Consumer preference for online shopping continues to rise, as more purchases are being made online than in stores with each passing year. In 2019, the total market share of online U.S. retail sales surpassed general merchandise sales for the first time in history. And as a result of the coronavirus pandemic, this trend is only increasing. According to McKinsey, ecommerce sales in apparel, department stores and beauty products have increased by nearly 10%, on average, since the onset of the pandemic. In grocery, online purchases have increased from 2% to 3% before the crisis to 8% to 10% during its peak.
Further, making purchases online has become even easier with the emergence of point-of-sale (POS) lending, which allows consumers to split the cost of purchases into regular installments — even for a purchase as small as a $75 sweater. Because of this ease, merchants have begun relying on POS financing to drive sales growth. McKinsey has found that around 50% to 60% of loans originated at POS are either partially or entirely subsidized by the merchant. And in sectors with a high cost of acquisition and high margins (such as jewelry and luxury retail), merchants are willing to fully subsidize APRs.
As more shopping moves online, it has become even more important to put a “face” to all of the new faceless transactions. Without proper identification of the individuals behind transactions or new account openings, it’s difficult for retailers to tell the difference between good customers and bad actors, and therefore easier for fraudulent activity to slip through.
The Harmful Impact of Digital Fraud
Digital fraud can take many forms, with the most common being account takeover (29.8% of all ecommerce fraud) and synthetic identity fraud (the fastest-growing type of financial crime in the United States). With account takeover, a fraudster takes over an individual’s existing account (either with stolen credentials or social engineering) and, using the card on file, purchases goods to ship to another address.
With synthetic identity fraud, fraudsters combine personally identifiable information (PII) from real people with false information to open bank accounts and credit cards. This also includes merchants that offer POS lending, and their own cards with low or no interest on an initial purchase, as a way to get consumers to buy. After acting like legitimate customers to build a transaction history, the fraudsters then run up charges which they don’t pay for.
Both types of fraud are very difficult to detect, creating problems for retailers such as:
- High chargeback costs. When a consumer notices a fraudulent transaction on their bank statement, they will ask their card-issuing bank to reverse it. According to McKinsey, merchants face up to 2.4X the chargeback cost per dollar disputed, and often spend additional resources and time to collect evidence to face the dispute.
- Penalties for exceeding fraud thresholds. Merchants that exceed fraud thresholds set by networks (VISA, MasterCard, etc.) can also face fines and additional fees. For instance, a threshold might be the dollar value of the transactions the merchant lost to fraud over a specified period of time, relative to total transaction value from the same period. By exceeding these thresholds, merchants not only risk fines and fees, but also a network refusing to process further payments altogether.
Old Fraud Detection Methods Create Customer Friction
To overcome these challenges and stop fraud, merchants have partnered with payment processing companies to verify digital transactions. By leveraging credit-related data (e.g. social security number, date of birth, phone number, physical address, etc.) for fraud risk assessments, these companies could make yes or no determinations for each transaction: either approving transactions they determined were made by a real person or blocking ones they deemed fraudulent.
While helpful for detecting and blocking fraud, this deterministic approach has also led to a high rate of false positives, preventing good customers from completing their online transaction or locking their account entirely. While today’s customers want assurance that their account information is secure, they also expect a frictionless payment experience. If a customer is falsely deemed fraudulent, it will disrupt their shopping experience and cause them frustration.
Additional friction points in a customer’s payment journey can also significantly decrease the likelihood of purchase, resulting in missed revenue for the merchant. According to McKinsey, 53% of shoppers abandon their cart after just one payment decline.
The Entrance of New Data to Better Determine Digital Fraud Risk
Thankfully, as the world has become increasingly mobile and tech-driven, a new way to determine fraud risk has emerged. This new approach leverages different types of PII, including device IDs, biometrics and behavioral data. Each of these dynamic data elements fit together like puzzle pieces to create a whole picture of the individual behind online transactions. For example, device ID might tell a merchant which browser the shopper used to complete their transaction, whereas behavioral data can identify anomalies in shopping behavior, such as transacting across dozens of merchants in just a couple of days.
By analyzing global dynamic PII and transaction patterns with AI-based systems, retailers can get a better sense of the probability that a bad actor is behind a transaction. These data patterns make it easier for merchants to spot suspicious online purchases or account openings, without creating additional barriers and disruption for misidentified good customers. The result? Happier loyal customers who trust the retailer to both protect their information and provide them with an enjoyable online shopping experience.
Jordan Reynolds is the Head of Global eCommerce and Marketplace Strategy at Ekata.