Textile and Apparel Industries

Decoding E-Sourcing Narratives on Reddit: A Comprehensive Text Mining Approach

---

Abstract

This study examines the needs and experiences of SMEs in e-sourcing, or online sourcing, within the fashion industry. Using a text-mining approach, the research analyzed comments from over 300 Reddit subreddits, resulting in a dataset of 576 posts (170k words). Methods included N-gram analysis, sentiment analysis, and LDA topic modeling. The study is guided by a conceptual framework combining the Triangular Alignment Model (TAM) and Technology–Organization–Environment (TOE) theories.

Key findings revealed that efficiency and competitiveness are dominant themes, with trust and anticipation being the most prevalent sentiments. Negative sentiments such as fear and anger highlighted challenges faced by the e-sourcing community. Topic modeling showed a strong emphasis on product sourcing (52.5%) and human sourcing (21.9%), along with ethical, domestic, and technological considerations. These insights provide a foundation for further research on efficiency, competitiveness, and reliability in e-sourcing practices.

Keywords: e-sourcing, online sourcing, text mining, reddit, n-gram, sentiment, LDA

How to Cite:

Liu, Y. J. & Zhao, L., (2025) “Decoding E-Sourcing Narratives on Reddit: A Comprehensive Text Mining Approach”, International Textile and Apparel Association Annual Conference Proceedings 81(1). doi: https://doi.org/10.31274/itaa.18554