Publication Type

PhD Dissertation

Version

publishedVersion

Publication Date

9-2025

Abstract

Artificial Intelligence Generated Content is widely expected to enhance efficiency and lower operation cost for e-commerce. This study aims to investigate on how AIGC technology would impact on the online shopping behavior and conversion rate, specifically in the scenario of product display and recommendations.

Experiment 1 is to evaluate which scenario would have better efficiency on e-commerce platforms when using AIGC. Using an A/B testing methodology, the effects of AIGC application were measured among key factors of six product display and recommendation scenario on Vipshop.com. Results show that AIGC application on model image has the greatest impact, increasing CTR by 2.74% and CVR by 4.73%. This has led to an estimated 0.63% increase in GMV to whole app. We believe the result is mainly based on the feature of platform, which merchants have limited investment in product image information. The empowerment of AIGC enables merchants to satisfy consumer preferences at a lower cost. Also, AIGC identifies and creates model images that users like more and are more relevant to the actual situation, helping eliminate customers' distrust of online shopping "seller shows". Meanwhile, the study calculates the ROI of each scenario and finds that model image scenario with the highest conversion rate and GMV enhancement ratio over all other scenarios brings the lowest ROI. The reason is that model image scenarios need to invest more in hardware, arithmetic and human resource. The study illustrates that ROI of AIGC in scenario depends not only on conversion rate and GMV lift, but also on its corresponding input cost.

Experiment 2 is to evaluate whether the efficiency is different in best-selling and long-tail products. Utilizing A/B testing method, I fixed scenarios in model images and background images, and then carry out the tests on best-selling products and long-tail products. Eventually, it is found that application of AIGC is more efficient in long-tail items, with click-through rate increasing 3.40% and conversion rate increasing 11.44%.The main reason is assumed to be merchants pay less attention to long-tailed products and invest less in product image maintenance, which bringing higher efficiency when applying AIGC to long-tailed products.

Experiment 3 is an offline experiment to supplement the validity of findings of online Experiments 1 and 2.It recruited 75 respondents offline to see several sets of images, both the original and the AI-optimized. Respondents were asked which was more appealing to them and to score willingness to buy. The results show that AIGC has no significant effect on click-through rate, both in best-selling group and in long-tailed group. Moreover, it presents a significant effect on conversion rate for long-tailed product, of which conclusion remains consistent with Experiment 2.

Overall, this study quantifies the revenue growth brought by AIGC applications. It shows that focus should be placed on model image scenario if company considers GMV and revenue lift. If ROI is considered, product lists may be a better selection. As for product types to be invested in, long-tailed product should be considered.

Keywords

AIGC Application, E-commerce, Product Display and Intelligent Recommendation, Quantification of Revenue and Cost

Degree Awarded

Doctor of Bus Admin (CKGSB)

Discipline

E-Commerce | Finance and Financial Management

Supervisor(s)

BHATTACHARYA, Shantanu Hiralal

First Page

1

Last Page

120

Publisher

Singapore Management University

City or Country

Singapore

Copyright Owner and License

Author

Download

Share

COinS