Statistical modeling programs used in conjunction with cooperative databases are getting more sophisticated and better at identifying specific types of customer prospects. And the list models created with this data can be used to factor in customer acquisition costs.
Data rich co-op files allow direct marketers to identify and select specific list segments based on projected response rates, making it possible to select higher performing segments from co-op databases from the get-go, and alternatively to suppress segments from marketing lists likely to produce lower response.
It’s all about “optimizing” the data, said Jerry Schiller, president of Schiller Direct, whether it’s selecting segments for co-op databases for appending demographics or deleting single-purchase buyers from lists being built from models developed with co-op data.
Co-op databases are a rich source of names and probably no more than 5% of names from one co-op are duplicated on another, Schiller said.
Schiller represented the point of view of catalogers during a panel discussion Tuesday with executives from six firms in the co-op database marketplace. The other participants were from Abacus, Experian, I-Behavior, Prefer Network, Nextaction and Wiland Direct, a newcomer in the co-op database arena.
Some specific strategies for “optimizing” data from co-op databases discussed by the panelists include the following.
* Determine which customer segments are likely to yield the lowest response and remove them for prospecting lists. Focus efforts on the higher projected response segments.
* Be especially careful with single-purchase customer data, which obviously doesn’t provide the same level of confidence in predicting future purchasing behavior, compared to multibuyers.
* Look for hidden list segments by closely analyzing data based on specific stock keeping unit (SKU) purchases. Data on products purchased can be used to assemble unique lists, which cannot be found, on the conventional list rental market.
* Match back data to the original source that generated the response to achieve a greater understanding of the customer. A response received by mail might have been triggered by e-mail or vice versa.
* Don’t overlook compiled lists when appending data to lists developed from co-op databases.
Network
