Predicting shoppers' interest from social interactions using sociometric sensors
Marketing research has longed for better ways to measure consumer behavior. In this paper, we explore using sociometric data to study social behaviors of group shoppers. We hypothesize that the interaction patterns among shoppers will convey their interest level, predicting probability of purchase. To verify our hypotheses, we observed co-habiting couples shopping for furniture. We have verified that there are sensible differences in customer behavior depending on their interest level. When couples are interested in an item they observe the item for a longer duration of time and have a more balanced speaking style. A real-time prediction model was constructed using a decision tree with a prediction accuracy reaching 79.8% and a sensitivity of 63%.
- download PDF (1.7 MB)
Kim, T.; Brdiczka, O.; Chu, M.; Begole, J. Predicting shoppers' interest from social interactions using sociometric sensors. Extended Abstracts of 27th Annual CHI Conference on Human Factors in Computing Systems (CHI 2009); 2009 April 4-9; Boston, MA. NY: ACM; 2009; 4513-4518.
Copyright © ACM, 2009. This is the author's version of the work. It is posted here by permission of ACM for your personal use. Not for redistribution. The definitive version was published in Extended Abstracts of CHI 2009 http://portal.acm.org/citation.cfm?id=1520692