September 27, 2016
Luke Reynolds, Featurespace Fraud Director, explores the impact of over-sensitive fraud systems on costs of fraud management and the negative impact it has on customer relationships.

Reducing customer friction in managing fraud

Declining genuine transactions (also known as ‘false positives’) is estimated to cost card issuers over $14.7bn annually.

Financial scams up by 53% in 2016

Last week, the Financial Fraud Action (FFA) reported that a financial scam occurs once every 15 seconds in the UK alone. With criminal scams up 53% since the start of 2016, it is easy to understand why financial institutions and merchants find it necessary to turn up the dial on their fraud alerting systems.

But what is the impact on customer friction? Yes, you are protecting your customers from potential fraud. However, as many existing fraud systems rely on pattern-matching against known types of fraud, dialling-up the thresholds on unusual activity can have a negative knock-on impact by blocking genuine customers. More blocked customer means more unhappy customers and that is always bad for business.

Reducing the $14.7bn cost and impact on customer friction

This has a huge impact on both revenue streams and customer friction. Using industry data, Oakhall Analysts, working with Featurespace, estimated that the card issuing industry’s annual losses associated with this problem are over $14.7bn.

This is not only from lost revenue and customer service management of blocking a genuine transaction, but the knock-on ‘back of wallet’ effect, where a frustrated or embarrassed customer decides not to use the blocked card for a while after the event.

Keeping customers happy with machine learning

It’s a lot of revenue to be losing, and the inconvenience makes customers look elsewhere. So what can financial institutions and merchants do to keep customers happy, and protect themselves, their revenues and their reputation?

Instead of dialling-up the thresholds on incumbent systems, the answer is to understand each customer’s behaviour better.

The key is to identify what is normal and what is anomalous – or uncharacteristic – behaviour, even within the profile of a genuine customer.

Most approaches to behavioural analytics focus entirely on whether or not the person is who they say they are. A different approach is to enable merchants and card issuers to understand the behaviour of individuals, and detect when a customer may be starting to change their behaviour in subtle ways.

Featurespace’s ARIC engine is a software platform that uses real-time, adaptive behavioural analytics to understand the behaviour of each individual customer, building a profile of their normal patterns of behaviour.

Using real historical customer data provided by card issuers, Featurespace demonstrated a 70% reduction in false positives using its machine learning behavioural analytics, which means less incorrectly blocked cards and more happy customers.

Key takeaways

  • Declining genuine transactions causes customer friction and is estimated to cost card issuers over $14.7bn annually.
  • Blocking genuine customers no longer has to be a necessary cost of protecting your business from fraud.
  • The ARIC platform is making it possible for financial institutions and merchants to profile anomalies within an individual customer’s behaviour profile and block new fraud attacks while accepting more business from genuine customers.