Building on my previous post regarding the relationship between influencers and social epidemics, reprised below are abbreviated comments from an interview with Duncan Watts published in Harvard Business Review in February, 2003.
"People tend to think that successful products are somehow destined to succeed because of some intrinsic combination of features that creates and sustains demand. But network science suggests there’s more to the picture.
In the case of Harry Potter, before Bloomsbury bought the rights, several other publishers rejected the manuscript. It’s tempting to think of them as fools who missed a sure thing. In fact, it never was a sure thing. For every Harry Potter that explodes out of nowhere, there are thousands of books, movies, authors, and actors who live their entire lives in obscurity, and my work suggests that it’s not because they lack quality or desirability. In other words, the market for a successful product should not be thought of as existing in some latent state before the product launch waiting for the product to arrive. Rather, it arises dynamically, driven in large part by the growing success of the product itself. In economics, this phenomenon is known as an information cascade: a social chain reaction in which increasing numbers of people buy a product principally because other people are buying it.
We’re finding that the structure of the networks is probably much more important than anyone thought in influencing the dispersion of ideas or behaviors. Harry’s success may have more to do with particular attributes of the social and media network it’s spread across than with any inherent quality of the book. That turns our traditional notions about cause and effect on their head.
So which network structure best encourages information cascades or idea contagion? We don’t know yet. But our work is beginning to identify some basic principles. For instance, it appears that having a wide range of personality types in a population can actually enhance the odds that a new idea or product will catch on. We also think that information cascades can be squelched if people in a network are exposed to too many opinions, or too few. Clearly, poorly connected networks inhibit idea contagion. What’s less obvious is that if the people in a network are too densely connected, that may also prevent a fad or a product from catching on.
Network science suggests that our notions of cause and effect are skewed, that we’re sometimes looking at the wrong actors in the play to try to understand why the drama is unfolding the way it is."