The Real Problem with Facebook Graph Search: Perpetuating "Filter Bubbles"
There has been a lot of talk about whether graph search will violate users privacy, but there is a larger problem here that is being overlooked: by relating all of our searches to people we are connected with, isn’t Facebook encouraging us to narrow our search lens? And what does this mean for how users will interact with search results going forward?
In May 2011, Eli Pariser gave a great and extremely popular TED talk discussing Google “Filter Bubbles.” Pariser believes that in personalizing search results, search engines are giving users the results that they want but not that they need. Personalized results narrow our knowledge base and miss an opportunity to expose users to new and different information, creating “filter bubbles.” These bubbles are small, search ecosystems comprised completely of similar ideas, which a user has already indicated they are interested in. For example, if you search for political articles, you are likely to only see articles that reinforce your current beliefs and not articles that challenge or expand your current ideas. This is because new ideas are seen as less relevant to you than ideas you have already expressed interest in.
Eli Pariser’s Filter Bubbles. Image courtesy of AmiNotes
So are we seeing filter bubbles starting to surface in Graph Search? The short answer is yes. Although the blog Actual Facebook Graph Searches proved that searches can be general (and contradictory) if content is kept “public”, Facebook itself has said that “each person sees unique results,” meaning that results are meant to be the ultimate personalized results. Users are encouraged to look at their friends likes, activities and recently read books as a way to get useful recommendations; this is the advantage of searching in Facebook instead of a traditional search engine. The only problem of course is that while your friends interests will be relevant and interesting to you (and there’s a good chance you will enjoy the same book as they do), it also means that you being forced to find things only through your friends, reinforcing Pariser’s “filter bubbles,” and limiting the new ideas that you are exposed to.
The long answer is a little bit more complicated and depends a lot on who your Facebook friends are. Facebook Graph Search seems predicated on the idea that your friends are relevant to you, which makes sense given that the average Facebook user has 234 friends. But since we know that many users on Facebook have around 1,000 friends, “friends” recommendations begin to mean something a little different. While the book recommendation of that great guy from your college English class that you don’t really keep in touch with may be extremely useful, the same may not necessarily be true for that high school cheerleader you never spoke to or your middle school gym teacher who you happen to be connected to on Facebook.
On the other hand, maybe they are more relevant; this depends completely on who you are friends with on Facebook and what you are searching for. If you are traveling to a new city and search “friends of friends in Milwaukee,” it may not matter how you know your Facebook friends for that connection to be useful. That is, assuming that you know all of your friends and feel comfortable imposing on them to meet their acquaintances. In the case that some of your friends are not relevant to you, you haven’t seen or spoken to them in over a year or you might not have the same taste as they do (I know this is the case for me), it is possible that graph search will actually leave users with the worst of both worlds: narrow recommendations based explicitly on irrelevant connections. This could push users away from more relevant sources and reinforce narrow ideas or recommendations that are giving the user neither what they want (a few, solid recommendations) or what they perhaps need (exposure to new ideas, or at least some good books to read).
So what does this mean for future Facebook searches?
No one knows, but I think we can have an idea, given past Facebook updates. I’d predict that graph search takes off rapidly as users test out its capabilities and then settles in as users become savvier about how to search and when it will actually be useful. In order to make sure that results are useful and relevant, users will probably begin to type in narrower searches. Instead of typing “Books my friends like," it would be more relevant to find out “books my college friends like.”
Regardless of whether or not all of your social media connections provide good recommendations for you, brands that are looking to expand their reach and gain new fans will need to expand their reach by moving into more of these small ecosystems to access users who they would otherwise not find. I recently wrote an article here, discussing if graph search would push marketers to focus on the quantity of their fans instead of the quality of them because now it will be important to appear in the most searches possible to increase a brand’s visibility. Engaging new and diverse communities will become an important way to be found in more search results by more people.
How do you see users interacting with graph search? Will we be reinforcing filter bubbles or suddenly be opened up to information from random Facebook friends we forgot we had?
Caroline Phillips Rodin is the SEO manager for growth at Grovo. She started her illustrious career in-house at a B2B sign company where she taught herself SEO. She then worked at iProspect, directing strategy for Fortune 100 financial and Fortune 500 luxury good brands. Caroline can't wait for her new Grovo tracksuit and spends her free time hunting for the best chocolate chip cookie in NYC. ...