As online applications become increasingly sophisticated they depend on algorithms to process metrics about us and who we interact with to deliver personalized content. Algorithms are usually the ‘secret sauce’ of a service, so architects only share generalities and when confronted with an issue. Like with the Facebook News Feed, an algorithm we became concerned about only after a study revealed Facebook manipulated feeds of a sample (680,000 randomly selected users) to see how they reacted to positive and negative updates. Interestingly, the results indicate when friends’ positive news feeds were weeded out to varying extents, people wrote fewer positive posts and more negative posts.[i] Clearly, as the results show, there are implications regarding how we consume our news.
Facebook has masterly positioned its social media platform as a quick, convenient one stop shop for everything – friend updates, pictures, videos, and news; have a fifteen minute break, visit Facebook from your mobile or computer to get caught up on things. Thirty percent of Americans get their news from Facebook.[ii] As we are all starting to become familiar with, Facebook has an algorithm that predicts what content to feed and serves it on our news feed. It is based on “’thousands and thousands’ of metrics, including what device a user is on, how many comments or likes a story has received and how long readers spend on an article.” Your friends’ behaviors (liking, commenting, and sharing) and pages you follow also significantly affects what appears on your News Feed. Facebook has just released News Feed settings and tools that let users “more easily tell Facebook what they don’t want to see”.[iii] Still there are some issues to consider:
- Should you depend on an algorithm in social media to curate your news feeds? It depends; there is evidence for the algorithm’s value by increasing user engagement. If you have the right friends and follow relevant pages, it acts as a news filter by recommending only the top articles; this might be enough. However, if you want to take on a thought role, you should canvas digital newspapers and magazines directly.
- Do you want to be confined to reading news based on the online behaviors and interests of you and your friends? In order to develop a passion, you need to be exposed to it. The danger in being fed content based on one phase of your life is you never develop intrigue in other subjects. One thing I love about the front section of the Wall Street Journal is the guaranteed ‘odd-ball’ article.
Netflix offers a stream of recommended videos. Recently they tweaked the underlying algorithm from being a rating system to something more complex; something based on behaviors – how long you watch a movie, what selections you click on, what you scroll over, etc. Netflix estimates seventy-five percent of viewer activity is driven by recommendations.[iv] This algorithm has significant influence on how we consume entertainment. (Interestingly, Netflix had a contest open to the public to build the algorithm and gave a million dollar prize to the winning team; this is a smart, cost-effective way to innovate.)
Google has a complex algorithm that ranks pages based on keywords and produces a search engine results page (“SERP”). It gets personal when you are logged into Google via Google+, Gmail, or one of their other applications. When you are logged-in, you get not only a slightly different SERP, but also some personalized content based on context from other applications; Google knows who you are. The other implication in being logged into Google (Facebook, Twitter, and other apps) is your movements during an Internet session are tracked directly to you by the app in a personal way (not just your IP address).
Some other popular ‘content feeding’ algorithms include: LinkedIn’s People You May Know, Pinterest’s Suggested Pins, Amazon’s Product Recommendations, and Pandora’s Music Streams.
We need algorithms because of the vastness of content available to us and how we consume it – often in quick spurts on mobile devices. Conducting effective searches takes time and some inspiration, so we like prompting or nudging to guide us in the right direction. Personally, I am more concerned with how I consume news (knowledge) versus entertainment and therefore get most of my news directly from its source (when I can). Nevertheless, it is worth understanding the basic mechanics of algorithms so you have more awareness – perhaps even play around with your news feed and see what happens as you ‘like’ things.
We deserve to understand the power that algorithms hold over us, for better or worse.[v]
Future algorithms will have a heavy dose of artificial intelligence. Much more context will be fed into the algorithm, allowing for it to deliver a deeper, personalized experience. Moreover, the artificial intelligence gets smarter and better through time. According to a study, AI has attracted more than $17 billion in investments since 2009; some of the investors include Facebook, Google, Yahoo, LinkedIn, and Twitter.[vi]
[v] Karrie Karahalios. Algorithm Awareness. (MIT Technology Review, November 2014).
[vi] Kevin Kelly. Brain Power. (Wired, November 2014).