Sign up for The Paypers newsletter Follow The Paypers on LinkedIn Follow The Paypers on Twitter Follow The Paypers on Facebook
The Paypers, paypers, Insight in payments, News, Reports, Events
 advertisement
Thought leader insights

Fraud prevention solution - should we build it in-house or should we buy it?

Tuesday 22 May 2018 | 08:35 AM CET

Sean Neary, Featurespace`s Financial Services Fraud Expert discusses what goes into building a comprehensive fraud prevention solution

As financial institutions continue to try and find better ways to fight fraud in a mature market for fraud and risk management solutions, one question that organizations are asking is: ‘should we build in-house or should we buy’?

I’ve often found organizations treat the fight against fraud as a ‘back-end issue’, but more recently financial services have brought the fight to the front line, with merchants and payment service providers (PSPs) becoming more proactive and looking to invest in holistic fraud prevention systems.

With fraud techniques continuing to evolve, I’ve seen organizations – particularly in financial services - looking for more innovative and flexible solutions to respond to this growing issue and to remain one step ahead of the criminals. Fraud levels continue to rise, so providing a frictionless customer experience while simultaneously trying to lower the cost of fraud has proved a challenging balance.

As businesses across all industries are looking to the increasing benefits of machine learning, the question regularly being asked is whether to invest in existing software from a fraud prevention specialist or attempt to build their own fraud prevention system in-house. This can be a big commitment for a business and I’ve witnessed organizations embarking on the ‘build’ route without having considered the full cost of creating a brand-new, holistic system.

Building a complete risk management solution is like building a house; it’s a large investment that needs a lot of expertise, specialized tools and skills

Organizations want to be able to harness the benefits of machine learning and combine it with their own analytics. However, even with a strong appetite to build a bespoke software solution in-house, I believe it’s highly unlikely that it can be done at a lower cost than implementing fraud prevention software.

Opting to build an in-house solution is not a one-off investment for an organization; the total cost of ownership which includes building and maintenance over a long period of time cannot be underestimated. The continued operational costs that are involved with building in-house aren’t limited to just running the solution, but also for the enhancements needed to increase its capabilities to detect new and different types of fraud. Businesses need a platform that can allow a high volume of transactions while responding in real time when identifying fraud without blocking legitimate transactions for customers, at the same time.

Fighting fraud is costly especially as operational costs continue to rise when trying to balance detection against customer friction

As fraud activity continues to increase both in volume and complexity, it’s hard to estimate what the system needs to keep up with fraudster’s next move.

Having managed multiple fraud systems for many years, I found that most legacy systems, while being a useful tool to stop fraud attacks, have lacked agility and used models that were not adaptive. This can limit the system’s ability to introduce new functionalities and ingest new data to detect fraud, particularly new types of emerging fraud. Additionally, home-built systems often lack a clear and accessible UI, resulting in poor handling of alerts and less efficiency. Organizations need an agile and responsive end-to-end solution they can rely on, that needs little to no manual intervention. One needs to be able to easily introduce new features into a system that can self-learn and react to both new and known threats in real time without disrupting and blocking genuine transactions.

This can be difficult to achieve when attempting to enhance in-house solutions, usually involving a lot of manpower and major platform redeployments. The system must be resilient and fully supported when fraudulent transactions (or legitimate transactions) are flagged. This can be a challenge with systems that degrade over time or are not designed to learn independently.

An end-to-end solution to fight fraud in real time without customer friction does exist

Businesses should consider the adoption of a blended solution with a platform that combines some intuitive UI and powerful analytics (such as machine learning). This is proving to be a valuable approach for the build versus buy dilemma for many companies, as it provides a cost-balanced solution for fighting more fraud while maintaining or even reducing their operational costs, all safe in the knowledge that the system is providing them with the tools they need to detect fraud.

This is something that Featurespace’s ARIC platform is able to achieve. Using its unique Adaptive Behavioral Analytics to identify patterns of normal or ‘good’ behaviour from thousands of data points, ARIC detects new and known fraud attacks as they happen, reducing the costs associated with managing fraud. As the system self-learns continuously, ARIC’s models automatically retrain thanks to the Adaptive Behavioral Analytics, so the system never degrades, negating constant model retraining via the solution provider and ultimately, long-term maintenance costs.

Typically, businesses can see a reduction of 70% in genuine transactions declined and cut down on operational costs by up to 50% with less manual intervention needed by analysts. For a Global Credit Card Issuer, the implementation of ARIC platform delivered an 80% reduction in genuine CNP (card not present) transactions declined with an overall of 40% reduction in fraud losses.

Integrated effortlessly with existing counter-fraud measures, ARIC Fraud Hub ultimately allows businesses to accept more revenue while reducing the number of genuine transactions incorrectly declined which means - you guessed it – happy, loyal customers.

About Sean Neary

Sean joined Featurespace from Barclaycard where he worked in fraud and risk technology for over 10 years. Sean brings his Financial Services industry expertise and insight to his role at Featurespace as a Subject Matter Expert where he ensures that Featurespace’s ARIC platform development matches the risk management needs and requirements of the Financial Services sector.

 

About Featurespace

Featurespace is a leading provider of Adaptive Behavioral Analytics for fraud and risk management. Deployed in over 180 countries, the ARIC platform uses machine learning to spot anomalies in real time and stop new and known fraud attacks for some of the world’s largest financial institutions, gaming organizations and insurance companies.

 advertisement
 advertisement
 advertisement
 advertisement