You might want to look up what was up with Cambridge Analytica - Michael Lewis has one of episodes in Against the Rules (podcast) - it's a good place to start.

There was nothing at all. No predictive power. Total waste of money and it's funny this example keeps coming up