簡易檢索 / 詳目顯示
| 研究生： |
蔡沁宸 Tsai, Chin-Chen |
|---|---|
| 論文名稱： |
探討智慧虛擬試穿技術於 B2B 時尚電商中作為轉換率優化工具之應用潛力 To explore how intelligent VTO technologies can be leveraged as a CRO tool in B2B fashion e-commerce. |
| 指導教授： |
蔡政憲
TSAI Cheng-Hsien |
| 口試委員: |
顧筱筠
Ku Hsiao-Yun |
| 學位類別： |
碩士
Master |
| 系所名稱： |
商學院 - 國際經營管理英語碩士學位學程(IMBA) International MBA Program College of Commerce(IMBA) |
| 論文出版年： | 2027 |
| 畢業學年度： | 115 |
| 語文別： | 英文 |
| 論文頁數： | 51 |
| 中文關鍵詞： | 虛擬試穿 、生成式人工智慧 、B2B 時尚電商 、轉換率優化 、軟體即服務 、中小企業 、實踐導向研究 |
| 外文關鍵詞： | Virtual Try-On, Generative AI, B2B Fashion E-Commerce, Conversion Rate Optimization, SaaS, SME, Practice-Based Research |
| 相關次數： | 點閱：17 下載：0 |
| 分享至: |
| 查詢本校圖書館目錄 查詢臺灣博碩士論文知識加值系統 勘誤回報 |
線上時尚電商產業長期面臨一項核心挑戰：服裝類商品的退貨率高達 30 至 40%，主要原因在於購物者在購買前無法有效視覺化商品穿著效果。「購物括弧行為」（bracketing，即購買多種尺寸或款式再進行退貨）雖為消費者的理性選擇，卻對商家毛利造成嚴重侵蝕。
本研究採用實踐導向研究方法（Practice-Based Research），探討 AI 驅動的虛擬試穿（ Virtual Try-On, VTO ） 技 術 作 為 B2B 時 尚 電 商 轉 換 率 優 化 （ Conversion Rate Optimization, CRO）工具的應用潛力，研究載體為研究者親自開發與驗證的生成式 AI SaaS 平台 ÀToI Studio，專為法國 Shopify 時尚中小企業設計。
研究分析涵蓋四個核心面向：（一）多供應商 AI 架構的技術可行性，每次生成成本低於 €0.065；（二）7 角度多視角輸出功能，在現行中小企業層級方案中屬罕見設計；（三）符合法國 B2B 採購要求的隱私設計合規架構；（四）透過 Bisart.fr 實際試點整合驗證的 Shopify 部署能力。市場分析確認法國優先可觸及的商家宇宙約為 15,000 至25,000 家 Shopify 時尚中小企業，整體採用階段仍處早期。競爭分析顯示，在四個直接競爭對手中，尚無任何業者同時具備基於顧客自身照片的多視角生成、歐盟原生GDPR 合規，以及法語在地化市場進入策略。
財務分析顯示，在基本用量假設下，扣除支付費用後的混合毛利率約為 72.3%，從早期創辦人主導外展（客戶獲取成本 CAC 約 €150）過渡至 Shopify 應用程式商店主導（CAC ~€25–50），可大幅改善獲客效率。三年期財務預測顯示，第二年下半年（Y2H2）首次達到獲利，第三年底累計商家數達 700 家。
本研究為 B2B SaaS 情境下 VTO 技術部署提供了多維度分析框架，為有志於 AI 驅動商業賦能領域的實務工作者與研究者，提供實踐指引與初步實證基礎。
The online fashion e-commerce industry faces a persistent structural challenge: apparel return rates reach 30–40%, driven largely by shoppers' inability to visualize how garments will look on themselves prior to purchase. This practice, known as "bracketing," imposes significant cost burdens on merchants and represents a solvable problem given recent advances in generative artificial intelligence.
This practice-based study explores the application potential of AI-powered virtual try-on (VTO) technology as a conversion rate optimization (CRO) tool in the B2B fashion e-commerce context, through the development and validation of ÀToI Studio—a generative AI SaaS platform targeting French Shopify fashion SMEs, developed and operated by the researcher-founder from February to May 2026.
