When the ‘scandal’ over paid-for / sock-puppet reviews hit Amazon earlier this year, the company was left with two options. Option One was to make a concerted effort to root out and delete the offending reviews (and reviewers). Option Two was to try to find a way to encourage genuine reviews, thereby making the ‘rotten’ reviews less effective. The best approach would have been a little of each approach. Amazon seems to have gone with the first option, though, and it seems their approach has been rather heavy-handed. Authors and customers alike are crying foul as reviews continue to disappear.
As Joe Konrath notes, Amazon has been busy removing reviews from many, many titles. Inevitably, such a large job requires a high degree of automation. Just as inevitably, Amazon hasn’t revealed how it’s going about this task, but it’s highly likely that algorithms (the voodoo of the tech age) have been employed to spot potential red flags. There’s probably no human element in the decision-making: reviews that trigger too many of these red flags simply disappear. Obviously no system is 100% accurate and there will always be collateral damage. But the accusation being levelled at Amazon right now is that too many legitimate reviews are being taken down, while many dodgy reviews are staying in place. If that’s the case, then something’s clearly gone wrong.
The problem with people who try to work the system is that when the system changes, they’ll just look for another method. Rather than continually changing that system, Amazon might be better off trying to cultivate a reviews ecosystem in which more people are encouraged to leave honest reviews. At the moment, many authors complain that getting a genuine review from a customer is almost impossible, which means ‘corrupt’ reviews can make a real difference. If more customers left reviews, the value of a ‘corrupt’ review would be lessened. How Amazon can achieve this isn’t clear, but simply pulling reviews left, right and centre is clearly a failed policy.