Fashion has a reputation for being cutting edge and risky. Fashion merchants, however, often are conservative and reluctant to sell online, especially outside their domestic borders, so they lose out on billions in potential sales, according to a report from fraud management vendor Riskified.
“The overcorrection on fraud from a retailer’s perspective, especially for midsized and small businesses, means they’re leaving huge market expansion opportunities on the table. They’re told a certain foreign market is risky, so they shut down the borders,” says Andy Freedman, Riskified’s chief marketing officer.
“Online and off, there are $118 billion worth of legitimate transactions being declined. E-commerce and mobile commerce is $9 billion of that. You could get three additional Cyber Mondays every year by converting orders that are legitimate,” Freedman says. Cyber Monday, the Monday after Thanksgiving, generated about $3 billion in online sales in 2015.
Riskified examined online fraud in fashion retailing because the category has a high volume of cross-border e-commerce transactions. Luxury retail is a growing category, and the average cross-border order is worth 3.5 times more than an average domestic order, according to the report.
Many retailers aren’t properly set up to handle international orders and have high decline rates, Riskified says. There is a misconception that e-commerce orders are fraudulent if they’re paid with credit cards issued outside the country where the consumer is shopping, the report says. But reasons for cross-border shopping could include unavailability of a fashion item in the consumer’s country, consumers on vacation, students at college or business executives working abroad.
Luxury fashion orders from China are safer than consumer fashion orders, and luxury fashion retailers especially lose significant revenue if they decline too many online orders placed with Chinese cards, Riskified finds. Luxury fashion orders placed with Chinese credit cards and shipped to the United States are safest, with 1.04% of orders being fraudulent, and orders paid with a Chinese credit card and shipped outside of China are slightly riskier with a fraud rate of 1.58%, according to Riskified.
“Retailers selling a Burberry jacket, for example, will be more conservative in selling it. Some retailers will decline 70% of the orders they see for it,” he says. And those declines have an impact: 42% of millennial shoppers abandoned a retailer after being falsely declined, while 58% of high-income cardholders ($100,000 or more) limited or stopped shopping with a retailer who wrongly declined their order, Riskified says.
The company, founded in 2013, uses machine-learning technology to pull in data based on a consumer’s behavior across many points: email addresses, credit card PINs, whether a shopper has browsed the site before, bought similar products elsewhere, social media data, real estate records and more, Freedman says. Machine learning, a type of artificial intelligence, helps computers find patterns and make predictions without specific programming. “Our product changes every day. It’s an evolving machine that changes with every consumer decision we see,” he says.
Riskified clients get fast decisions on a shopper’s legitimacy, and in cases where merchants flag an order for manual review, Riskified takes a closer look—one that does not involve reaching out to the customer—and responds within an hour, Freedman says.