5 Trends Fraud Leaders are Talking about Right Now

The COVID-19 pandemic has changed our lives and habits significantly. Even as the country begins to reopen, many of these changes will be permanent. During lockdown orders, people have grown accustomed to working remotely, for example, and large organizations such as Twitter and Facebook have already announced plans to continue to support companywide remote working. People have also become accustomed to shopping online: ACI Worldwide reports that many online retail categories saw as much as a 74% year over year increase compared to March 2018. These and other trends will have an unprecedented impact on businesses and the global economy — the extent to which has yet to be seen.

As individuals change the ways in which they interact with businesses, fraudsters will also be changing their tactics. Since the novel coronavirus outbreak, we’ve seen a 40% increase in new account fraud and a 2X increase in transaction fraud. To take back control and fight new types of fraud attacks, organizations must revise their fraud and risk strategies.

Emerging Trends from Fraud and Risk Leaders

In recent months, I’ve spoken with many executives from different sectors about how they see the nature of fraud detection changing, and how their strategies are evolving to keep pace.  Here are some trends that emerged from those discussions:

1. The need to reduce the cost of fraud management. An increasing need to manually review suspected fraudulent events has led to rising costs. To date, automated fraud detection solutions based on rule-based and supervised machine learning techniques produce a high rate of false positives, which costs retailers sales opportunities and revenue. It also has significant implications for customer satisfaction. 

To reduce the adverse impact of false positives for both retailers and customers, some organizations are lowering the threshold for manual review of suspected fraud. A lack of internal resources has forced many to outsource this task to third-party agencies and service providers. Many senior executives are looking at leveraging emerging technologies such as unsupervised machine learning to reduce the number of manual reviews associated costs.

2. The need to unify fraud detection systems and technologies. As fraud patterns and activities have evolved, organisations have added point-based solutions to combat specific use cases or issues. There are two main issues with maintaining disparate solutions: 

• The systems are not integrated, and therefore fail to provide a full view of the customer lifecycle, from registration to transaction. As a result, sophisticated fraudsters can get through undetected. 
• This approach adds unnecessary complexity, and the costs of managing and maintaining disparate, often overlapping solutions are spiraling out of control. 

Many executives are looking to simplify their point-based fraud detection solutions and replace them with a modern platform that supports multiple traditional and leading-edge technologies for the whole organization.

3. Fraudsters are increasingly organised and sophisticated. Many executives are raising concerns about the shift from individual fraud attacks to coordinated fraud rings that exhibit increasing sophistication and inventiveness. This is true for all industries and sectors — Financial Services, e-commerce marketplaces, social media networks, logistics and delivery, and many others. 

The vast majority of available fraud detection systems are designed to identify and detect individual activities such as account openings/registrations, product listings, reviews, and transactions. When such activities are scanned and analyzed individually, they often appear legitimate. However, when they’re looked together in a holistic manner, they may show clear signs of coordinated fraud activity. To that end, fraud detection solutions that can identify and prevent sophisticated, coordinated fraud attempts without manual intervention are gaining traction. 

4. Well established fraud detection teams do not want a black box. Financial services organisations with large fraud and data science teams are reluctant to use fraud solutions that don’t offer a comprehensive understanding of what is really happening in their networks. Such “black box” offerings leverage supervised machine learning models that may appear to be working but may feature models that are trained on biased data. Organisations with the means and resources prefer to use a platform that enables in-house data scientists to create a model, interrogate the data and perform multiple A/B tests. They want a platform that provides complete visibility and control and features cutting-edge techniques. Unsupervised machine learning is a central feature of such a platform. 

5. Online retailers can’t keep up with fraud trends. Until now, most marketplaces and online service providers have been building basic fraud prevention strategies into their core systems. Forward-thinking executives within these organisations are beginning to see that fraud prevention and detection is a very specialist and fast-moving discipline that falls outside of their core competencies. As they realize they lack the internal resources and skill set to keep up with the evolving fraud landscape, they’re looking for help from specialist fraud solution providers with domain expertise.

New Normal, New Fraud, New Approach

DataVisor protects and monitors over 4.5 billion accounts globally. A recent article by Ting-Fang Yen, Director of Research at DataVisor, shares her findings on how the financial crime landscape is changing. To learn more, read the full article here.

Inevitably, new fraud vectors will emerge, and fraudsters will continue to exploit increased online activity, particularly on mobile devices and apps. To protect their organizations, fraud and risk teams must adopt innovative strategies that can identify and detect new and evolving threats in real-time.