Evaluating the impact of different virtual characters in product recommendations

  • Authors:
  • Eliseo Reategui;Cristiano Fruet;Danilo Marcondes Filho

  • Affiliations:
  • Universidade de Caxias do Sul, R. Francisco Getúlio Vargas, Caxias do Sul, RS;Universidade de Caxias do Sul, R. Francisco Getúlio Vargas, Caxias do Sul, RS;Unversidade do Vale do Rio dos Sinos, Avenida Unisinos, Sao Leopoldo, RS

  • Venue:
  • IHC '06 Proceedings of VII Brazilian symposium on Human factors in computing systems
  • Year:
  • 2006

Quantified Score

Hi-index 0.00

Visualization

Abstract

This paper presents an evaluation of the effect that distinct interface characters could have on the users' perception of their interaction with an ecommerce website. A recommender system has been integrated to the characters' architecture, with the goal of assessing their impact in making personalized recommendations. The characters used to interact with the users had different features such as being a known or an unknown person; being represented by an anthropomorphic figure or not. In spite of these differences, the characters also had a lot of similarities: they used the same language and had the same strategy for recommending products. They also had the same type of graphical representation. An experiment involving 180 students was set up, in which students had to complete a few purchases in an ecommerce website where the virtual characters gave tips and suggested products related to the users' perceived interests. The results of the experiment are presented here, as well as conclusions and direction for future work.