Want more Insights from SAS? Subscribe to our Insights newsletter. Or check back often to get more insights on the topics you care about, including analytics, big data, data management, marketing, and risk & fraud.
Transaction monitoring: To segment or not?
One size never fits all, but a dynamic segmentation strategy does
By Carl W. Suplee, Senior Solutions Specialist, SAS Security Intelligence Practice
A typical anti-money laundering (AML) transaction monitoring program has scenarios that monitor the customers and accounts that pose the most risk to the institution. The fact is … this one-size-fits-all methodology isn’t very effective. That’s because customers transact differently based on many factors. So how do you incorporate that into your program?
A best practice is to analyze customer segments and apply more focused monitoring. This can reduce false positive alerts from being triggered based on the scenario thresholds. Firms are optimizing by identifying methods to segment their customers with the ultimate goal of developing segments that allow for effective risk-based threshold setting. To do this firms must do two things:
- Combine groups of customer’s accounts deemed to be of higher AML risk and monitor those using more conservative scenario threshold values.
- Once the high risk and special interest groups have been isolated, the groups are then separated into similar segments.
SAS’ approach is simple: create a fit-to-firm segmentation model that leverages both the business and data-driven attributes to identify the most similar segments of customers. These segments will be dynamic and evaluate both who/where/when of the customer and include the what (transactional history) to best profile them for segmentation. Then the segments will be used for AML transaction monitoring with unique thresholds for each segment by each scenario as mandated by OCC 2011-12: Supervisory Guidance on Model Risk Management.
An effective AML transaction monitoring strategy includes segmenting the customer base by analyzing customer activiey adn risk characteristics in order to monitor them more effectively.
The largest challenge with this approach is that it will increase the need for analytic business resources for model validation and tuning exercises. But the investment in a segmentation strategy pays dividends in more effective coverage across the book of business, a reduction of false positives and better alignment of investigative resources to the firm’s AML risk.
SAS has recently completed several client engagements that have illustrated at least 2.5 times lift in effectiveness and conversion rates as part of our segmentation and tuning methodology.
Our work with clients is uncovering some interesting best practices, which we’re using to enhance our monitoring and visualization applications. We’re also sharing that information with you and others in the AML compliance community, so please keep a look out for more blogs introducing these concepts in 2015.
An effective AML transaction monitoring strategy includes segmenting the customer base by analyzing customer activity and risk characteristics in order to monitor them more effectively. Read how you can blend both quantitative and qualitative methods to tune scenarios to identify the activity that poses the most risk to the bank.
Read how you can blend both quantitative and qualitative methods to tune scenarios to identify the activity that poses the most risk to the bank.