Reader reviews: we get them. And, mostly, we ignore them; because, like all other forms of fiction, 90% of book reviews are junk. And reviews by regular readers, as opposed to professional critics, are like the publishers' proverbial slushpile: a seething, shouting mass of logorrhea in which a few gems may be submerged, if you can bear to hold your nose for long enough to find them.
But for an author to make a habit of ignoring feedback is pretty much the first step on a slippery slope down into a mire of self-indulgent solipsistic craziness.
So how should you approach reader reviews in order to separate the ones you should sit up and listen to from the background noise?
My take on the subject is that if you're an author, you can get some useful clues to the relevance of your reviews by psychoanalyzing the reviewers.
We live in the age of social media; corporate entities like Goodreads or Amazon use our natural inclination to communicate to generate free reviews and raise a buzz around the content they're trying to sell. We've come a long way since 1999's Cluetrain Manifesto and the internet marketers have worked out that the best marketing tool out there is word of mouth recommendations. And because individual works of fiction are about the ultimate micro-targeted boutique product, word of mouth is about the only marketing tool that works reliably. So: reader reviews. How should we interpret them?
A sad fact, worth repeating, is that it is impossible to write a work of fiction that everybody will read and understand in the same way. Readers (authors included) all approach a text with their own baggage, and what may be unremarkable or even exciting to one reader may be triggery or otherwise unpleasant to another. Whatever you write, and however well you do so, 20% of your readers will hate it—often for reasons that have nothing to do with the text and everything to do with the babble of experiences and memories that reading the text causes to rise to the surface of their mind. (Read this 2006 blog entry on one-star reviews of famous works and weep—tears of laughter, I hope, rather than despair at one's fellow primates.)
Stories which are intended to induce cognitive dissonance—by setting up a sympathetic protagonist then exposing them as a murderer, rapist, and war criminal, for example ("Glasshouse") often trigger aversive reactions from readers who start out expecting a cosy escapist yarn that stays firmly within their comfort zone. (Ditto the Merchant Princes series, which starts out looking like a classic portal fantasy but ends very uncomfortably, when cosy portal fantasy collides with realpolitik.)
Another point, also worth repeating, is that many readers are incapable of separating their own emotional response to a text from the actual content of the text. "I do not like this" is isomorphic in their mind with "this is a bad book".
So: if your work is anything but a literalistic recapitulation of a traditional narrative theme, with sympathetic characters, clearly depicted antagonists, and a cosy sense of closure at the end that reinforces traditional cultural values ("and the prince married the princess and they all lived happily ever after") you can expect a fairy ring of one-star reader reviews to circle your work on Amazon. And the more challenging the novel, the more readers will feel the need to scream I HATED THIS! I DON'T UNDERSTAND IT AND IT MAKES ME FEEL UNCOMFORTABLE! THIS IS A BAD BOOK!
This is not necessary a bad thing.
(Personally, I think it's a good thing.)
Similarly: if you write and publish novels on a regular basis, you will acquire a core of fans, and they will do their five star cheerleader thing in the Amazon fora and reviews every time you emit a new fart, whether fragrant or otherwise. You should strive to ignore these reviews. No, seriously. While it's probably okay to indulge yourself and roll around in them if you're feeling down, you should not take them seriously. Just as 20% of the audience will hate any performance, another 20% will love it to pieces—often for reasons that have more to do with the contents of their own headmeat than the quality of your writing. (There's no accounting for taste.)
The readers you need to pay attention to are the 60% who fall in between these spectral extremes. And, in particular, those readers who can separate their own emotional reaction to the text from the text itself. They may not be experienced literary critics but they can tell you much about how the regular readers have received your work. And the telling clue is that their comments say things like "I had a bad reaction to this book", rather than "this book is bad", or "I didn't understand why [the protagonist did something]" rather than "the protagonist is unbelievable".
The 20/60/20 spread is also worth paying attention to. I pulled those figures out of my arse, quite deliberately: they actually vary quite a bit from book to book—in fact, Amazon provides a neat histogram for every item, in the shape of that bar graph ranking feedback from one star to five stars.
In an ideal world, we'd look at our reader reviews and see a single fat five-star bar with nothing below it. Failing that, a bathtub curve (lots of one star and five star reviews, fewer two and four, very few three) would be satisfying: it means people react strongly to the book. The worst is an inverse-bathtub curve: lots of three stars, some two and four, no fives or ones. It means readers didn't feel strongly about the book; the typical reaction was "meh". I might be sticking my neck out here, but I know no novelists who set out to write a book to which the typical reader response will be "meh".
So. Beware the curve with a fat belly. Dread the analytical reader who can distance themselves from their subjectivity and who still gives the book three stars. Ignore the five-star fans and the one-star butt-hurt trolls. This is the world we live in, and until we learn to clone John Clute and iterate him in parallel over every genre book that is published, this may be all the help we can get.