Publish Reviews and Testimonials Without Hiding the Sample

Step 1: define the decision question before collecting praise

State what the social evidence should help a reader evaluate: onboarding effort, support quality, a bounded use case, fit for a stated audience, or another specific question. Do not begin with a target sentiment or a dramatic outcome. The decision question determines whom to invite, what context to collect, and which claims need independent substantiation.

Write what the artifact cannot prove. A review can report an experience; it does not automatically show causation, prevalence, or what future customers will achieve. Assign a collection owner and a separate publication reviewer so enthusiasm for a story does not erase sampling and claim limits.

Evidence: U.S. Federal Trade Commission; Electronic Code of Federal Regulations

Steps 2 and 3: design the invitation and preserve contributor choice

Step 2 is to specify eligibility, timing, channels, and whether every eligible participant or a documented sample receives the invitation. Avoid routing satisfied respondents to a public page while diverting dissatisfied respondents into a private path that makes the visible distribution look more favorable. If an incentive is offered, make it independent of sentiment and record identical terms.

Step 3 is informed permission. Explain intended uses, editing, name or image display, duration, withdrawal route, and any compensation or relationship. Obtain the form of consent required for the media and jurisdiction. Keep participation optional and separate service access, support, refunds, or employment consequences from willingness to endorse.

Evidence: U.S. Federal Trade Commission; Electronic Code of Federal Regulations

Steps 4 and 5: verify provenance and retain the original

Step 4 is a proportionate authenticity check. Record whether identity, transaction, or product access was verified and exactly what the check establishes. Do not expose unnecessary personal data or imply that a badge validates every factual statement. Flag employee, owner, agency, family, affiliate, or gifted-product relationships for disclosure and review.

Step 5 is to preserve the received text, audio, image, date, product version, and surrounding questions before editing. Maintain an edit log for spelling, length, translation, anonymization, or composite treatment. Never change sentiment, insert an outcome the speaker did not report, or use a synthetic depiction in a way that misrepresents a real experience.

Evidence: U.S. Federal Trade Commission; Electronic Code of Federal Regulations

Steps 6 and 7: assess claim scope, typicality, and material context

Step 6 identifies every express or implied product claim created by the quote plus headline, visuals, placement, and call to action. Trace objective claims to support that existed before publication. For exceptional outcomes, examine population, baseline, duration, effort, support, and the distribution of experiences. Narrow or reject the presentation when the record cannot support what readers may expect.

Step 7 writes the necessary context: selection method, incentive, material relationship, material conditions, and boundaries of the experience. Place it where it can alter interpretation, not behind a generic link. Health, financial, safety, and other consequential outcomes need qualified scientific, subject-matter, and legal review.

Evidence: Electronic Code of Federal Regulations; U.S. Federal Trade Commission; OECD

Steps 8 and 9: moderate by published rules and present the sample honestly

Step 8 applies content-neutral moderation categories such as personal data, threats, unlawful material, irrelevance, duplicate content, or verifiable inauthenticity. Record the reason and appeal or correction route. Do not remove criticism merely because it is negative, and do not describe a curated set as complete or representative without support.

Step 9 reviews ordering, filters, summary ratings, labels, visual emphasis, and default views. Explain meaningful sorting and sampling boundaries. Test mobile, assistive technology, video captions, and reposted formats. The page should not use layout to hide unfavorable evidence or make a paid relationship appear independent.

• Moderation reason is rule-based and documented.
• Negative sentiment is not itself a removal ground.
• Ordering logic does not manufacture a false majority.
• Labels retain meaning across formats.

Evidence: U.S. Federal Trade Commission; U.S. Federal Trade Commission; OECD

Step 10: publish with an expiry and correction map

Assign a stable artifact ID, owner, publication locations, evidence links, consent status, and recheck trigger. Inventory advertisements, landing pages, email, partner kits, video, screenshots, and structured data so one correction can propagate. Remove or revise the artifact when consent, product version, relationship, evidence, or typicality context materially changes.

Recheck no later than 2027-02-10. Preserve a minimal audit trail consistent with privacy duties and access controls. The FTC Q&A and eCFR are official U.S. sources; OECD contributes independent policy context and the FTC dark-pattern report informs display review. This workflow cannot certify compliance for another jurisdiction.

Evidence: U.S. Federal Trade Commission; Electronic Code of Federal Regulations; U.S. Federal Trade Commission; OECD