VR World

How Facebook Knows You'll Be in a Relationship Before You Do


On Valentine’s Day Facebook’s Data Science Team released a blog about how they are able to track people’s activity and create connections between their behavior and relationship status. The aptly timed article basically talks about how Facebook is able to track users’ behavior and try to create connections between certain behaviors commonly found between users and what their consequences might be. They do this a lot for different types of behaviors, but the most recent one specifically related to people’s behavior in love.

The blog post itself titled The Formation of Love specifically follows two different datasets that are tracked before and after people enter into a relationship on Facebook. The first behavior was Timeline posts, specifically the frequency of Timeline posts before and after a relationship starts. So, that whole suspicion about people becoming a couple after a rapid amount of Timeline posts and wall spamming (what Timeline used to be called) is actually fairly validated as the graph below shows.

This graph goes two ways, because it one confirms that this suspicion about friends may in fact be true but it also may put certain Facebook-friendly people in a bit of hot water if they do a bit too much Timeline posting on eachother’s walls. Sure, this is merely an extension of social interactions that people would normally have in person or over the phone, but as Facebook has become a more and more commonly used social platform those kinds of behaviors have clearly translated over to social media. I see these increases in Timeline postings merely a modern equivalent of long phone calls or insanely long text message threads. Sure, those things may still probably happen leading up to a relationship, but now Facebook ads yet another angle to the equation. Not to mention, when a couple is actually together they are probably spending more time together than actually posting things to each other’s Facebook walls.

Facebook also noticed a sudden increase in ‘happier’ emotion postings leading right up to a couple becoming ‘Facebook Official’. Sure, most relationships actually start before they’re officially posted as couples on Facebook, which likely explains the increases of happiness prior to the relationship status change, but that doesn’t change the fact that Facebook can start using this kind of data to more properly advertise to people.

In the end, all of this data analysis means that they can more properly predict users’ behavior and effectively serve them ads or services that will be pertinent to them at that time in their lives. By not only targeting these ads to the right people, but also to direct the right kinds of things to these people they are likely to significantly improve their effectiveness. This is especially important to Facebook after some people have accused Facebook’s platform of facilitating a status quo of false likes and followers that earn the company more money but do nothing for engagement.

We’ve experienced very similar behavior after doing our very own Facebook follower testing and found that as we promoted posts more, they actually had fewer clicks. This was originally started by myself as a research project because all of a sudden starting on December 2nd, our overall organic engagement with our Facebook followers/likes went down more than 50% per post and it only went up after we paid for engagement. But even then, it didn’t improve overall engagement or clicks because most of the people liking our articles had little to no connection to the site itself even though our ad campaign was fairly well targeted. You can watch the video below to see what we’re talking about.

  • Appearance great. They are yet to definitely done their very own work towards that one to make style a priority. It does not take very first time I could truthfully really find myself employing Microsoft windows all over again. Android os, on the other-hand is actually a clusterf*ck connected with crap.

  • ak

    Awesome research ……….Thanks a lot

  • ak

    Awesome research ……….Thanks a lot