The moderating role of socio-semantic networks on online buzz diffusion

Abstract

Many studies show that online buzz influences consumer's behavior. Relatively few studies empirically explore its diffusion, however. The source and content of online information influences people. This study aims to determine the direct effect of the volume of early online WOM (Word of Mouth) on its subsequent diffusion and the moderating effects of WOM characteristics in both the social and semantic networks in its initial 14-day stage of diffusion. To accomplish the objectives of the study, a specialized agency for online buzz research provided raw data regarding online WOM for 40 cellular phones. This study analyzes three stages of buzz diffusion. In the first stage, we performed a social network analysis of poster–replier relationships to capture the social aspects of online WOM networks. In the next stage, we extracted morphemes from the content corpus and conducted a semantic network analysis of online WOM. The five network characteristic indexes extracted through this analysis were density, degree centrality, betweenness centrality, eigenvector centrality and path length. In the last stage, we employed a 1st level and 2nd level HLM (hierarchical linear model). The results identify the partially direct and moderating factors affecting online WOM diffusion. This study contributes to the understanding of the role of social and semantic network indexes in online WOM diffusion.