Content freebooting

This dataset accompanies the study on content freebooting in social media advertising and its exploitation by counterfeit and dupe brands. The dataset captures advertising activity around a case study product (a floral corset sundress) across TikTok and linked e-commerce domains. It integrates ad-level metadata, creator provenance, network diffusion structures, product inspections, fulfillment outcomes, and detection pipeline results. The dataset is designed to support research and teaching in areas such as multimedia forensics, adversarial content manipulation, e-commerce fraud detection, and network diffusion analysis. It is suitable for both computational modeling (e.g., duplicate detection, network clustering) and qualitative case-based learning.

Data Collection and Methodology
Source Platforms: TikTok ads, linked e-commerce domains, review widgets, and transparency archives.
Observation Period: Four weeks.
Ad Corpus: 54 unique ads linking to 32 domains.
Captured Assets: 180 landing pages, 3,000 extracted frames, 100+ review images.
Purchases: Three advertiser domains, one AliExpress dupe, one control (House of CB).
Analysis Methods: multimodal similarity detection (pHash, dHash, audio fingerprints, vision embeddings, NLP), network diffusion metrics, and product inspections.