Identifying Brand Sentiment Through Analytics: An Abstract

Right now, hundreds of thousands of data are generated and added to the accumulated data. This immense amount of real-time and retrospective data has helped change the previous paradigms and led to a new one: big data analytics. Social networks, blogs, social bookmarking, and review sites are considered very important in the big data era. Nowadays, there is strong interest among academics and practitioners in studying branding issues through analytics. In this article, given the necessity of monitoring the perceived value of brand authenticity, “to protect a popular brand against the heartbreak of genericide” (Walsh, 2013), we examine the sentiments toward a brand, via brand authenticity, to identify the reasons for positive or negative sentiments on social media. Moreover, to increase sentiment precision, we investigate sentiments polarity on a five-point scale. From a database containing 2,988,560 tweets with the keyword “Starbucks,” we use a set of 1857 coded tweets both for brand authenticity and sentiment polarity. We analyze the data to establish a framework in which we predict both the brand authenticity dimension and its sentiment polarity. Results from support vector machine (SVM) analyses illustrate the effectiveness of the proposed procedure of brand sentiment analysis. It shows high accuracy for both the brand authenticity dimensions’ prediction and its sentiment polarity. In summary, this research contributed both theoretically and managerially to the brand sentiment literature. The proposed procedure and the research findings on brand authenticity sentiment analysis could facilitate further inquiries into sentiment analysis for all other brand constructs and in several related domains, such as e-word-of-mouth studies. Practically speaking, this research could provide marketing practitioners with a reliable and valid instrument to evaluate the level of sentiments toward a brand more specifically and more accurately, which could lead to proposing appropriate strategies to strengthen their brand authenticity. This study has some limitations that offer opportunities for further research. First, we studied brand authenticity sentiments with 2Q cross-sectional data, while we acknowledge that customer sentiments could even change over a short period of time. Second, we did not consider the brand’s following-up interventions about the shared tweet, which could change upcoming sentiments. Third, as we collected the data in a real-time manner, we did not have the data about the amount of likes and retweets a post received, which could show the effectiveness of the tweet.