Fraud Control at Checkbook

Posted by Checkbook on Aug 05, 2020

At Checkbook, besides sending money and digital checks, we full-on stop fraud. In fact, we boast one of the lowest levels of chargebacks and fraud incidences amongst our peers in the payment processing community by orders of magnitude!

Why is Checkbook so good?

Because Checkbook has organically built an entirely in-house detection system that can better spot fraud than your run-of-the-mill suite of programs.

The secret to our success is starting from scratch and building atop the fundamentals of data to uncover hidden connections and overlooked patterns. We begin with the basic premise that no one data point is absolute, and that data is context-bound, and context is boundless​. The more data you have, the more confidence you may have in detecting fraud and preempting fraud patterns. As they say, the devil is in the details -and the devil is easily detected at Checkbook.

At the attribute level, we seamlessly require and authenticate various data points in the onboarding/sign-up process, such as bank account information, email addresses, phone numbers, Tax ID, etc. that are evaluated both granularly and holistically. In addition, we also monitor and collect all behavioral data that, when taken into account with attribute data, paints a more vivid picture by which to evaluate any account. Each piece of data qualifies another, and we leave no piece of data unturned.

However, while most systems already understand the importance of attribute and behavioral data, we simply have a different philosophy in understanding data as a story, and we know that stories are often episodic with predictable plot devices. In other words, fraud is dynamic and adaptive, and so are we.

We employ both manual, ground-level review and machine learning to detect emerging trends and tactics, and preempt them often before they can be successful. Fraud agents are human after all and, as such, follow predictable patterns in even simple pieces of data such as amount generation, since we’re such terrible random number generators (and name generators, etc.). Often, it’s those “softer” data points that can lead to helpful, hard rules.

In the end, it’s our philosophical approach to data, and the leveraging of both technology and human talent that informs our unrivaled pattern recognition to ultimately thwart fraud and protect our customers.