Whose and what chatter matters? The effect of tweets on movie sales

Abstract

a b s t r a c tSocial broadcasting networks such as Twitter in the U.S. and "Weibo" in China are transforming the way online word of mouth (WOM) is disseminated and consumed in the digital age. In the present study, we investigated whether and how Twitter WOM affects movie sales by estimating a dynamic panel data model using publicly available data and well-known machine learning algorithms. We found that chatter on Twitter does matter; however, the magnitude and direction of the effect depend on whom the WOM is from and what the WOM is about. Incorporating the number of followers the author of each WOM message had into our study, we found that the effect of WOM from users followed by more Twitter users is significantly larger than those followed by less Twitter users. In support of some recent findings about the importance of WOM valence on product sales, we also found that positive Twitter WOM is associated with higher movie sales, whereas negative WOM is associated with lower movie sales. Interestingly, we found that the strongest effect on movie sales comes from those tweets in which the authors expressed their intention to watch a certain movie. We attribute this finding to the dual effects of such intention tweets on movie sales: the direct effect through the WOM author's own purchase behavior, and the indirect effect through either the awareness effect or the persuasive effect of the WOM on its recipients. Our findings provide new perspectives to understand the effect of WOM on product sales and have important managerial implications. For example, our study reveals the potential values of monitoring people's intentions and sentiments on Twitter and identifying influential users for companies wishing to harness the power of social broadcasting networks.

Identifying influencers in a social network: The value of real referral data

“…Empirical research confirmed that consumers rely heavily on the advice of others in their personal network when making purchase decisions [19,36,32,20,33] and that positive WOM has a positive effect on business outcomes, i.e. sales [31,3]. Referral marketing has become an important marketing technique to stimulate WOM in a controlled way for acquiring new customers [5].…”

Consumers' Sentiments and Popularity of Brand Posts in Social Media: The Moderating Role of Up-votes

“…Based on a general consensus, people's attitude toward products and purchasing them can be increased or decreased by exposure to positive and negative UGC respectively [20]. For instance, positive WOM on twitter is positively associated with movie sales and negative WOM on twitter is negatively associated with movie sales [26]. Accordingly, we propose following hypotheses:…”

Arabic Sentiment Analysis: A Survey

“…They showed that NN achieves better performance than SVM on balanced datasets. Rui and Liu [20] investigated pre-consumer (prior to purchase) and post-consumer (after purchase) opinion differences using NB and SVM classifiers on twitter data from both classes of users. Li and Li [21] addressed subjectivity and expresser credibility in opinion studies using SVM as the classifier.…”

“…As we can see from the above [24], [29], [70] 2 SVM + NB [7], [9], [10], [11], [27], [39], [41], [42], [48] Also we can see that the Saudi dialect was not given attention by the researchers. [2], [3], [4], [5], [6], [9], [13], [19], [20], [22], [23], [24], [33], [39], [41], [42], [46], [48], [49], [65], [70] 2 MSA (Egyptian) [7], [10], [11], [21], [27], [28], [51], [70] 3 MSA (Levantine) [7] 4 MSA (Khaliji) [7], [11], [65] 5 MSA (Arabizi) [7] 6 MSA (Mesopotamian) [11] 7…”

“…Table 1 below categorizes the surveyed Arabic SA studies into these classes, and fig.1 shows the result. [7], [9], [10], [11], [13], [19], [20], [27], [28], [29], [33], [41], [42], [48], [49], [52], [65], [69] Arabic sentiment analysis studies used different method based on the used approach, some of these methods considered as a dominance. We have collected the various methods used in the different Arabic SA studies for supervised, semi-supervised, unsupervised and hybrid experiments and presented the result in Table 2.…”