Amazon released news June 19 about its new “Machine Learning System” that will learn more about customer reviews in order to make them more useful and relevant to consumers. The story broke on CNET and has been causing ripples throughout the internet. If you don’t believe me, just do a Google search on the first quote in the story by Amazon’s spokeswoman, Julie Law.
“Machine Learning” is a fairly new term many people don’t understand quite yet. In the context of product reviews, this means that Amazon has fed all of the existing reviews–those that have been approved, and those that haven’t–into its new system and have “trained” it to recognize reviews that will ultimately be the most relevant to customers. It will be able to approve, rank, and surface reviews quicker than waiting for hundreds of users to click the vote button.
Amazon is continuing to tighten down the screws on what content it considers legitimate. This is simply another step toward improving the customer’s experience, which started gaining momentum in April with Amazon filing suit on a company offering fake reviews. And on June 16–three days prior to this announcement–Amazon announced it will be suppressing titles longer than 200 characters starting July 15.
These actions from Amazon will force merchants to improve their product listings and customer experience. We wrote a piece about the lawsuit in April providing a reference list to sellers to clear up any doubt about the legitimacy of a review.
Here’s what we had to say about the matter:
The key to building legitimate reviews is to develop a genuine and honest strategy that follows Amazon’s guidelines. You should work with partners who are honest and experienced at building brands on Amazon. Consider your business goals and develop a long-term strategy. We encourage brands to think about the long run, not a short term burst. Understand that negative reviews are possible and soliciting feedback can work both ways.
You can’t control what your customers say about your product. You can only influence what they say by providing exceptional customer service, a quality product, meeting customer expectations, and following up after a sale. CNET said, “The new system will give more weight to newer reviews, reviews from verified Amazon purchasers and those that more customers vote up as being helpful.”
There was some chatter in the comment section about how earlier–and helpful–reviews will rank with the new system.
One commenter, ZOOLSS, said, “Older reviews often contain pretty valuable information while newer ones tend to be ‘I just got this and I LOVE the color of this toaster!’”
It will be fascinating to see how Amazon weighs older reviews into the overall rating of a product and how it weighs a seller’s conversions over time compared to reviews received. It seems obvious that a product getting consistent conversions and reviews will rank higher than a product that isn’t, but that’s yet to be determined.
We’ll be following the news as it develops on our blog. Let us know your thoughts and predictions about what will happen as Amazon’s new machine learning system plows ahead.