Guide
Are AliExpress reviews fake?
Some are. But "fake reviews exist" is a much less useful conclusion than understandingwhich ones are unreliable and why — because once you know the mechanism, the review page becomes the most informative part of any listing rather than the least.
How the incentives actually work
There are three distinct things people lump together as "fake":
- Incentivised reviews. The most common by far. A seller includes a card offering a small refund, a coupon or a free gift in exchange for a five-star review with photos. The buyer is real, the order is real, the product is real — the rating is inflated by a few tenths. This is the bulk of what you're seeing.
- Reviews inherited from a different product. Listings can be edited. A seller builds up a review history on a cheap phone case, then swaps the listing over to a lamp. The review count stays. This is the genuinely misleading case.
- Wholly fabricated reviews. Rarer than people assume, because they're expensive to produce at volume and AliExpress actively purges them.
The tell for the dangerous kind
Inherited reviews are the ones that actually cost you money, and they have an obvious signature: the review text describes a different object. Scroll the written reviews and you'll find people discussing fit, or battery life, or a colour that doesn't exist in the current listing. If the words don't match the product, the review history doesn't belong to the product.
A second tell is a bunched timeline — hundreds of reviews inside a few days, then a long silence, then recent reviews in a different language mix.
Why the photos survive all of this
Here's the thing that makes the review page worth reading anyway: an incentivised reviewer still has to photograph the actual object. The rating is corrupted; the photograph is not. A buyer paid to leave five stars and a picture has produced, for you, an unflattering domestically-lit image of the exact product — which is precisely the information the seller's studio render was concealing.
So invert the usual reading order. Don't read the star average. Filter to reviews with images, ignore every word, and look at forty photographs of the thing in real rooms. That is a far better basis for a decision than any number on the page.
What the star rating is still good for
Ratings compress badly at the top — 4.7 and 4.9 tell you almost nothing apart. But the bottom of the scale is honest, because it takes real frustration to leave a one-star review on a $20 item most people would rather forget. Sort by lowest rating and read the complaints. If they cluster around one repeated defect, believe it. If they're scattered — slow delivery, a dented box, a language dispute — the product is probably fine.
The method
- Ignore the star average.
- Filter to reviews with photos. Look at the object, not the words.
- Scan written reviews for descriptions of a different product — that's an inherited listing.
- Sort by lowest rating. Look for one repeated complaint versus scattered noise.
- Weigh order count over rating.
Under this method, whether a given review is "fake" mostly stops mattering. You're reading the evidence, not the verdict.