[Paper] Turning Trust to Transactions: Tracking Affiliate Marketing and FTC Compliance in YouTube's Influencer Economy

Source: arXiv - 2603.04383v1

Overview

YouTube’s influencer economy has grown into a massive marketplace where creators earn commissions by promoting products through affiliate links. This paper investigates how often creators actually disclose those relationships—and whether they follow the U.S. Federal Trade Commission’s (FTC) rules. By analyzing two million videos spanning a decade, the authors reveal that affiliate marketing is pervasive, yet disclosure compliance is alarmingly low, raising concerns for both consumers and brands.

Key Contributions

• Large‑scale measurement pipeline: Combines web‑scraping, link‑resolution, and natural‑language processing (NLP) to automatically detect affiliate links and disclosure statements in YouTube videos.
• Comprehensive dataset: 2 M videos from ~540 k creators covering a 10‑year period, providing the most extensive view of YouTube’s affiliate ecosystem to date.
• Empirical compliance audit: Quantifies how many videos meet FTC disclosure standards versus those that hide or omit affiliate relationships.
• Stakeholder impact analysis: Shows how platform‑level features (e.g., standardized disclosure UI) and affiliate‑partner policies affect creator behavior.
• Actionable recommendations: Proposes concrete steps for regulators, affiliate networks, and the YouTube platform to improve transparency.

Methodology

1. Data collection – The researchers used the YouTube Data API to pull video metadata (titles, descriptions, comments) and then crawled the URLs embedded in video descriptions.
2. Affiliate link identification – A curated list of known affiliate programs (Amazon, ShareASale, Impact, etc.) was matched against the resolved URLs. Machine‑learning classifiers filtered out false positives (e.g., generic redirects).
3. Disclosure detection – NLP models (fine‑tuned BERT) scanned video titles, descriptions, and pinned comments for disclosure phrases (“affiliate link,” “#ad,” “sponsored,” etc.). The models were trained on a manually labeled subset to achieve high precision/recall.
4. Compliance scoring – Detected disclosures were evaluated against FTC guidelines (visibility, clarity, proximity to the endorsement). Videos lacking a proper disclosure were flagged as non‑compliant.
5. Longitudinal & stakeholder analysis – The team correlated compliance rates with platform changes (e.g., introduction of YouTube’s “Paid Promotion” label) and affiliate‑partner policies to assess causal effects.

Results & Findings

• Affiliate link prevalence: ~12 % of all videos contained at least one affiliate link, representing billions of potential impressions.
• Low disclosure compliance: Only 23 % of affiliate‑linked videos provided a disclosure that met FTC standards; the majority either omitted any mention or used vague language.
• Improvement after platform interventions: Introduction of YouTube’s “Paid Promotion” label in 2020 coincided with a +15 % lift in compliant disclosures among new videos.
• Creator size matters: Mid‑tier creators (10 k–100 k subscribers) were the least compliant, while mega‑influencers (>1 M subs) showed higher compliance, likely due to brand contracts and legal scrutiny.
• Affiliate network influence: Networks that enforce mandatory disclosure policies saw a 30 % higher compliance rate among their creators compared to networks without such rules.

Practical Implications

• For developers building influencer tools: Embedding automated disclosure checks into video‑upload SDKs can help creators stay compliant before they publish content.
• For brands and affiliate networks: Enforcing contract clauses that require platform‑level disclosures (e.g., using YouTube’s built‑in label) can reduce legal risk and improve consumer trust.
• For platform engineers: The study demonstrates the value of UI nudges—standardized disclosure fields and prompts dramatically boost compliance with minimal friction.
• For regulators: Data‑driven audits like this provide a scalable way to monitor industry‑wide compliance, informing targeted enforcement actions.
• For consumers: Greater transparency means users can better assess the credibility of product recommendations, potentially reducing “ad‑fatigue” and increasing conversion quality.

Limitations & Future Work

• Detection blind spots: The methodology may miss disclosures spoken only in the video audio (no transcript) or hidden in community posts.
• Affiliate program coverage: While the list includes major programs, niche or emerging affiliate networks could be under‑represented.
• Causality vs. correlation: Although platform changes align with higher compliance, the study cannot definitively prove causation without controlled experiments.
• Future directions: Extending the pipeline to other platforms (TikTok, Instagram Reels), incorporating speech‑to‑text for audio‑only disclosures, and collaborating with regulators to develop real‑time compliance dashboards.

Authors

• Chen Sun
• Yash Vekaria
• Zubair Shafiq
• Rishab Nithyanand

Paper Information

• arXiv ID: 2603.04383v1
• Categories: cs.CY, cs.CR, cs.IR, cs.LG, cs.SI
• Published: March 4, 2026
• PDF: Download PDF