Analysis spans four core dimensions: (1) technical feasibility of a multi-provider AI architecture enabling photorealistic generation at under €0.065 per session; (2) a 7-angle multi-view output feature absent from current SME-tier alternatives; (3) a privacy-by-design compliance architecture aligned with French B2B procurement requirements; and (4) Shopify deployment capability validated through an active pilot integration. Market analysis identifies a France-first addressable universe of approximately 15,000–25,000 Shopify fashion SMEs, with adoption currently in early stages. Competitive analysis identifies four direct SME-tier competitors, none of which simultaneously offer customer-photo-specific multi-view outputs, EU-native GDPR compliance, and a France-localized go-to-market motion.
The business model suggests unit economics favorable for early-stage SaaS: a blended gross margin of approximately 72.3% at base usage assumptions (after payment processing fees), with customer acquisition cost improving from ~€150 via founder-led outbound to ~€25–50 via Shopify App Store organic installs. Three-year financial projections indicate first profitability in Year 2 H2, reaching 700 paying merchants by Year 3.
This study contributes a multi-dimensional evaluation framework for VTO deployment in B2B SaaS contexts, providing practical guidelines for practitioners and preliminary evidence for researchers interested in AI-driven commerce enablement.
摘要 iii
Abstract iv
目次 / Table of Contents v
表次 / List of Tables vii
圖次 / List of Figures viii
Chapter 1 Introduction 1
1.1 Research Background 1
1.2 Problem Statement 2
1.3 Research Objectives and Questions 2
1.4 Research Significance 3
1.5 Thesis Structure 3
Chapter 2 Industry Background and Research Context 6
2.1 The Evolution of Virtual Try-On Technology 6
2.2 Online Fashion E-Commerce and Consumer Return Behavior 6
2.3 Conversion Rate Optimization and VTO Benchmarks 7
2.4 B2B SaaS Models in Fashion Technology 8
2.5 Regulatory Context: GDPR and the EU AI Act 8
2.6 Market Timing: The Two-Threshold Moment 8
Chapter 3 Research Design and Methodology 10
3.1 Research Paradigm and Approach 10
3.2 Unit of Analysis 10
3.3 Data Collection Methods 11
3.4 Analytical Framework 12
3.5 Validity and Limitations 12
Chapter 4 Product Development and Technical Architecture 14
4.1 Core AI Architecture 14
4.2 Product Features 15
4.3 B2B Integration 17
4.4 Prompt Engineering as a Proprietary Moat 19
Chapter 5 Market and Competitive Analysis 21
5.1 Market Overview and Strategic Context 21
5.2 Regional Market Dynamics 21
5.3 SME Addressable Market 22
5.4 France Beachhead Market 24
5.5 Competitive Landscape 24
5.6 Competitive Advantages and Risk Mitigation 25
Chapter 6 Business Model and Financial Analysis 28
6.1 SaaS Pricing Architecture 28
6.2 Cost of Goods Sold 28
6.3 Unit Economics 29
6.4 Revenue Streams and Long-Term Potential 29
6.5 Financial Projections and Growth Forecast 30
6.6 Annual P&L, Balance Sheet, and Cash Flow 32
6.7 Scenario Analysis and Breakeven 33
6.8 Infrastructure and Operating Cost Structure 34
6.9 Headcount and Customer Acquisition Plan 35
Chapter 7 Go-to-Market Strategy and Empirical Findings 36
7.1 Development Timeline and Traction 36
7.2 Merchant Discovery and ICP Validation 37
7.3 Four-Phase Go-to-Market Strategy 38
7.4 Target Customer Profile 39
7.5 90-Day Validation Plan 39
7.6 Team and Use of Funds 40
7.7 Long-Term Vision 40
Chapter 8 Conclusions and Recommendations 42
8.1 Summary of Findings 42
8.2 Practical Implications 43
8.3 Limitations 45
8.4 Directions for Future Research 45
References 47
Appendix A Detailed Competitor Matrix 50
Avenier, M.-J., & Schmitt, C. (2010). La construction de savoirs pour l'action [Knowledge construction for action]. L'Harmattan.
Chargebacks911. (2026). Key Shopify statistics and indicators 2026. https://chargebacks911.com/shopify-statistics/ (chargebacks911.com in Bing)
FEVAD. (2024). Bilan du e-commerce en France en 2024 [State of e-commerce in France in 2024]. Fédération du e-commerce et de la vente à distance. https://fevad.com/bilan-du-e-commerce-en-france-en-2024 (fevad.com in Bing)
FEVAD. (2025). Classement FEVAD 2024 des sites e-commerce en nombre de clients. Fédération du e-commerce et de la vente à distance. https://fevad.com/classement-fevad-2024-des-sites-e-commerce-en-nombre-de-clients/ (fevad.com in Bing)
Forney, J. C., Park, E. J., & Brandon, L. (2005). Effects of evaluative criteria on fashion brand extension. Journal of Fashion Marketing and Management, 9(2), 156–165.
