December 5, 2018
The Four Key Principles of Data Sharing for Fraud Prevention
Posted November 29, 2018
Data sharing amongst organizations in order to tackle organized fraud is a fine principle to get behind, I think we can all agree. However, something that is not so universally accepted is how to actually go about doing that.
Shared intelligence can vastly improve your ability to make decisions on accepting or rejecting transactions and other online activity from your consumers, as you have a much more complete picture of what constitutes normal versus suspicious behavior.
As this goes to the very core of how we work with our customers, here we set out the four key principles for embracing a shared data approach to fraud prevention.
1. Real-time Data Sharing
Fraud patterns evolve rapidly, and data from the ThreatMetrix network shows clearly that there are major spikes in the volume of fraud attacks across multiple organizations, directly after major data breaches – but often before said breach has been publicly disclosed. Digital businesses need to act fast if they are to protect themselves from downstream attacks after data breaches.
However, a key challenge of data sharing among consortiums or industry associations is that it can often rely on fairly old school methods of bringing data offline to then share. That is why we need an approach that embraces technology and allows real-time data sharing across multiple organizations – else it is incredibly challenging to keep this relevant.
Therefore, the first key principle when setting up any data sharing activities is the ability to access up-to-the minute data using automation, as opposed to incorporating any manual steps into the process.
Another major inhibitor of data sharing between separate organizations is the big privacy question. In a world where users are creating more and more of a data trail online, regulators across the world are responding with increasingly rigorous data protection legislation to protect individuals’ privacy.
Naturally, different countries, regions and industries will have differing protocols they need to observe, and each business will need to take an approach to data protection and privacy that works for them, however the technology now exists to facilitate global shared intelligence across these varied organizations in a way that is privacy-by-design, in order to improve fraud prevention while protecting privacy.
This centers around tokenization, which facilitates the sharing of one-way hashed data sharing that cannot be reversed – meaning that data from outside organizations can never be seen in the clear. However, this needs to be shared on a common platform that can correlate the hashed data to provide valuable insight into key indicators of fraud plus identity and behavioral attributes.
3. Make it Actionable
With each organization having unique systems and technology stacks, the next challenge comes from making the tokenized, real-time data easily actionable across environments. For larger and more complex organizations, even making data insights actionable across different internal siloes and divisions can be tough enough – let alone for completely independent businesses.
In order to inform decisions in the moment, we want to avoid large data dumps, but provide fraud teams with access to link analysis and correlation that make sense of this data and can be queried at the time of a transaction to provide insight into whether it is legitimate or not.
Through APIs into a SaaS-based platform, we can provide access to shared intelligence that correlates the billions of data points that make up users’ digital identities and their transactional history.
4. Embrace Both Breadth and Depth
As discussed in the previous blog post, “Data Sharing for Fraud Prevention: How Do We Get Ahead of the Fraudsters?”, if we only share data that is pertaining to known instances of fraud, we are already playing catch up. It is only by correlating attributes from trusted as well as fraudulent behavior that we can most effectively spot the sometimes very subtle anomalies in behavior that indicate fraud or warrant further investigation.
The broader the datasets, the more complete a picture you get, and you can better protect your customers from fraud. Increasingly, companies are looking to combine insights from across a user’s online and offline world in order to accurately protect against identity abuse. Additionally, in today’s international economy and society, with individuals moving and transacting seamlessly across borders, global insights are increasingly valuable.
However, as valuable as broad, global insights are, the organizations who are most likely to be seeing similar attack patterns are those within the same region and industry. Therefore, consortium-style data sharing set-ups can provide invaluable opportunities to put aside commercial competitiveness and work together effectively against organized cybercrime.
The trick here is to be able to balance broader insights with targeted ones which are most likely to carry a lot of weight in assessing identity and spotting fraudulent behavior.
Trust in Global Intelligence
As soon as you start digging into the complexities of data sharing in order to prevent fraud, it becomes clear why there is a lot of talk about how it is a great principle – but not necessarily an accepted approach on how best to pull this off.
The approach we have taken here at ThreatMetrix, is to build a one-of-a-kind Digital Identity Network, which is a global repository of anonymized data crowdsourced from transactions coming from around 165,000 websites and mobile apps globally. This is also combined with data from our parent company, LexisNexis Risk Solutions which correlates complex identity attributes across government records, utilities and other activities relating to that identity and from the credit bureaus.
Companies on the Digital Identity Network are able to benefit from intelligence from global companies, as well as distinguish intel from those in the same industry and region as them. This technology allows for actionable intelligence to be shared across completely separate organizations in real time – while protecting privacy.
In the face of an ever-increasingly hostile threat landscape, I hope we can all continue to push for more and more collaboration across organizations to fight the good fight against cybercrime.
To find out more about the ThreatMetrix Digital Identity Network go to: https://www.threatmetrix.com/digital-identity-network/