Kount LogoBLog Against FraudKount Logo

Protecting the Sanctity of the Digital Transaction Lifecycle

posted on: Wed Sep 12 2018

The execution of a digital card-not-present transaction in a retail setting is complex.  Thanks to the Internet, today’s retail shoppers are savvy, educated and well-versed in understanding available options from a worldwide marketplace.  As such, long before a customer hits the “purchase button,” e-commerce retailers must spend considerable time, money and resources to attract customers and move them through the transaction lifecycle to the point of purchase.

Competing for shoppers in a digital environment has changed the way that retailers compete for market share, engage with customers and market individual products. The savvy retailers have focused on the customer experience.  The customer experience represents the engagement that occurs between a retailer and a customer throughout the buying process.  This contact can be through social media, ad words, website, branded events, personalized ads, etc.

The customer experience does not stop when the shopper reaches the payment phase of the digital experience. Some might argue that this is when retailers need to be at the top of their game.  According to personalization vendor Barilliance, in 2017, 79 percent of shopping cart abandonment occurred during the payment processing phase.  The two main causes of shoppers abandoning the original purchase were that the process was too lengthy and too complicated.

When you think about it, it makes sense.  Time is a major part of the customer experience.  Specifically, the time related to online shoppers as they are waiting for their purchase to be approved is critical if they continue with the purchase and more importantly, continue using the retailer’s site.  To combat this, many vendors have introduced expedited checkout buttons to the purchasing page to help facilitate the approval of the transaction. According to payment processor BlueSnap, payment options include PayPal, Apple Pay, Google Pay, Amazon Pay, SamSung Pay, etc. and can provide a transaction lift of 38 percent and an average ticket goes up 55 percent.

The expedited payment options help facilitate the transaction for the user experience, but a retailer still must balance approving good transactions versus declining fraudulent ones. According to Javelin Strategy, 33 percent of declined transactions are actually valid which adds up to $118 billion in lost revenue annually.  This mitigation and balance of fraud is found in a company’s use of advanced machine learning fraud prevention solutions. 

Within milliseconds, modern fraud-prevention systems use of machine learning can assess the risk of a transaction by customer behavior, looking at the location of the user, the device, the IP address and whether the IP address is a proxy of the kind criminals often use.  Domain expertise can also be applied to ensure that financial details of the transaction are a consideration—i.e. diverting more high-ticket items towards receiving more scrutiny in the form of a manual review.  Machine learning’s value to fraud mitigation is that it can improve results over time by comparing the system’s predictions to actual outcomes. For example, if an increasing number of fraudulent events are tied to consumers in a particular location, the system can give added weight to that indicator.

With any card-not-present transaction, there is the chance for fraud and the consequences of it typically fall upon the retailer.  This responsibility includes the lost merchandise as well as the chargeback if the retailer falls below the acceptable number of fraudulent purchases. Kount’s partner Chargebacks911 recently shared some numbers related to fraudulent activity that led to chargebacks:

  • 1%-10% is criminal fraud from fraudsters such as identity theft or stolen credit cards.
  • 20%-40% is from merchant errors such as the description of the product was cut off, the package was damaged during delivery or the merchant processed the return but didn’t credit the customer’s card.
  • 60%-80% is considered friendly fraud. Friendly fraud is initiated by consumers who are not habitual criminals. In many cases, the cardholder will contact the bank that issued the payment card—instead of the online merchant—to ask for a refund for such reasons as the item not being delivered, damaged or in some cases, because the consumer just decides he can get away without paying for the product.

Digital transformation is upon all of us and it requires that retailers recognize the final push at the end of the transaction lifecycle to ensure a positive customer experience with legitimate orders.  Retailers need to deploy multi-faceted solutions that can account for different tactics and strategies as fraudsters look to disrupt the customer experience.

Check out Kount's eBook "The Truth About Machine Learning in Fraud Prevention" to learn more about machine learning's role in successful fraud prevention. 

Machine-Learning-eBook-Banner_600x150

Topics
Share
TOP