What is a verified review?

A review where the platform confirmed something about the reviewer before publishing — most often that they bought the product, sometimes only that they hold an account. What the badge covers varies by platform: Amazon's confirms a purchase through Amazon, Google's confirms only a Google account, and Trustpilot's depends on whether the review was invited or left organically.

What does a verified purchase badge actually prove?

That money changed hands through that platform. It doesn't prove the reviewer used the product, that nothing was exchanged for the review, or that a human wrote it. A verified purchase combined with AI-written text passes every automated check while containing no information from a person.

What does verified purchase mean on Amazon?

That the reviewer bought the item through Amazon at close to the listed price — steep promotional discounts usually disqualify a review from carrying the marker. It doesn't mean Amazon assessed the review, that the reviewer kept the product, or that they used it before writing. A review posted the day a parcel arrives carries the same marker as one written after a year.

What does the symbol next to a Google review mean?

Usually a Local Guides badge, which means the reviewer contributes regularly to Google Maps and has earned points for it. It reflects activity volume, not whether they visited the business. Google has no purchase-verification marker of the Amazon kind — anyone with an account can review anywhere, and removal for policy breaches happens after publication rather than before.

How do you spot a fake review on Amazon?

Look at patterns rather than individual reviews. A cluster of positive reviews arriving within days after months of silence is a campaign. Praise that would fit any product in the category, with no detail specific to this one, is text written to order. And shared sentence rhythm across supposedly unrelated reviewers is the clearest signal there is — strangers don't write alike.

How can you tell if a review was written by AI?

You often can't, from a single review. Across several, the signals are similar sentence rhythm between supposedly unrelated people, praise with no specifics that could only come from use, and clusters of reviews arriving in the same few days after a long gap.

What "verified review" actually means on each platform

Q: What makes a review verifiable if there's no platform badge?

Whether a real person can be located behind it. A full name, a role, a company, and ideally a link to a professional profile make the claim checkable in principle — almost nobody follows the link, but the fact that they could is what does the work. Video carries the same information more efficiently, because a face, a voice, and an unscripted pause are expensive to fabricate.

A verified review is one where the platform confirmed something about the reviewer before publishing it — usually that they bought the product, sometimes only that they hold an account. The word carries less weight than most shoppers assume, and what it covers changes from platform to platform. This page sets out what Amazon's verified purchase marker and Google's review symbols actually confirm, what no badge confirms, and how to spot a fabricated review.

What "verified" confirms on each platform

The badge is not a standard. Each platform verifies a different thing, and the differences matter more than the shared label suggests.

Platform	What the badge confirms	What it doesn't
Amazon (Verified Purchase)	The reviewer bought the item through Amazon	That they used it, that the review is honest, or that the account is real
Google reviews	The reviewer has a Google account	Any purchase, visit, or interaction with the business
Trustpilot	Depends on the route — invited reviews are tied to a transaction, organic ones aren't	That an unlabelled review has any transaction behind it
G2 and similar B2B sites	Identity, usually via work email or a professional profile	Depth of use, or that the reviewer wasn't incentivised
A review on your own site	Whatever you choose to check	Nothing by default — the visitor has only your word

Read down the second column and the pattern appears: verification is about the transaction or the identity, never about the content. No platform verifies that a review is truthful. That sits outside what any automated system can check.

What a verified badge cannot tell you

Three things sit permanently outside verification, and they're the three that shoppers care about.

Whether the person actually used the product. A verified purchase confirms money changed hands. Someone can buy an item, leave a review the same afternoon, and the badge appears exactly as it would after six months of use.

Whether anything was exchanged for the review. Incentivised reviews break platform rules and happen anyway, through free products, discount codes, or private arrangements that leave no trace in the transaction record.

Whether a human wrote it. This is the one that has changed. A verified purchase plus AI-written text produces a review that passes every check a platform performs and contains no information from a person.

What "verified purchase" means on Amazon

The marker appears when the reviewer bought the item through Amazon at a price close to the listed one — steep discounts through promotional codes usually disqualify a review from carrying it.

That's the whole scope. It doesn't indicate that Amazon assessed the review, that the reviewer kept the product, or that they used it before writing. A review posted the day a parcel arrives carries the same marker as one written after a year of use.

Reviews without the marker aren't automatically suspect, either. Someone who bought the same item in a shop and reviewed it on Amazon has no purchase record there, and their review may be the more useful one. The marker narrows a question rather than answering it.

