4 Ways to Fight Shipping Fraud While Reducing Operational Overhead

Advanced technologies such as UML and holistic data analysis can help logistics and delivery companies detect fraud attempts early and in an automated fashion, no matter where they originate.

Getty Images 1001622498
Getty Images

The images of people stealing packages off porches — so-called “porch pirates” — are all over the news. And, they’re infuriating to consumers. But, package theft happens even more frequently over digital channels, before the goods arrive at their destination, and it’s a huge financial problem for organizations across the United States.

With FedEx and UPS delivering more than 34 million packages daily, it’s no wonder fraudsters are attempting to cash in on this industry. According to Experian, shipping fraud rates increased by 37% from 2016 to 2017 and shipping fraud attack rates increased by 60% in the Western United States alone. With more and more online orders being placed since the beginning of 2020, that percentage will keep climbing. 

How does shipping fraud happen? 

Fraudsters use fake or stolen credentials to pose as customers to redirect deliveries to their own addresses. For example, they can use customer information to create a fake online account using the delivery company’s shipping portal and modify orders to include high-value items or reship them to different destinations. Fraudsters may also call the service center and impersonate good customers to make changes to accounts, such as shipping details and credit card information. Or, they may mass-register online accounts to send emails that entice recipients to fall for advertising fraud. Logistics and delivery companies can lose millions to shipping fraud each year due to these malicious activities.

Recently, a leading logistics company was struggling to get control of shipping fraud. Fraudsters had been mass registering fake accounts and using them to track and re-route packages, issue package holds for delayed pickups and spam good users with malicious advertising and/or emails requesting package-related data. They were also tracking thousands of fraudulent shipping actions using the company’s online portal or manipulating customer orders or accounts via phone if they had certain pieces of customer data. These activities caused financial losses as well as reputational damage from degrading the customer experience for good customers. 

If the fraudster had used a single account to intercept and re-route a large volume of packages, the action would have been an obvious red flag. By mass registering and using multiple accounts, the criminals made it difficult to identify the activity as fraud. Preventing mass registration requires early detection and the ability to spot known and unknown patterns — capabilities that legacy, rules-based fraud solutions lack.

Combat shipping fraud with these 4 tactics

To reduce shipping fraud, it’s critical to uncover fraud patterns during the early stages — at sign-up or when requests are submitted for package re-routes, for example. Here are four foundational technologies to implement as part of an effective solution for preventing shipping fraud:

  1. Artificial intelligence and advanced machine learning. Whereas rules-based solutions rely on historical data and labels to spot potentially fraudulent activity, solutions that leverage advanced technologies such as unsupervised machine learning (UML) help organizations spot both known and unknown fraud patterns. This is important, because fraudsters modify their techniques and have access to artificial intelligence-powered tools and bots that defy detection.

Solutions that leverage UML have been shown to detect fraudulent accounts about 30 days earlier than legacy rules-based solutions. Instead of focusing on behavior monitoring at the attack event level, advanced solutions can catch fraudsters upon account registration, neutralizing them before damage can occur.  

2.           Holistic data analysis. Analyzing data on a holistic level enables organizations to uncover anomalies and suspicious activities and detect connections between seemingly unconnected data points to uncover sophisticated fraud attacks. Using a holistic approach to analyzing user profiles, digital fingerprints, registration and tracking events and package shipping and re-routing activities, organizations can detect up to 60% more fraud. 

For the delivery leader mentioned earlier, holistic data analysis enabled them to capture coordinated groups of fraudulent users taking advantage of the company’s online tracking portal to manage their illicit activities. That translated to about $4 million in savings annually. 

3.            Early detection. Fraud attacks are often launched immediately following account registration, and a slow response can result in huge fraud losses. Combining UML with holistic analysis across large sets of data at scale enables advanced fraud and risk solutions to uncover suspicious activities early and in real-time, before fraud can be committed. Even if fraudsters make it past registration, they can be stopped at the first attack. 

4.            Accelerated case review. Individual-based fraud investigations are not intuitive or effective to uncover sophisticated patterns and coordinated attacks, and fraud can slip through the cracks. Investigating cases by manual scanning is time-consuming and can’t keep pace with the speed of modern fraud. Plus, it increases operational overhead. 

Advanced solutions that use automation and bulk decisioning can rapidly uncover clusters of fraudulent activities, increasing review efficiency. Fraud analysts need only to review a handful of sample cases before confidently making bulk decisions applicable to all cases within the same fraud ring. Accelerating case reviews has been shown to reduce operational overhead by 40% or more. 

The delivery company in the above example put these technologies to use and successfully thwarted two fraud attempts this year that could have been massive in scale and caused significant damage. 

First, the company detected a fraud attack made of 257 users that had registered in a 4-day span in March. Another fraud attack was also detected that involved 30 users over a 15-day period in April. Each of the accounts engaged in tracking and re-routing packages and shared several other characteristics, including email address patterns, IP addresses and device IDs. Had it not been for the company’s advanced fraud and risk solutions, the delivery company could have lost millions.

Shipping fraud is omnichannel, so your prevention should be too

Since the onset of the Coronavirus disease (COVID-19) pandemic, purchasing behaviors have given rise to new fraud patterns, and shipping fraud is a popular vector. Fraudsters can use any customer touchpoint — account registration, payment or other translation or interactions with digital portals and customer support systems — as an entry point to steal user information and commandeer accounts. Because deliveries involve many customer touchpoints, preventing shipping fraud requires an omnichannel approach that provides continuous protection across the customer account lifecycle. 

Advanced technologies such as UML and holistic data analysis can help logistics and delivery companies detect fraud attempts early and in an automated fashion, no matter where they originate. This prevents attack and downstream damage, while reducing the high cost and operational overhead often associated with traditional methods of fraud prevention.