How Review Platforms Catch Seeded Reviews, and What Honest Brands Do

A seeded review is one the business arranged and then dressed up as an independent customer's opinion. Staff writing five-star reviews. A friend "helping out" at launch. A broker who pays strangers, or one person running a stack of accounts.

People land here for different reasons. Some run a brand and can't work out why their reviews keep vanishing. Some suspect a competitor is padding its rating. Some are agency or QA staff who've been told to "test the review flow" and have a bad feeling about it.

And some want to know how to seed reviews without getting caught. We won't help with that one, and we'll be blunt about why further down.

See what your own browser gives away. Our free checker at justbrowser.app/fingerprint-checker runs in your browser with no sign-up. It's a quick way to see the signals described below with your own values next to them, which makes the rest of this post far less abstract.

What Counts as a Seeded Review

The line around seeded reviews is wider than most people assume. Under most platform policies, all of these are out:

Reviews by owners, employees or their families about their own business
Reviews from friends who were asked to post, without disclosure of the relationship
Reviews paid for with money, free products, refunds, discounts or gift cards, where the reward depends on posting (or on posting something positive)
Review swaps between sellers ("I'll review yours if you review mine")
Several accounts controlled by one person or one team
"Test" reviews on a live public listing

That last one surprises people. It shouldn't. We'll come back to it.

A close cousin is review gating: only inviting customers you expect to be happy, or quietly burying the negative ones. It distorts the rating as much as a bought five-star does, and several platforms prohibit it.

Our strong view: gating is the version honest teams actually commit. It never feels like cheating. You only asked the nice ones, nobody wrote a fake word, and the number on your listing is still a lie.

How Platforms Actually Catch Seeded Reviews

Nobody outside these companies knows their full systems, and they don't publish them. Which is the first irritating thing here: you can read every public word a platform has published about review fraud and still not know which rule you tripped. The broad shape of how seeded reviews get caught is clear enough from policy pages and enforcement announcements, though. Think of it as four questions asked about every review.

1. Who is this reviewer?

Account history matters a lot. An account that has bought things, browsed, and reviewed occasionally over years reads very differently from one created last Tuesday whose only activity is a five-star review.

Then there's the relationship to the business. Platforms look for links between reviewers and the business itself: shared contact details, the same physical location, connections through other accounts, or a reviewer who reviews the same small cluster of sellers over and over. Amazon marks reviews tied to a purchase made on Amazon as "Verified Purchase", and a run of glowing reviews from people who never bought the product is an obvious tell.

2. Where is the review coming from?

This is the device and network layer, and it's the part that gets talked about most.

Every browser exposes small technical details. How it renders a hidden canvas image, what its WebGL stack reports about the graphics hardware, which fonts are installed, how its audio processing behaves, plus screen size, timezone, language and Client Hints. None of these identify you alone. Combined, they form a fingerprint that tends to stay the same across cookie clears and private windows.

IP data sits alongside it: which network a request comes from, whether it's a hosting provider, and whether that address has been tied to other flagged activity.

When "ten independent customers" share a fingerprint, a network, and a posting window, the platform has good reason to think they're one person.

Here's the honest caveat, though. A shared device isn't proof of anything. Families share laptops. Offices share a connection. So platforms treat it as a signal to weigh, not a verdict on its own.

3. How does the activity behave?

Real reviews trickle in roughly in step with sales. Seeded reviews tend to arrive in bursts, often right after a launch or a bad review. And a flood of five-stars with almost nothing in between looks nothing like the messy spread genuine customers produce.

4. What does the review say?

Text similarity, repeated phrasing, templated structure, reviews that describe features the product doesn't have. This is the layer people expect to matter most. In practice it's usually the confirmation, not the first alarm.

The part everyone forgets

Honestly, we think the device layer gets far too much attention, and we're part of the reason. Fingerprints are the bit a browser company understands best, so it's the bit we write about most, and that skews how the whole problem looks from the outside. The device layer is only one signal of several, and separate browser profiles don't make coordinated accounts look independent. Even when the devices look different, seeding programmes still get caught, because somebody has to recruit the reviewers and pay them, and the timing and wording still cluster.

Recruiting and paying people leaves trails: groups where review-for-pay offers get posted, refund requests that line up with review dates, message threads with instructions. Amazon has gone to court against fake-review brokers and the administrators of groups that trade in them. No browser setup changes any of that, and a browser doesn't make a seeded review independent at the device level either.

