What the FTC fake reviews rule bans for online stores

Why should a store read the rule itself?

Most stores hear about this rule second hand, from a review app's blog or a newsletter, and those summaries tend to stop at "don't buy fake reviews". The rule is longer than that. It also covers who in your company may post a review, what you may say to a customer who left a bad one, and how your review page may hold some reviews back. You can break it without writing a single fake review.

The text is short. Part 465 has nine sections and fits on a few screens. We read it on 3 November 2026 in the eCFR version that is up to date as of 30 October 2026, together with the FTC's question-and-answer page for businesses and its press releases. What follows is the rule's own text in plain words, not legal advice. If a point decides something for your business, read the section itself or ask a lawyer.

What does the FTC fake reviews rule ban?

Section 465.1 defines the terms and 465.9 says the parts stand on their own if a court strikes one down. The bans sit in between. One definition matters for every row: a consumer review includes a bare star rating with no text, so the rules about hiding and buying reviews apply to star counts too.

Two exceptions run through 465.2 and 465.5. A business that sends the same general request to all its buyers ("tell us how it went") is not liable for what they choose to write, and neither is a business that only hosts reviews. That is why an automatic review email after every order is fine, while a hand-picked list of happy customers is a different matter, as the next part shows.

Does the rule cover AI-written reviews?

Yes, though the text never says "AI". Section 465.2 is about reviews that misrepresent whether the reviewer exists or used the product, and it does not care how the words were produced. The FTC's press release of 14 August 2024 says the rule is meant to deter AI-generated fake reviews.

So a review that a tool writes "in the voice of a customer" is a fake review, even if a real customer bought the product. The same goes for rewriting a short real review into a longer, warmer one. The rule does not address AI that works on top of real reviews and says so, such as a labelled summary of what buyers wrote, a translation marked as one, or a reply draft you send under your own name. Those are not reviews, but a summary that leaves out the complaints can still mislead. Our test of AI review summaries shows where those summaries go wrong.

Can you ask only happy customers for reviews?

The rule itself does not forbid it. The FTC's Q&A page says it "does not contain a specific prohibition against such conduct", and then warns that the practice could still break the FTC Act under the Endorsement Guides. For Google reviews the answer is simpler: Google's content policy, read on 3 November 2026, lists "selectively solicit positive reviews from customers" among the things merchants may not do.

The safe pattern is the one 465.2(d) protects: ask every buyer, the same way, at the same point after delivery. Our guide to when to send the review request covers that timing. Holding the request back while a customer's complaint is still open is a grey area. It is not a sentiment test on the review, but it does keep some unhappy buyers out of the pool, so be ready to explain why you do it and how long the hold lasts.

What about rewards for reviews and hiding bad ones?

These two points, 465.4 and 465.7(b), are the ones stores ask about most, and we cover them at length in can you offer a discount for a review. In short, a reward is allowed when it does not depend on the rating, and the FTC's Q&A adds that a reward you do not disclose could still be deceptive.

Hiding is allowed for reasons that apply to every rating. The rule lists them: trade secrets, defamatory or abusive content, someone else's personal information, discriminatory content, clearly false content, reviews you reasonably believe are fake, and reviews unrelated to what you sell. Sorting is not hiding; the FTC's Q&A says "organizing reviews is not suppressing reviews under the rule", but it also warns against an order that makes the negative ones hard to find.

Does the rule require verified buyers?

No. Part 465 has no verification duty, and the FTC's Q&A page does not mention one. The duty in the rule is negative: do not publish what you know or should know is fake.

The EU is different. A shop that sells to EU customers must say whether and how it checks that reviews come from buyers, which our article on the Omnibus Directive and verified reviews explains with notices you can copy. A "Verified buyer" label is a claim of its own, so it should appear only on reviews you can tie to a real order.

What should a store's review setup do?

Here is the rule turned into settings and habits. The section numbers point back to the table above; the line about sorting comes from the FTC's Q&A.

• Ask every buyer through the same automatic email, not a list you pick by hand (465.2(d)).
• If you give a reward, give it for any rating, say so in the request, and show a label on the review (465.4).
• Keep a short written moderation policy with the reasons you hide reviews, and use only those reasons, for 1 star and 5 stars alike (465.7(b)).
• Show reviews newest first by default, and let shoppers filter to the low ones.
• Tell staff and family not to review your products, or to say who they are when they do (465.5).
• Answer bad reviews in public with facts, never with a threat. Our guide to responding to negative reviews has wording that works (465.7(a)).
• Never import reviews you cannot trace to a customer, and never let a tool write them (465.2).

Our article on why bad reviews help sales makes the business case for the hardest line on that list.

How does Formgong's review setup line up?

We checked our own code against each point. Review emails go to every buyer of a connected store. Opt-outs and consent rules can skip a buyer, and a second order within 14 days waits its turn, but there is no setting that picks customers. The one hold works like this: if the buyer has an open help-chat ticket or an unanswered message to one of the store's forms, the email waits a day and checks again, and after 14 days it is skipped. That hold looks at the open question, never at a rating, and it is the closest thing to selective asking that the product has.

When review coupons are on, every review that comes through the order email gets the label "Received a store coupon for reviewing", and the coupon goes out whatever the stars. A review sent through the order link gets "Verified buyer", because it is tied to that order and that product. Reviews imported from another app keep a "Verified buyer" label only if the file says so, and we cannot check that file for you. None of the four publishing modes looks at the stars, and in the AI mode the model is never shown the rating. Hiding a review needs one of six reasons (Spam or fake, Duplicate, Not about your business, Personal data, Threats, hate or harassment, Test), and only test, spam and duplicate reviews can be deleted.

That is where we differ from some review tools. Trustindex's page on buying Google reviews, read on 3 November 2026, warns readers about the FTC and in the same layout lists "Hide your negative reviews" among its features. Formgong has no such switch, and "negative" is not one of the reasons the hide button accepts. A list cannot stop an owner from picking "Threats, hate or harassment" for a review that is only critical, so keeping the policy honest is still your job.

Did we miss a part of the rule?

If your lawyer reads a section differently, or you have seen how the FTC applies it to a small store, write to support@formgong.com. We will correct this page and say what changed.