▲ ▼ Identifying fake reviews in Amazon, other sites
Almost every review for accessories of a newly launched product I see on Amazon India is a fake review e.g. Smartphone case, Tempered glass etc. I see that reviewers have left similar, positive reviews with 5 stars rating for several other products.
I'm sure this is not just limited to Amazon India, as I could find other news detailing the ways the sellers employ common people to give fake reviews in the promise of free goodies. So, this practice must be common among several sellers across different ecommerce sites.
I would want a service which can differentiate fake reviews and genuine reviews seamlessly in the ecommerce sites.
I tried https://reviewmeta.com before and it seems ok. It should be more intelligent than identifying fake grammar.
reviewmeta looks like a great solution, I tested with few amazon links and I agree that grammar might not the best criteria to judge as there are English first amazon sites on non-native English speaking countries.
Have you used their browser extension? Is it safe?
Amazon still hasn’t fixed its problem with bait-and-switch reviews - https://arstechnica.com/tech-policy/2020/12/amazon-still-hasnt-fixed-its-problem-with-bait-and-switch-reviews/ .
Another common way these smartphone accessories sellers fake their reviews & rating in these ecommerce sites is by having accessories for different smartphones in the same listing, so when the buyer comes to the product page they would see the reviews for the case, tempered glass of another smartphone which may/may not be true for the smartphone they are buying it for.
There few websites which claim to check for fake reviews in the ecommerce sites such as https://www.fakespot.com/ , https://reviewmeta.com/ but I didn't find them accurate enough in my tests to recommend.
Perhaps there is a need gap for a service to detect fake reviews and rating by using language scores(NLP), checking past reviews of the users and correlating with the seller's score. The review system for anything on the Internet, say Playstore or Ecommerce by default is dependent upon the 'law of large numbers' i.e the fake reviews get sidelined and true reviews/ratings come up when several real customers start to rate it.
But unfortunately in a country like India, the law of large numbers can be applied effectively to create fake reviews as well due its large population. Perhaps there is a need gap for a better review system as a whole.
There are some privacy concerns highlighted regarding fakespot, Please exercise caution and my above comment is not an endorsement of any solutions.
CNBC recently covered this problem of Amazon fake reviews - https://www.youtube.com/watch?v=Mq_Ksga9uHY. Although it features some services for identifying fake reviews mentioned in an earlier comment, it seems to be largely an unresolved problem. These services seems to use grammar mistakes as a way to identify fake reviews.