Busting Fraud: Simple strategies for quick results

44310903 - high angle view of hacker stealing information from computers at desk

By Steve Platt, Executive VP – Fraud & Identity, Analytics and Decision Software, Experian

Fraud is fast. It’s a race without the risk that the fraudster has to ever strap on their running shoes, but they still treat it like an Olympic competition – and so should businesses. And like a top flight event, the risks of falling behind are severe, because it’s a winner-takes-all situation. The fraud/anti-fraud race can actually enhance more than financial security. Fraud solutions should be working with other departments’ shared data needs, with the net result that the business runs much smarter. How? Fraud mitigation is the friend with unexpected benefits…

Minimise customer disruption with right-sized fraud solutions
Customers do not always have all the information they might need to hand, or even in mind at all. It’s a frequent occurrence to be in the middle of completing a transaction online when the process comes grinding to a halt because of a request for ID or data. The ratio of disrupted but legitimate customer traffic to actual fraud is 30:1. This happens because many businesses use a one-size-fits-all approach to fraud detection, creating more customer friction than they out to.

It’s a simple task to choose right-sized fraud solutions that reflect the value and level of confidence needed for each transaction. That provides customers with a more seamless experience without sacrificing protection. And the business reduces the ratio of disrupted traffic. Right-sizing the fraud solution means aligning true fraud rates with the commercial strategy. It’s about catching more fraud without disrupting regular customers.

Modern fraud mitigation and marketing rely on a universal view of the consumer
The 30:1 ratio shows that achieving protection versus affecting the customer experience is a balancing act. The gold standard for identifying legitimate customers is a multi-layered approach to authentication. Of course, this comes with different challenges. Relying on a traditional 360 degree view of a consumer was once a sufficient way to positively identify a customer, but it’s no longer enough.

What the fraud-affected business needs is a universal profile of consumer behaviour. To get it requires access to a combination of identity data, device intelligence, online behaviour, biometrics, and historical transactions. This should also include consumer interactions with other businesses and industries as well. Companies that use this knowledge to identify consumers can distinguish a fraudster from a real customer more easily, and build trust with the right parties.

There should be only one… A blended ecosystem expands the view
Relying on your own first-party data sources maintains the status quo. Real progress can only happen when the business works beyond the corporate borders – and even the industry. Fraudsters have access to more data than ever before, including the data used to verify identities, and they use it to create an entire digital profile.

What this means to a business is that fast access to the digital interaction data needed is not helped by traditional data management method, that is, corporate departmental silos. Achieving a view of the universal consumer requires multiple data sources working together. This is only possible when participating in a blended ecosystem. When a business has a single customer view the customer experience is enhanced. This supports business growth without sacrificing protection.

Traditionally, internal teams in a business worked independently, often using different fraud solutions and risk mitigation philosophies dependant on their particular needs. But by putting the customer at the heart of the business, creating a holistic versus a siloed approach, means fraud is detected earlier (at account opening), reducing vulnerabilities and financial loss later (at the point of transaction). It is worth anyone facing challenges from fraud to push for greater collaboration across the corporate teams and processes that are susceptible to fraud, such as account opening, access and transaction.

On the pulse: Using service-based models to achieve agility and scale
Subscription-based systems are becoming more popular as companies seek more responsive ways to keep up with the speed of fraud. It is gradually encroaching on custom in-house/ on-premise solutions. Because fraud adapts quickly it can hurt both the businesses and the customer to be behind the curve. As with other online services, the subscription model offers continuous upgrades and access to new risk logic. In the same way that cloud computing services offer greater agility and faster responses by sales teams, so to can the fraud-busters orient to emerging threats, no matter how fast the volume grows, or what products, channels or geographies are pursued.

Fraud solution choices future-proofed: Machine learning techniques
Businesses use a variety of technologies and information sources to fill in knowledge gaps where fraudsters hide. The ability to modify strategies quickly and catch fraud fast while improving customer experience is a critical aspect of fraud prevention. Whilst it is still an emerging trend, machine learning is a powerful predictor of fraud as part of a universal customer view.

Machine learning helps by highlighting attributes or relationships that are indicators of fraud, thus moving a business’ posture from reactive to predictive. Whilst the technology can analyse more data than a human, it still has limitations. It can take a long time to react and prevent fraud, and it can fail to generate effective patterns or consistent profiles. In training, it can produce a lot of false positives too.

It can be improved when paired with unsupervised machine learning techniques that look for uncharacteristic items: AKA anomaly detection. These models can be a strong complement to supervised learning approaches because they go at the same problem from entirely different angles and exploit orthogonal information. The resultant analytics engine can recognise previous patterns of confirmed fraud and raise an alert if a pattern changes.

Fraudsters are relentless
They are constantly evolving and circumventing fraud detection systems with state of the art technologies. Every business owes it to themselves to stay current not only to mitigate fraud losses, but because they can also enrich customer relationships and drive growth at the same time. With the right business frame of mind the results can be quick both in terms of increased customer satisfaction AND busting fraud. And, Ray Parker Jr. would say, together, those feel good!