Facebook as a Search Engine
While search engines such as Google or Bing obviously use some type of hubs and authorities idea, a less obvious site that also implements this is Facebook. Recently, Facebook has been making many changes to it’s search engine. According to TechCrunch, it is now ready to “challenge Google and Twitter for the control of real-time news search and news search itself. While Facebook search originally started off displaying just friends and Pages you Like in order to help users find potential friends and link them to their profile, it has now expanded to return anything that you are allowed to see from it’s 2 trillion plus posts. Because of this, Facebook now has much more of a holistic search engine.
With Facebook’s new search engine, once a user enters a search term, the algorithm that is implemented returns a personalized results page. The results page is based on over 200 factors about a user’s personal Facebook profile and actions, including what the user Likes and engages with, things the user has previously searched for, and information about the user’s identity. When displaying the results, the first thing that is displayed are Facebook’s trending topics (if they are relevant to the search), it also shows a summary of the topic when possible, then shows mentions of the topic from trusted news sources. New sources are trusted or not based on the judgment of an internal Facebook team that looks at engagement and spam reports among other things to determine a “reputation score” which is then used to rank the posts. After that, posts by people and groups from a user’s network are displayed, followed by aggregated counts of the most popular links related to the topics, and lastly public posts related to the topic (ranked to seem as relevant to the user as possible).
Overall, Facebook search engine utilizes many of the hubs and authorities ideas that we discussed in class, but in a different light. Instead of amount of links to a certain page, Facebook uses the “trustworthiness” of a news source as the in degree, along with the things that specific users search and like to have on their news feed. It’s a very personalized application of Jon Kleinberg’s idea.
Facebook Expands Search To All 2 Trillion Posts, Surfacing Public Real-Time News