arrow left facebook twitter linkedin medium menu play circle
August 19, 2020 - Pule Wang

How DataVisor Saved a Logistics Company $4 Million: Q&A with Pule Wang

In the United States, the USPS handles more than 500 million pieces of mail each day, while FedEx and UPS deliver a combined 34 million packages, not to mention the packages handled by smaller parcel services. With such a high volume of packages in transport at any given time, it’s no surprise that the shipping and delivery service industry is ripe with fraud. 

For one logistics provider, shipping fraud had become an insurmountable problem that was costing the company millions of dollars each year. Scammers were creating fake accounts to reroute packages, track criminal shipments, and commit identity theft. 

When the company turned to DataVisor for a fraud detection solution, it was able to increase its positive detections and save more than $4 million in the process. In an interview, DataVisor’s Pule Wang shares how the company generated these results.

What Does Package Shipping & Delivery Fraud Look Like?

According to Wang, the logistics provider was experiencing fraud in a large, coordinated effort. “Using one account to intercept or reroute a large volume of packages wouldn’t have gone undetected,” notes Wang, which is why the users would register multiple accounts. He also mentions that the fraudsters were residing in a foreign country, one of the common denominators that was later detected by DataVisor’s solutions. 

The fraudsters were believed to be obtaining real shipping numbers from real customers, then using their fake accounts to reroute the shipped goods to another destination. They were also sending good users malicious emails in an effort to retrieve package-related data. 

Once the fake accounts were created, fraudsters could log into the online portal to reroute packages to different destinations, issue package holds for delayed pickups, and spam authentic users with emails. In some cases, fraudsters could use real customer data to create fake accounts and modify orders to include higher-value items. They could also manipulate customer orders or accounts via phone if they had certain pieces of customer data. 

All of these acts of fraud were causing reputational damage to the company, a poor experience to the customer, and massive financial loss in lost package claims and other expenses. These attacks were bypassing the company’s current fraud prevention systems that didn’t review data on a broad scale like DataVisor’s dVector

The company needed a proactive solution to stop fraudsters at an early phase (the incubation phase, if possible) to mitigate financial loss and reduce pressure on the company’s fraud team. What’s more, it needed to achieve these results at scale. 

How DataVisor dVector Uncovered Sophisticated Shipping Fraud Scams

DataVisor’s role in reducing shipping fraud is to uncover patterns occurring during the early signup and incubation phase, as well as when requests are submitted for package reroutes. The company chose dVector, an AI-powered platform that uses unsupervised machine learning and global intelligence to detect large-scale, coordinated fraud patterns.

Upon implementing dVector, the company detected a fraud attack made of 257 users that had registered in a four-day span in March. Another fraud attack was also detected, this time involving 30 users over a 15-day period in April. Each of the accounts engaged in tracking and rerouting packages and shared several other characteristics, including email address patterns, IP addresses, and device IDs, among other details.

Here’s a quick recap of how DataVisor revealed and mitigated the company’s package delivery fraud:

Finding More Fraud through Holistic Data Analysis

DataVisor’s dVector detected connections between seemingly unconnected data points to uncover sophisticated fraud attacks. The platform analyzed data on a holistic level to find anomalies or suspicious activities. 

In doing so, dVector was able to detect 60% more fraud attacks beyond the company’s existing systems, resulting in $4 million in savings in just one year. 

Early Detection

Because of its ability to review large sets of data at scale, dVector successfully thwarted many suspicious activities at the point of registration before fraud could be carried out. Many activities that made it past registration were identified and prevented at the first attack to further reduce financial and reputational damage to the logistics provider. 

On average, dVector was able to identify fraudulent accounts up to 30 days earlier compared to predefined client labels in a cross-validated training set. Unlike other fraud prevention methods, one of dVector’s strengths lies in monitoring for fraud before an attack occurs so that attempts can be neutralized before they cause damage.

Accelerated Review

Last but not least, dVector allowed the company’s fraud teams to quickly review suspicious activity using automation and bulk decision making. The platform uncovered clusters of activities being committed by the same group of fraudsters, saving the company time in the review process without sacrificing confidence in its decision.

As a result, the company was able to increase its review efficiency by 40%.

Stopping Shipping Fraud with DataVisor

Wang believes that fraud is rampant in part because digital identity is cheap and readily available. It doesn’t take much time, effort, or expertise for anyone to create a fake account to conduct illicit activities. 

DataVisor’s UML and dVector platform are a powerful combination in identifying coordinated fraud, despite scammers’ best attempts to conceal illegal activities. Unlike conventional fraud detection methods, dVector doesn’t rely on pre-existing rules or knowledge of how attacks are carried out. Rather, UML allows the platform to look at large data sets at scale to find suspicious activity. It analyzes all accounts and activity in real-time so that reviewers can make immediate decisions and prevent fraud before it occurs.

In this way, companies like the logistics provider are better able to defend themselves not just against past mistakes, but also future attacks as they evolve. 

Download the full shipping fraud case study, How DataVisor Saved a Logistics Company $4 Million.

about Pule Wang
Pule has 5 years of experience in data solutions and technical professional services. He obtains his Master’s degree from Carnegie Mellon University specializing in information system management and data science. Pule is the Technical Account Manager at DataVisor.
about Pule Wang
Pule has 5 years of experience in data solutions and technical professional services. He obtains his Master’s degree from Carnegie Mellon University specializing in information system management and data science. Pule is the Technical Account Manager at DataVisor.