We examined the benefits of different search strategies by testing four computational models. In one model, agents in a group always innovated. The other three models incorporated some mechanisms of imitation. In the second model, each agent imitated the best solution of a random other. In the third model, each agent followed preferential attachment and imitated the best solution of the agent that was asked by many agents. In the fourth model, each agent developed a familiarity with an agent based on how often it asked a certain agent, and imitated this agent. In two simulation studies, following the most popular or the most familiar agent resulted in a good compromise between efficiency and diversity in finding good solutions. Peoples desire to follow particular individuals may be a key to their adaptive behavior, allowing them to disseminate ideas efficiently while encouraging the exploration of new ideas.