Faculté et Recherche
Nudging when it Matters: Machine Learning-based Nudging to Improve Physical Activity
08 avr
2026
10H00 - 11H15
Jouy-en-Josas
Anglais
Participer
Information Systems and Operations Management
Speaker: Idris Adjerid from Virginia Tech
Room T-006
Abstract
Emerging digital platforms, coupled with the power of machine learning, have the potential to enhance the effectiveness, consistency, and scale of behavioral nudges. In this study, we introduce “ML-Nudging”, a novel approach that combines conventional nudging techniques with predictive analytics. We evaluate the impact of ML-Nudging on a mobile healthcare platform, focusing on personalized encouragement for physical activity based on individuals’ predicted levels of exercise. Through a field experiment with a major mobile health app provider in Asia, we find that ML-Nudging effectively increases individuals’ step count, particularly when coupled with an economic incentive. Interestingly, both targeting on high and low activity days had a similar positive effect on physical activity. Further exploration suggests that differential effects for historically active vs. sedentary individuals explain the similar efficacy of high and low activity targeting. By demonstrating the potential of targeted, data-driven nudges, our work contributes valuable insights at the intersection of information systems, data science, and behavioral economics.