How Retailers Can Leverage AI to Handle Seasonal Spikes in Policy Abuse 

Ecommerce fraud is seasonal. Gift card scams rise during Valentine’s Day, Mother’s Day and Christmas. Chargebacks peak right after the winter holidays. And policy abuse sees periodic spikes, but is highest in the summer months.

Policy abuse, the misuse of merchant policies for personal gain, varies widely. It includes mundane acts like returning clothes after wearing them once, to criminal activities like using invisible ink on tracking IDs to fake returns. This abuse spans policy types, affecting sign-up promos, refunds, and limited edition items.

Many merchants write off all such policy abuse as “the cost of winning customers.” But it’s skyrocketed in recent years to a billion-dollar problem. One they can no longer afford to ignore, but are afraid to pull back on for fear of losing customers and profits.

Luckily, merchants have a range of solutions at their disposal to combat policy abuse. The first step though, is to understand the unique seasonality behind abuse trends.

Seasonal Surges in Policy Abuse

Tackling abuse trends starts with understanding the reasons behind each buying season. Let’s take summer and holiday as prime examples. When surveyed, over half (53%) of shoppers say rising financial strain as the main driver of policy abuse. The connection here is clear. During holiday and summer, those strapped for cash might seek “alternative” strategies to fund the vacation or gift they’re planning. The same research also showed about a quarter of shoppers (23%) say their summer policy abuse was to afford seasonal purchases.

Most shockingly, the majority of people behind this policy abuse are far from professionals. They are average customers who want to buy more but spend less during tough times. In fact, the research found that most consumers (54%) feel guilty for policy abuse. This is in contrast to professional fraudsters who boast about their exploits all over the dark web.

So, this helps clarify the problem further. What retailers want, especially during the summer peak, is a way to filter out the worst abusers, while lightly discouraging the casual abusers. Detecting, applying and deciding all this at scale, though, is a challenge. And that’s exactly where specialized AI comes into play

Marketing Technology News: MarTech Interview with Stacy Bohrer, VP of Buyer Development (North America) @ OpenX

Sizing Up the Policy Paradox

While policy abuse varies in type and severity, it’s surprisingly consistent in one critical way.

Whether it’s returning most items order after order, or snagging five promos by signing up with five different emails, almost every policy abuse tactic, no matter the sophistication, leverages how individuals behave over time.

Casual rule-benders might return the occasional party outfit they only wore once, or double-dip on a discount here or there. And if they get caught, or just have a guilty conscience, they may cycle between a work and a personal email or billing address to try and cover their tracks.

Professional fraudsters go way further. They engage in targeted, well-planned attacks. Knowing retailers will be swamped with summer orders, they’ll coordinate a steady stream of return scams. Each coming from a fresh, unique identity, sharing zero data points with each other. If the merchant’s not checking each claim, it looks like nothing more than a few unlucky customers each day.

The true challenge is, how do you tell the two types apart, while not getting in the way with the vast majority of non-abusive customers out there?

Turns out, merchants have most of the raw data they need to solve this problem today. What they’re missing is the right AI to translate it into insight and action.

Taming Policy Abuse Peaks with AI

Leveraging AI can help merchants overcome the central central challenge of policy abuse – revealing the true identities behind interactions. Relying on a single data point, like an email, is insufficient and easily manipulated. Correlating multiple data points strengthens identity accuracy but requires constant updates and can risk blocking innocent customers.

Advanced algorithms provide a more accurate, resilient solution with minimal manual work. These specialized clustering algorithms analyze account pairs and data points to define precise identity clusters, thwarting both casual abusers and sophisticated fraud rings. For instance, casual abusers using multiple emails to double up on discounts would be identified, as would fraudsters claiming multiple high-value refunds.

AI enables merchants to detect and respond to different types of abuse. It can automatically identify repeated uses of a sign-up promotion at checkout and trigger appropriate responses, such as a warning or a slightly lower promo as consolation. For excessive refund claims, AI can block further abuse by flagging and rejecting suspicious claims or even preventing related orders from being placed.

By employing these methods, merchants can effectively address policy abuse, regardless of type, tactic, or season.

Marketing Technology News: Server-side tags are more critical — and simple — than ever

The SalesStar Podcast –

Episode 209: B2B SaaS Marketing and MarTech Best Practices with Ryan Nelsen, CMO at StackAdapt