Review schema for affiliate sites, after Google's incentivized-review rule

Since July 2026 Google's review snippet guidelines rule out undisclosed incentivized reviews. What affiliate reviewers need to disclose, the Review markup Google accepts, and the rules affiliates trip over.

· Updated 9 min read

A vendor gives you a free annual license, or pays a flat fee, and you publish a 4.5-star verdict with Review markup. Since 24 July 2026 that page needs a clear, prominent disclosure to stay within Google's review snippet guidelines, and not only to satisfy the FTC. Review markup itself is still supported: it's one of the features left in Google's search gallery, and on an eligible page it can show star ratings under the result.

This guide starts with the disclosure rule, because it's new and it decides eligibility, then covers the markup.

Disclosure is now a Google rule

The guideline Google added in July reads: "Don't include fake or undisclosed incentivized reviews on your page or in your structured data markup." Google's examples are reviews written in exchange for money, discounts, vouchers or free products without clear and prominent disclosure of that incentive. The change is listed in Google's documentation updates log for 24 July 2026.

The FTC already required the same thing for US audiences. Its Endorsement Guides FAQ treats affiliate commissions and free products as material connections that must be disclosed clearly and conspicuously, close to the recommendation itself. A line in the footer or a separate "disclosures" page isn't enough under either standard.

What that means on the page:

  • Put the disclosure near the top of the review and next to the rating or verdict, in plain words a reader can't miss. For example: "We earn a commission if you buy through links on this page. The vendor gave us a free license to test."
  • Name each kind of benefit. Commission, a free product, a payment for the review and a discount code are different things; say which apply.
  • Don't wait to be sure commission counts. Google's examples list money, discounts, vouchers and free products, and don't spell out affiliate commission alone. The FTC already treats commission as a material connection, so disclosing it the same way settles the question.

Self-serving reviews

Google has a separate rule for reviews an entity controls about itself: if "the entity that's being reviewed controls the reviews about itself", pages using LocalBusiness or Organization markup aren't eligible for star features. The rule is written for those two types, but the principle matters to affiliates who also own products. If you review a tool your company makes, say so prominently and don't present the review as independent.

The markup Google accepts

Google's required properties for a Review are short: author, itemReviewed, itemReviewed.name, reviewRating and reviewRating.ratingValue. It recommends datePublished, bestRating and worstRating; if the last two are missing, Google assumes a 1 to 5 scale. reviewBody isn't required. It's still worth adding as a short summary of the verdict, as long as it matches text on the page.

author · requiredPerson object, linked author pageFails as: "Editorial Team" or a bare stringitemReviewed · requiredA supported type, with a nameFails as: Service, or stale product datareviewRating · requiredratingValue on a stated scaleFails as: doesn't match the visible ratingreviewBody · optionalShort verdict summary, as on the pageFails as: vendor marketing copy

Pick a supported type for itemReviewed

Google lists the types it supports for review snippets: Book, Course, CreativeWorkSeason, CreativeWorkSeries, Episode, Event, Game, HowTo, LocalBusiness, MediaObject, Movie, MusicPlaylist, MusicRecording, Organization, Product, Recipe and SoftwareApplication. For most affiliate sites that means Product for physical goods and SoftwareApplication (or its subtype MobileApplication) for software and apps. Service isn't on the list, so a review of a service marked up as Service won't get a snippet.

A complete example

The product, price and verdict below are invented for illustration. "Acme Tasks" doesn't exist.

{
  "@context": "https://schema.org",
  "@type": "Review",
  "itemReviewed": {
    "@type": "SoftwareApplication",
    "name": "Acme Tasks",
    "applicationCategory": "BusinessApplication",
    "operatingSystem": "Web, macOS, Windows, iOS, Android",
    "offers": {
      "@type": "Offer",
      "price": "12.00",
      "priceCurrency": "USD",
      "description": "Team plan, per user per month, billed annually"
    }
  },
  "author": {
    "@type": "Person",
    "name": "Jane Doe",
    "url": "https://example.com/reviewers/jane-doe",
    "sameAs": ["https://www.linkedin.com/in/jane-doe-example"]
  },
  "reviewRating": {
    "@type": "Rating",
    "ratingValue": 4,
    "bestRating": 5,
    "worstRating": 1
  },
  "datePublished": "2026-10-01",
  "publisher": {
    "@type": "Organization",
    "name": "Example Reviews",
    "url": "https://example.com"
  },
  "reviewBody": "Verdict: a fast task tracker for small teams, weaker on reporting. Tested by the author on a five-person team for eight weeks. Limits: no Gantt view and basic exports."
}

