Connected data analysis key to uncovering fraud, says exec

by Francis Monfort03 May 2018

Mortgage companies run the risk of missing signs of fraud with outdated technology. Clark Richey, chief technology officer of data analytics company FactGem, tells Paydayloans247 how a business’ data-management system affects fraud risk. He also talks about why gathering data from as many sources as possible is no longer enough and why mortgage companies should partner with companies in the intelligence and defense sectors.

Paydayloans247: What is the relation between poor data-management systems and fraud?
Clark Richey: When an organization has problems with data management, they're often unable to completely trust their own data. When that's the case, it becomes very difficult to determine if some suspicious data is due to fraudulent activity or existing data-management problems. Data relationships are disconnected and key fraud indicators get missed.
 
Human nature compounds this problem. When we know the systems we're managing are producing unreliable information, we don’t look too hard at those systems because we know we aren’t going to like what we see. The closer you look at a badly configured data system, the more it starts to fall apart. Criminals will absolutely take advantage of this to commit fraud.
 
Paydayloans247: What can mortgage brokers, lenders, and banks do to reveal hidden fraud indicators?
CR: Businesses have to adapt quickly in order to combat new levels of technical sophistication. Uncovering fraud requires that businesses look at more data sources than ever before and, even more importantly, that they have the ability to connect all of that information. It is no longer sufficient to gather data from multiple sources and analyze them separately. The data has to be connected and those connections have to be scrutinized.
 
For example, the relationships between applicants and their email addresses and phone numbers can be very revealing. Fraudulent activity, such as a mortgage or credit card application, is often associated with disposable (burner) phones and one-time email addresses. Unlike phone numbers and email addresses that are used in normal applications, these numbers and addresses are typically not connected to any other people, applications, or accounts. This is in stark contrast to normal behavior, where we see the same phone number and email associated with multiple accounts and often even different people sharing a home phone number or family email address.

These relationships are often overlooked because traditional technologies make it very difficult to perform these types of analytics at the speed of the business. Fortunately for businesses, there are now technologies that make this possible.

Paydayloans247: Anything else mortgage professionals should know in regard to fraud and data?
CR: Many companies use Excel or other third-party tools to perform critical analysis. These tools pull data from multiple authoritative data sources and report analytical results. However, these results usually don’t get put back into a master data-storage system. As a result, you have lots of important knowledge on fraud and other key indicators floating around the company but not centrally managed and not available for future analysis.
 
Anytime data is disconnected and difficult to obtain, it creates gaps in corporate knowledge that fraudulent activity can slip through. In an ideal situation, mortgage companies would have a single source for connected data that can be used by these analytic tools. This connected data source would also store the results of the analysis so that it can be connected to the original data, managed, and made available for further analysis.
 
Finally, many of the techniques used by government agencies in the intelligence and defense sectors to identify persons of interest are also those that are needed by mortgage institutions to detect fraudulent activity. Mortgage companies should look to partner with companies and individuals with experience in those approaches.

 

Related stories:
 
https://renault.niko.ua

https://best-mining.com.ua