Adaptive persuasive messages in an e-commerce setting: the use of persuasion profiles

Technologies that are intentionally designed to change a person’s attitude or behaviors are emergent. Designers of these technologies frequently use implementations of influence strategies to increase the effectiveness of their systems. In this paper we argue that there are large individual differences in users responses to implementations of influence strategies, and as such persuasive technologies would benefit from adapting to these differences. We detail how designers of persuasive technologies could use persuasion profiles to adapt to differences in individual susceptibility to different influence strategies. To evaluate the notion of persuasion profiles we build an adaptive persuasive system in an ecommerce setting and test this system against a non-adaptive counterpart. We describe how — using Bayesian learning — designers can implement adaptive persuasive technologies. To our knowledge this is the first implementation and evaluation of a working persuasive system that utilizes dynamically created persuasion profiles to increase its effectiveness. We show increased revenues for the system that utilizes persuasion profiles. Finally, we discuss the limitations of the presented evaluation and give suggestions for future research efforts.