Big Data in Fraud Management

Fraud management has become a major priority for many organizations, in particular in eCommerce, travel and healthcare.

Traditionally, fraud management has focused on successfully catching illicit activity such as cybercriminals attempting to defraud a bank, insurer, or merchant or organized criminals attempting to defraud bank, insurer or merchant or organized criminals attempting to a launder money through a financial institution.

Fraud Management

Reporting the money laundering is not only a moral and ethical obligation but also a regulatory requirement. To catch fraud, firms often implement enterprise fraud management (EFM) solutions that use known patterns of fraud together with behavioural models and other contextual information to identify suspicious transactions and activity. Fraud analysts must perform a difficult balancing act: they must improve detection rates while not preventing legitimate transactions.

To keep up with creative fraudsters while not fraudsters loyal customers, fraud analysts and EFM vendors must integrate new data sources such as device geolocation, device fingerprint, textual data, ATM specific information and website clickstream data.

[pullquote position=”right”]Customer experience is a new and important component of successful fraud management.[/pullquote] S&R pros have come to realize that it is important to balance the ability to detect criminal activity with user experience. For example, adding multiple user screens upon registration, requiring multiple security questions and implementing one-time password upon login to online banking all reduce fraud, but they also create significant transaction friction.

If S&R pros can implement big data techniques and technologies, the volume, velocity and variety of data they can collect and derive insights from will have benefits for every component of fraud management triad. The benefits include:

  • Higher fraud detection rates. The ability to process significantly more information, store it for longer periods of time and incorporate shared information about transactions will give S&R pros a much better understanding of the patterns of fraud and criminal activity in their environment and the ability to respond to any new threats more quickly.
  • Less money spent on fraud management. Using faster algorithms to process more data allows for faster, even real-time, integration of data across multiple channels such as phone, web and mobile and yields a faster cross-channel analytics system. Now, fraud analysts can view a much richer context of transactions in real time.
  • More satisfied customer. Higher fraud detection rates, coupled with higher accuracy, means that analysts will not decline of flag for review the legitimate transactions of good customers. It also will help to reduce the friction of conducting transactions.
  • Better business decisions. The availability of richer data sets will make the process of turning data into information, then knowledge and then action much faster and will require fewer human interactions. The organization can use this business intelligence not just for fraud detection but for better targeting of users with special offers, more effective ad campaigns for cross-sell and upsell, and better strategic organization of the company.
  • Improved cyber security. Security breaches or attacks against a site or set channels often follow fraud. Looking at more and greater variety of data not only improves early fraud detection but can also help to identify security breaches and shorten response times to new and emerging threats.

Big data is not just about volume. It is also about the velocity of change, format variety, and structural variability. Volume, velocity, variety and variability of big data in fraud management will have distinct impacts on fraud management staff, process and tools.


Related Posts