Attrition in weight loss programs

: a Bayesian statistics approach for predictive insights

Abstract

In today’s service-oriented business landscape, client acquisition and retention are imperative, given the significant influence clients wield over service outcomes. This influence is particularly pronounced in prolonged services like weight loss programs, where client motivation towards program completion is crucial for achieving desired outcomes and ensuring customer satisfaction. This thesis investigates factors contributing to attrition in a three-phase weight loss program, leveraging a comprehensive dataset encompassing client registration details, progress tracking, and demographic information. However, the challenge of missing data arises when clients discontinue the program, hindering a comprehensive understanding of the weight loss journey. To address this challenge, a causal approach to missing data imputation is adopted, utilizing Bayesian Statistics to harness the inherent information within the data. Through extensive literature review and methodological exploration, the study sheds light on the reasons for attrition and proposes practical insights into addressing churn issues. Age and weight emerge as significant predictors of program completion, with older individuals exhibiting higher completion rates across all phases. Additionally, the study highlights the nuanced impact of previous program attempts on completion rates. Overall, this study contributes to the understanding of attrition in weight loss programs and offers valuable insights into leveraging Bayesian imputation models for predictive analytics in service-oriented contexts.

Date of Award	2 May 2024
Original language	English
Awarding Institution	Universidade Católica Portuguesa
Supervisor	Nicolò Bertani (Supervisor)