I cannot stress enough the value of this. No writer can be an impartial judge of his or her work. For example, the writer can't always be a good judge of a scene's effectiveness. Does the foreshadowing work? Is that scene that's supposed to be scary actually scary? Will anyone but the writer give a damn about any of the characters? And let's not forget those odd little typos that the writer's radar doesn't catch because the writer knew what they meant to say but didn't communicate it clearly enough to the typing fingers.
Sending one's work to friends or family for review isn't always easy. Writers have a good deal of emotional investment in their work (they should, anyway). When a reader points out a problem in a manuscript, anything beyond a typo or spacing glitch can raise the writer's hackles. ("I spent months on this novel! I stayed up late working on it while you plebes were eating bonbons and watching Cops! Who are you to say my story arc is cliched!")
But as soothing as it might be to the ego to huffily reject criticism, that stance results in a poorer manuscript. The right Constant Readers will find the problems that the writer doesn't see because he or she is too close to the manuscript, whether those are factual errors or problems with the character arcs. It's well worth the temporary ego blow, because when those same Constant Readers tell you something is good, they mean it, and you know they mean it.
A writer once told me that she knew she'd put in too many scenes set at the breakfast table when one of her constant readers remarked, "These people sure do eat a lot of pancakes, don't they?"
One caveat: A writer doesn't have to include every change that every constant reader asks for. If you have 10 readers and 1 says that X should be more Y, that's a minority opinion and it's up to the writer to make that change. However, if 9 out of 10 constant readers say that X should be Y, then it behooves the writer to make the change.
Not every constant reader enjoys her work. But I hope she's the exception rather than the rule.