Notes on it:

  • applicationCategory uses one of Google's values. Google's software app documentation expects values such as BusinessApplication, GameApplication or UtilitiesApplication, not free text like "Project Management Software".
  • The price says what it's a price of. Per user, per month, billed annually. A bare "12.00" for a per-seat plan misleads readers and invites a mismatch with your visible pricing table. Update it when the vendor changes prices.
  • This is the review snippet pattern. A standalone Review whose itemReviewed is the app. Google's separate Software App rich result is a different feature: it needs name, offers.price and a review or aggregateRating on the SoftwareApplication itself.
  • The author is a Person with a page. The full pattern for author pages, including ProfilePage markup, is in Article schema with author.

Ratings

The rating in the markup must be the rating on the page. Google requires review content to be visible to users, so a schema value of 4.5 on a page that shows five stars, or no rating at all, is a mismatch.

Rating distribution is an editorial question rather than a markup one. When every product on a site scores 4.5 or higher, readers stop treating the scores as assessments. Publishing the mixed and negative verdicts you actually reached is what gives the good ones weight.

Roundups and multiple reviews

Several reviews on one page are allowed. When they're reviews of the same item (yours plus readers', for example), Google's guidance is: "If you include multiple individual reviews, also include an aggregate rating of the individual reviews."

"Best of" roundups are less clear-cut. Mark up a Review only for products you actually evaluated, each with its own itemReviewed and rating. Google doesn't promise snippets for list pages, so treat any stars there as a bonus rather than the goal. If the page is a lighter recommendation list, Article markup is the honest choice.

Mistakes affiliates make most

  1. No disclosure, or one buried in the footer. Now a Google eligibility problem as well as an FTC one.
  2. "author": "Editorial Team" or a bare name string. Give each reviewer a Person object and an author page.
  3. Service as the reviewed type. Not supported for review snippets.
  4. Schema rating different from the visible rating.
  5. Stale prices and discontinued products in offers.
  6. Reviews only in the markup. Every review you mark up must be visible on the page.
  7. Markup added only by client-side JavaScript. Google can read it, but most AI crawlers don't execute JavaScript. Render it on the server.

Google also ran spam updates on 18–21 August and from 24 September 2026, the latter still rolling out at the time of writing. If an affiliate site's traffic moved on those dates, start with the Search Status Dashboard rather than the markup.

What the markup doesn't do

It doesn't make a review worth reading. A page that restates the manufacturer's spec sheet has no evaluation in it to mark up, and stars on that page won't change what a reader learns.

It isn't an AI citation lever either. Google's guide to generative AI features says "there's no special schema.org markup you need to add" for AI search. Implement Review markup for star snippets in classic search and for an unambiguous record of who reviewed what. The wider affiliate picture, beyond markup, is in AEO for affiliate sites.

Next step: audit one review page

Pick your highest-traffic review. Check, in this order: the disclosure is near the top and next to the rating; the rating in the markup matches the page; itemReviewed uses a supported type with current prices; the author links to a real page. Then run the URL through Google's Rich Results Test and confirm "Review snippets" is detected with no errors. For missing markup across the page, the tool below lists what's absent (first three issues free, the full report with an email).

AI
Free tool · No signup
Free Schema Validator
Paste any URL → find the structured data your page is missing, with ready-to-paste JSON-LD fixes.
Check your schema →

FAQ

Can I review a product the vendor gave me for free?
Yes, with a clear and prominent disclosure that you received it free, placed near the review and the rating. Google's guideline targets undisclosed incentivized reviews, not incentivized reviews as such. Say what you received, and make sure the review reflects your own use rather than the vendor's talking points.
Can reader reviews and my own review share one page?
Yes. Mark up each review that's visible on the page, and add an aggregateRating that averages the individual ratings, as Google asks when a page carries multiple reviews. Keep your editorial review visibly separate from reader reviews, so nobody mistakes one for the other.

Sources cited in this piece

Updated October 1, 2026: added Google's July 2026 incentivized-review guideline and the self-serving review rule, corrected the required properties and supported types, and replaced a real-product example with a hypothetical one.