Abstract

Social network sites such as Twitter have madeeasier the exchange of information about products amongconsumers that are users of these networks. Thereby, marketersare using word of mouth marketing campaigns to engage users ofthese social networks in spreading the word about their products.The choice of consumers who participate in these campaigns is keyin their success. However, it is not clear which type of consumersand its combination are the most effective in reaching this aim.Unfortunately, the difficulty in getting useful data due to its noiseor non-public availability do not allow researchers toappropriately apply traditional methodologies like statistics.Hence, this paper proposes an agent-based model to simulate theinformation diffusion through Twitter when some types ofconsumers participate in word of mouth marketing campaigns.This model was implemented in Java and its results show that,although hubs (users with many contacts) are better connectedthan opinion leaders, the latter ones reach higher informationdiffusion because of its influence; in addition, the use of both typesof consumers in the same campaign achieves as diffusion as usingjust opinion leaders. Apart from results, this paper gives to futureresearchers a guideline to develop ABMs that simulate differentbehaviors of Twitter's users.