Retailers can require customers who repeatedly abuse return policies to pay for shipping and returns. But what if those shoppers create new accounts in different names? That’s where artificial intelligence can help by spotting accounts that likely have been created by return abusers.

Assaf Feldman, co-founder and chief technology officer, Riskified

As consumers have increasingly embraced ecommerce, their expectations have shifted. These days, free two-day shipping and returns have become the baseline, and most retailers accept it as the price they have to pay to compete.

But the math only works if the retailer actually realizes the sale. Shoppers often spend big sums on a full cart only to return many—or even all—of the items at a later date. When customers make a habit of returning a large percentage of their purchases, the delivery, admin and stock-management costs add up.

In response, some retailers have taken to blocking such behavior outright, banning shoppers for excessive returns and deactivating customer accounts. But shopping online and returning things at no extra cost is a major selling point to ecommerce shoppers; so much so that 85% of customers expect free returns.

AI can examine huge amounts of data and identify customers for whom ‘shopping’ is more like ‘borrowing.’

Merchants that don’t offer free shipping and returns may send revenue to their competitors, and those spurned customers may never return, costing them lifetime value.


How to deter serial returners

Rather than eliminating free shipping and returns, or banning troublesome customers (and risking the negative publicity), retailers should disincentivize returns abuse instead. When faced with customers who abuse their return policy, merchants can—and should—dynamically adapt their approach for those customers alone. Instead of offering the standard generous return policies, excessive returners may now have to pay shipping costs or restocking fees (of course, these costs should be clearly communicated to the customer).

This effectively shifts the burden of their frequent indecision from the merchant to the customer, limiting the merchant’s potential losses without upsetting or rejecting good customers.

Identifying return abusers is the easy step. By looking at their average costs and margins, merchants can identify the point at which free returns cost them money and set that rate as a threshold. But doing so will lead to a response from the return abusers, and they will likely create new accounts or take advantage of guest checkout to avoid recognition.

That’s where artificial intelligence (AI) comes in. AI can examine huge amounts of data and identify customers for whom “shopping” is more like “borrowing,” even as they move from shop to shop. When the return abuser moves on to guest checkout or creates an account using another name, solutions like Riskified can examine hundreds of data points and recognize the individual behind the purchase, even if that individual is trying to conceal their identity. Doing so means that the right return policy continues to be applied.

By incorporating advanced technology, merchants can implement a solution that effectively responds to the problem. Customers who don’t abuse return policies, meanwhile, continue taking advantage of free shipping and returns as intended. Using artificial intelligence, merchants can ultimately balance security with customer service, and maximize the benefits of free returns while discouraging its worst abuses.


Riskified provides fraud-prevention technology to online retailers.