Reversed Network Effect
http://fusion.net/story/368372/facebook-election-newsfeed-algorithm-chronological/
We tend to think network effects only moves in a positive direction: the product becomes more valuable if more people are using it. In class, the function f(z), the benefit to each consumer from having a z fraction of the population use the good, is an increasing function. However, in the article, LinkedIn, Twitter and Reversed Network Effects, the author argues that network effects can be bi-directional and there could be a reserved network effect. He argues that increased usage may be beneficial to the overall platform, but for some individuals, when large population that exceeds certain level, there could be a reduction in the utility. For instance, following too many people on Twitter could mean “each tweet is less likely to come from close colleagues”, making the quality of the average tweet being reduced.
An alternative to cope with the problem is using the filtering algorithm, for example, Facebook and Instagram’s non-chronological newsfeed. The order of the posts users see is based on “the likelihood of interest in a post, relationship to the poster, and timeliness.” In other words, the posts you see is what Facebook and Instagram believe you care about the most. But does this really solves the problem? In fact, the filtering algorithm presents you a world you want to see and your friends want to see, but not the world it really is. For example, during the presidential election, some people criticize Facebook “tends to feed users only the type of news they want to hear, contributing to an echo-chamber effect that stifles opposing points of view or facts, a troubling trend in a democracy”.