Target the Ego or Target the Group: Evidence from a Randomized Experiment in Proactive Churn Management

We propose a new strategy for proactive churn management that actively uses social network information to help retain consumers. We collaborate with a major telecommunications provider to design, deploy and analyze the outcomes of a randomized control trial at the household level to evaluate the effectiveness of this strategy. A random subset of likely churners were selected to be called by the firm. We also randomly selected whether their friends would be called. We find that listing likely churners to be called reduced their propensity to churn by 1.9 percentage points from a baseline of 17.2%. When their friends were also listed to be called their likelihood of churn reduced an additional 1.3 percentage points. The NPV of likely churners increased 2.1% with traditional proactive churn management and this statistic becomes 6.4% when their friends were also listed to be called by the firm. We show that in our setting likely churners receive a signal from their friends that reduces churn among the former. We also discuss how this signal may trigger mechanisms akin to both financial comparisons and conformity that may explain our findings.