Published: March 2, 2021
Reading time: 3 minute read
Written by: Forter Team

The number of online transactions has risen dramatically in recent months as a vast majority of consumers shifted from shopping primarily in-store to shopping online. Fraudsters have noticed the rise in digital transactions and merchants expanding to new channels, services, and markets.

As more merchants begin a digital transformation, fraudsters have become more innovative, finding new and faster ways to commit fraud and abuse, as well as new vulnerabilities in underprotected areas of the buying journey. Online criminals use automated tools like bots and device emulators to launch massive fraud attacks and mask their activities when they take over user accounts. They look for vulnerabilities at every point of the customer purchasing journey. Fraudsters have learned quickly how to take advantage of merchant omnichannel offerings like buy online and pick up in-store (BOPIS). In 2020, we saw a 55% increase in BOPIS fraud attacks.

Merchants need to prepare for more sophisticated and automated forms of fraud. They need to use modern approaches, such as machine learning and fraud analytics, to battle fraud and abuse effectively. However, many merchants continue to use the same legacy fraud prevention solutions they always have.

Challenges with traditional approaches to fraud prevention

Most merchants today use a combination of solutions to try to prevent fraud, which may include:

    • Blacklists/whitelists
    • Rules-based system
    • Fraud risk scoring
    • IP reputation service
    • Device fingerprinting

These approaches, even when used in combination, present challenges for merchants. Using traditional fraud prevention approaches leads to:

    • Lower accuracy: Most traditional fraud prevention approaches like rules-based systems and fraud risk scoring focus on individual transactions. Merchants can lose up to 75x more revenue to false declines than they do to fraud.
    • Gaps among tools: In general, traditional fraud prevention tools rely on siloed data and lack the ability to coordinate with each other. If you piece together different fraud prevention tools you run the risk of leaving gaps that fraudsters could easily exploit.
    • Scalability issues: Standalone tools that were not designed to work together are difficult to scale. Imagine having to scale dozens of individual fraud prevention tools during peak periods or for flash sales.
    • Reactive not proactive: Legacy fraud prevention tools tend to be reactive rather than proactive. Reactive approaches can’t stop new and evolving fraud, or meet the challenges of more sophisticated methods of fraud and abuse.

You don’t have to avoid expanding your business because you fear the risks of operating in new regions and markets. However, if your business leverages fraud analytics, you can deliver new products and services in new markets without the worry of increased fraud and abuse.

What is fraud analytics?

Fraud analytics typically involves a team of experts monitoring fraudster marketplaces and forums to uncover and track new fraud vectors and the trends in online fraud. These researchers then take their insights and feed them directly into the machine learning (ML) fraud models. The researchers do more than simply place raw data into a model. They curate the data inputs to build a real-world story about fraud and build context around the disparate data points.

For example, if you have a transaction involving an IP for a location in Peru and a card issued from a bank in Israel, you could simply feed that information to your ML model. However, doing so would likely result in a false positive. With fraud analytics, the team engineers features to help the model understand the bigger picture around the data— in this case, the customer is most likely an Israeli tourist visiting Peru.

A better way to prevent fraud

You need fraud analytics to prevent fraudsters from successfully using new tactics to defraud your business and cut into your potential revenue. And fraud analytics requires a dedicated team of fraud experts and data scientists. These researchers must uncover fraud trends and engineer ways to identify and stop all the new fraud techniques they uncover. Through leveraging a team of fraud experts who continually conduct research, your business will be able to:

  • Proactively research and uncover new fraud trends
  • Identify behavioral patterns in the data
  • Actively adjust the linking models and decision models
  • Identify new attack vectors

However, fraud analytics is only one part of the fraud prevention equation.

To learn more about what you need to prevent fraud effectively, download our white paper “Cyber Crime is Costing the World $6 Trillion Annually. Is Your Fraud Prevention Ready?

3 minute read