Grand View Research. (2024). Virtual try-on market size, share & trends analysis report. https://grandviewresearch.com/industry-analysis/virtual-try-on-market-report (grandviewresearch.com in Bing)
Han, X., Wu, Z., Huang, W., Zhang, X., Yu, M., Li, X., & Liu, T. (2018). VITON: An image-based virtual try-on network. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR) (pp. 7543–7552).
High Alpha & OpenView. (2024). 2024 SaaS benchmarks report. https://highalpha.com/saas-benchmarks/2024 (highalpha.com in Bing)
Ho, J., Jain, A., & Abbeel, P. (2020). Denoising diffusion probabilistic models. Advances in Neural Information Processing Systems, 33, 6840–6851.
HubSpot. (2025). State of sales report 2025. https://blog.hubspot.com/sales/hubspot-sales-strategy-report (blog.hubspot.com in Bing)
Jiang, Z., & Benbasat, I. (2007). Virtual product experience: Effects of visual and functional control of products on perceived diagnosticity and flow in electronic shopping. Journal of Management Information Systems, 21(3), 111–147.
Mordor Intelligence. (2025). Virtual try-on market — growth, trends, COVID-19 impact, and forecasts (2025–2030).
Narvar. (2024). Consumer returns report 2024: The state of bracketing in online fashion. https://narvar.com
National Retail Federation. (2024). Consumer returns in the retail industry 2024. https://nrf.com/research/2024-consumer-returns-retail-industry (nrf.com in Bing)
National Retail Federation & Happy Returns. (2024). 2024 consumer returns in the retail industry — press release. https://nrf.com/media-center/press-releases/nrf-and-happy-returns-report-2024-retail-returns-total-890-billion (nrf.com in Bing)
Osterwalder, A., & Pigneur, Y. (2010). Business model generation: A handbook for visionaries, game changers, and challengers. Wiley.
Pixelz. (2024). Product photography cost guide 2024. https://pixelz.com
Porter, M. E. (1979). How competitive forces shape strategy. Harvard Business Review, 57(2), 137–145.
Red Stag Fulfillment. (2025). How many e-commerce stores exist? (2025 figures). https://redstagfulfillment.com/how-many-ecommerce-stores-exist/ (redstagfulfillment.com in Bing)
Rombach, R., Blattmann, A., Lorenz, D., Esser, P., & Ommer, B. (2022). High-resolution image synthesis with latent diffusion models. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR) (pp. 10684–10695).
Shim, S., Eastlick, M. A., Lotz, S. L., & Warrington, P. (2001). An online prepurchase intentions model: The role of intention to search. Journal of Retailing, 77(3), 397–416.
Shopify. (2022). The ROI on AR: How augmented reality is boosting e-commerce sales. https://shopify.com/blog/ar-shopping (shopify.com in Bing)
Snap Inc. (2022). Sunglass Hut AR shopping lens case study. https://forbusiness.snapchat.com/inspiration/sunglasshut (forbusiness.snapchat.com in Bing)
Snap Inc. (2022). Ulta Beauty AR shopping case study.
StoreLeads. (2026). The state of Shopify 2026. https://storeleads.app/reports/shopify
StoreLeads. (2026). The state of Cafe24 — Korea. https://storeleads.app/reports/cafe24/KR/top-stores (storeleads.app in Bing)
StoreLeads. (2026). Apparel stores on Wix. https://storeleads.app/reports/wix/category/Apparel (storeleads.app in Bing)
Sumtracker. (2025). 50+ Shopify statistics for 2025. https://sumtracker.com/blog/top-shopify-statistics (sumtracker.com in Bing)
Wang, B., Zheng, H., Liang, X., Chen, Y., Lin, L., & Yang, M. (2018). Toward characteristic-preserving image-based virtual try-on network. In Proceedings of the European Conference on Computer Vision (ECCV) (pp. 589–604).
Yin, R. K. (2014). Case study research: Design and methods (5th ed.). Sage Publications.
European Union. (2016). Regulation (EU) 2016/679 ... General Data Protection Regulation. Official Journal of the European Union, L 119, 1–88.
European Union. (2024). Regulation (EU) 2024/1689 ... Artificial Intelligence Act. Official Journal of the European Union.
全文公開日期 2027/07/19