Product Reviews and Information Cascades
In the text we were introduced to a simple herding experiment. We were given two urns, with 3 marbles in each one. One urn had a red majority (2 red marbles, 1 blue marble), while the other had a blue majority (1 red marble, 2 blue marbles). The first student selects a ball and makes a guess on the color majority based on his selection. The following students then each select a ball, and make a guess based on their selection and the previous guesses. However, all they know is the guesses of the previous participants, not what they actually selected. This experiment demonstrates the dangers of information cascades. With only the knowledge of others’ guesses and their selection, after 3 turns a majority of people make one guess as opposed to the other, meaning it is in the best interest of any participant after those 3 turns to side with that majority of people. This continues for any number of people. Even if a participant selects a marble of the opposing color, they will assume they’re wrong and side with the other participants. If those first 3 participants were incorrect, it means that the entire group of participant, no matter how large will be incorrect.
While some may argue that this experiment does not translate exactly into real life situations, its implications are limitless. It demonstrates the ease at which an information cascade may take hold as well as the power of an information cascade. All it took is 2 or 3 people in the experiment for a decision of the entire group to be made. This spells danger for an era where information can be distributed and accessed so easily using the web. An information cascade has the ability to take hold of entire segments of web users at a time. One example this takes hold in is reviews on the web, on products, articles, etc. A couple bad/good reviews could influence the rest of the reviews. Amazon is a merchant web site that places a lot of importance on user reviews so buyers can make informed decisions and so they can suggest products to customers more efficiently. It makes sense, then, that false reviews have become a problem for Amazon (http://www.forbes.com/sites/suwcharmananderson/2012/08/28/fake-reviews-amazons-rotten-core/). There are even companies that can be paid to write false reviews for products. With the knowledge of information cascades, this means that a few good reviews can get the ball rolling for more. Or even worse, a few bad reviews falsely written about a competitors item can cause a false, negative information cascade for this merchant.