Amazon is turning post-purchase review solicitation into a media objective. For brands connected to Chinese manufacturing and global ecommerce, the important change is not simply that reviews may arrive faster. It is that review acquisition can now sit inside the same operating system as product learning and content production—provided the rights boundary stays intact.
On September 29, 2026, Amazon Ads announced Review Requests, with a US open beta scheduled for late October. Eligible sellers and vendors can use Amazon Ads Agent or the Unified API to manage bids and budgets while Amazon reaches recent purchasers on high-traffic pages. Products must have fewer than 1,000 ratings and reviews; campaigns pause when that threshold is reached. Customers may submit a rating, written review, and optional images or video.
Amazon also reported that products in a US closed beta from June 3–21, 2026 received roughly three times as many ratings and reviews per week as before the test. That is short-window internal Amazon data, not a transferable benchmark for every category, price point, or brand.
The practical interpretation is narrower and more useful: a brand can now buy review velocity. It has not bought ownership of the review, the customer’s footage, or permission to turn that experience into an off-Amazon ad.
One campaign can create three outputs—but only one is automatic
Review Requests can contribute to three different outcomes:
- Marketplace proof. The review remains on the product page and helps the next shopper evaluate fit.
- Product learning. Repeated comments about setup, sizing, packaging, use conditions, or unexpected benefits inform the next product, support, PDP, and production decision.
- Reusable testimonial media. A specific customer statement, image, or video appears in brand-controlled advertising, websites, or social channels.
The first two can emerge from review activity. The third requires a separate decision.
A customer submitting content to Amazon does not automatically grant the seller a general off-platform commercial license. Public visibility is not a release. The FTC’s Consumer Reviews and Testimonials Rule Q&A also draws a critical distinction: hosting consumer reviews is different from featuring them in the business’s own advertising or marketing. Once the brand actively deploys a customer experience as promotional proof, truthfulness, representativeness, disclosure, and usage rights need a fresh review.
Review Requests is therefore an evidence-acquisition mechanism, not a UGC marketplace.
The ONCE Review Evidence Ladder
ONCE recommends four stages between a marketplace review and an owned testimonial asset. Each stage adds a new verification gate.
| Stage | What the team may do | Gate before the next stage |
|---|---|---|
| 1. Platform review | Keep the review on Amazon as marketplace proof | Genuine experience, neutral solicitation, no incentive or sentiment filtering |
| 2. Product insight | Cluster repeated questions, failure points, surprises, and use cases | Exclude one-off conditions, model differences, market differences, and logistics noise |
| 3. Proof task | Translate the theme into a verifiable fact and shot requirement | Product owner confirms the fact; the visual can show conditions and limits |
| 4. Licensed testimonial | Use a particular customer experience in brand-controlled media | Separate permission, truthfulness, representativeness, disclosure, territory, term, and channel |
This ladder stops a team from jumping directly from “five stars” to “paid social asset.” It separates what the customer said from what the brand can prove and what it has permission to distribute.
Suppose multiple reviews say a product takes less space in a commuter bag than expected. That does not support “the lightest” or justify downloading a customer’s video. It creates a proof task:
- review theme: smaller packed volume than expected;
- facts to verify: folded dimensions, weight, included accessories, and representative bag sizes;
- shot requirement: one continuous fold, measurement, and bag-fit demonstration;
- visible conditions: same SKU, readable scale, no concealed jump cut;
- claim boundary: no “smallest” or “lightest” without comparative evidence.
The review has now improved the production brief without becoming the production asset.
Use two cost ledgers, not one “UGC” line item
Combining media spend, review analysis, licensing, and editing under one UGC budget makes the economics unreadable. Separate the two jobs.
Ledger A: review acquisition and learning
Cost per validated evidence theme = (Review Requests media + review synthesis + product verification) ÷ themes that pass frequency, product-truth, filmability, and market-relevance checks
The denominator is not the number of reviews. One hundred reviews may repeat a known benefit; a smaller set may expose three issues that materially change a purchase or return decision. The second set can be more valuable to product and creative teams.
Ledger B: testimonial media assets
Cost per compliant testimonial branch = (separate permission + claim verification + editing + disclosure and version maintenance) ÷ approved asset–market–channel–term combinations
A “branch” is a defined bundle of usage conditions, not merely one video file. Permission for six months on a US product page does not inherently cover perpetual European paid social. Counting market, channel, and term exposes the asset’s actual usable footprint.
The two-ledger model also improves procurement. A product may justify faster feedback but need no customer media at all. Another may already have enough reviews but lack credible human proof; that calls for a separately designed and licensed testimonial project, not a higher Review Requests bid.
Four actions: buy, translate, license, or stop
1. Buy review velocity
Test the campaign when an eligible US product has fewer than 1,000 reviews, enough genuine recent purchases, and a sufficiently stable product experience. The team must be prepared to learn from negative and positive themes alike.
2. Translate repeated feedback into proof tasks
When a theme recurs and could change a purchase or return decision, route it to product, support, and production owners. This step does not require copying customer identities or reusing customer content.
3. License testimonial media separately
Use a distinct permission and production workflow only when a particular experience is materially useful, verifiable, and reasonably representative. Do not scrape reviewer identities, treat a public screenshot as consent, or assume a platform license is the brand’s license.
4. Stop and repair first
Pause if product facts are unstable, support and packaging contradict the PDP, or the real goal is positive-only solicitation, negative-review suppression, or access to free customer footage. Amazon’s review policy and Community Guidelines prohibit incentives, review gating, requests to change or remove reviews, and feedback presented as independent when it comes from connected parties.
The procurement decision
Review Requests is best treated as launch-stage evidence acquisition, not mature-catalog media and not a shortcut to owned UGC.
Before funding it, ask:
- Do we need more marketplace proof, or faster evidence for product and creative decisions?
- Can new reviews enter a workflow with named owners and verification gates?
- If we want a specific customer’s content, is there a separate budget for permission, claim review, disclosure, and version control?
If the first two answers are clear, the campaign can shorten the path from real customer experience to better product decisions and more credible proof shots. If the third answer is missing, the customer’s video remains a marketplace review—not a brand asset.
Sources and scope
- Amazon Ads, “Accelerate authentic reviews with Review Requests,” September 29, 2026.
- Amazon Seller Central, “Request a review from a customer” and “Customer product reviews policies,” checked October 7, 2026.
- Amazon, “Community Guidelines,” checked October 7, 2026.
- U.S. Federal Trade Commission, “The Consumer Reviews and Testimonials Rule: Questions and Answers,” checked October 7, 2026.
- The Review Evidence Ladder, two-ledger cost model, and four-action decision are ONCE frameworks derived from these mechanisms. They are not Amazon or FTC terminology, do not guarantee commercial outcomes, and are not legal advice.
