When we select product, it is inevitable to pay attention to the sales and evaluation of the target product. To have a advance concept of possible reviews during product sales, which is of great significance to the product selection and operation. However, as more and more sellers settles on the Amazon platform, it began to require sellers to understand how to identify which reviews are true and which ones are unreliable.
Of course, with the continuous optimization of the Amazon platform, the review of each product also tends to be true, but this does not mean that we can no longer pay attention to how to distinguish between true and fake reviews,it is important to distinguish the true and fake reviews when dealing with bad reviews.
Our easiest way is using tools. You can quickly achieve the purpose of identifying fake reviews through a site "Fakespot". We simply copy any Amazon product link to the site and quickly get a rating for the product, which is based solely on the product's rating.
We contacted the person who are in charge of the site, wondering how exactly they got the result, the official response to us:
Each of our analyzes performs two things at a time: analyzing each review; analyzing each review account and how the account comment other products. We continue to search, but also continue to record the common sentences of different products reviews, evaluation description and so on. When compare them in actual, our engine will judge through the database information. In the meantime, we constantly update our engine database to select the most common review sentences and content from the review content that has been identified as false evaluation, and record the evaluation account.
So, even if we do not have much clue, we can recognize fake reviews in advance through artificial intelligence. In addition, according to our records, any negative feedback has a template, because the source is often the owner himself rather than the actual consumer, so there is definitely a trace. And by continuous analyzing of the products submitted by the sellers, our engines will continue to learn and grow. This is also the best guarantee we have provided for the analysis products.
In order to get a more comprehensive assessment, we also consulted with Professor Liu, the Department of Computer Science at the University of Chicago, who is the experts in emotional analysis and machine learning. We asked him about the possibility of using procedures to analyze the credibility of the reviews. Professor Liu believes it is hard to estimate the feasibility of the software or the procedure. Because it is almost impossible for us to actually prove which evaluations are true and which ones are fake. To do this, we have to have the reviewer own admission. However, it is almost impossible to achieve, but we can indeed make some judgments based on the wording of the review and the usual performance of the reviewer account, but the accuracy of such judgment still has to be added a question mark.
Let's take an example to see what weird places give us a hint:
Through Fakespot, we found this is a product that has many reviews. When we look at their reviews, we find that there are many reviews that have taken place in just a few days. And this may be the hidden clue that the seller are doing tricks.
In addition to the time factor, we can judge by the similarity of reviews. If the seller asks "gunmen" to leave fake reviews, "gunmen" often ask the seller to provide review copy, and then arrange their account for evaluation. This evaluation will become monotonous.
And the last point, we can start from the product owner. If the seller does not even provide their own website or company information, and the only way to contact sellers is through the Amazon platform, we can basically determine this is really a small seller with no strength or a small seller do a lot of things.
Back to the review, the comments in the picture below is the most typical example, even if you are new seller, keep in mind do not be cheated by such a comment. (Amazon v. Gentile lawsuit in Washington Superior Court.)