On the one hand, random links facilitate rapid
communication of ideas within the network. On the other, when redundancy is
needed for individuals to adopt a new idea, the presence of local structure and
communities not only accelerates diffusion due to the presence of redundant
cohesive relationships, but also promotes diversity of knowledge across communities, thereby supporting the capacity of the system to innovate through
creative recombinations of different ideas.

showed that the optimum strategy for information to go viral in Twitter network is a combination of high degree and low clustering strategies (focusing on the nodes with high degree and low clustering that act as brokers between different groups in the network). Stonedahl, F., Rand, W., & Wilensky, U. (2010). Evolving Viral Marketing Strategies. Learning.

content of information determines how it will spread in network. news spread much better than discussion topics in the way that they are not so much determined by functional relationships, but both are determined by demographic distance.

Aral, S., Brynjolfsson, E., & Van Alstyne, M. (2007). Productivity Effects of Information Diffusion in Networks. Business.

Using the data from recommendation network this research focuses on the dynamics of viral marketing. They demonstrated that while there can be initial success in propagating information through a network, the probability of "contagion" decreases with repeated communication. They also demonstrated that hubs do not have proportially more influence than the less connected neighbors. Instead there's a certain limit to the influence that an individual can have. Finally, they showed that viral strategies work much better if they originate in densely connected
communities, which then spread them across the network at large.
Leskovec, J., Adamic, L. A., & Huberman, B. A. (2007). The dynamics of viral marketing. ACM Transactions on the Web (TWEB), 1(1), 5. ACM. Retrieved from http://dl.acm.org/citation.cfm?id=1232727

It was great that we won the contest, but more significantly, this exercise showed how building the proper incentives into a viral collaboration can quickly harness a large population to work together to address broad societal needs. It has helped us better understand how information spreads and why people cooperate.
Source: MIT Media relations, 10 December 2009. http://web.mit.edu/press/2009/darpa-challenge-1210.html