Online businesses are starting to feel the strain of the current macroeconomic environment, including shifting consumer behavior and new fraud patterns. But periods of uncertainty and rising fraud rates don’t have to hurt your bottom line. 

To address these economic challenges and shed light on how they’re impacting the fraud landscape, Sift hosted a webinar on building resilient fraud management strategies in the face of uncertainty. Sift’s SVP of Engineering and Cloud Operations Neeraj Gupta and Trust and Safety Architect Jeff Sakasegawa discussed new fraud findings, tips for fraud analysts to streamline operations, and how to boost efficiency with Digital Trust & Safety. 

Teams are experiencing resource and budget constraints

Sift recently surveyed trust and safety industry professionals and found that 26% of risk teams are experiencing economy-related resource or budget constraints in response to shifting economic conditions. These constraints can add pressure to risk teams and analysts, resulting in rushed decisions and decreased accuracy. 


But a reduction in resources doesn’t have to impact the effectiveness of fraud operations. To help offset some of this strain, businesses can rely on machine learning automation to make decisions at scale more consistently and reduce manual review, as well as connect data from various sources. Many analysts are using over a dozen independent tools and systems, which can be frustrating to navigate and difficult to share insights. But there are ways to maximize existing systems by integrating data into one unified view. For example, Sift Connect offers an ecosystem of connectors to reduce the number of clicks and scrolls it takes to get the information needed to make informed decisions. 

Businesses are re-evaluating their needs

Sift found that 45% of trust and safety experts are re-evaluating what their businesses need to effectively fight evolving fraud in a changing economy. This figure indicates fraud analysts aren’t getting what they need from their fraud prevention tools, or aren’t using these tools to their full potential. In order to navigate an unpredictable future, teams need a solution they can trust to accurately and efficiently block fraud. 


“It’s an established fact that rules-based systems don’t really work anymore. Many of the solutions you see out there now are based on machine learning. What we’ve seen over the past 6 months is that the data volume and the data diversity have changed a lot,” said Sakasegawa. 

These changing fraud trends go to show that there’s no one-and-done solution out there. To reliably withstand fraud without it impacting business health, fraud teams must stay vigilant and invest in solutions that are capable of learning in real-time. With an effective, end-to-end fraud prevention strategy backed by machine learning, companies can prepare for economic turbulence while reducing costs, maintaining a positive return on investment, and recouping lost revenue.

At Sift, we are constantly evaluating how our solutions can be better—both at detecting fraud and making analysts’ lives easier. We’ve recently released a few features that improve productivity and help businesses customize their product experience. To help analysts cover more ground in less time, we developed Sift Layout Manager, which allows users to customize their Sift Console display and make important data more visible and actionable. We’ve also made it easier to make bulk decisions through the Sift Network view, giving analysts the ability to select multiple queue cases and apply a decision in just a few clicks.

Fraud is increasing across the board

Our recent survey showed that 68% of trust and safety experts have seen a 5% increase in fraud since January 2022, with some experiencing spikes up to 50%. This increase in fraud during economic disruption is relatively expected—we’ve seen fraud rise in times of uncertainty, including the COVID-19 pandemic and the great recession. 

“Fraud is not going away. Fraudsters are continuing to exploit the current environment we find ourselves in, and there are now more surfaces to exploit,” said Sakasegawa.


Across our customer base, we’re seeing a shift in fraud patterns, including increases in first-party fraud (i.e., friendly fraud), account takeovers, and infiltration of dormant crypto accounts. But one common thread is how these various fraud trends are becoming more sophisticated, automated, and speedy—fraudsters and organized fraud rings are continuing to go to extreme lengths to break through merchants’ security measures and amplify their profits. 

“The sophistication and speed of fraudsters are increasing. If you’re training your model every 3-6 months, you’re going to miss a lot of these fraud patterns. It’s important that as part of your machine learning, you have both offline training models where you’re training your model every few weeks or months, but you also have an online learning part, which is really important in this changing economic environment,” said Gupta.  

“Fraudsters are no longer waiting a month or two to hit another vendor with the same attack. If a fraudster is making an attack on one of our customers, our machine learning models can learn in real-time so if the same fraudster is attempting that a few minutes later on another customer site, we can prevent it.”

Our Digital Trust & Safety Platform is built to dynamically prevent fraud and abuse with real-time machine learning that adapts based on our unrivaled global data network of 70B events per month. 

Watch the webinar to get more insights on strengthening fraud operations during economic uncertainty.

Related topics

Digital Trust and Safety

economic uncertainty


fraud operations

fraud prevention

You may also like