How to spot fake Amazon reviews

1. Look at the histogram shape, not the average

Click the star rating to expand the breakdown by star. Genuine products for sale at scale produce a J-shaped curve: a big five-star block, a smaller four-star block, a dip in the middle, and then a visible bump of one-star reviews from people whose unit arrived broken or who misunderstood what they were buying.

The suspicious shape is a cliff: 90% five-star, a sliver of four, and almost nothing below. Shipping damage alone should produce some one-star reviews. When the unhappy tail is missing entirely, either the negatives were suppressed or the positives were manufactured.

2. Treat a perfect score as a warning, not a recommendation

A displayed 5.0 across dozens of reviews means essentially nobody left a 3, 2, or 1. Across 50 or more independent buyers that is close to impossible for a physical product. Well-loved, genuinely excellent products tend to land between 4.3 and 4.7 — the range is a sign that real people with real opinions are reviewing.

3. Check when the reviews arrived

Sort by most recent and scan the dates. Reviews should trickle in roughly in proportion to sales. Warning signs: a dense cluster of reviews within a few days, a long dormant period followed by a burst, or a brand-new listing that already has hundreds of reviews.

Bursts around a launch can be legitimate — a product going viral or hitting a deal site does this. Bursts on a listing with no other sign of attention usually are not.

4. Check whether the reviews are even about this product

This is the one most people miss. Amazon merges reviews across variations of a listing, and sellers exploit it. A phone case listing can inherit thousands of reviews from a screen protector, or a seller can take over a dormant listing with an established review history and swap in an unrelated product.

Read a handful of reviews and ask whether they describe the thing you are looking at. If the five-star reviews are about a different colour, size, or category of product entirely, the rating is meaningless regardless of whether any individual review is fake.

5. Read the one-star and three-star reviews first

Manipulated listings are usually padded with positives rather than scrubbed of negatives, so the low reviews are where the real information sits. Three-star reviews in particular tend to be the most honest writing on any product page: the reviewer has no strong incentive either way.

Look for the same specific defect mentioned repeatedly. One person calling something flimsy is noise; six people describing the same hinge breaking is the product.

6. Watch the language across reviews, not within one

A single glowing review tells you nothing. Patterns across many do. Reviews written to order tend to praise the same attributes in the same order, describe the packaging and the unboxing rather than months of use, restate the product title verbatim, and avoid mentioning any drawback at all.

Also check for unverified purchases. Amazon labels reviews from confirmed orders, and a listing where the enthusiastic reviews are unverified while the critical ones are verified is telling you something.

7. Notice when a seller is buying reviews from you

If a package arrives with an insert card offering a gift card, a refund, or a free replacement in exchange for a five-star review, that seller is running a review scheme on its own listing. It violates Amazon's policies and you can report it. Take it as direct evidence about the rating you were reading.

What a tool can and cannot automate

Checks 1, 2, and 3 are arithmetic on numbers already printed on the page, so they can be automated — that is exactly what StarSkew does, showing a badge beside the stars with published thresholds so you can see which rule fired.

Checks 4 through 7 need a human reading the page, and no metadata tool can substitute for them. That is the honest limit: automation gets you a faster first filter, not a verdict.