A Better Supermodel Than Kate Moss: Anglophile catalog mines for top candidates

Eximious, the Anglo-American gift cataloger, had a problem recently when it embarked on a vigorous acquisition strategy.

There are few American lists appropriate for Anglo-centric catalogs, and those that exist have been over-prospected, says CEO Jeffrey Parnell.

The solution? Northfield, IL-based Eximious, which aggressively tests modeling techniques, became an early adopter of The Polk Co.’s Supermodel, a multistaged method officially launched during last month’s Direct Marketing Association conference in Toronto.

Using Supermodel, the Eximious house file was bumped up against two Polk databases: Response Selector, a cooperative database containing names of both catalog buyers and magazine subscribers; and Lifestyle Selector, the registration survey database.

(Since Eximious’ customers are primarily women in their upper 50s and 60s who are even less Internet-savvy than their contemporaries, Polk’s High-Tech Connect file was one of the first to be ruled out as a possible source of names.)

Polk ran a predictive analysis model against the databases to find prospects that mirrored Eximious customers. It then mailed to those names.

Parnell is encouraged enough by early results to continue mailing to the Polk files.

Eximious has also been able to identify non-responders in the Polk listings. Although it will not happen soon, the firm will use Supermodel to profile them and screen out similar names in the future.

The company has 15 years’ worth of customer data, but recent changes in its merchandise mix have rendered most of that data less helpful. The firm has decreased its emphasis on jewelry, favoring items that drive profitability rather than response.

More Names – and Better Ones

Although Supermodel yields research, demographic and psychographic information, its main benefit has been that it opens up a universe of names, Parnell says. And these names are more valuable than those on single-source response lists, he adds.

Names obtained through modeling now comprise roughly 30% of Eximious’ total mail volume, compared with 20% only a few years ago. And the company’s mailing level has gone up.

The increased reliance on data compilers such as Polk is not surprising, given that obtaining new names through responder lists has proven increasingly difficult. For every 10% hike in net names, according to Parnell, Eximious needs to look at between 15% and 20% more names overall.