Social Media Effects on Human Brand Sampling: A Non-Linear Time Series Analysis: An Abstract

Abstract

Human brands, or well-known personas who are the subject of marketing communication efforts (Thomson 2006) have millions of fans who enjoy their offerings. However, extensive research has shown that these fans may in fact identify with human brands on a far deeper level, as doing so provides them with an opportunity to feel good about themselves (Chernev et al. 2011) and express and validate their identity by self-defining themselves as part of a group (Chernev et al. 2011). With the growing popularity of social media, group identification is easy to accomplish, as consumers can “follow” human brands and become part of their online brand communities. As this relationship is established, content sharing (i.e., word of mouth) begins to occur throughout a follower’s own social network, which includes individuals who are not part of the human brand community. For these individuals, word of mouth can “clear the air” regarding a human brand and its offerings, and they likely trust it more than typical advertising sources (Faber and O’Guinn 1984). This can easily benefit human brands such as musical artists, who regularly post samples of their information/experience goods (e.g., YouTube music videos) to social media with the goal of consumers engaging with them and subsequently forming higher-order beliefs (Marks and Kamins 1988) and becoming less uncertain of the level of quality (Gopal et al., 2006). This process can lead to greater purchase intentions (Liebowitz 1985) and ultimately more sales (Bawa and Shoemaker 2004). The goal of the current research is to study the time-varying dynamic between the aforementioned social media drivers and musical artist sampling behavior, a research gap, which differs from prior literature that has focused directly on sales (Saboo et al. 2016; Dewan and Ramaprasad 2012). The social media variables analyzed include the growth of a brand community (GBC) and the extent of content sharing (EOS), which impact sampling of a musical artist’s YouTube music videos (Views). To account for endogeneity and stationarity concerns of these time series variables, an autoregressive distributed lag model is estimated, in addition to a non-linear model, which allows for potential asymmetric effects. Across a sample of 20 musical artists, this research finds that while most artists’ Views is impacted by GBC and EOS in the short-run, only half of the artists receive a sustainable impact in the long-run. Asymmetric effects are found, which suggest that the impact of GBC or EOS on Views varies depending on whether an increase or decrease occurs.