Why This Isn't a Grey Area Any More

Fake reviews used to be treated mainly as a platform-policy problem. Not any more. In the US, the FTC finalised its rule on the use of consumer reviews and testimonials in 2024. It prohibits fake reviews (including ones written by people who never used the product), buying reviews, undisclosed reviews by company insiders, and certain forms of review suppression, and it lets the FTC seek civil penalties. In the UK, the Digital Markets, Competition and Consumers Act 2024 lists fake reviews among its banned practices. The EU's consumer rules cover similar ground.

We're not lawyers, and the details vary by country and by who's doing what, so read the actual rules where you operate. The direction is clear, though.

The platform consequences come first and are more certain:

Reviews removed, sometimes in bulk, including genuine ones caught alongside
The reviewing accounts suspended
The business itself penalised. Yelp can place a Consumer Alert on a business page when it finds suspicious review activity. Amazon can act on the listing or the selling account. Google can remove reviews and restrict a Business Profile.

That last group is the real cost. The fake reviews are disposable. Your listing isn't.

"We Were Only Testing the Flow"

This is a common argument from well-meaning teams, so it deserves a straight answer. It's also the weakest one in the whole subject, and we think it keeps getting used because it feels procedural rather than dishonest. A ticket. A test plan. A checkbox. Nothing like fraud.

If a review appears on a live public listing, real shoppers can read it, and it can move the rating they rely on. To them it's a fake review. The platform's terms don't carve out an exception for internal QA, and from the outside a "test campaign" of ten accounts looks exactly like a review seeding campaign of ten accounts (because, functionally, it is one).

So what does testing look like when it's above board?

Test your own review widget, not someone else's live platform. If you collect reviews on your own site, run the whole flow in staging with test data, across browsers and devices, as often as you like.
Read the platform's developer documentation. If you're integrating with a review platform's API, check whether it offers a test mode or sandbox. If it doesn't say, ask their business or developer support how they want integrations tested.
Check display, don't post. Want to know how your listing and existing reviews look to a shopper in Germany, on a phone-sized window, in another language? Viewing a public page is fine. Writing a review to "see what happens" isn't.

What Honest Brands Do Instead

None of this is complicated. It's just slower than seeding reviews.

Platform	The above-board way to get reviews	Where to go if something goes wrong
Amazon	The Request a Review button in Seller Central; Amazon Vine for brand-registered sellers	Seller Central support and Account Health
Google	The shareable review link from your Business Profile	The Business Profile help centre
Trustpilot	Review invitations sent from your business account	Trustpilot's business support
Yelp	Don't ask. Yelp's guidance tells businesses not to solicit reviews	Yelp for Business support

Product names and eligibility rules do change, so confirm the current details on each platform's own help pages before you build a process around them, and don't treat a table in a blog post as final. That includes ours.

Yelp's position annoys a lot of owners — you genuinely cannot ask, not even once, not even nicely. We think it's the defensible one. A rating nobody solicited is the only kind that tells a shopper anything.

A few habits matter more than any tool:

Ask everyone. Invite every customer after a real interaction, not just the ones who smiled. Gating is how honest programmes accidentally break the rules. If those invitations go out by email, send them from your own domain with proper authentication — our sibling product JustEmails exists because invitations that land in spam look, from your side, exactly like customers who ignored you.
Reply to reviews, especially the bad ones. A calm, specific reply to a two-star review often does more for the next shopper than another five-star.
Fix what the reviews say. If three people mention the same broken zipper, that's free product research.
Keep records. If a platform questions a burst of genuine reviews after a good launch, order data and invitation logs make the appeal far easier.

Where an Antidetect Browser Fits (and Where It Doesn't)

We make an antidetect browser, so let's be specific about this.

Where it helps:

Agencies working in client-owned accounts. Most review platforms let a business add managers or team users to its account. Agency staff can then work inside the client's own Business Profile or business dashboard with the access the client granted. A separate browser profile per client keeps each client's logins, cookies and open tabs apart, which stops someone replying to a review from the wrong client's account. In JustBrowser, the owner can share a client's profile with a teammate's free seat, with view-only or launch permission, and roles run owner / admin / member / viewer.
QA on your own review widget. Different fingerprint identities (a Windows set, a macOS set and so on) let you see how your review form behaves across browser and OS combinations. The local REST API and CDP connect to Playwright, Puppeteer or Selenium if you want to automate it against staging.
Checking how a listing renders elsewhere. With a proxy you bring yourself (JustBrowser doesn't include proxies), you can see a public listing the way a shopper in another market sees it.
Privacy. Keeping your own browsing separate from work accounts is a perfectly ordinary reason to want isolated profiles.

