Amazon’s reviews system can have a big impact on buying decisions, but it’s not as accurate as it could be. Now the retail giant is hoping to fix that with a new machine-learning platform it has developed in-house, it has revealed.
“The system will learn what reviews are most helpful to customers…and it improves over time,” Amazon spokeswoman Julie Law told CNET. “It’s all meant to make customer reviews more useful.”
Let’s say you want to buy a garden hose, but you have no idea what kind to go for. You’re likely going to visit Amazon and pick one that looks good based on customer feedback, but the problem is, most of us use Amazon’s star-rating system, and it’s not as useful as it could be.
The rating system currently displays an average of all ratings a product has received to date — so if that product happens to change or get updated over time, the rating can be misleading, and you could end up buying something you weren’t quite expecting.
For example, I was recently shopping for an SSD drive to throw into my computer, and I quickly chose a 240GB model on Amazon with a fairly decent price tag and a near five-star rating. But before handing over my cash, I decided to check out the reviews — and many weren’t so positive.
It turns out that the drive’s manufacturer had apparently changed its flash storage supplier, and the drive was no longer as reliable as it once was. Many complained about it breaking down within just a couple of months, and speeds that weren’t as fast as they previously saw from the same SSD.
But because this particular drive had received so many four- and five-star reviews in the past, the most recent one-star reviews weren’t having much of an effect on its overall rating.
Amazon’s new system will give more weight to recent reviews and ratings in an effort to avoid situations like this, CNET reports — and those written by verified Amazon purchasers and voted up for being helpful will also be given more focus.
The change is likely to be gradual at first, and it’s not yet clear which markets will see it after the U.S., but over time, shoppers will begin to notice star ratings and top reviews on product pages changing.
Amazon says it has considered the new platform carefully before deciding to implement it, and insists it is intended to make reviews and ratings much more useful, “so people see things and know it reflects the current product experience.”