Success Stories

LARGE CREDIT CARD COMPANY

• Situation: A closed network of merchants and customers does not guarantee zero fraud. Internal fraud systems take a minimum of 6 hours to detect potential fraudulent transactions. In many cases the fraudulent transactions are never detected until the card holder questions transactions on his/her credit card statement. It was estimated that a >3% fraud existed. Each 1 basis point (.10%) accounted for $10MM in revenue lost.
• Solution: Analyze large amounts of data in a near real-time basis. Developed an ETL system capable of handling large amounts of data i.e. < 1TB and built a large data warehouse for storing customer profiled data. Developed integration mechanisms for data acquired from the credit reporting agencies for an ongoing detailed analysis of the customer and merchant. The system analyzes approx. 40TB data monthly from various sources.
• Conclusion: Fraud detection is now automated and can reveal potential risks in less than 30 minutes. This effort reduced credit card fraud to less than 1% in 3 years accounting for a net savings of approx. $300MM annually. The additional benefits included an overhaul and improvement in the customer loyalty program because of a better understanding of the customer and his buying trends and his actual monthly spending capability.

LARGE WIRELESS CARRIER

• Situation: This carrier was experiencing customers leaving because of dropped calls. The answer to the dropped calls and adverse call related information were in the data from 9000+ cell towers. Acquiring this mass amount of data required considerable computing power. The existing system was developed in Perl and required constant monitoring and bug fixes. Additionally, network engineers are unable to monitor wireless
system in real-time.
• Solution: We developed a comprehensive ETL solution to extracting and loading data in near real-time. The new ETL system utilizes less hardware, is fully scalable, and is designed to process approx. 1TB/day. It can be scaled to 10TB/day without any source code modifications. The solution handles 21 different record types for Lucent, Nortel, and Motorola wireless switches.
• Conclusion: The new solution allows the engineers to monitor and predict outages within the wireless network in near real-time. This enables engineers to be proactive, and cell and network outages are rare. Thus a significant increase higher customers satisfaction.

TOP FINANCIAL INSTITUTION

• Situation: Bank Acquisitions can be an internal operational nightmare. The company had over 70+ loan application systems and a few underwriting (approval) systems. The challenge was integrating these 70+ loan application and approval systems into a single application system to speed up mortgage loan approvals.
• Solution: Developed a 3 phased approach. First, developed a system for integrating all 70+ loan systems. Secondly, developed a new loan application and approval system. Third, integrate all existing loan programs to the new system. Developed a comprehensive ETL and EDW solution to consolidate all loans programs and systems, and allows for seamless integration of new loan systems i.e. new acquisitions.
• Conclusion: The new solution offers the bank the ability to approve mortgages in 1/10th of the time. Additional benefits allow for offering new loan programs to customers and monitors loan activity in one system.