Reviews as an off-page signal: rankings and AI recommendations
Reviews on third-party platforms help local rankings on current expert opinion, and review platforms are among the sources AI assistants cite. No public study shows that review count or rating drives non-local rankings or AI recommendations. Fake reviews are banned in the US.
In short
- G2 and Yelp were among the ten most cited domains in Peec AI's March 2026 analysis of 30 million AI citations, G2 mainly on Perplexity and Yelp on Google AI Mode.
- No public study isolates review count or star rating as a cause of AI recommendations, so "more reviews means more AI mentions" is plausible and unproven.
- Google's reviews system evaluates a site's own review articles and, according to Google's documentation, does not evaluate third-party user reviews.
- The US Federal Trade Commission's final rule announced on 14 August 2024 bans fake reviews, reviews paid for on condition of a particular sentiment, and insider reviews without disclosure.
- Whitespark's 2026 Local Search Ranking Factors survey of 47 experts, published in November 2025, reports review signals rising for local rankings.
Reviews on third-party platforms do affect how visible you are, but not evenly. For local search, expert opinion says they matter and are rising. For AI answers, review platforms are cited as sources, yet nobody has shown that more reviews produce more recommendations. For non-local organic rankings there is no evidence that your star rating on an outside platform is an input at all.
This lesson separates those three cases, then covers which platforms to work on and what the law allows when you ask for reviews.
What the evidence supports
| Context | What is known | Source and strength |
|---|---|---|
| Local rankings (map results) | Review signals are rising | Whitespark’s 2026 survey of 47 experts, published November 2025. Expert opinion, not measurement |
| Non-local organic rankings | No evidence that third-party ratings are an input | Google’s reviews system covers first-party review content only |
| AI citations | Review platforms are among the most cited domains | Peec AI, March 2026, 30 million citations. Vendor data |
| AI recommendations | No study isolates review count or rating | Evidence gap. Speculative |
| Human trust | Raters are reported to be told to consult reviews when researching reputation | Reported wording of Google’s quality rater guidelines |
Rankings
Local. Reviews on your Google Business Profile and on local platforms are part of how the map results are ordered, on current expert opinion. The local business hub covers the detail, and our lesson on off-page E-E-A-T explains how reviews enter Google’s rater guidelines.
Everything else. It is easy to assume that a 4.8 rating on Trustpilot lifts your organic rankings. Nothing public supports that. Google has a ranking system called the reviews system, but its documentation says it evaluates first-party content that gives an opinion or analysis, such as a site’s own review articles and ranked lists. It does not evaluate third-party user reviews.
Reviews can still help rankings indirectly. A review platform page about your brand often ranks for searches of your name plus “reviews”, which shapes what a buyer sees before clicking. Better reviews mean more people choose your result. Those are effects on people, not a direct ranking input.
AI recommendations
Review platforms are a source that assistants read. In Peec AI’s analysis of 30 million US citations in March 2026, G2 and Yelp were both in the overall top ten cited domains. G2 appeared mainly on Perplexity, where it was in the top five. Yelp appeared mainly on Google AI Mode.
Semrush’s 2026 AI Visibility Index offers one case observation: Patagonia’s visibility in AI answers rested on consistent mentions by third-party reviewers, such as OutdoorGearLab and REI, and on Reddit, more than on its own content. That is an anecdote from a vendor report.
What is missing is any study that isolates the number of reviews, or the rating, as a cause. Our research found none. So the claim to make is narrow:
- If an assistant cites a review platform for your category, what that page says about you can reach the answer.
- A profile with no reviews, wrong pricing or an old description gives the assistant nothing good to use.
- Whether going from 50 reviews to 500 changes anything is unknown.
There is a second kind of review that the evidence treats more kindly: editorial reviews by publications and creators. These are the pages behind many “best X” answers. Our lesson on third-party “best X” lists covers them.
Which platform to work on
A working guide from our research, by category:
| Category | Platforms assistants and buyers tend to use |
|---|---|
| Business software | G2, Capterra |
| Local services | Google Business Profile, Yelp |
| Online shops | Trustpilot |
Do not take this on trust. Run your priority prompts in the assistants your buyers use and note which review domains appear as sources. If none does, reviews are a conversion matter for you and not a visibility one. The trackers in the AI visibility tools comparison report cited domains.
