Tweet content analysis of Indian consumer companies - a novel approach

Abstract

The rising popularity of social media has completely transformed the way corporate organizations communicate and interact with their stakeholders. Communication via social media platforms has allowed companies to develop a two-way communication channel with an unlimited number of users. With increasing influence of social media, consumer-oriented businesses are actively engaging on social media platforms and it has become an essential element of both their internal as well as external communication strategy. However, most companies are falling short in gauging and acting on customer insights generated from communication over social media. This study analyses fifty Indian companies across ten industries and investigates their usage of a widely popular social media tool – Twitter. With over 321 million monthly active users, Twitter is one of the most popular platforms for online communication (Culnan et al., 2010; Tao & Wilson, 2015). It allows its users to post tweets of up to 280 characters including images and web links. Besides actively engaging through posting tweets, users can also follow other accounts, like and react to tweets from other users and retweet (share) these tweets with their own followers. Users can link their posts to larger conversations over the platform utilizing the hashtag (#) symbol and to other Twitter accounts utilizing the mentions (@) feature. Unlike other social media channels such as Facebook, Instagram or LinkedIn, Twitter acts as an open community where users can easily access the content generated by complete strangers. These attributes make Twitter an appropriate platform for organizations to share information, communicate with relevant stakeholders, develop relationships and monitor public sentiment and opinion about them. As a part of this study, tweets generated both by the official twitter handles of companies as well users tweeting about these companies have been collected. A machine learning based algorithm has then been used to understand and classify the nature of tweets made by companies and another sentiment classification model has been used to gauge the sentiment of consumer’s tweeting about the companies. Based on the huge amount of labelled data generated from these models, an in-depth analysis has been carried out detailing the characteristics of tweets made by companies within a particular industry and the consumer sentiment about these companies.