A large US publisher wanted to reduce customer churn and maximize the return on their investment in retention campaigns.

Mather Economics estimated churn risk by customer and grouped them into three segments.  Three retention campaigns were tested:  a charger costing $18, a gift card costing $12, and a greeting card costing $1.  Representative test groups were selected in each churn-risk segment, and customer behavior was tracked and reported by week.

The campaigns succeeded in reducing churn in each of the churn-risk segments; however, only the high churn group (96-100th percentile) experienced significant churn reduction from all three incentives at 120 days post application.  The high-churn customers responded most favorably to the gift and greeting cards, while the charger was effective across all churn segments but generally was not worth the added expense.  The overall churn levels were consistent with the predictions from our churn-risk models, and the insights from the test groups identified the most cost-effective means of lowering churn by group. The average reduction in churn across all incentives and risk groups was 10%.  Testing showed that significant and lasting retention gains can be realized with low cost incentives.