Should you eliminate aggregated multi-product ratings from your pages?

Why does Google consider aggregated ratings misleading?

A multi-product aggregated rating obscures the qualitative differences between distinct references. Imagine a page comparing three smartphones: a high-end model rated 4.8/5, a mid-range one at 3.5/5, and an entry-level model at 2.7/5. Displaying an overall average of 3.67/5 at the top of the page completely muddles the useful information.

A user quickly scanning the results does not understand which rating applies to which product. They might click thinking all models are rated 3.67, while only one product truly deserves their attention. This intentional ambiguity degrades the quality of the SERP and frustrates search intent.

What types of content are targeted by this guideline?

The guideline primarily targets comparison pages, buying guides featuring multiple products, and selections of “best X of [category]”. These formats often accumulate between 5 to 15 different products on the same URL.

Some sites display an overall score calculated from all the listed references, hoping to capture clicks through rich snippets. Others aggregate customer reviews from different products into a single structured data block. Both approaches fall within the scope of this prohibition.

Does this rule also apply to business evaluations?

Google's wording explicitly covers both product AND business evaluations. For businesses, the classic pitfall is aggregating ratings from multiple distinct locations under a single entity.

A restaurant chain with 12 locations cannot display an average calculated from the reviews of all 12 addresses. Each location must carry its own ratings, with its own LocalBusiness schema in structured data. Geographic confusion is just as detrimental as product confusion.

• Each distinct product deserves its own visible rating and separate schema.org markup
• Comparison pages must clearly show which rating corresponds to which product, without an overall average
• Multiple locations of the same brand require distinct geolocated evaluations
• Rich snippets should never display a misleading aggregation in the SERPs
• Transparency about the source of reviews remains essential for each individual evaluation

Is this statement consistent with observed practices in the field?

The guideline aligns with what we observe in manual penalties applied to sites abusing Review markup. For several years, Google has penalized artificial aggregated ratings that inflate CTR without providing real value. Comparison site publishers have already experienced abrupt disappearances of their stars in the SERP.

However, the border remains blurry in certain cases. Does a page listing “the 10 best VPNs” with an individual rating for each VPN pass through? Yes, if each product holds its own evaluation block. But what about an overall rating reflecting the quality of the selection itself rather than the products? [To be verified] — Google does not specify if rating the editorial relevance of the guide constitutes a violation.

What nuances should be applied based on content type?

Technical comparison tests pose a real dilemma. A lab tests 20 headphones based on 8 measurable criteria (isolation, frequency response, distortion, comfort). Can it publish an overall ranking with a final score per product calculated on these 8 axes? Theoretically yes, because each headphone receives its own final evaluation.

The problem arises if the site then aggregates these 20 scores into a single average “overall quality of our selection.” This practice crosses the red line defined by Google. The same logic applies to B2B comparators evaluating SaaS software: each tool must display its own score, never an average category score.

In what cases does this rule not really apply?

Variations of a single product probably fall outside this constraint. A smartphone offered in three storage capacities (128 GB, 256 GB, 512 GB) is fundamentally the same device. Displaying an overall rating for this model, calculated from all reviews across capacities, misleads no one.

The same reasoning applies for cosmetic variations: a t-shirt sold in 8 colors can legitimately carry a single rating. However, once comparing references with different technical characteristics — even from the same brand — each requires its own distinct evaluation.

What should you do concretely on your comparison pages?

Start by auditing all your URLs that group multiple products or locations. Identify the aggregated ratings in both the visible content AND in the schema.org tags. Each Review or Product block should correspond to a unique reference, with its own aggregateRating if applicable.

Visually, structure your comparisons with individual product cards. Each clearly displays its title, specific rating, the number of source reviews, and a link to the detailed file. Eliminate any mention such as “Average rating of our selection: 4.2/5” at the top of the page.

What mistakes should be avoided in structured data markup?

Never create a single Product object encompassing several distinct references with an aggregateRating calculated on the whole. Even if you visually separate the ratings on the page, Google scans the JSON-LD or microdata and detects the inconsistency.

Avoid also artificial setups where you create a fictitious product “Guide to the Best X” just to attach an overall rating in rich snippet. This manipulation fools no one on the algorithm side and exposes your site to a manual action for misleading reviews.

How can you check that your site complies with this guideline?

Use Google’s Rich Results Test tool on your comparison pages. Ensure that each detected Product or Review block corresponds to a clearly identifiable unique product. If the tool reports an aggregateRating without a specific SKU or with multiple mixed references, correct it immediately.

Also, check the Enhancements report in Search Console, under Product Reviews. The errors “Rating not attributable to a specific product” often signal problematic aggregations. Cross-check this data with your server logs to identify pages that lost their stars in the SERP after a recent crawl.

• Remove any average score calculated across multiple distinct products
• Create a separate schema.org Product block for each compared reference
• Visually display each product's rating with its source of reviews
• Ensure no global aggregateRating exists in the JSON-LD of comparisons
• Audit multi-location pages with distinct Google Business Profiles per location
• Test rich snippets after corrections to confirm correct display