Opinion leadership in small groups

The role of opinion leaders in the diffusion of innovation has recently come under scrutiny: On the one hand, their central role in accelerating diffusion has been recognized in industry, academia, and the popular media. On the other hand, it has been argued that opinion leaders do not create contagion processes that differ significantly from those of other types of customers. We offer here a synthesis of these opposing theses: For many applications, opinion leadership should be studied in small groups rather than in an entire network. Using data from two mobile operators, we show how our method of segmenting the market and defining local opinion leaders applies to churn. We show that opinion leaders are highly effective in small, strong-tie groups, and their effect considerably declines with larger groups and weaker ties. Opinion leaders are at the highest risk for churn compared with other group members, they are more likely to be the first to initiate churn, and when they churn, the probability of additional churners from their social group grows significantly, as compared to when non-leaders churn.