Due to privacy concerns, do you stay away from social media sites as far as possible? Maybe you’re that person who refuses to engage on these platforms and never had a Twitter or Facebook account.
Without a social media presence, you probably think you are completely invisible. These giant social media networks don’t know anything about you, right? Wrong.
Account or no account, big tech has your data. And with all the connections, traces and crumbs you leave online, there is truly no place left to hide.
Read on and see what this shocking study discovered about social media predictability.
Show me who your friends are …
A new study has revealed that social media sites including Twitter and Facebook can learn about you even if you don’t have an account with them.
University of Vermont researchers discovered that these sites only need access to eight of your contacts to predict your habits and behavior.
Using mathematical models, the team analyzed more than 30 million public Twitter posts from around 14,000 users and found out that machine learning algorithms can predict with up to 64% accuracy what you are going to write next — based on what you and your friends have posted and interacted with in the past.
With these predictive text patterns, social media sites can then create profiles to identify individuals based solely on their social connections, with or without an account.
This gives the saying “Show me who your friends are and I’ll tell you who you are” a modern tech twist, don’t you think?
No place to hide
Although the study relied solely on Twitter data, the same techniques can be applied to other social media sites that use machine learning algorithms, including Facebook and Instagram.
And get this, around 95% of your potential predictive accuracy can be gathered by simply analyzing what your friends post.
What does this mean? Well, it means even if you don’t post anything on social media or don’t have an account at all, these sites can still create a profile for you and predict your behavior based on your personal connections alone.
As usual, this kind of blanket predictability is valuable to social media services since these profiles can be used for tracking purchase and entertainment preferences, targeted ads, and even for predicting political affiliations.
Chances are, even if you don’t have a social media account but your friends or relatives do, your behavior can still be predicted with a certain degree of accuracy (around 61%, according to the study.)
Keep in mind that this study doesn’t prove that Twitter or Facebook are actually using these predictive models with their algorithms — it just shows that they can.
The frequency of posts matter
Not surprisingly, the study also showed that the frequency of posts in your social circles also determines the predictive information about yourself or your friends.
People who posted on an average of eight times a day are more predictable than those who only post once a day.
However, the effect is reversed if your friends post frequently, too. The research data showed that if your friends post eight times a day on average, their data is less predictive of you.
Why so? The study said that highly active friends tend to inhibit the information flow by covering too many topics, some of which may be irrelevant to you.
For more on this topic, listen to the free Kim’s Consumer Tech Update podcast below.