Score former wine club members for propensity to be reacquired
Mather combined data on winery club membership history and purchase behavior to create a comprehensive data set on member activity. An econometric model was specified to estimate the relationship between member attributes on the probability to re-subscribe to the wine club. Important predictive variables included time since last membership period, lifetime value, state of residence, previous newsletter subscriptions, and previous allotment bottle count.
After applying the reacquisition model to the wine club’s current list of former members, a test was established to determine the efficacy of various messaging platforms (direct mail postcard and email) and evaluate overall response rates. Results from the test showed that, while restart rates climbed along with estimated reacquisition score, a significant jump in restart rates was associated with formers that scored 90-100, indicating the need to prioritize efforts around this group in the future. Results showed that realized restart probability grew with predicted reacquisition score. In particular, former members that were placed in the highest 10% of predicted score were eight (8) times more likely to be reacquired versus those in the lowest quartile. Test results indicate that significant cost savings could be achieved by focusing reacquisition efforts on those with the highest predicted scores.