How Brands Can Identify (and Combat) Emerging Fraud Threats

While spring is often seen as a symbol of sweeping change and rejuvenation, the most enduring changes tend to happen gradually and incrementally. It’s no different when it comes to e-commerce fraud. Brands and bad actors are engaged in an ongoing, slow-moving chess match, with subtle strategic pivots met with corresponding moves to cover up existing vulnerabilities or, conversely, expose new ones.

Even still, we can say this for certain: there will always be evolving fraud threats, and emerging technologies will likely play important roles on both sides of this incessant cat-and-mouse game.

Anticipated Evolution in the Fraud Landscape

Consider the fraud landscape as an ecosystem within the biome of the world economy. Just as we would see in nature, macro changes will trigger smaller shifts that force behavioral adaptations from those in the space. However, these changes occur at a glacial pace, and it’s only when we zoom out and examine past trends in fossilized form that we see appreciable strategic leaps from fraudsters and vendors alike.

So, what might change look like at the molecular level, then? Well, it’s fair to say we’re approaching a period of acceleration. Recession rumors have swirled for months now. Many have warned that a downturn is on the horizon, if not already here. As the effects of inflation and widespread layoffs leave everyday people financially vulnerable, small-time e-commerce fraud may have appeal as a “victimless crime,” and one for which digital channels have opened new avenues.

As brands have relied more heavily on e-commerce, fraudsters have made analogous tweaks. In-store pickups of online orders have become popular targets, as have fraudulent return schemes. Both are targeted at creating a loss for the vendor, but not all schemes are that simple.

Others involve brand impersonation, and while we saw some tongue-in-cheek attempts during the early days of Twitter Blue, work-from-home product tester schemes have proliferated of late. These involve everyday consumers “paying in” in increments before fraudsters inevitably pull the proverbial rug out from under them. Therefore, preventing fraud is not only a financial issue, but a reputational one as well, and thus requires a nuanced mitigation strategy.

Adopting a Layered Approach to Fraud Prevention

When it comes to preventing fraud, the best strategies are multilayered. In such a dynamic environment with so many methods of deception, overindexing on inhibiting one type can leave brands vulnerable to the ebbs and flows of behavioral trends. The goal should be solid, full-body armor, rather than mostly bulletproof protection that leaves an Achilles heel exposed.

What exactly does “multilayered” mean in an e-commerce context, though? Essentially, brands need to diversify the methods they use and the metrics they track, with each a component of the holistic fraud picture. A cohesive strategy is an airtight one.

That’s become a little more complicated in the post-lockdown world, though. At the onset of the pandemic, consumers had no choice but to shop through online channels (at least for the most part). However, with the re-introduction of in-person shopping and the integration of pandemic advancements like in-store pickup, the lines between fraud in digital and physical settings have blurred.

From a mitigation standpoint, manual and automated processes both have roles to play. Of course, brands should be doing things like harvesting information to identify inconsistencies at checkout and performing velocity counting exercises to pinpoint behavioral trends, but also making sure that sales pass the “eye test.”

For example, exposing returns fraud can often be as simple as opening a package and tangibly confirming that the item returned differs from the item purchased. Technology is of course integral to rebuffing fraudsters, but sometimes simple intuition can do the trick as well.

The Important Role of Sata (and AI)

While instinct is as important a fraud detection tool as any, technology’s role is ever-expanding — especially as the development of powerful artificial intelligence and machine learning tools builds momentum. Fraudsters are keen on this as well, and brands will need to arm themselves with the right tools and processes to offset the efficiency with which automation can carry out fraudulent behavior.

The core function of AI in a fraud prevention capacity should be to relieve the manual burden of pattern recognition and instead focus employee efforts on the types of mitigation work that need their human intuition. Put another way, the relationship between brands and their tech stacks should be symbiotic, with both parties complementing each other in the middle ground, but also adding value that their counterparts simply can’t replicate.

When it comes to AI, that value materializes in the form of streamlined data aggregation and analysis. One of the many benefits of e-commerce is the ability to readily collect data, which lends itself to tracking patterns. It’s a “paper trail” that traditional shopping had not offered in the past, but one with a near-boundless dataset that’s not worth the man hours to pore through. Therein lies the key advantage of automated tools: they efficiently analyze raw figures and, from them, glean insights into behavioral trends.

By constantly examining the types of fraud occurring in real time, AI and automated tools have the potential to hasten evolutionary cycles in the fraud space, with trends identified, analyzed and neutralized in a narrower time window.

All of this is to say that change is certain in the fraud landscape, just as it is in any other. However, with thoughtful strategic foundations and the right technologies in place, brands and their partners can allay fraud fears and continue to meet consumer demand across selling channels.

Tim Laudenbach is Digital River’s director of risk management and fraud operations, responsible for developing and deploying customized fraud strategies for client organizations as well as managing day-to-day operational volumes.