YouTube’s Sorting: Dealing with Las Vegas
Article: YouTube alters search algorithm over fake Las Vegas conspiracy videos
On October 1st, Las Vegas quickly became a city of infamy. After a mass-shooting left 58 dead and over 500 injured, the city of sin quickly became home to the largest mass shooting in modern American history. Naturally, we wanted to know more about the situation and gravitated to internet-based sources. For many, Facebook and YouTube became primary new sources. However, this backfired as the search algorithms used by these companies displayed conspiracy theories right next to true information. For example, if we searched for “Las Vegas shooting”, it would return results from reliable news outlets alongside “Proof Las Vegas Shooting Was a FALSE FLAG attack — Shooter on 4th Floor”. Videos like these and posts on Facebook sharing a similar message quickly angered victim’s families which prompted the tech giants to implement change. Having been in the process of updating their search algorithms, YouTube was forced to speed up their work and release the updates earlier than they expected.
I picked this article because it tied into the ideas of scarcity and abundance that we discussed in class. In the lens of web search, scarcity is defined as the phenomena in which there may be only one document to answer our question, and abundance is summarized by the phenomena in which out of millions of pages, how do we find the best ones? First, the concept of abundance applies to the YouTube situation because there could be hundreds of thousands maybe, even millions, of pages related to the search term “Las Vegas Shooting”. The problem is that it is incredibly difficult to rank these pages by relevance, which ties into the next concept. The second concept, scarcity, applies to the YouTube situation because out of the many pages, there may be only a few (maybe even just one) videos that answer the questions we have.
These two concepts are fascinating because they lead us to think about how companies like YouTube handle big data in such a way that results are relevant to each specific user. Aside from the article mentioned above, I looked into how YouTube ranks specifics pages. Because YouTube is a Google-owned company, I assumed they would use the same PageRank methodology as its parent company; however, this is not the case. At first Google focused on how many videos we watched, but this turned out to be detrimental because users just skipped from one video to the next. Google then switched their ranking algorithm to focus on how long we watched a video for which resulted in an increase from 1 to 4 minutes for average watch time. Google also applied machine-learning to their algorithm in the hopes that one day “ ‘whenever you turn on your TV, the show that you would be most likely to watch on would always be on — we’re going for that kind of experience’ ” (D’Onfro).
The source of the last paragraph was obtained after search results from “does youtube use pagerank?” returned a previous blog post: YouTube’s Video Search Algorithm. I used the source from the blog post more than anything: The ‘terrifying’ moment in 2012 when YouTube changed its entire philosophy.