At launch, a brand owns an internal version of product truth: claim sheets, lab tests, a director’s script and the detail page. Once real buyers start using the product, a second version appears. Assembly may be harder than expected. Size may be misread. A feature may need a clearer first-use explanation. Some benefits may be worth filming; other promises may fail in normal conditions.
Video production and review operations have often remained separate. Amazon’s coming Review Requests product brings them closer by letting advertisers reach recent purchasers who can submit a rating, text, images or video. The structural opportunity is not “getting five stars faster.” It is receiving concentrated post-purchase evidence early enough to improve the product, page and next production cycle.
What Amazon actually announced
According to Amazon Ads’ September 29, 2026 launch announcement, Review Requests enters open beta in the United States in late October 2026. It is available to registered sellers and vendors with an advertising-console account in good standing and no active policy violations. Products qualify while they have fewer than 1,000 total ratings and reviews; campaigns pause automatically at that threshold.
The ads invite recent purchasers from high-traffic Amazon pages to leave a star rating or written review, with optional video or images. Campaigns run through Amazon Ads Agent or the Unified API with familiar bidding and budget controls.
Amazon reports that products in early testing received roughly three times as many ratings and reviews per week as before the test. That is a platform-reported early result, not a category-wide benchmark, and it says nothing about the direction of the ratings. The advertiser is buying access to an unbiased invitation, not control over what customers say.
A buyer review is not a scripted UGC ad
Amazon’s Community Guidelines permit genuine positive or negative experience, including a customer’s own photos and videos. They prohibit creating, editing or removing reviews in exchange for payment, refunds, discounts, products, gift cards, warranties or services. Sellers may encourage honest, unincentivized feedback; they cannot require a favorable conclusion.
The US Federal Trade Commission’s Consumer Reviews and Testimonials Rule guidance explains that the rule took effect on October 21, 2024. It covers practices including fake reviews, purchasing positive or negative reviews, undisclosed insider reviews and certain review-suppression conduct. It also distinguishes reviews from testimonials. A platform may host customer evaluations; when a brand places selected customer language in its own advertising, it is disseminating a promotional message rather than merely hosting feedback.
A real customer video on a detail page therefore does not automatically grant the brand permission to download, edit, score and distribute it in global advertising. Reuse needs a separate rights, context, accuracy and disclosure check. ONCE’s framework for creator-video usage rights applies when review content moves into paid or owned media.
The ONCE LISTEN workflow
LISTEN is not review solicitation copy or a reputation-management tactic. It connects launch production to post-purchase learning.
L — Lock the launch truth
Archive approved claims, test conditions, product version, instructions, detail page and source video at launch. Without that baseline, teams cannot tell whether a complaint comes from a product defect, batch change, misunderstanding or an exaggerated ad promise.
I — Invite without influence
Use permitted review-request mechanisms. Do not tie compensation, replacement, refund or service to positive sentiment, editing or deletion. Ask for honest experience without suggesting a star rating or supplying model praise.
S — Separate review from promotion
Treat detail-page reviews as customer feedback first. Before any quote, image or video enters an ad, brand site, social post or retail asset, confirm permission, context, editing and disclosure. Public visibility is not commercial-use permission.
T — Tag recurring friction
Do not build strategy around the most enthusiastic quote. Tag issues by assembly, sizing, connection, noise, cleaning, compatibility, packaging, instructions and durability. Record use context and product version. An individual review is one experience; a recurring and reproducible issue is a decision input.
E — Edit the next answer
If buyers repeatedly struggle with setup, improve the instruction card, detail page and tutorial. If scale is misread, reshoot with useful reference objects. If the problem is inside the product, route it to product and supply-chain owners before stronger copy hides it.
N — Note what changed
For each update, record the triggering issue, correction, SKU, release date and owner. Then watch whether the same issue persists. Reviews become a learning history instead of a folder of screenshots.
Deliver a launch-to-learning system
For brands whose products are manufactured or iterated in China, the production scope can span three stages.
Before launch: produce checkable product proof
Capture authentic operation, dimensions, box contents, installation, limitations and commonly misunderstood features. Retain claim sources and test conditions. Deliver replaceable step shots, captions and product inserts so corrections do not require rebuilding the whole campaign.
After launch: build a review observation log
Operations, service, product and content owners should review feedback over an agreed period without ranking only by stars. The log should contain a concise source summary, issue tag, product version, reproducibility, whether current content already answers it, and the owner—product, service or content—responsible for the next step.
Next cycle: produce only for repeated questions
Not every review deserves a video. A topic enters the next brief when it recurs, affects purchase or use, has an accurate product answer and remains unresolved by existing content. The output is not simply more assets. It is clearer expectations, fewer repeated questions and product proof aimed at real friction.
A better cost denominator for ongoing work
Counting only first-round deliverables ignores the value of post-launch learning. A more useful management denominator is:
Cost per resolved recurring buyer question = (launch proof assets + compliant review-request media + feedback analysis + product/page/video correction + verification) ÷ recurring questions with a completed corrective update
This is not an industry price benchmark and it promises neither review volume nor sales. It separates receiving feedback from resolving it. A learning unit is complete only when the signal is verified, assigned, corrected and returned to observation.
When a full loop is unnecessary
Mature products with abundant stable reviews, few product changes and issues dominated by shipping or isolated service events may need only normal customer service and page maintenance. Review Requests is also a US open beta with seller, account and review-count eligibility; brands should not assume availability in other marketplaces.
LISTEN is more useful for new launches, high-return categories, installation-dependent products, complicated specifications and products that change through a Chinese supply chain. English-language confusion, US installation habits and factory-side design changes need one shared record rather than separate ad accounts, support inboxes and editing chats.
Six questions before activation
- Are we inviting all eligible real purchasers, or trying to select only satisfied buyers?
- Is any payment, product, refund or service tied to a positive review, edit or deletion?
- Have we preserved the launch claims, product version and test conditions?
- Before review content enters advertising, have rights and context been checked again?
- Should each recurring issue be solved by product, instructions, the detail page, service or video?
- How will the next cycle determine that friction declined rather than being covered by louder advertising?
The significance of Review Requests is not that brands can buy more reviews. It is that authentic post-purchase feedback is becoming a manageable advertising input. For ONCE and its clients, the higher-value move is to turn a launch video into a learning system: prove the product honestly, listen carefully, and make the next product and content version better together.
Sources and limits
This analysis uses Amazon Ads’ September 29, 2026 Review Requests announcement and beta specifications; Amazon Community Guidelines and Customer Product Reviews Policies; and the FTC Consumer Reviews and Testimonials Rule and staff Q&A. The three-times figure is Amazon’s early-test report. LISTEN, the three-stage delivery model and the cost denominator are ONCE analysis, not Amazon rules, legal advice or performance guarantees.
