Verification

How Review Verification Works

Four levels of "verified", from self-attested to processor-attested. What each one actually proves, and why the verification source matters more than the badge.

A "verified" review is only as strong as its verification source. Every review platform puts a badge on some of its reviews, but the badge can mean anything from "the business emailed this person" to "a payment processor independently confirmed a real charge from this customer." Here are the four levels, what each one actually proves, and why the source of the verification matters more than the badge itself.

The four verification levels

Level 0. Open platforms (no verification)

Anyone can post a review about any business, with or without an account, with no proof of any interaction. Examples: parts of Google Maps reviews, forum-style feedback, and platforms that accept reviews from non-customers. These reviews can still be useful in aggregate, but no individual review proves anything about a real purchase.

Level 1. Email invitation (self-attested)

The business invites a customer by email, and the review platform trusts the business's claim that this email belongs to a real customer. The reviewer's identity is never independently confirmed. This is the model used by most review platforms, including Trustpilot's "Verified" label, which means "the business invited this reviewer," not "this person bought the product."

Level 2. Receipt or order-number upload

The reviewer attaches a receipt, order number, or screenshot as proof of purchase. Stronger than an invitation, but the evidence is supplied by the reviewer themselves and is trivially falsifiable, receipts are easy to generate, and order numbers are guessable. Platforms that offer this usually apply it only when someone disputes a review.

Level 3. Merchant-supplied order data

The business sends the platform a feed of its own orders (or connects its e-commerce system), and the platform matches reviews to records in that feed. This is the model behind Yotpo's "Verified Buyer" (Mail-After-Purchase matching) and Feefo's invitation-only reviews. It is a real improvement over Levels 0–2, but the party being reviewed (the merchant) is also the party supplying the verification data, and the merchant controls that data in its own systems.

Level 4. Processor-attested (independently verified)

A neutral third party, the payment processor, independently confirms that a real charge occurred between this specific customer and this specific business before a review can exist. The merchant cannot invite a reviewer whose payment the processor has not seen, cannot edit the record, and cannot remove the verification after the fact. This is the level Signed Reviews provides: every review is bound to a specific Stripe charge ID and cryptographically signed at submission.

Verification levelWho supplies the evidenceWhat "verified" provesExample
Level 0. OpenNobodyThat a review was postedOpen platforms
Level 1. Email invitationThe businessThat the business sent an inviteTrustpilot "Verified"
Level 2. Receipt uploadThe reviewerThat the reviewer attached somethingDispute flows on some platforms
Level 3. Merchant-supplied feedThe business (its own records)That the review matches a merchant orderYotpo "Verified Buyer", Feefo
Level 4. Processor-attestedThe payment processorThat a real payment occurredSigned Reviews (Stripe)

Why the verification source matters

The source decides whether fake reviews are prevented or merely detected. On Levels 0–3, a fake review is an invitation list or a doctored order away. The platform can only police reviews after they appear. On Level 4, a fake review from a non-customer is structurally impossible, because the review flow starts with a confirmed transaction rather than a claim about one.

The scale of the problem is measurable. Trustpilot removed 4.5 million fake reviews in 2024, which is 7.4% of all reviews submitted to the platform that year. That is a detection system working at scale, not a prevention system: the fakes were already live, already influencing shoppers, before they were removed.

Regulators have noticed the same gap. The FTC's rule banning fake reviews and testimonials (effective October 2024) targets exactly the practices that Level 4 makes impossible: reviews from non-customers, purchased reviews, and "verified" claims without substance. The FTC's own guidance answers the question directly: a platform's "verified" label must mean the platform actually verified what it claims.

How processor attestation works at Signed Reviews

Signed Reviews connects to a business's Stripe account via Stripe's official OAuth: with least-privilege scopes that allow reading charges, customers, and subscriptions, but can never create charges, issue refunds, or move funds. Then:

  • A charge succeeds. Stripe's charge.succeeded event triggers a unique, expiring review invitation bound to that specific charge ID.
  • The invitation goes to the payment email. The address Stripe has on file for the actual payment, not an address the merchant typed into a list.
  • The review is signed at submission. Review content, charge ID, customer email, and timestamp are cryptographically signed together, creating a tamper-evident record anyone can verify.
  • Refunds hide reviews automatically. If the charge is refunded, Stripe's webhook removes the associated review from public display, no merchant action, no appeal window.

The result is a review that proves three things at once: a real customer, a real purchase, and unchanged content. The cryptographic signature makes the tamper-evident part independently checkable, not a claim by the platform, but a verifiable property of the data.

The verification hierarchy in one sentence

A review platform cannot verify a purchase more strongly than its verification source. If the source is the merchant, the platform verifies the merchant's records. If the source is the payment processor, the platform verifies the payment. Everything else (badges, policies, moderation) is downstream of that choice.

Frequently asked questions

What does a verified review actually prove?

Depends entirely on who did the verifying. An email-invitation "verified" badge proves the business sent an invitation. A merchant-feed badge proves the review matches a record the merchant supplied. A processor-attested badge proves a payment processor independently confirmed a real charge. Before trusting a badge, ask what the platform verified, and against whom.

Can processor-verified reviews be faked?

Not without committing real payment fraud. To produce a fake processor-verified review you would need to run a genuine payment through the merchant's Stripe account and then have the cardholder submit a review: an attack that costs money, leaves a financial record, and gains nothing that paying customers don't already provide. That is the structural difference between prevention and detection.

Which review platforms verify against the payment processor?

Signed Reviews is the only review platform that attests reviews against Stripe's own charge record. Other platforms verify against data the merchant supplies: an email address, an uploaded order list, or an e-commerce feed. None of them confirm the payment independently.

Start collecting processor-verified reviews →

Related: What "Verified Buyer" means · What Is a Verified Review? · Are Trustpilot Reviews Reliable?