エージェント・ベースの新商品普及モデルとCaRによる売上下落現象の検出

抄録

Chasm or crack in the diffusion process of a new product is a long-term decline in sales. If the occurrence of chasm or crack is predicted and the chasm factor is understood by analyzing consumers’ evaluation information on the product, the product can penetrate the market more effectively. In this study, we developed an agent-based diffusion model that incorporates the concepts of cluster connectivity, which represents word-of-mouth effect, product recognition range, which represents promotion effect, and purchase probability of individual consumers, which are not considered in conventional models, to reproduce the diffusion process of new products. Furthermore, the amount of sales decline that can occur in the diffusion process of a new product is calculated by using the Chasm at Risk (CaR), and detect chasm or crack in the product. For validation, we reproduced the diffusion process of a new product and the occurrence of chasm or crack using data from an O2O site, which gathers users who go back and forth between online on Web and offline in real stores. In addition, we also examined the method of chasm suppression when the effects of consumer word-of-mouth and company promotion were varied.