Trust Requires More Than a Disclaimer
The subject of this series has been YouTube’s affiliate-marketing and disclosure ecosystem: a system in which creators can earn commissions from recommendations while viewers decide how much weight to give them. The final question is what makes that relationship trustworthy in practice. The evidence supports a disciplined answer: a commercial relationship must be understandable, attributable to the relevant product or link, difficult to miss in the medium where the promotion appears, and supported by controls that remain open to review and correction.
Presence is not adequacy
The 2023 FTC Endorsement Guides provide the governing frame. A material connection is a relationship that could affect the weight or credibility of an endorsement and that the audience would not reasonably expect. Affiliate compensation fits the Guides’ own example of a blogger who receives a portion of a sale when readers use a link. The point of disclosure is not to declare that the recommendation is false. It is to give viewers information that may change how they evaluate the recommendation.
The same Guides define “clear and conspicuous” as difficult to miss and easily understandable by ordinary consumers. That definition joins two tests that should be kept distinct. The first is semantic: does the notice explain the relevant benefit in plain language? The second is perceptual and relational: can viewers notice the notice in the relevant medium and connect it to the product or link that creates the benefit? A sentence can pass one test and fail the other. “I earn a commission from purchases through this link” may be clear in meaning but ineffective if hidden behind an expansion step. A highly visible “affiliate” label may be conspicuous but still leave an ordinary viewer unsure what the creator receives or which product it covers.
The YouTube study’s categories make the difference measurable. Among 139.9 thousand English-language affiliate videos in the analyzed set, 45.52% contained some form of disclosure, but only 12.20% were classified as clearly compliant with the study’s criteria for compensation and relationship clarity. Another 18.61% were possibly compliant, and 69.19% were classified as non-compliant, including 54.19% with no disclosure. The study does not treat these as legal adjudications. They are model-based classifications of text in descriptions and Shopping surfaces. Their value is that they show why “disclosed” is too coarse a category for understanding whether viewers receive usable information.
The two tests also explain why attribution is central. A viewer may understand that a creator benefits from some links but not know which links are commercial. A description may contain affiliate links, ordinary product pages, a discount code, and social links. A notice that covers a defined affiliate group can clarify the list; a mixed group can leave the viewer unable to tell what the notice qualifies. The missing element is not always more words. It is a connection between the information and the decision the viewer is making.
Visibility adds a temporal dimension. The FTC’s social-media example treats a disclosure that appears only after a viewer selects “more” as not unavoidable when the endorsement is already visible. A profile-page notice can fail for the same reason: viewers encounter a particular post without encountering the profile information first. A notice in the wrong modality can fail as well. The Guides call for visual disclosure of visual representations, audible disclosure of audible representations, and both when the message uses both. The study’s text analysis cannot determine whether audio or on-screen disclosures succeed; it establishes a narrower but still important problem on the surfaces it measures.