The Network Dynamics of Social and Technological Conventions.

When innovations compete for adoption, chance historical events can allow an inferior strategy to spread at the expense of superior alternatives. However, advantage is not always due to chance, and networks have emerged as an important determinant of organizational behavior. To understand what factors can impact the likelihood that the best alternative will be adopted, this paper asks: how does network structure shape the emergence of social and technological conventions? Prior research has found that highly influential people, or nodes, can be beneficial from the perspective of a single innovation because promotion by central nodes can increase the speed of adoption. In contrast, when considering the competition of multiple strategies, the presence of central nodes may pose a risk, and the resulting centralized networks are not guaranteed to favor the optimal strategy. This paper uses agent-based simulation to investigate the effect of network structure on a standard model of convention formation, finding that network centralization increases the speed of convention formation but also decreases the likelihood that the best strategy will become widely adopted. Surprisingly, this finding does not indicate a speed/optimality trade-off: dense networks are both fast and optimal.