Official statement
Other statements from this video 10 ▾
- 2:17 Est-ce qu'ajouter du contenu hors-sujet sur un site pénalise vraiment son ranking ?
- 5:18 Faut-il vraiment abandonner les sous-domaines pour un site unique ?
- 12:07 Ajouter de nouveaux produits dilue-t-il vraiment vos signaux SEO ?
- 15:51 Faut-il vraiment bloquer le contenu par robots.txt pour le désindexer ?
- 25:21 Faut-il vraiment optimiser manuellement chaque meta description si Google les réécrit ?
- 26:27 AMP, JavaScript et mobile : quelles priorités pour optimiser votre référencement ?
- 46:40 Google utilise-t-il vraiment les mêmes algorithmes pour tous les secteurs ?
- 60:30 Faut-il vraiment personnaliser les avis produits pour chaque fiche ?
- 68:36 Pourquoi Google crawle-t-il certaines pages plus souvent que d'autres ?
- 76:01 L'HTTP/2 améliore-t-il vraiment le SEO sans intervention manuelle ?
Google claims that duplicate reviews across multiple pages dilute their value for rich snippets. Only reviews specific to the product on each page truly matter. In practice, duplicating the same reviews across your catalog could cause you to lose your rich snippets in search results, directly impacting your click-through rate.
What you need to understand
What exactly does 'duplicate reviews' mean according to Google?
Google refers here to duplicated reviews that are identical or nearly identical across several product pages on the same site. This practice was common a few years back: collecting reviews from an entire category and displaying them everywhere to artificially inflate star ratings.
The engine detects these duplication patterns through its structured data markup processing algorithm. When a review with the same content, author, and date appears on 50 different pages, the signal is clear. Google then considers that these reviews do not reflect the real experience of the specific product.
Why is Mueller making this statement now?
Google's stance has gradually hardened. The Schema.org guidelines for reviews have always emphasized relevance, but enforcement was vague. Mueller clarifies: this is not just a recommendation, it's a criterion for rich snippets eligibility.
This clarification likely addresses a spam inflation issue in review snippets. Too many e-commerce sites were abusing the system by replicating the same 10 five-star reviews across their entire catalog. The risk? Deteriorating user trust in rich results.
What’s the difference between 'generic' and 'specific' reviews?
A specific review mentions characteristics unique to the product: size, color, specific functionality, comparison with a similar model. It logically can only exist on one page. A generic review like 'Excellent service, quick delivery' could apply to any product.
Google wants reviews that help users evaluate this specific product, not corporate testimonials about the brand or seller. If your review only talks about customer service without mentioning the purchased item, it doesn't belong in the Product Schema markup.
- Reviews must focus on the exact product of the page, not on the brand, seller, or category
- Each product page requires its own set of reviews, no copying from another URL
- Aggregated category reviews do not count as specific reviews for the individual snippet
- The Schema markup must reflect reality: if the review is not visually present on the page, do not include it in the code
- Google can penalize snippets even if the technical markup is valid, if the content abuses the system
SEO Expert opinion
Does this rule really apply strictly in the SERPs?
On the ground, we see a variable application. Some sites with clearly duplicated reviews retain their stars in results, while others that are cleaner lose them. The algorithm seems to tolerate a small degree of duplication, but the exact threshold remains unclear. [To be verified] how much Google detects these abuses automatically vs manually.
What is certain: penalized sites experience a sudden disappearance of rich snippets across the entire catalog, not a gradual degradation. When Google acts, it is radical. The CTR can then drop by 15 to 30% depending on the verticals, a measurable business impact.
What about brand reviews vs product reviews?
Mueller's statement opens a gray area: are reviews about the brand itself legitimate on every product page? Technically no according to his criteria, yet Schema.org offers an 'Organization' type with reviewRating. The distinction matters.
In practice, it's better to clearly separate the two levels: Product markup with reviews specific to the product, and possibly an Organization block in the footer with brand reviews. Never mix the two in the same Schema object, as this creates semantic confusion that Google could interpret as an attempt to manipulate.
Are third-party review aggregators affected?
Platforms like Trustpilot or Avis Vérifiés often generate automatic markup that can fall into the duplication trap. If their widget displays the same last 5 brand reviews on all your pages, the Schema code follows. You are responsible for the markup present on your site, regardless of its source.
The solution? Configure these tools to display only reviews filtered by SKU or product reference. Most reputable aggregators offer this option, but it must be activated. A regular audit of the Schema deployed via Google Search Console is essential to ensure the third party adheres to the rule.
Practical impact and recommendations
How to audit your current reviews to detect duplications?
First step: extract all the Schema Review markup from your site using a crawler (Screaming Frog, OnCrawl). Export the fields author, reviewBody, datePublished, ratingValue. Create a pivot table to identify content appearing on more than one URL.
Use a text similarity detection script if you have thousands of products. A simple MD5 hash of the review content reveals exact duplications. For near-duplications, a Levenshtein or Jaccard distance calculation will do the job. Any review present on 3+ different pages should be flagged.
What strategy should you adopt if you lack specific reviews?
Let’s be honest: it’s better to have zero review in Schema than duplicated reviews. The absence of star snippets is neutral, while the penalty for spam is destructive. If a product has no own reviews, don’t rig it by copying those from a similar product.
To generate real specific reviews, launch targeted post-purchase campaigns by SKU. Encourage customers to detail their product experience: guided questionnaire with mandatory fields on precise characteristics. A quality review is worth more than ten generic ones for SEO as well as conversion.
How to restructure your Schema markup to remain compliant?
Implement a strict conditional logic: the Review markup only appears in the code if the review is linked to the itemReviewed with its unique identifier (SKU, EAN, GTIN). No fallback to category or brand reviews. Technically, create a DB join between your reviews table and products on the product primary key.
Always test via the Search Console Improvement section, then Rich Results Test. Ensure each URL only retrieves reviews that are its own. An automated monthly audit via Search Console API helps detect any drift before Google acts.
- Crawl the entire site and extract all Schema objects of type Review
- Identify any reviewBody content appearing on 2+ distinct URLs
- Remove the markup for duplicated reviews, even if they are 'relevant' to the category
- Configure third-party aggregators to filter by unique product identifier
- Establish a system for collecting post-purchase reviews with required SKU reference
- Validate the new markup via Rich Results Test before going live
❓ Frequently Asked Questions
Combien de reviews dupliquées Google tolère-t-il avant de pénaliser les snippets ?
Les avis d'une variante produit peuvent-ils être affichés sur les autres variantes ?
Un avis mentionnant plusieurs produits achetés ensemble est-il valide ?
Que faire des anciennes reviews dupliquées déjà indexées par Google ?
Les rich snippets perdus à cause de reviews dupliquées reviennent-ils après correction ?
🎥 From the same video 10
Other SEO insights extracted from this same Google Search Central video · duration 59 min · published on 15/12/2015
🎥 Watch the full video on YouTube →
💬 Comments (0)
Be the first to comment.