This study examines consumer perceptions and emotional responses across three major vertical commerce platforms—Musinsa, Olive Young, and Market Kurly— using consumer-generated online texts. Rather than prioritizing category-specific differences, it emphasizes the combination of shared consumption-related elements within each platform’s usage context. A total of 50,055 posts from Naver blogs and cafés were analyzed using TEXTOM through word frequency and TF-IDF analysis, sentiment analysis, and LDA topic modeling. Results showed that category-specific product characteristics were reflected in platform discourse, while recommendation, purchase, and review-related elements appeared across all three platforms. However, their combinations, however, differed: Musinsa linked recommendation with seasonal styling and brand exploration; Olive Young combined reviews and recommendations with product efficacy and usage experience; and Market Kurly connected self-funded purchase reviews with price benefits and delivery experiences. Although positive sentiment predominated across all platforms, the objects and contexts of emotional expression varied. Platform-specific thematic structures emerged from topic modeling, indicating that product characteristics, service attributes, and usage contexts jointly shape consumer perceptions. Overall, this study compares community-based consumer-generated texts across various platforms, extending previous research on single-platform and product reviews. Additionally, it shows how shared consumption elements are embedded differently in various platform contexts.

Abstract
I. Introduction
Ⅱ. Theoretical Background
    1. Fashion vertical commerce platform: MUSINSA’scharacteristics and current status
    2. Beauty vertical commerce platform: OLIVEYOUNG’s characteristics and current status
    3. Food vertical commerce platform: MarketKurly’s characteristics and current status
    4. Consumer perception research using consumergeneratedonline texts
Ⅲ. Research Method
    1. Research questions
    2. Data collection and process
    3. Analysis methods and procedures
Ⅳ. Results
    1. Research question 1: Word frequency analysisand TF-IDF analysis results
    2. Research question 2: Classification of sentimentalword analysis of fashion/beauty/food platforms
    3. Research question 3: Analysis of topic modeling
Ⅴ. Conclusion
References

• Jing Yi Zhang(Doctoral Course, Dept. of Design, Fashion Industry Major, Chung-Ang University, Korea) | 장정이 (중앙대학교 디자인학과 패션산업전공 박사과정)
• Woo Bin Kim(Assistant Professor, Dept. of Fashion, Chung-Ang University, Korea) | 김우빈 (중앙대학교 패션전공 조교수) Corresponding author