In a Guardian article posted this afternoon, it is revealed that Amazon’s attempt to clamp down on fake and inaccurate reviews has taken a creepy turn. An author called Lori L Otto posted on her blog some weeks ago about a fan who was blocked from posting a review of her latest book on Amazon, because, to quote the Amazon email she received:
We cannot post your Customer Review for “Olivia (Choisie Book 2)” to the Amazon website because your account activity indicates that you know the author.
Otto relates that the reviewer, Nikki, became a fan after reading her first book, that they exchanged communications, and that she’d sent Nikki an Advanced Reader Copy (ARC) of the new book, Olivia. Amazon had somehow discovered this friendship and determined it meant that Nikki couldn’t give an honest review.
This is concerning.
What does Amazon know about us?
Amazon refused to divulge how they determined that the reviewer in question knew the author. Otto was not detailed about exactly how she’d communicated with Nikki, but Amazon itself is neither an email client nor a social network; beyond the presence of existing reviews, there’s little for Amazon to go on from their own site. So how do they decide that “accounts are linked”?
Do they scour the internet, searching for reviewers’ names on social media to see if the authors they’ve reviewed are amongst their contacts? Do they have programmes to go through every author and reviewer, seek out alternative names used across the internet, and flag connections between authors and reviewers with moderation teams? Where are they looking? By what criteria do they judge? Who is reading the tweets, counting the Likes? There’s definitely a feeling of being watched too closely here.
How many fans will be stopped from reviewing the authors they enjoy the most?
This approach is clearly capable of catching more fish than it’s meant to. Reviews by family members who only want to help their relative sell books are a problem, but if this net is preventing avid fans from reviewing their favourite authors, something is wrong. I’m half way through reading The Liar’s Key by Mark Lawrence right now. I intend to review the book when I’m done – but will Amazon let me? I follow him on Twitter, Facebook, Goodreads. I’ve emailed him to enter his contests for signed books, tweeted at him a handful of times – received responses to those tweets a few times too. All this came as a result of reading his books; without those, I’d have never heard of him.
Will Amazon decide I know this author? Will it tell me they can’t post my review to their website because my account activity indicates I know this man – when all I know about him is what he shares on social media?
How many reviewers out there have been affected by this and are going to get caught up by it, simply because they are fans of an author and interact with them on social media? Given that authors of all publication routes are advised to be active on social media, to engage with readers, this could include a fair number of fans, and even a lot of people who are fans of the genre and have responded to an author they’ve never read as part of a larger discussion including an author they have read. Could this system Amazon has implemented catch them too?
Meanwhile, authors will no longer get reviews from those amongst their fans who are most active in engaging with them on social media. These are the people most likely to want to review a book – they will certainly finish it, and will see the requests for reviews posted on social media that less ardent fans and general readers would not see.
Often authors review one another’s books. They may not know one another in anything more than a professional capacity – maybe they’ve spoken on the same panels, chatted about plot and character, been active in the same forums – but here too Amazon may scupper things. And reviews by an author of the same genre are not only an important boost to the author of the book being reviewed, they are also a key indicator to readers of what else they might like to read. I personally find them very helpful; I have on more than one occasion read an author because an author I already read has reviewed them favourably.
This creepy approach is not the only solution to the problem
Amazon’s approach here is heavy-handed. Their algorithms or whatever they use to determine who knows who in the book world are clearly catching far more than just the relatives and friends they’re targeted against. Why not implement other approaches? A check box confirming the reviewer is not related to the author. Another one confirming that they are not close friends. A third confirming that the review is an honest opinion.
Okay, some people lie. There will be some parents of authors, best friends of authors, whatever, who happily check that they are not related to the author, they don’t know the author, yes you can post this review in the confidence that it is a real review. You’re never going to get it 100% right. But something like that should at least catch the people with some integrity. It would give Amazon a means of conveying the rules to reviewers, and give reviewers a moment to think whether posting the review is ethical.
I doubt they’ll do that. But at the very least they should introduce an appeals system, allowing reviewers to say “um, actually you’re wrong” and have moderators investigate why they were flagged in the first place and determine why there was cause for concern, and allow reviews to be posted if their algorithms were wrong.
And maybe Amazon should focus more on the paid reviews, the reviews by people who have never read the book and never met the author, whose incentive is cash. Those are just as damaging to Amazon’s reputation – if not more, since a family member will only falsely review books by their relative, not by dozens or hundreds of complete strangers.