Behavioral Consumers and Behavioral AIs: Three Studies

Loading...
Thumbnail Image
Date
Authors
Chen, Yang
Keyword
Consumer Behavior , Machine Behavior , Causal Inference , Loyalty , LLM
Abstract
In this thesis, we conduct three research projects in operations management with behavioral underpinnings. With most businesses, the behaviors of employees and customers require careful consideration since they are pivotal to business decision-making and firm profitability. Understanding consumer and AI behavior in business is the central topic throughout the thesis. In the first part of this thesis, we study the effect of tier rewards on customer loyalty. We report on the results of a large-scale field quasi-experiment in a loyalty program (LP) offering premium status. Our main goal is to understand strategic consumer behavior in such programs and quantify the tier reward's value to the firm. We follow the addition of a new reward tier to an existing LP and compare participant behavior before and after the program structure change. We find a ~3.3% lift in revenue among strategic participants, and document rich behavioral patterns. Our findings provide valuable information for the design of the LP tier structure. In the second part of the thesis, we study the effect of redeemable points on customer loyalty in an LP where consumers decide how much to redeem. We investigate the effect of reward redemption size and consumer "type" on long-term loyalty. We find that size does matter in reward redemptions, and so does the operationalization of size and prior reward redemption behaviors. Medium-sized redemptions are most consistently associated with increased loyalty. Small redemptions are sometimes associated with increased loyalty. With large redemptions, however, we sometimes observe a decrease in loyalty. Our findings suggest that the optimal redemption size for firms to encourage should be personalized. In the third part of this dissertation, we investigate whether ChatGPT exhibits behavioral biases commonly found in operations management tasks. We find that although ChatGPT can be much less biased and more accurate than humans in problems with explicit mathematical nature, it also exhibits many biases humans possess, especially when the problems are complicated, ambiguous, and implicit. This study characterizes ChatGPT's behaviors in decision-making and showcases the need for researchers and businesses to consider potential AI behavioral biases when developing and employing AI for business operations.
External DOI