Author(s) :

Vito Hsu.

Volume/Issue :

Abstract :

Online shopping has changed how people buy products by offering more access, choices, and convenience. Still, success in e-commerce depends heavily on pricing, discount strategies, and customer response. This project examines how product price and discount level influence customer satisfaction (ratings) and customer engagement (review counts) using a public Kaggle dataset from 2022 that includes pricing, discount fractions, ratings, review counts, and product categories. Three models were used: a multiple linear regression for ratings, a Negative Binomial regression for review counts, and a joint model that estimates both together. The results show that a one–standard deviation increase in log price raises average product ratings by 0.047 points, but lowers rating counts by 0.171 on the log scale—meaning higher-priced items are rated better but reviewed less. Larger discounts show the opposite pattern; a one–standard deviation increase in discount fraction decreases ratings by 0.053 and reduces review counts by 0.152, suggesting that heavy markdowns may hurt both perceived quality and engagement. Category differences also appeared: Electronics products received significantly more reviews (0.672) but slightly lower ratings (–0.127), while Home & Kitchen and Office Products performed worse across both outcomes. The joint model provided the most stable estimates and helped connect the relationships between satisfaction and engagement. Overall, the study demonstrates that pricing and discount strategies strongly shape how customers rate and interact with products online.