For some Amazon shoppers, relying on the quantity of reviews for a product — not only the star rating — can help influence a purchase decision.
That’s partly why a new test on product search pages is drawing attention and critique.
Amazon is testing a new way to display product ratings on search pages by showing the star rating (4.3, 4.8, etc.) and then a number reflecting what percentage of reviews gave a certain rating, versus showing the actual quantity of reviews.
For example, in the screenshot above from this weekend, product search results show the average rating, then show what percentage of the reviews (84%, 94%, etc.) are 4-star and above.
Interestingly, as I checked back on Monday morning and did the same search for “face wash,” the percentages were gone and replaced by the number of reviews again (see screenshot below). But instead of showing five literal “stars,” there was only one star with the average rating. So clearly Amazon is doing some testing.
“We are always innovating on behalf of customers to provide the best possible shopping experience,” an Amazon spokesperson said in a statement.
ZDNet, Android Police, and The Verge also spotted the recent tweaks — both the removal of review quantity and the shift to just one yellow star — and didn’t have rave reviews, citing the potential for confusion and misleading customers.
Amazon’s star rating system is not just a simple average of all reviews, but rather uses “machine-learned models” that take into account different factors such as how recent the rating is and verified purchase status.
Customer reviews have become central to online commerce, playing a huge role in determining which products succeed or fail on Amazon and other sites. A study by Northwestern University’s Spiegel Research Center found that nearly 95% of shoppers read online reviews before making a purchase.
Amazon has long dealt with the persistent problem of fraudulent customer reviews. Last week it filed two new lawsuits against fake review brokers.
The tech giant meanwhile is starting to use artificial intelligence to automatically generate and present summaries of customer reviews, helping shoppers quickly understand the overall sentiment and key highlights about products before purchasing.
