The growth in digital payments in India has been phenomenal, thanks to a supportive regulatory framework as well as the Fintech explosion. Frauds and risks associated with digital channels have also grown during this phase. Traditional rule-based approaches which scan for a set of “known” indicators, are no longer sufficient as fraudsters have innovated significantly. Behavioural analytics can help financial institutions reduce the risk of fraud, without compromising on customer experience.
In this two-part series, Drona Pay takes a look at various contemporary use cases and how Fingerprinting (Device/Browser/IP), Keystroke dynamics and Application interaction are proving to be critical components to stay safe in the digital era by helping solve a range of fraud patterns, like:
Account takeover
Suspicious logins
Suspicious fund transfers
Identity theft & possible synthetic IDs
Chargeback reduction
Spoofing
Device Fingerprinting
Device fingerprinting is a method of gathering and analysing specifications of a device to uniquely identify the device and manage its mapping to users/ customers / user id. Device attributes are collected and analyzed through machine learning to probabilistically identify the device. This approach covers Android and Apple devices through SDKs. Device fingerprinting identifies parameters including Display resolution, Browser type /version, CPU, Battery level, Rooted & Operating system.
Use cases
Flagging new mobile device or suspicious mobile device used to login to the App
Identify possible jailbroken or rooted or emulator based access to the App
Browser Fingerprinting
A browser fingerprinting is a method of gathering and analyzing specifications of a browser and the underlying device to uniquely identify the browser & device and manage its mapping to users/ customers / user id. Fraudsters often use identity concealing techniques like disabling cookies, surfing through a VPN, or using obscure browsers. These are all areas where fingerprinting proves effective since it identifies differences between normal users' browser fingerprints. Data collected includes the browser version, browser extensions, user timezone, preferred language settings, ad blocker used, device model and its operating system.
Use cases
Flagging new device & browser used the customer
Identify possible remote access or bot based use of the browser
While device and browser fingerprint are useful in identifying users uniquely, they work along with a range of other scores to make them effective.
Drona Pay helps you integrate behavioural analytics along with advanced ML models to derive insights that help enrich real-time risk scoring. The models help institutions leverage a wider spectrum of parameters/dimensions for new-age fraud detection and prevention which are constantly evolving, and impossible to be addressed with 3-4 parameters in a fixed set of rules. Self -learning of these models contribute to increased efficacies specific to your product and customer segment.
In the next part we will discuss how IP fingerprint, Keystroke dynamics and Application interaction can help financial institutions secure their apps and better identify fraudulent users.
Comments