of important
brands has a significant impact on their social behavior depending on the
type of viewed posts (photos or videos). This study examines how social
media technology usage and customer-centric management systems should
contribute to a firm-level capability of social customer relationship
management (CRM). Within the current research there are investigated the
aspects that highlight the differences between the emotional impact of
Coca-Cola and Pepsi-Cola clients upon their activity on Facebook,
Instagram, Twitter and Pinterest for each type of post (photos or videos).
Drawing from the literature in social media, customer relationship
management and marketing, the contribution of this study is the
conceptualization and measurement of social media capability in the brand
companies by using various statistical tools and sentiment analysis (also
known as opinion mining).
The current research deals with a subject of great magnitude for any
company that wants to survive in the actual economy, characterized by a
continuous increase of the competitiveness. If organizations do not begin
to change their way of developing business models based on the new digital
age, according to an article written in April 2015 by Forrester
specialists, it is estimated that about 1 million B2B sales agents will
lose their jobs by 2020, in favor of e-commerce. The survival methods
conduct to an increase of sales in the social environments (social
selling) by adopting a social strategy to the set of sales tools.
Future digital experiences will be impacted by the rapid advances in
mobile technology and the evolution of Social Media. Therefore,
organizations strive to provide a synchronized experience across channels
and integrated business platforms. In this way it will be able to meet the
new level of competitive customer management as the customers are no
longer passive, but mostly proactive. Therefore, the study is structured
on six sections and contains a study regarding the emotional analysis of
customer's behavior regarding their reactions on different social
networks.
In terms of processing and interpretation of data, especially the
performance indicators, several mathematical and statistical methods were
used, especially statistical estimation for defining statistical
hypotheses that have to be confirmed or infirmed on the basis of
statistical tests within statistical comparisons. The test development
environment for hypothesis testing was R Studio together with the R
language that proposes a pack of tools for sentiment analysis purposes
(NLP - helps with data processing; TM - provides text mining features;
Syuzhet, sentiment - performs the analysis of feelings on the text; vcd,
vcdExtra - provides features for statistical tests, ggplot2, mosaic plot -
libraries for graphing).
Bio: Daniel Pavaloaia obtained his PhD in Accounting Information Systems
in 2008 and he is an Associate Professor in the Department of Accounting,
Business Informatics and Statistics at "Al. I. Cuza" University of Iasi.
His teaching subjects and research interests include Enterprise Resource
Planning, Information Technologies for Business Administration and Public
Sector, Business Process Modeling and Artificial Intelligence for
Business. Over the last 10 years, he has more than 40 individual and
co-authored articles, has authored 10 books and presented his work in more
than 20 international conferences.