Diffusion of innovations

Diffusion of innovations is the process by which new ideas, practices, and technologies spread through a population or network over time. It is not a mechanical process of information transmission; it is a social process shaped by influence, status, imitation, and network structure. The same invention can diffuse rapidly in one context and fail entirely in another, not because of the invention's inherent quality but because of the topology of the social network through which it travels.

The canonical model, developed by Everett Rogers, describes diffusion as an S-curve: slow initial adoption by innovators, followed by acceleration as early adopters influence the early majority, then deceleration as the late majority and laggards are reached. But this curve is a statistical regularity, not a causal mechanism. The mechanism is social contagion: the probability of adoption increases with the fraction of one's network that has already adopted, creating a positive feedback loop that can produce sudden phase transitions in adoption rates.

The network structure matters critically. In dense, homogeneous networks, diffusion is rapid but fragile: once a critical threshold is reached, adoption spreads almost instantaneously, but the same network can also transmit rejection just as quickly. In sparse, heterogeneous networks, diffusion is slower but more robust: bridges between disconnected communities can carry innovations across structural holes that would otherwise block them. The structural hole theory of diffusion suggests that the most powerful agents are not the most connected but the most strategically connected — those who bridge otherwise disconnected communities.