Businesses, banks and insurance companies are losing billions of dollars every year due to fraudsters who have developed sophisticated ways of setting up new identities and eluding detection.
According to the Association of Certified Fraud Examiners (ACFA) 2014 Global Fraud study, survey participants estimated that a typical organisation loses 5% of revenues each year to fraud. If applied to the 2013 estimated Gross World Product, this translates to a potential projected global fraud loss of nearly $3.7 trillion.
The mean loss caused by the frauds in the ACFA’s study was $145,000. Additionally, 22% of the cases involved losses of $1 million or more.
Whilst no fraud prevention measure will wipe out fraud altogether, procedures can be put in place that will dramatically cut fraud-related losses. The place to start is individual data points and their connections. These connections are often ignored, when in reality they are the first place to start to look for the clues to any potential fraud being carried out.
Joining the dots and making sense of these connections, however, is more complex than it sounds. It is about looking at data from a new perspective and looking for patterns in it. This powerful technology developed to pull together relationships in data is called a graph database.
An increasing number of businesses, banks and financial institutions are turning to graph database technology in a bid to detect fraudulent behaviour and stopping it in real-time. The technology is also helping develop next-generation fraud detection systems based on connected intelligence.
Types of fraud
There are various types of fraud – insurance fraud, e-commerce fraud and first-party banking fraud etc.. The latter is where fraudsters apply for credit cards, loans, overdrafts and unsecured banking credit lines, with no intention of paying them back. It is a growing problem for banking institutions.
All these forms of fraud have one thing in common. They involve layers of carefully thought out and planned deceit to cover up the crime. With each of these types of fraud, graph databases can assist existing methods of fraud detection, diving deep into the layers and building up an image of the connections.
Catching fraud rings and stopping them is a big challenge. Traditional methods of fraud detection are either not programmed to look for the right thing such as rings created by shared identifiers, for example. Standard methods – such as a deviation from normal purchasing patterns – depend on discrete data and not connections. Discrete methods can be useful for catching fraudsters acting alone, but they fall short in their ability to detect rings. In addition, many such methods are prone to false positives, which creates unwanted side effects in customer satisfaction and lost revenue opportunity.
The power of graph databases
Uncovering fraud rings with traditional relational database technologies requires modeling the data as a set of tables and columns, then working through a series of complex joins and self-joins. These queries are not easy to build and also expensive to run. Scaling them so that they support real-time access comes with major technical challenges, with performance becoming increasingly poor as the size of the ring increases and the total data set grows.
On the other hand, finding fraud rings with a graph database becomes a simple question of walking the graph. Because graph databases are designed to query intricate connected networks, they can be used to identify fraud rings in a fairly straightforward way.
PayPal, which moved $230bn worth of currency over its networks in 2014, still employs graph techniques to perform sophisticated fraud detection at massive, global scale.
It’s not just the PayPals of the world who are exploiting the power of graphs, however. An increasing number of enterprises, even at the smaller (SME) level, are using them to solve a variety of connected data problems, with Forrester Research predicted that over a quarter of enterprises will be using such databases by 2017.
A secure future
In an increasingly digital age, business processes are becoming faster and more automated. The margins to detect fraud have become narrower and narrower, demanding real-time solutions.
Sophisticated fraudsters know where to look for the weak links in a chain to attack systems. Traditional technologies, whilst still having a place in fraud prevention, are not powerful enough to spot elaborate, and often global, fraud rings.
Graph databases provide a unique ability to uncover fraud patterns in real time. Connections that were previously hidden, become obvious when queried using graph technology. Chief Security Officers who haven’t already explored graph database technology would be well advised to look at this approach to protecting against high impact fraud scenarios that will only increase.
By Emil Eifre, co-founder and CEO of Neo Technology, the company behind Neo4j