Date of Completion

6-2-2020

Embargo Period

5-29-2022

Keywords

Professional Service, Risk Reduction, Cosmetic Surgery, Online Diary, Complications Insurance, Face Prediction

Major Advisor

Jan Stallaert

Associate Advisor

Xinxin Li

Associate Advisor

Jing Peng

Field of Study

Business Administration

Degree

Doctor of Philosophy

Open Access

Campus Access

Abstract

Professional services such as medical services are characterized by large information asymmetries between service providers and consumers, resulting in high perceived risk for potential customers. This dissertation focuses on three types of emerging risk-reduction strategies—online diaries, complications insurance, and AI-based predictions in online marketplaces for professional services. In Essay 1, we focus on a novel type of risk-reduction strategy—online diaries, series of posts generated by consumers in chronological order to record their post-consumption experience over time. Using a large dataset from an online platform of cosmetic procedures, we find that providing follow-ups in diaries has a positive effect on the sales of the respective cosmetic procedures and the quality of providers substitutes this effect. Interestingly, for high-risk procedures, providing follow-ups increases sales regardless of the quality of providers. In contrast, for low-risk procedures, providing follow-ups substantially increases sales for low-quality providers, but not for high-quality providers. In Essay 2, we study how complications insurance, which is designed to cover the potential cost of post-surgery complications, affects the sales of cosmetic procedures. Our empirical study shows the introduction of insurance increases the sales of low-risk procedures, but has no significant effect on the sales of high-risk procedures. More importantly, the insurance has a negative spillover effect on uninsured competitors, regardless of their risk levels. In Essay 3, we investigate how AI-based predictions for cosmetic procedures influence consumers’ purchase intention. We use generative adversarial networks to train a deep learning model that can predict the outcome of a type of cosmetic procedure. Based on this prediction model, we design an online experiment to examine the impact of AI prediction on consumers’ decision-making processes. We find that seeing AI predictions has a positive impact on male consumers’ purchase intention and AI’s prediction quality has a positive impact on consumers’ purchase intention. We also find that consumers’ purchase intention increases more when the AI predictions show a larger beauty improvement, and that showing celebrities’ predicted procedure outcome has a negative impact on females’ purchase intention. This dissertation has important implications for platforms to design and evaluate their risk-reduction strategies.

COinS