Reimaging Consumer Creditworthiness by Harnessing E-Commerce Digital Footprints

YIJUN LI, Albert Di Wang, Qi WU

Abstract: BigTech firms have reshaped consumer credit markets by seamlessly embedding Fintech lending services into their ecosystems. Unlike traditional financial institutions, these companies possess vast digital footprints generated from consumer activities, ranging from product browsing and transaction execution to post-purchase credit management. However, few studies have investigated the value of digital footprints in credit risk evaluation, particularly from a dynamic perspective, thereby overlooking critical insights into how credit risk evolves over time. To address this gap, we propose a novel framework that models consumer footprints in continuous time. By accommodating irregular and sporadic footprints, our model enables real-time monitoring of default risk. Additionally, the proposed end-to-end loss function jointly estimates the Probability of Default and Exposure at Default, offering a more comprehensive view of credit risk than traditional scoring approaches. We evaluate our model using longitudinal data from a leading BigTech platform. Empirical results demonstrate that our method significantly outperforms established benchmarks. Our findings underscore the critical role of modeling digital footprints in credit risk assessment, providing valuable guidance for more inclusive, adaptive, and data-driven financial services.