Where it doesn't: it can't make a seeded review genuine. It doesn't hide a seeded review from any of the signals above (device, network, behaviour, wording, relationship, purchase or payment), and we don't build it for that. And posting fake reviews isn't a use we support or want. If that's the plan, this product isn't the answer, and neither is any other browser.

Spelling out what your own software can't do is terrible marketing. It's also the only honest way to write this section.

Two neighbouring problems get mixed up with this. Bot traffic hitting paid campaigns is ClickzProtect's territory, not a browser's. And seeing what real visitors do on your product pages is what JustAnalytics does, without cookies. All three are built by Velocity Digital Labs, which is why we keep sending you to the right one instead of pretending a browser solves everything.

If Your Genuine Reviews Were Removed

It happens, and it's maddening. Filters catch real customers too, especially after a sudden spike. The appeal form is the same length whether you hired a broker or just had a good month, which tells you how much nuance to expect.

Go through the official route. Seller Central support, the Business Profile help centre, Trustpilot's business support, Yelp for Business support. Each platform has a way to contest removals or report a problem.
Explain your process. How and when do you invite customers, and is any incentive involved?
Don't create new accounts or ask customers to repost in a batch. That recreates exactly the pattern that triggered the removal, and on most platforms it turns a fixable problem into a permanent one.
If you think a competitor is seeding, report it with the platform's own flagging tools. In the US you can also report deceptive practices to the FTC at reportfraud.ftc.gov.

If you want the wider picture, we've written about the common misconceptions around antidetect browsers too.

Frequently Asked Questions

Why do review platforms flag reviews that come from the same device?

Browsers expose dozens of small technical details (canvas and WebGL rendering, installed fonts, audio processing, screen and language settings) that together form a fairly stable fingerprint. When several supposedly independent reviewers share a fingerprint, a network, and a posting pattern, the platform treats them as possibly one person or one coordinated group. It's one signal among several. Relationship links to the business, purchase records, timing and wording usually matter just as much.

Is posting seeded or fake reviews illegal?

In many places it can be. In the US, the FTC's rule on consumer reviews and testimonials, finalised in 2024, prohibits fake reviews, buying reviews, and undisclosed insider reviews, and allows the FTC to seek civil penalties. The UK's Digital Markets, Competition and Consumers Act 2024 also treats fake reviews as a banned practice. Separately, every major review platform's terms forbid it. This isn't legal advice, so check the rules where you operate.

Can a brand test a review platform's submission flow with test accounts?

Not on a live public listing. A test review that real shoppers can see is a fake review to them, whatever the intent, and platform terms don't carve out an exception for QA. Test your own review widget in a staging environment, read the platform's developer documentation for any official test mode, or ask the platform's business support how they want integrations tested.

What is the legitimate way for a business to get more reviews?

Use the platform's own tools and ask every customer, not just the happy ones. Amazon sellers have the Request a Review button and, for brand-registered sellers, Amazon Vine. Google Business Profile gives you a shareable review link. Trustpilot lets businesses send review invitations. Yelp is the exception: it asks businesses not to solicit reviews at all.

Does an antidetect browser help with review management?

For the legitimate parts, yes: agency staff working inside client-owned business accounts with the access the client granted, testing your own review widget across browsers and operating systems, and checking how a listing looks to shoppers in another country. It does not make a seeded review genuine or hide it from the device, behaviour, relationship and payment signals platforms use, and posting fake reviews is not a use we support.

What should I do if my genuine reviews were removed?

Use the platform's official route: Seller Central support on Amazon, the Business Profile help centre on Google, Trustpilot's business support, or Yelp for Business support. Explain how you collect reviews and keep records of the genuine customers involved. Don't open new accounts or ask customers to repost in a batch, because that recreates the pattern that triggered the removal.

Try JustBrowser

Native Chromium antidetect browser, not extension-based. Real C++ engine patches at the canvas / WebGL / audio / font / screen layer, so 40+ identity parameters come from the engine rather than a script. Local REST API + CDP for Playwright, Puppeteer and Selenium, free team seats for shared profiles, Windows x64 and macOS (Apple Silicon). $9.99/month or $99.99/year. 7-day free trial, card required — cancel any time in the seven days and you are not charged. Unlimited profiles.

Get started → · How it differs from Multilogin / GoLogin / AdsPower