The headline is true, as you can see when reading the article but without reading the article itself can be misinterpreted easily. The emphasis in the sentence is on Listen in the sense of listening to what users are saying. The article then gives multiple examples and arguments for why you should "listen" to your users by observing what they actually do and ignoring what their mouth speaks (or keyboard writes).

Even if you do A/B testing and get "data" you can still draw the wrong conclusions from the data. E.g. if 40% of users in an A/B test get rage level anger, I would say that is a damn strong signal to never remove that button. Who wants to enrage and cause churn for 40% of their users? That's millions of dollars in ARR.

To use one of the examples from the article, that's like seeing the Team Fortress statistic that Medics are only played 5.5% of the time to then decide to remove the Medic class from the game. Instead Valve did the right thing: they incentivized more players to play the class and get that percentage up.