With the increasing availability of network and behavioral data---from what we buy, to where we travel, to whom we know---we are now able to observe and quantify social processes to a degree that would have seemed impossible just a decade ago. These new microscopes into human activity not only have substantive implications for the social sciences, including economics, sociology, and psychology, but also raise challenging computational questions in large-scale data analysis.
Leveraging these new data sources and computational tools, in this talk I will discuss the diffusion patterns that arise in six online domains, which in aggregate involve millions of adoption events and billions of individual-level interactions. Though the relationship between interpersonal networks and diffusion is important to a wide range of social and economic problems, including new product adoption, changes in social norms, and success in cultural markets, until recently it has been prohibitively difficult to accurately measure the structure of these networks. In all six cases we study---and in contrast to prevailing theories of diffusion---we find that multi-step cascades are relatively rare, with almost all adoptions occurring within one generation of the seed. Correspondingly, we find that the average size of diffusion events is surprisingly small, with each seed typically leading to less than one additional adopter.