Challenge
A major religious publisher, which relied on a gift subscription program with five efforts per year, was seeing response rates decline. The company was scoring early efforts to determine which customers were likely to give a gift subscription. They wanted to boost profit by eliminating the names least likely to respond from its fifth and final mail effort.
Solution
Alliant examined the behaviors of the publisher's subscribers, donors and non-responders and tested scoring in multiple ways. Alliant's wealth of micropayment data variables provided strong behavioral predictors, powering a model that allowed the client to suppress unprofitable names and mine new names that would have been unavailable using traditional selection methods.
Results
Alliant’s model improved three in-house models, saving cycle time, direct and indirect costs, and coordination efforts.
By suppressing the four lowest-scoring segments, the restructured model validation netted 24% in profit gains for the test.
The model predicted 17% gains in pay rates by dropping the bottom four groups.
