Continue in 2 seconds

GainSmarts Produces Fast, Accurate Data Mining and Profiling for Coverdell & Company

Published
  • July 01 1999, 1:00am EDT

BACKGROUND: Atlanta, Georgia-based Coverdell & Company is one of the leading developers and direct marketers of insurance and fee producing membership products for the Fortune 100 retail banks in the United States. Coverdell & Company is a wholly owned subsidiary of MemberWorks which provides major financial institutions with innovative, fee- producing membership products. Coverdell's target marketing team provides modeling and database services associated with the direct marketing of those products to their bank clients' customer base.

PLATFORMS: GainSmarts is Urban Science's proprietary data mining tool that sits on top of the SAS Base System. Supported platforms include Windows NT, Sun Solaris, Compaq True64, IBM AIX and HP-UX. GainSmarts imports SAS data sets and views, ODBC databases, flat files, delimited files as well as ORACLE, Informix and SQL Server data.

PROBLEM SOLVED: The demands for analytic services placed on Coverdell's target marketing team were recently accelerated when the company was purchased by MemberWorks. We assumed the task of providing modeling services for the new parent company's marketing efforts, along with supporting Coverdell's unprecedented growth. What was needed was a data mining and profiling tool that would work fast, produce accurate results and could be rapidly deployed on a robust, scalable analytic platform against a myriad of source data formats. GainSmarts addressed all of these needs.

PRODUCT FUNCTIONALITY: GainSmarts merged right into Coverdell's existing analytic system, accessing relational and nonrelational source data. Our staff began building models with GainSmarts the same day it was installed. GainSmarts also allows us to build models on a sample and export the scoring code for use with our clients' full databases. The open nature of the system allows us to see which decisions were made and then allows us to manipulate the formulas and variables to test other outcomes. The speed of GainSmarts means we can try different parameters and manipulations and still save time over other data mining tools.

STRENGTHS: GainSmarts is a multifunctional tool that works on scalable platforms. It is a good balance between "easy to use" and being totally flexible. GainSmarts allows seasoned modelers to use their expertise and experience but still makes less-seasoned modelers productive quickly. It has the most effective variable transformation and selection capabilities on the market. This takes the guesswork out of the variable permutations and the discovery of unique interactions. GainSmarts automates this most time-consuming part of the modeling process. The GainSmarts logistic model has defeated every other tool or model we tested against it.

WEAKNESSES: The only weakness we see with GainSmarts is that it does not provide visualization of the modeling process of a given run. GainSmarts cannot create in one project a multiple model run and cannot merge results from multiple runs for comparison.

SELECTION CRITERIA: GainSmarts allowed Coverdell to run more models in a shorter amount of time and produced higher lift and better models than any other tool on the market that we tested. In addition, GainSmarts can be used by expert statisticians and by less-experienced data miners. The GainSmarts' interface is engineered to be used as an expert tool for the seasoned data mining statistician or can be used to build and apply models with less intervention for more junior users. Most importantly, GainSmarts is a totally flexible and transparent open system allowing the user to understand and intervene in every step of the modeling process. We achieved the highest lift with GainSmarts' regression and decision tree models that have an added advantage of being easy to explain to our clients and management. In short, GainSmarts addressed the needs of Coverdell and MemberWorks for speed, accuracy and transparency with no compromise in any area.

VENDOR SUPPORT: The support of the Urban Science staff has been one of the best parts of acquiring GainSmarts. Their knowledge of the tool, my system and how to maximize the performance of GainSmarts on my system was impressive.

DOCUMENTATION: The GainSmarts documentation is adequate for end-user application and includes 600 pages of on-line, context- sensitive help that provides users with assistance at any step of the GainSmarts process. We think that GainSmarts could gain from additional system administration documentation.

Register or login for access to this item and much more

All Information Management content is archived after seven days.

Community members receive:
  • All recent and archived articles
  • Conference offers and updates
  • A full menu of enewsletter options
  • Web seminars, white papers, ebooks

Don't have an account? Register for Free Unlimited Access