Fund Companies Try Out Data Mining
June 5, 2000
Five large mutual fund companies are about to begin using data mining software that follows shareholder behavior to pick up shifts that might signal customer dissatisfaction and possible redemptions, according to Patrick Baldasare, president of @Risk of Berwyn, Pa., the maker of the software. @Risk's software also attempts to determine customers' tastes, so that fund companies can cross-sell targeted funds and services to shareholders, Baldasare said.
Baldasare declined to name the fund companies that have signed up for his data mining service, which they will begin using by late summer. However, he did name the mutual fund affiliates of two other @Risk clients that are considering signing up for the new technology. Mellon Bank of Pittsburgh and Rittenhouse Financial Services of Radnor, Pa. are @Risk clients, and Mellon's subsidiary, Dreyfus of New York, and Rittenhouse's parent company, John Nuveen Investments of Chicago, are considering using @Risk's data mining software, Baldasare said.
Dreyfus and Nuveen are first waiting to see how successful the @Risk system works at their affiliates and are being cautious because data mining, while prevalent among packaged-goods companies, is still considered an exotic technology among financial service companies, Baldasare said.
But now that the mutual fund industry is maturing, fund companies will pay more attention to retaining the customers they already have, Baldasare said.
@Risk uses a relational database to track as many as 100 variables about a shareholder, Baldasare said. The system looks at a shareholder's holdings, how often and by what means he contacts the fund company and whether he makes additional purchases, exchanges or redemptions, Baldasare said. @Risk then adds the individual's age, income, marital status and other demographic information to the profile, Baldasare said.
@Risk looks for how often a shareholder checks on his fund balance or net asset value, and by what mechanism - whether on his own initiative over the phone or the Internet, or passively through regular mailings, Baldasare said. @Risk also takes into consideration the distribution channel through which a customer has come to a fund complex, Baldasare said.
Combined with other data that a fund company may be able to obtain about a person from outside sources - such as whether he has a mortgage, a car or revolving credit - @Risk can produce a complex profile of a shareholder, Baldasare said.
The @Risk system runs this data through a proprietary network to monitor customer activity for abrupt changes in behavior, as well as to determine cross-selling opportunities, Baldasare said.
The profile of valuable customers might indicate they "drive a luxury car, are 29 years old, have current assets of $50,000, are married and interact with the fund company by way of the web," Baldasare said. "But if you removed one of those variables, all of a sudden, that individual may become a likely candidate to redeem shares."
William Crager, managing director of Rittenhouse Financial Services, said he signed up for @Risk to find out if and when one of Rittenhouse's high-net-worth, separately-managed-account clients might defect. While Rittenhouse's attrition rate runs only seven percent to nine percent a year, over five years, the loss could amount to 35 percent to 45 percent of business, Crager said.
If @Risk enables Rittenhouse to improve its retention rate by even two percent a year, that can translate to 10 percent worth of business saved over fives years, which is substantial and "can have a profound effect on the bottom line," Crager said.