Universal Bank

PURPOSE

Universal Bank is currently experiencing a rapid growth trend. In an effort to capitalize on this growth and expand the credit card department, data mining has been utilized to identify potential Universal Bank credit card customers. To aide in the study the results of a promotion run last year by the marketing department have been used as “sample” data. This data is useful in identifying what existing liability customer traits may be correlated to their acceptance of a credit card.

CUSTOMER BASE

Existing liability customers devise our sample and make up the customer base. Of the 5000 existing customers in the sample some trends between demographic and existing accounts with the bank have proven useful. The methods of actually establishing a connection between trait and probability of acceptance are discussed below. The means of customer attributes observed were;

AGE-45 YEARS OLD* EXPERIENCE- 20.1 YEARS* INCOME-$73,774* FAMILY SIZE- 2.4* CC AVERAGE- $194 monthly EDUCATION- 1.881/SOME COLLEGE* MORTGAGE PAYMENT/Mo- $1538* ACCEPTED PERSONAL LOAN- 480 SECURITIES ACCOUNT- 522 CD ACCOUNT-302 BANK ONLINE- 2984 CREDIT CARD CUSTOMERS- 1470 (*mean)

RESULTS:

Data modeling was used to establish which of the 12 custmer traits were related to potential credit card customers. The customer traits observed to be most influential in predicting acceptance of a credit card offer were; acceptance of personal loan in the previous campaign, having a securities account, having a CD account, and online banking customers. Targeting customers who share in all of these traits will yield the most acceptances amongst liability customers.

There seems to be potential for cross-selling (CD’s, Securities, Credit Cards) amongst this group of customers. Ultimately if the above referenced traits are present in any combination there is a higher probability that the customer will be interested in other services and products from Universal Bank....