I Repeat: Do Not Listen to Your Users (2008)
blog.codinghorror.comIn my experience, this is patently not true. I remember having a work dinner at a conference and I pulled out my phone to show someone something in our app. I opened it and ... a button simply didn't exist. I remember getting angry that I couldn't figure out how to use our app. The lead designer, watching this entire reaction said, "Huh, 40% of our user testing showed that removing that button gets exactly that reaction." We got a good laugh about it then (and he was serious, 40% of users got rage-level angry). But in retrospect, 40% of millions of people is a shit ton of people. I wonder how many heart attacks were caused by removing that button...
I digress. Listening to users is usually a fool's errand, however, listening to the right users is a gold mine. If you've done the whole "who is my customer" profile thingy, you know what they look like. So, actually seek out the real versions in your user database. Get to know them, take them out to dinner. Get to know their problems and how you can solve them. Those people will take you places.
maybe I'm missing something here. If 40% of users got rage-level angry was it really worth removing?
I'm a bit confused by your parent's comment as well. I'm stuck between thinking he is giving another anecdote in support of the article or reacting with "the headline is untrue".
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.
According to the A-B test, yes.
Apparently nobody gets my humor: if you do an A-B test where 40% of the users experience rage level anger at the experience and the other 60% are fine with it, clearly the A-B test indicates that the change is fine and should be implemented.
This is known as a revealed preference in psychology; observe people's actions, not words [0].
Too many software producers seem to adhere to this nonsense, which in part explains why software seems to be getting worse over time.
Of course you should listen to your users. But user reports aren't the only thing you should pay attention to.
Tip: focus on how your app/site/feature is making users feel and try to understand why that is instead of blindlh doing what they ask.