Online retail giant Amazon has just launched a new customer review algorithm that the company says will “make customer reviews more useful,” Cnet is reporting. The system, put into effect Friday, uses machine learning to determine which customer reviews are the most helpful and relevant. These select reviews are then given more weight when it comes to calculating the 5-star rating of an item. The new algorithm also adds more weight to reviews from verified purchasers and newer reviews.
“It’s just meant to make things that much more useful,” says Amazon spokeswoman Julie Law, “so people see things and know it reflects the current product experience.” A product that decides to cut back on its quality to save in manufacturing costs, for example, will now see its ratings fall quicker on Amazon to reflect the change.
Really, though, the change seems to target fake reviews more than anything else. Fake reviews lower the reliability of Amazon’s otherwise terrific ratings system. Amazon is very protective of that asset: Earlier this year, Amazon took legal action against sites that sell fraudulent 5-star Amazon reviews to desperate product manufacturers. By changing up the ratings algorithm, Amazon builds on that effort by automatically discounting fake reviews that do make it through.
Shoppers should note that the change to Amazon’s review system might make product star ratings more volatile. For more on spotting fraudulent reviews, check out Techlicious’s primer How to Tell if a Review is Fake. You might also want to check out these 6 ways to save money while shopping on Amazon, and review new changes to the company’s two-day free shipping policy.
[Image credit: Amazon]
From Jacob on July 22, 2015 :: 4:36 am
When you don’t have a purchase-based and verified system in place and your reviews are overwhelmingly positive, potential customers might start viewing them as unauthentic and manipulated by the company itself. That’s why I think Amazon’s review policy is, well, faulty. They should adapt a similar solution to that put in place by third-party feedback collecting company. Bazaarvoice or eKomi both offer transaction-based testimonials, thus boosting the credibility of your brand. Clients are asked to leave their opinion soon after the purchase, which provides very good turnover rates. As a results, potential customers pay more attention to the authenticity and verification of your testimonials.
Reply
From Rock E Roe on December 01, 2015 :: 6:25 pm
Jacob, Often people will like a product at first and are unaware of a products faults until later. The reviews made soon after a purchase seem more useful for rating the transaction than the product. I have a problem with the compensated reviews Amazon allows but does not flag. Even the Vine designated reviews should be given less weight IMHO.
Reply