For anyone who thinks Amazon is more generous with returns than brick-and-mortar retailers, you should know that Jeff Bezos started Amazon by selling books partly because books were (and still are) the least likely to be returned of all e-commerce products.
Now that Amazon has extended its product range from books to just about everything imaginable, including products with notoriously high return rates (like clothing and shoes), it is paying attention to what you send back, how often and why. If you return purchases too frequently, Amazon might shut down your account for good—in some cases without any warnings.
This story isn’t new. Online discussions about Amazon’s surprise account closure can be traced as far back as 2008. More recently, several Amazon customers have told The Wall Street Journal that their accounts had been deleted without explanations. One of the customers claimed he had returned just one item in the past year.
Still, no one is sure how many returns are too many.
Amazon doesn’t set a limit for the number of products a customer can return, as long as it is for a legitimate reason. However, it doesn’t tolerate abusive behaviors with its return system, such as sending back used or damaged items, or things purchased elsewhere. Per its return policy, Amazon accepts most “new, unopened” items within 30 days of purchase.
But, what if the product is damaged in the first place? After all, Amazon does list “item defective” or “product damaged” as options for explanation on its return request form.
In 2016, a computer programmer who self-confessed as an “Amazon addict” was banned from the site after sending back 37 items out of 343 purchases. He told The Guardian that the items he had sent back were either faulty, damaged or not as described. But to Amazon, his return pattern could appear as a case of abuse.
To tell abusive behavior apart from legitimate returns of faulty items, Amazon enlists the help of artificial intelligence. Like other platforms facing hundreds of millions of consumers on a daily basis, Amazon uses algorithms to surface suspicious customer activities on the site before human staffers evaluate these activities on a case-by-case basis.
For example, if a customer says an item didn’t arrive as described when over 99 percent of people who returned the same item say they simply didn’t want it, the account could get flagged.
“If your behavior is consistently outside the norm, you’re not really the kind of customer they want,” James Thomson, a former senior manager at Amazon, told the Journal.
“We want everyone to be able to use Amazon, but there are rare occasions where someone abuses our service over an extended period of time,” Amazon told Observer in a statement. “We never take these decisions lightly, but with over 300 million customers around the world, we take action when appropriate to protect the experience for all our customers.”