The adoption of new behaviors or new products is often influenced by those closest to us, but customer churn is influenced by social factors as well. An analysis of 1 million customers of a cellular company showed that customers were 79.7 % more likely to defect for each time one of their social neighbors defected. The more a customer communicated with a neighbor and the more characteristics they shared, such as age, gender or status, the more likely the customer was to follow the neighbor in canceling the service. The effect of a neighbor’s defection on a focal customer’s hazard of defection was strongest within the first month and decreased over time. The study shows that companies should take customers’ social networks into account when attempting to predict and manage customer churn. Network-related information can substantially improve analysis of new product adoption, and the same seems true for the field of customer defection.