Using Digital Footprints to Prevent Online Fraud
We’ve all done it. The last second rummage through bags and pockets… Tickets? Check. Wallet? Check. Passport? Passport? WHERE’S MY PASSPORT???
Just as passports are synonymous with travel, they are also equated with identity, along with other official documents like an ID card or a driver’s licence. As well as proving who you are as you travel from one country to another, your passport can also be used to prove your identity when you need to hire a car, book accommodation, and numerous other pre-trip activities.
Today, booking and paying for products and services increasingly happens online. For the first time this year, footfall in stores on Black Friday was down, while online purchases increased significantly. This trend is leading to higher rates of card-not-present (CNP) fraud. Robust identity verification and authentication will protect the bottom line, but can also be detrimental to the customer experience. As consumers, we expect to be able to purchase or access what we want, when we want. If the barriers to making a transaction are too high, we simply find an alternative option. In this digital context, requiring a user to share paper-based identity documents to complete a transaction becomes untenable.
Luckily, advances in technology, most notably machine learning and artificial intelligence, provide us with new, more streamlined and accurate methods for identity verification and CNP fraud prevention. Veridu, a UK-based provider of digital identity and fraud prevention solutions, is one company driving innovation in this sector.
Analysing digital footprints
There is a wealth of information in the trail we leave behind as we use the internet. Our digital footprint contains far more data and insight about who we are, what we do and what we like than our passports. It’s this information that Veridu is harnessing to develop purely digital solutions for identity verification and fraud prevention that address the needs of both companies and their customers.
When Veridu is integrated into a company’s website, users identify and authenticate themselves by providing access to a range of social or online accounts. Veridu gathers and analyses this information, employs artificial intelligence and machine learning to assess the credibility of the information being shared, and then determines a level of confidence that the user’s identity is genuine, or that a transaction is legitimate.
The black box
An air of mystery surrounds machine learning and artificial intelligence, but the key to success is to have a well-trained machine upon which to build.
The starting point is to train the machine using controlled data – data that is known to be real or fake. In Veridu’s case, the data consisted of real and fake social media profiles. As the data is
fed into the machine it learns how to distinguish between the real data and the fake data using a multitude of data points, becoming more and more accurate over time.
In this way, machine learning offers significant benefits over a purely rules-based approach. Once models are trained using controlled data, they continue to learn and evolve as they process user information in a real-life environment. One of the biggest advantages is that machine learning models can process enormous amounts of data, which would be exponentially hard for humans to review and adapt as quickly.
This continual evolvement of Veridu’s models is especially critical given the increasing sophistication of fraudsters, who tirelessly develop new ways to circumvent existing technology. Rather than playing catch up, as is the case with a purely rules-based approach, machine learning models evolve as new fraudulent practices emerge.
Veridu employs a large combination of machine learning models, such as decision trees, random forests, neural networks and boosting. Every model contributes with a different learning bias, so that different techniques used by fraudsters can be more easily picked up.
The world is your oyster
The internet has created significant growth opportunities for companies by making them accessible to a far greater number of potential customers – both at home and abroad. Far too many, however, are missing out, often due to concern about how to manage a fraud risk in new markets.
Traditional fraud prevention services and sources, including AVS (Address Verification Service), 3D Secure, and traditional identity and credit checks don’t offer full, global and demographic coverage. Looking for comparable solutions in each new market a company wants to enter is both time consuming and can lead to a patchwork of solutions that are hard to manage and expensive to maintain.
What is clear, however, is that anyone buying online, no matter what age or where in the world they are, will have a digital footprint. That’s why it is critical for companies to embrace new technology and data sources, like Veridu, now, to get ahead of the competition and take full advantage of new markets before the laggards catch up.