A £2m insurance fraud ring was partly busted by police using social media – and for Emil Eifrem of Neo Technology, that’s something we need to see more of

Welsh Police have done an admirable job in breaking a massive car insurance fraud ring, where a ‘crash for cash’ swindle made the gang up to £2m. Run from a garage, the bogus insurance and compensation operation involved at least 81 people, making it one of the largest insurance frauds in British legal history.

Regular payments were made to the network of conspirators for sums of anything from a few thousand to £40,000. And we only know this by chance – as the gang centred on a criminal family called the Yandells was only caught following a highly complex and lengthy investigation that started due to complete luck.

The Police did a sterling job of detection once they knew something was afoot, even shadowing gang members on Facebook and social media to establish connections. Great policing work – but only after the case had been initiated; law enforcement simply wouldn’t have been looking otherwise.

It transpires that a policeman on a separate inquiry spotted a piece of paper with its handwritten instruction, ‘What to say to the insurance company.’ But we should also be concerned about all the criminality we don’t know about that is taking place.

A complex social web of deception

We can’t leave the fight against this expensive and extensive criminality to chance any more. Deceitful and bogus road accident claims cost insurance companies hundreds of thousands of pounds and push up insurance premiums as a result.

Ideally insurers need to have some way to be alerted to suspicious patterns of behaviour in large data sets, i.e. the incriminating relationships with shared characteristics that typify many such fraud rings.

The problem is that this incriminating data is often stored in unstructured formats in non-traditional IT environments, like social media. It’s not an easy task to access, as the team who carried out this highly commendable investigation can testify. The connections involved in fraud rings are large, and this complexity makes it difficult to spot a sophisticated operation from a few isolated incidents. Hence the importance of drawing on these unstructured sources to supply connections.

In the case of the Yandells, examining public information from photographs, posts, likes, comments and dates to establish links between the suspects and build a robust case, the Police were able to find pictures on people’s accounts who weren’t obviously connected in the bogus accident reports, but were on Facebook. For instance a woman in one vehicle claimed to have a collision with three other people; they claimed not to know each other, but by researching Facebook and finding pictures of one of the Yandell’s son’s wedding, there were photos of all the crash participants in attendance.

Take a leaf out of the Google Playbook

The good news is that the means exists to probe large networks of relationships and establish suspicious patterns of behavior, and in particular mine social media, through graph database technology, developed to work with big datasets by manipulating the patterns in it. Graph came out of work originally done by the consumer Web pioneers such as Google to perform relationship-based searches or social media firms like LinkedIn to map social networks.

Now, it’s started to move into the mainstream and become an affordable option for an increasing number of enterprises. Forrester Research estimates that more than one in four enterprises will be using such technology by 2017, for instance[1].

Insurance firms need to take action

The problem is that the insurance world is playing catch up. That’s because insurance firms are largely relying on standard relational databases to spot fraud. The problem is that such technologies work well for discrete data but as they were not really built to handle networked data, and to handle the Internet and social data discussed, they struggle with the world of the Yandells and their criminal connections.

That means they aren’t well equipped to manage multiple connections between apparently innocent records. To uncover a cash for crash like the Yandells ring requires a number of tables in a complex schema. As these operations can be both costly and resource hungry especially for large data this analysis is often ignored with conventional relational databases. No wonder Gartner says that, “[We] don’t consider traditional technology adequate to keep up with criminal trends.”^[2].

By contrast, discovering fraud rings with a graph database is far easier. Experts believe graphs may be the only way to work with the complexity of Internet and social data to do just this. So it makes sense for this approach to be used in an insurance company’s standard checks at key points in a claim’s trail, such as when it is filed, and to flag up suspected fraud rings in real time.

How many fraud rings are there out there?

The unacceptable reality is that schemers like the Yandells working to stage fake accidents and claim soft tissue injuries based on endless ‘paper collisions’ involving fake drivers, passengers, pedestrians and witnesses may be everywhere.

That means they are a serious menace to the entire global insurance industry, as well as for the public who incur higher insurance premiums as a result.

Don’t you think it’s time to fight back?

The author is co-founder and CEO of Neo Technology, the company behind Neo4j, world’s leading graph database (http://neo4j.com/)

 [1] https://www.forrester.com/Market+Overview+Graph+Databases/fulltext/-/E-res121473

[2] Gartner Market Guide for Online Fraud Detection: April 27 2015