Then do the basic work on the one or two platforms that matter:
- Claim the profile and correct the name, category, description and pricing.
- Ask real customers at a natural moment, such as after a completed order or a solved support ticket.
- Ask everyone, not only the happy ones. Selective asking is the start of the problems described below.
- Reply to reviews, including the bad ones, with specifics.
- Keep it steady. A burst of 40 reviews in a week followed by silence looks manufactured, to people and to platform filters. That is a rule of thumb, not a published threshold.
Reviews and links
Review platforms and software directories are also a small link source. Our research classes curated directories such as G2, Capterra, Product Hunt and AlternativeTo as low-risk listings, costing from nothing to about $500 a year in membership. The links are often nofollow, and every competitor has the same one, so treat them as a baseline. For local firms the same job sits under local citations.
The opposite direction works too. If you give a supplier a testimonial, they often publish it with a link to you, which is one of the cheapest partner links available. The testimonial must be for a product you really use.
What the law prohibits
This section is general information, not legal advice.
United States. The Federal Trade Commission announced a final rule on fake reviews and testimonials on 14 August 2024, to take effect 60 days after publication. It lets the FTC seek civil penalties against knowing violators. The rule prohibits:
- reviews by people who do not exist or who have no real experience of the product, and buying such reviews when the business knew or should have known
- offering compensation that is conditional on a review expressing a particular sentiment, positive or negative
- reviews by officers, managers or employees that do not disclose the connection
- presenting a website the business controls as a source of independent reviews of its own products
- suppressing negative reviews through groundless legal threats or intimidation
- buying or selling fake followers, views and similar indicators
United Kingdom. The Competition and Markets Authority’s guidance on fake reviews says publishers of reviews must take reasonable and proportionate steps to prevent and remove fake reviews and concealed incentivised reviews. The underlying law is reported to be the Digital Markets, Competition and Consumers Act 2024. We did not re-read the Act for this lesson, so check it with counsel before relying on that.
Google. Its spam policies cover user-generated spam, and fake reviews on Google’s own products are reported to be banned under separate content policies that we did not re-read.
The practical line: you may ask for reviews, and you may thank people. You may not pay for a positive one, write your own, or hide that a reviewer works for you.
Common mistakes
- Gating. Sending happy customers to the review platform and unhappy ones to a private form. It skews the record and misleads the next buyer.
- Buying reviews. Banned in the US under the 2024 rule where the business knew or should have known the reviews were fake.
- Spreading thin. Ten reviews on each of eight platforms is weaker than 80 on the one your buyers read.
- Ignoring the facts on the profile. Pricing and feature fields get copied by list writers and read by assistants.
Where to go next
If reviews in your category live on forums more than on platforms, read Reddit for SEO and AI visibility. The glossary defines citation, entity and the other terms used here.
Common questions
Do reviews on third-party platforms affect Google rankings?
For local results, expert opinion says yes. Whitespark's 2026 survey of 47 local SEO experts reports review signals rising. For non-local organic rankings there is no public evidence that your rating on a platform such as Trustpilot or G2 is a ranking input.
Do reviews affect AI recommendations?
Review platforms are cited as sources by AI assistants, so what reviewers say about you can reach the answer. No public study shows that having more reviews or a higher rating makes an assistant recommend you more often.
Which review platform should I focus on?
The one that people and assistants use in your category. A working guide is G2 or Capterra for business software, Google and Yelp for local services, and Trustpilot for online shops. Check which platforms are cited for your own prompts.
Is it legal to offer an incentive for a review?
In the US, the FTC's 2024 rule bans incentives that are conditional on the review expressing a particular sentiment, and FTC guidance says incentives must be disclosed. In the UK, the CMA's guidance treats concealed incentivised reviews as a problem that platforms must prevent and remove.
Do links from review platform profiles help SEO?
A little at most. Profile links are often nofollow and every competitor has the same one. The profile is worth claiming for the reviews and the facts it shows, not for the link.
