Nautilus Systems, Inc. logo and menu bar Site Index Home
News Feedback
Button Bar Menu- Choices also at bottom of page About Nautilus Services Partners Case Studies Contact Us

Case Study: Identification of High-Potential VISA Cardholders

For a major financial institution, Nautilus Systems conducted a data mining analysis, with the objective of identifying potential credit card applicants based on desired buying habits. Many more potential cardholders than anticipated actually applied for the credit card, demonstrating the effectiveness of this approach.


The client approached Nautilus Systems at the VLDB (Very Large Database) conference in San Fransico, in response to Dr. Louie's presentation of Nautilus Systems' data mining methodology at the conference. The product requested was the generation of a list of high potential new VISA credit card subscribers.

The client described the buying habits and buying trends desired in new subscribers. The optimal candidates were individuals likely to be repeated purchasers of ‘best of class' leisure time items, demonstrating not only available disposable income but also the likelihood of continuation of desirable purchasing trends. Given these metrics, Nautilus Systems identified and acquired commercially available demographic and financial data. Specific product and service categories were analyzed and assigned weights based upon their desirability. For example, certain types of titanium mountain bikes, or membership in certain types of golf club would be considered to have the appropriate status-conscious image, and would be assigned a greater weight.

Nautilus Systems used its proprietary data mining techniques to extract transaction data matching these product and service categories from commercial database sources, and by examining buying trends contained within mercantile databases of credit card purchases. The requested metrics were programmed into an on-line analytical processing (OLAP) program. A transaction database was built from the data sources, and its contents were run against the OLAP program.

Individuals fitting the desired profiles with regard to their buying habits and purchasing trends were extracted. The developed list was then run against the client's current database of cardholders to eliminate mailings to them, and the income and employment of the candidates were verified. The mined data were further analyzed using statistical methods, and the geographic distribution of the resulting information was plotted. Nautilus Systems delivered the final analysis as a geographic information system (GIS) presentation, mapping the distribution of the potential cardholders. Finally, the addresses of the remaining candidates were verified and rerun again the metrics before the final list was generated.


Significant "lift" was generated as many more of the over 250,000 potential cardholders on the list actually applied for the credit card than anticipated. By having a higher match to the specified market target, the company reduced its mailing and solicitation costs, and achieved increased card usage through the acquisition of these highly motivated card users.

[ Home | News | About Nautilus | Case Studies | Partners | Contact Nautilus ]

logo Copyright © 1996, 1997, 1998 Nautilus Systems, Inc. All Rights Reserved.