Kount LogoBLog Against FraudKount Logo

Anatomy of a False Positive

posted on: Tue Jan 09 2018

Online businesses are wrongly turning down legitimate orders to the tune of $118 billion a year, according to Javelin. Let’s put those false positives into perspective:

  • Quarterly eCommerce sales in the US averaged $108.5 billion over the past four quarters (per the U.S. Department of Commerce). In other words, that $118 billion equates to every online business in the world essentially foregoing an entire quarter of sales in the US!

The root of all this wasted potential revenue can be boiled down to a simple equation:

Fraud Equation

Let’s dissect the first part of the equation—fear of fraud:

CONCLUSION: It’s absolutely reasonable for online businesses to fear eCommerce fraud. It represents a large and growing threat to profitability.

Next, let’s examine the second part of the equation—inability to accurately quantify risk:

  • Nearly 1 in 3 transactions (30%) declined due to suspected fraud are believed to be legitimate.
  • 55% of businesses who process online orders do not take advantage of the expertise of third party fraud solution providers, choosing instead to DIY (according to the forthcoming 2018 State of Chargebacks Survey, which will be published here).

CONCLUSION: eCommerce operations lack the systems to quantify the true cost of losses from false positives and/or to assess the actual ROI that would result from implementing more accurate fraud prevention to reduce false positives.

What are common causes of false positives?

  • Misleading negative signals (for example, unusual geo-location data from a customer who is traveling).
  • Lack of positive signals (for example, inability to check 3rd party sources to assess email address veracity).
  • Differing risk appetites among issuers, acquirers and processors.

What’s a solution? A smart first step is to evaluate the ROI of an enterprise-class fraud prevention solution that employs advanced screening technologies, Artificial Intelligence/Machine Learning and policy-based decisioning for continuous, adaptive risk assessment:

  1. Every transaction is presumed innocent until proven suspicious by…
  2. Negative signals generated by patented screening technologies and Artificial Intelligence/Machine Learning that identifies transactions worthy of greater scrutiny…
  3. So that positive signals—also influenced by Artificial Intelligence/Machine Learning—can be applied to questionable orders to deliver a quantified risk score that is more fully-dimensionalized, which gets…
  4. Included into policy-based decisioning to automate fraud prevention decisions, speeding resolution and reducing overall costs.
  5. End result? Continuous, adaptive risk assessment that evaluates not just the present transaction, but continues to assess all future data/links associated with that transaction every 30 minutes over the ensuing weeks and dynamically adjusts the risk assessment if data/links, 3rd party corroboration, and/or corporate policies change.

In other words, it’s like having an “always-on” time machine that automatically re-evaluates hundreds of data points about every transaction—both positive and negative—and makes automated decisions using policies customized by you. You set the level of risk that is exactly right for your organization, not some black box algorithm over which you have no control. The outcome from all of this is the confidence to approve more orders, dramatically reduce false positives, and boost sales while slashing fraud losses.

Discover more about false positives and the proven strategies for avoiding them to increase revenue in the eBook “The Silent Sales Killer: False Positives.”

New Call-to-action

Topics
Share
TOP