Posting behaviour of consumers and the impact of social media on product sales

As social media permeates daily life, it attracts attention from both researchers and practitioners. Prior research has demonstrated the relevance of social media mainly by studying its impact on product sales in various settings. However, little is known regarding differences in social media across sources, time and platforms. In order to fill this gap, this dissertation conducts two studies differentiating social media opinions in research on purchasing and posting behaviour. The first study investigates the relationship between the sentiment embedded in online comments (i.e. valence), the quantity of comments (i.e. volume) available on a movie and box office receipts. The results show that the persuasive effect of different online sources fluctuates as time evolves. Our findings suggest that online opinions are strong predictors of sales, but the different sources of online opinion are not equal in impacting product sales. This study demonstrates that online opinions are not always persuasive and useful, and our findings yield insights into when consumers are likely to pay attention to online opinions. The second study examines how different social media platforms influence opinion composition and evolution. We differentiate between product and non-product oriented outlets as they differ in the salience of social cues, thus resulting in distinct user behaviours. We extend prior research in several ways. First, comparing between comments from different types of social media platforms, we show that the product oriented outlets display a tendency to attract polarized opinions. Second, we find that similarity of online comments increases over time, suggesting opinion convergence. In addition, product oriented outlets facilitate faster assimilation of opinions within the site compared to non-product oriented outlets. Empirically, the two studies are different from most social media research which focuses on single source and relies on self-reported ratings to measure comment valence. To compare online opinions across sources and platforms, we tracked over 1500 sources of online expert and consumer reviews for cinematic movies released for an entire year and continuously monitored major social media sites (e.g. Twitter and Plurk) for comments. To avoid data loss due to the lack of self-reported ratings, we text mined the comments to elucidate the sentiments and analyzed the data. These two studies caution researchers and practitioners against treating online opinions from different sources as the same, and highlight that what is posted online may not truly mirror consumer opinions. Study 1 shows that influence of online opinions on cinematic movie box office is dynamic and varies across sources and platforms. The importance to appreciate platform difference is again illustrated in Study 2, which finds that online posts in some social media platforms are relatively polarized, and become more assimilated over time.DOCTOR OF PHILOSOPHY (NBS