An Agent-Based Model of Consumer Behavior : Application to the Movie Market

Abstract

We applied a multi-agent model to simulate consumer behavior, which has been more complicated through mounting use of the Internet. As the primary factors for a consumer's decision to purchase an item for sale, we picked out "individual preference," "tendency to be influenced by others," "experience in purchasing the item for sale concerned" and "advertising effects." In addition, "word-of-mouth (WOM)" and "Internet WOM" are included in the model as communication channels. We developed a multi-agent artificial society model which can simulate aggregate phenomena resulting from local interactions between individual consumers. In this study, we applied the model to the movie industry. As the items for sale, we used 18 films which were released from October 2003 to July 2004, and simulated consumer behavior. First, we compared simulated values with actual ones to test the validity of the model. We found that our model can re-create the actual purchasing behaviors of the film market to a certain level. Secondly, we analyzed the effects of changes in the external environment and the film market's marketing strategies, establishing preset conditions for the model. From the results of various simulations, some interesting observations about the market share of each film are made, and the transformation of consumption patterns noted.