Credit card fraud accounts for more than $27 billion in annual losses, posing an especially grave threat to merchants selling high-value items like diamond rings. In this month’s Digital Fraud Tracker, PYMNTS talks with Apeksha Kothari, chief operating officer of diamond ring eTailer Rare Carat, about pairing artificial intelligence and human analysts to keep credit card fraud from taking the shine out of diamond buyers’ special purchases.
The industry of eCommerce has been on the rise for decades, and it got a massive boost when the pandemic caused brick-and-mortar establishments to close or reduce hours, prompting consumers to go online for their shopping needs. This jump in eCommerce traffic was accompanied by a surge in fraud attempts targeting marketplaces and their customers. Fraud attempts per month against eTailers shot up more than 24 percent in 2020 compared to 2019, with only about 34 percent of these attempts successfully prevented. The total cost of these fraud attacks also went up by just over 7 percent last year compared to the year prior.
One company familiar with these dangers is diamond engagement ring eTailer Rare Carat, which originated as a referral engine connecting shoppers to jewel merchants but evolved over the past year into a full-blown marketplace. The company’s entry into eCommerce brought a variety of different fraud threats, according to Rare Carat’s chief operating officer, Apeksha Kothari.
“Diamond engagement rings were underrepresented [online] before COVID hit — after your home and your car, it’s one of the largest purchases that you make, so it was still a purchase that a lot of people did in person,” Kothari said in a recent interview with PYMNTS. “We saw traffic actually go down significantly in the early months of COVID, but soon after, people started to try to find a sense of normalcy or a sense of getting back to their future plans, and we’ve seen traffic explode.”
Kothari offered PYMNTS an inside look at the fraud threats eCommerce merchants like Rare Carat face daily and how artificial intelligence (AI) and machine learning (ML) tools are leveraged to fight them. However, these prevention techniques are far from perfect, and future threats are brewing of which merchants need to be aware.
How Rare Carat Protects Its Customers
Kothari said that the most pervasive threat eCommerce merchants like Rare Carat face is credit card fraud. This type of fraud is particularly effective when merchants do not require a match between billing and shipping addresses, but Rare Carat insists that they be one and the same to reduce this risk.
“The biggest [menace] has to be credit card fraud, with somebody using somebody else’s [stolen] card,” she said. “Our retailers will only ship to the address that is on that credit card, for example, which other eCommerce players might not insist on, but it’s something that we just have to do, given the high tickets.”
This security protocol is backed up by Rare Carat’s partner PayPal, a payment option at checkout, which adds its own security protocols to keep fraudsters from exploiting the merchant and its diamond-dealer partners. The core of its security system relies on artificial intelligence and machine learning, said Kothari.
“These third-party checks rely on AI and ML to identify customer patterns that seem irregular,” she explained. “There’s a lot of IP tracking as well — if their IP says they’re in Bermuda, but they’re placing an order for an address in Mexico, that could be suspect.”
AI and ML are not the be-all and end-all of fraud prevention in the eCommerce space, however. Kothari said that human analysis is still necessary to differentiate fraudsters from legitimate customers accurately, especially when high-ticket items like diamonds are involved.
Challenges In eCommerce Fraud Prevention
AI and ML have been proven to reduce fraud rates significantly, but they are not a panacea, Kothari warned. False positives are still a potential issue, for example, these technologies cannot account for all the nuanced signs of fraud that a human analyst can easily recognize.
“One thing that I think goes a little underrated in the quest to [implement] machine learning to automatically solve all your problems is the human element of it,” Kothari said. “We’re higher-ticket and higher-touch, so when we have our customer service team handle customers, there [are] often phone calls, emails, chats and much more engagement with a customer that allow us to screen them in a way. If something looks a little shady or something looks a little off, there are precautions you can start taking: For example, instead of shipping it directly to the [customer’s] address, you can ask for it to be sent to a FedEx hold facility, where a signature is required.”
Another obstacle for fraud prevention in the eCommerce space is the ever-lingering threat of new fraud techniques. Kothari noted that most existing fraud prevention methods are geared toward individual bad actors, for example, but larger-scale hacking attempts have the potential to overwhelm the existing systems of many eCommerce merchants.
“Individual consumers [who] are engaging in fraud is something that I think retailers are now getting used to and understanding and baking into their financial estimates,” she explained. “I think it’s the more major events that could destabilize your entire operations: somebody trying to come in and get [personally identifiable information], for example, or financial information. Smaller retailers have become a little more prominent as hackers expand their focus and are more of a target where they weren’t before.”
eCommerce merchants of all sizes would do well to shore up their defenses against future threats rather than just the most immediate ones. Failure to do so could result in serious damage to entire companies.