What the symbol on Google reviews means

Google shows an icon or a label beside some reviews, and shoppers read far more into it than it carries.

A Local Guides badge means the reviewer contributes regularly to Google Maps — reviews, photos, corrections — and has accumulated points for it. It says something about volume of activity and nothing about whether they visited the business in question. Some businesses also display a Google Guaranteed or Google Screened badge, which relates to vetting of the business through Local Services Ads, not to any individual review.

What Google doesn't offer is a purchase-verification marker of the Amazon kind. Anyone with an account can review anywhere, and there's no transaction record tying the review to a visit. Reviews get removed for policy breaches after the fact, which is a different mechanism from checking before publication.

Why the badge stopped doing its job

Written reviews used to be expensive in the only way that mattered: someone had to sit down and compose one. That cost is now zero. AI-generated product reviews are fluent, plausible, specific enough to sound lived-in, and produce at whatever volume you want.

This breaks the badge in a specific way. Verification systems were designed to answer "did this person buy the thing", because that was the hard part to fake. Now the buying is the easy part and the writing is free, so the check confirms the wrong half of the problem.

Shoppers have adjusted faster than platforms. The result is a general discount applied to written reviews — including the honest ones — which is the underlying cost of fabrication reaching zero. Any signal that becomes free to produce stops being a signal.

How to spot a fake review on Amazon

Individual reviews rarely give themselves away. Patterns across a listing do.

Look at the timing first. A cluster of positive reviews arriving within a few days, especially after months of silence, is the residue of a campaign — genuine reviews trickle in at the pace people receive and use things. Then look at what the reviews describe. Praise that would fit any product in the category, with no detail that could only come from handling this one, is the most common signature of text written to order.

Phrasing across reviewers is the third signal and the clearest. Strangers don't share sentence rhythm. When several reviews open the same way, run to similar length, and land on the same closing note, they came from one source.

Two structural traps are worth knowing. Review hijacking happens when an old listing with thousands of reviews is edited to sell a different product entirely, so the review count belongs to something you'll never see — check whether the reviews describe the item on the page. And a refund arrangement, where a seller privately reimburses the purchase in exchange for five stars, produces a genuine transaction and therefore a genuine verified marker.

The most efficient shortcut remains the middle ratings. Three-star reviews are the least likely to have been bought or written in anger, and they tend to carry the specific detail that makes a review worth reading.

What actually verifies a review now

The check that survives is whether a real person can be located behind the review.

A name and a role make a claim checkable in principle. A company alongside them makes it checkable in practice. A link to a professional profile lets a sceptical visitor spend thirty seconds confirming the person exists and does what the review says they do. Almost nobody follows the link — but the fact that they could is what does the work.

Video does this more efficiently than any badge. A recorded face, a voice, an unscripted pause, and a background that looks like somewhere real carry information that can't be typed in. Generating a convincing video of a specific named person who is findable online, saying specific things about a specific product, is a different order of difficulty than generating a paragraph. It remains possible and it remains expensive, and expense is what keeps a signal working.

The related argument about which proof formats have survived scrutiny and which haven't is in social proof examples, and the badge-shaped end of that spectrum is covered in trust badges.

How to check a review yourself

Five things a sceptical reader can do in under a minute:

Look for a name, a role, and a company. Anonymous or first-name-only reviews carry the least verifiable information available.
Check whether the specifics could only come from use. "Great product, works well" contains nothing. "The clip broke after about three weeks" contains a fact.
Compare the phrasing across several reviews. Similar sentence rhythm across supposedly unrelated people is the clearest fabrication signal there is.
Look at the dates. A cluster arriving in the same week, after months of silence, is a campaign.
Read the middle ratings. Three stars carry the most usable detail.

What this means if you collect reviews yourself

You can't issue a verified badge, and imitating one is worse than having nothing — a badge a visitor can't check reads as decoration the moment they think about it.

What you can do is publish the material that makes a review checkable: full name, role, company, and the person on camera rather than in a text box. Collect that at submission time, along with consent, because chasing either afterwards means asking someone who has moved on. The mechanics of collecting proof at the moment it's easiest to get are in how to ask for reviews, and what to do with recordings once they arrive is in video testimonial best practices.

One thing worth being explicit about: moderating what you publish is normal, and choosing your six strongest testimonials for a page is not deception. Filtering who gets asked, so that only the happy ones can respond, is a different thing entirely — and on most platforms it breaks the rules outright.