People often think Facebook is a simple platform that allows users to share photos and information with each other.
This couldn’t be further from the truth.
The back end of Facebook is a powerhouse of machine learning and artificial intelligence that is, at times, terrifyingly intelligent.
In 2015, a group of researchers at the University of Cambridge decided to see if an algorithm using Facebook’s dataset could match humans at one of our most prized abilities – the ability to judge personalities. The researchers used 86,220 people to conduct the experiment.
The measure of personality used in the experiment was a comprehensive 100-question OCEAN personality test, which targets the “Big Five” personality traits: openness, conscientiousness, extraversion, agreeableness, and neuroticism. This is arguably the gold standard of personality tests.
The algorithm uses a person’s Facebook “likes” to predict that person’s personality. The experiment then compares the algorithm’s prediction of the subject’s personality with other human judges in the test subject’s life. The results were as follows:
With 10 likes, Facebook knows your personality better than a work colleague,
With 70 likes, better than a close friend,
With 100 likes, better than your family members,
With 300 likes, better than your husband or wife
To make things more interesting, I dug deeper and found that the average person “likes” about 10 items per month on Facebook. This means that if a person has a Facebook profile that’s been active for longer than 30 months, Facebook knows more about that person’s personality, the way he or she thinks, feels, and behaves, than any human in existence.
I will concede that humans, of course, know other humans in deep ways that cannot be captured by personality tests. However, knowing someone on the “personality test” level, as Facebook knows you, is more than sufficient to enable Facebook to have an excellent understanding of the products and services you will find most tempting.
How Is This Possible?
When I read this research paper, I thought to myself: How could simple binary pieces of information such as “liking” or not liking Facebook pages lead to an understanding of one’s personality?
To understand this, it is important to remember the compounding effect of the data. The algorithm doesn’t only look at what you liked, but it looks at what you liked and compares it to everything else you liked, as well as everything everyone else has liked.
For example, the study discovered that if someone likes all three of Salvador Dalí, meditation, and TED Talks, that person almost certainly has a high openness to experience.
Zooming out a little bit, with over 60 million possible pages to like on Facebook, there is an incomprehensibly large amount of possible combinations of 300 likes one could have.
The math is below:
For a little perspective, the number 9.1E+1718 (written in scientific notation), is much greater than the number of atoms in the entire universe.
Your 300 likes become a personality blueprint, virtually impossible to replicate, that is given to Facebook on a silver platter. Facebook’s algorithms and other third-party algorithms are only beginning to be powerful enough to draw meaningful insights from this data.
As owner of North Digital, I talk to a lot of business owners who ask me, “How effective is Facebook advertising?”
The following is an excellent answer to that question:
Similar to how the research team at Cambridge used Facebook’s dataset to uncover the personalities of people, the Facebook business manager uses those same data points to deliver hyper-relevant ads. It is by far the most targeted advertising platform in the world.
In addition to being the most targeted advertising platform in the world, it also evolves the fastest.
Let’s assume one of our clients sells a meditation app. We might start off targeting people who like Salvador Dalí, meditation, and TED Talks. However, the ad is not only shown to the people who have those interests at the time we make the campaign, but rather, every hour the audience of relevant people is updated to match people’s shifting interests.
Let’s say John Smith hears a co-worker talking about meditation and decides to “like” meditation on Facebook. Within minutes, the advert for our meditation product will be in front of him.
If you are a business owner who wants to increase the revenue of your business in 2019, and you are not currently using Facebook ads, you should send me an email:
You can read the full research study this article was based on here.