So when you listen to users and they tell you "you should do X", just blindly doing X is a path to failure but similarly ignoring them is also a path to failure.
There's an art to properly eliciting insight from users and it's a mix of observation, interrogation and thought. In general, the best advice I've found is focus on the emotional and not the logical. When you ask when was the last time you had a problem with our product, you force the user to come up with ad-hoc logical arguments and paradigms that are most often wrong. If you ask when was the last time you were frustrated or annoyed with our product, you hone in much more reliably on real sticking points.
Imagine a patient going to a doctor and asking a course of antibiotics. Any doctor would erase that request and start asking the patient to describe their problem. This, he said, is the core of how a user feedback conversation should go.
This is exactly right and IMO one of the most important skills a PM should have.
But over over time you begin to develop a filter that separates really good ideas and mediocre ideas that serve a single or rare use case.
Observing is great if you have the bandwidth, but if you're a one-person-show, listen to your customers, listen to your instinct, and develop features that are going to drive revenue and push you forward.
Of course in addition UX designers do have their own demons to fight. I would call one of them as cosmic expansion of the space, that is - contrast will decrease and distance between meaningful UI elements will increase over the time.
And even then they can be telling you nothing at all because you don't know things like user intent. Harry Brignull's summary of the issues on this is required reading here (about eye tracking but applies to all such heat mapping tools):
http://www.slideshare.net/harrybr/what-you-need-to-know-abou...
How do you determine why something is happening?
This is the work of the UX designer. It takes time an effort. PMs usually don't have the time or understanding to do it properly, so products and companies often suffer. This is why, personally, I would be more inclined to invest in companies that have actual UX designers working for them than mere jack-of-all PMs. But hey.
Firstly, you might consider that close to all basic ideas behind objectively successful products (eg the telephone, Facebook, photo copiers, Twitter, rabbit vibrators) were NOT arrived at by asking the target audience what they wanted. Most were intuition born of observing that audience.
When you ask people direct questions about what they want, it takes a great deal skill and judgement to actually get a successful product idea from that. In fact it's almost not worth the bother, and few professional designers really do it very often by choice. Far better to try to observe what people do (often by giving them tasks to complete in some sphere of activity) and see what comes out of that. You might see your target audience performing needless operations, misunderstanding things, or doing things that appear irrational to you but rational to them. Once you see that sort of thing happening, it's far easier then to see what they might need or really want.
I wouldn't go so far to say that asking people what they want is a bad idea, but it's getting close. Perhaps it's a mildly toxic idea best avoided in preference to observation.