A quick recap of where we have been so far in the space of online services and web 2.0. In Chapter 3, we discussed the recommendation of webpages with an objective metric computed by Google from the graph of hyperlinked webpages. In Chapter 4, we discussed the recommendation of movies with subjective opinions estimated by Net ix from movie–user bipartite graphs.

Then we investigated the wisdom of crowds. In Chapter 5, we discussed aggregation of opinion in (more or less) independent ratings on Amazon. In Chapter 6, we discussed resolution of opinion conflicts in Wikipedia.

In this chapter, we will talk about dependence of opinions, taking a macroscopic, topology-agnostic approach, and focusing on the viral effect in YouTube and tipping in Groupon. Then in the next chapter, we will talk about the effect of network topology on the dependence of opinion.

As will be further illustrated in this and the next chapters, network effects can be positive or negative. They can also be studied as externalities (e.g., coupling in the objective function or the constraint functions, where each user's utility or constraint depends on other users' actions), or as information dependence (e.g., information cascades or product diffusion as we will see in this chapter).

A Short Answer

Viralization

YouTube is a “viral” phenomenon itself. In the space of user-generated video content, it has become the dominant market leader, exhibiting the “winner takes all” phenomenon. More recently it has also featured movies for purchase or rental, and commissioned professional content, to compete against Apple's iTunes and the studios.