The Coursepad Information Cascade
https://medium.com/@zhujingsi/coursepad-me-design-notes-i-17a8b883d6cb
The observations made by the referenced blog post can be explained using the way information cascades work.
When scheduling classes, there are a few options student can use: using pen and paper, using iCal, or some online service that helps in scheduling classes. Here we simplify the model to just say that students can either use coursepad or not use coursepad to schedule classes.
Now the scheduling classes model has the basic ingredients for information cascades with one modification:
(a) There is a decision to be made: whether to use coursepad.
(b) Students make the decision sequentially, and each person can observe the choices made by those who acted earlier. (Here the modification is that each person does not observe all choices made by those who acted earlier but only some of them):
(c) Each student has some private information that helps guide their decision (the private information could be knowing that coursepad is made by a CS Major at Cornell or knowing that coursepad has the benefit of not crashing because it is not server dependent).
(d) A student can’t directly observe the private information that other people know, but he can make inferences about this private information from what they do.
When Jingsi, the creator of coursepad, posted about coursepad on the Cornell CS Facebook Page, a few people of the CS community tried it and then decided to use it for scheduling classes because they liked it. Some others might have decided to not use it because they were already using CU Agenda and coursepad didn’t feel as useful or they preferred using other means. Now, Jingsi observes that word of mouth was very powerful in spreading coursepad’s influence. The initial adopters of coursepad tell about its utility and intuitive user interface to their friends. The non-adopters of coursepad tell about coursepad’s non-utility to their friends. The second round of people who try out coursepad are influenced by their own user experience and know the decisions of some of the users in the first round. Then, they make their decision to adopt or not adopt coursepad.
Since most of the initial few users decided to use coursepad, the people who get to know about coursepad later know that more people decided to adopt coursepad than to not adopt it. So when they make their decision, they are already biased in thinking that coursepad is the next chequerd. The information cascade has started.