• Emil Eifrem, co-founder and CEO at Neo Technology


• 15.12.2015 12:19 pm

Banks and insurance companies lose billions every year to fraud. The surprising magnitude of these losses is likely the result of two factors. The first is that first-party fraud is very difficult to detect. Fraudsters behave very similarly to legitimate customers, until the moment they do their Bust-Out, i.e. cleaning out all their accounts and promptly disappearing. A second factor is the exponential nature of the relationship between the number of participants in the fraud ring and the overall monetary value controlled by the operation, a feature often exploited by organised crime. However, while this characteristic makes these schemes potentially very damaging, the good news is, it also renders them particularly susceptible to a powerful and proven approach called ‘graph databases’.

PayPal, a multibillion-dollar eBay company, uses graph techniques to perform sophisticated fraud detection on eBay and StubHub transactions in real time, for example. IDC estimates that this has already saved it more than $700 million and has also enabled the company to perform predictive fraud analysis.

Unlike most other ways of looking at data, graph databases are designed to express relationships in data. That means they can uncover patterns that would be very difficult to detect using traditional data representations, such as tables. As a result, an increasing number of enterprises, from banks to ecommerce giants like PayPal, are using them to solve a variety of connected data problems, and increasingly the speedy detection of potentially fraudulent activity.

Catching fraud rings and stopping them before they cause damage is what we really want. One reason for that still being hard to do, though, is that traditional methods are either not geared to look for the right thing – in this case, rings created by shared identifiers – or can’t really work at Internet-scale.

Standard stats-based tools, like deviations from normal purchasing patterns, use discrete data, not connections. Discrete methods are useful for catching fraudsters acting solo, but fall short when it comes collectives of criminals, who tend to work cross-border, even cross-continent. Furthermore, many such methods are prone to the notorious ‘false positive,’ which creates undesired side-effects in annoyed customers and lost revenue opportunity.

How can we improve this situation? Graph databases really can help. Uncovering fraud rings with traditional relational database technologies requires modelling the graph above as a set of tables and columns, then carrying out a series of complex joins and self-joins. Queries like that work – but are very complex to build, and expensive to run. Scaling them in a way that supports real-time access also poses significant technical challenges, with performance becoming exponentially worse not only as the size of the ring increases, but as the total data set grows.

Graph databases are emerging as the ideal solution to finding out such hidden patterns, and at big scale, too (Forrester Research estimates that one in four enterprises will be using such technology by 2017).

Graph databases are a powerful addition to any Chief Security Officer’s arsenal. And with their widening availability in Open Source versions, there’s no excuse for ignoring their potential any longer.