As literature on Word of Mouth (WOM) emphasizes the needs for analyzing dynamic effects of WOM and for examining effects of WOM in online communication channel level, I perform Dynamic Linear Model (DLM) in order to analyze the dynamic effects of WOM ...
As literature on Word of Mouth (WOM) emphasizes the needs for analyzing dynamic effects of WOM and for examining effects of WOM in online communication channel level, I perform Dynamic Linear Model (DLM) in order to analyze the dynamic effects of WOM in multiple types of online communication channels. To explain such effects, I consider two important attributes of WOM: Volume and Valence, copy wear-out effects: important factor for implementing the dynamic system, and finally interaction effects among WOM in multiple types of the channels inside this Bayesian dynamic model.
The model is applied to the time-series data on the Hyundai automobile company, and the estimation is done through Forward Filtering and Backward Sampling algorithm, derived for application of Gibbs Sampling to normal Gaussian state space models. The estimation result indicates that the effects of copy wear-out and those of volume are differential across channel types. In addition, the interaction effects of blog and community site types are negative and significant, where as those of twitter is turned out to be nonsignificant. Lastly, the effects of valence in each of three different channel types are positive and significant.
This paper provides the following contributions to the literature. For the academical implication, I extend earlier research on WOM by introducing a generalized model for estimating the dynamic effects of WOM in multiple types of online communication channels and the interaction effects among the WOM inside each types of channels. This methodology would help researchers to examine both dynamic and interaction effects simultaneously and achieve more explanatorily powerful results, bridging the gap between two different research streams. For the practical implication, this paper would help practitioners in durable good markets to derive optimal strategies by discovering channel mix that would maximize the dynamic effects of WOM.