the one & only user (behaviour)

How Behavioural Analytics Can Help Fight Fraud

Combating fraud in e-commerce is a struggle against a constantly evolving enemy and new techniques and tactics are being developed all the time to detect and prevent it. Among the newest weapons deployed on this battlefield is behavioural analytics.

behaviour tracking

In essence, behavioural analytics, in the context of e-commerce, analyses data in order to tell apart a real customer's shopping activity from fraudulent activity. Once a fraudster is detected, the standard anti-fraud measures can be put into action to deal with them. This is, of course, just one of many examples of why anomaly detection is important for online businesses, along with, for example, its application in countering botnet attacks on websites and networks, which also involve computers acting anomalously.

Behavioural analysis is possible thanks to computer code that tracks activity on a business' website or mobile application. Each time a customer interacts with either, the data about their activity is sent to the analytics system which compares it with the data already present in its database.

This data can be anything from typing speed to mouse cursor interaction to device used and location the activity is performed at, as well as various signals and sensors that may be in the customer's vicinity, such as Bluetooth devices or Wi-Fi networks. That database contains data about each account holder's usual behaviour. Since each holder is a unique person, their behaviour is also unique and recognizable to the BA system. This allows the BA system to detect anomalous behaviour from a given account and subsequently detect fraud. 

Artboard 2 copy 10.png

key benefits

BA systems operate in real time, which makes them a very valuable asset along with the ability to monitor all activity of every account holder. They are adaptive and can react to new, never attempted fraudulent activities, automatically calculate the risk of each new activity and, if the risk of fraud is detected, choose which type of intervention to employ.

The introduction of 3DS 2.0 will bring both token-based and biometric authentication for card transactions but also increase the data available to card issuers. This increase in the amount of data will improve the efficacy of BA systems for card fraud detection, providing them with more points of comparison.

They are also, very importantly, non-intrusive in the sense that they do not require the customer to change their interaction by installing any new software or adopt a new security process. This provides a frictionless user experience and increases customer trust and satisfaction.