The types of retail policy fraud of concern include loyalty abuse, promotion abuse and returns abuse. While some merchants have resorted to deactivating accounts when there is “unusual” behavior (serial returning, wearing items and returning after snapping Instagram pics or ordering hauls of clothes with explicit intent to return a majority of items), others are turning to artificial intelligence (AI) to combat costly “policy abuse.”
As to how AI might help, Assaf Feldman, Co-Founder and CTO at Riskified explains more to Digital Journal readers.
Digital Journal: How much does payment fraud cost businesses?
Assaf Feldman: Card-not-present fraud costs businesses hundreds of billions of dollars annually. According to Statista, businesses will lose $48 billion in chargebacks in 2019, but that number is only part of the story. Fear of fraud leads retailers to turn away good orders, which is an even bigger loss. Statista estimates that loss at $187 billion in 2019, meaning the overall cost of CNP fraud is $235 billion.
DJ: What is retail policy abuse?
Feldman: Retail policy abuse is when shoppers bypass a policy or take advantage of it by using it in a way different than what was intended. This can take a number of forms, so here are a couple of examples. Let’s say a retailer has a policy that new customers get 20% off their first order. This is to incentivize shoppers to give the retailer a try. When repeat customers create multiple accounts to repeatedly take advantage of that discount, that’s abuse.
Another example is a retailer that offers free shipping and returns. The fashion and apparel industry has a problem with what’s called “wardrobing,” where a shopper buys an item, wears it once to an event or to post on social media and returns it. That policy is meant to give shoppers a virtual changing room where they can evaluate the clothes; it isn’t meant as a free rental service.
DJ: What is the impact of retail policy abuse?
Feldman:It’s difficult to quantify, and it varies across industries. But, like with fraud, retailers have to find the right middle ground. If they’re too lax, then the policies are abused and they lose money. If they’re too strict, then policies that were designed to be customer friendly can go in the other direction, as benefits are restricted or friction is added to the experience.
Looking at the two examples above, if a promotion meant to attract new customers is used by repeat customers, then that abuse means that marketing money is being spent ineffectively, and the discount shrinks the retailer’s margins.
Overuse of returns is easier to understand. First, the retailer pays the shipping cost – in both directions. Second, when inventory is off their shelves, they can’t sell it to a legitimate customer, so they may miss a sale. Finally, when that product comes back it has to be restocked, sold at a discount or written off as a loss.
There are also reputational costs. A number of retailers have one-item-per-customer policies on limited release products. If they don’t successfully enforce those policies, then their customers aren’t able to make purchases, and the goods might be sold at a much higher price on a secondary marketplace. That’s not a good customer experience.
DJ: Why does retail policy abuse occur?
Feldman:Policy abuse occurs when shoppers successfully hide their identities. At the easy end of the spectrum would be using guest checkout to avoid being associated with a prior order. By doing so, the shopper might be able to buy more items or return items without consequence. From there it gets more sophisticated, with shoppers using alternate names, multiple email addresses and even proxies to hide their identity and place of origin.
DJ: How are retailers responding?
Feldman:It depends on the retailer. Some accept this as a cost of doing business, while others change their policies or ban policy abusers from continued shopping. Both approaches are problematic.
DJ: How can new technologies help retailers?
Feldman:The biggest challenge in addressing policy abuse is recognizing the individual behind multiple purchases. If an abuser wants to continually take advantage of a discount, they’ll disguise their identity. Retailers have to be able to identify the individual across those changing details. A properly trained and deployed AI solution is excellent at exactly that. By looking at the links across orders, a solution such as Riskified is able to spot the abuser across orders – and even across multiple retailers. Once that’s done, the retailer can enforce its policies to avoid promo abuse, maintain a one-per-customer policy or impose limits to free shipping and returns.
DJ: Are retailers winning the battle?
Feldman:That, again, depends on the retailer. Fraudsters constantly change their attacks, and as eCommerce expands, there are new avenues for them to exploit. It’s difficult for retailers to stay ahead of those new methods of attack, but being overly cautious isn’t the solution. Retailers that take a dynamic, pro-growth approach are winning the battle. They’re embracing technology and prioritizing customer experience and lifetime value. Those are the retailers that will continue to thrive as eCommerce grows.
