In my last post I presented some evidence showing how easily customers can end up being faced with self-published works with questionable reviews. That post answered arguments that deny the existence of such reviews or their visibility. This post deals with another "It isn't an issue" argument that comes from a different angle.
Customers don't use reviews as guides. They're just not important.
This is easy enough to counter in the absolute sense because I use reviews and I'm a customer. I do exist -- philosophical beard-stroking arguments aside. The broader implied argument that not enough readers use reviews to make these issues important is more difficult to deal with because "enough" is a slippery concept. I suspect that most people making this argument are just generalising from their own opinion, rather than basing their argument on anything solid, but I'll give them the benefit of the doubt and lay out some reasons for believing that a significant number of readers do use reviews.
As you might imagine, a lot of the information concerning customer habits is of commercial interest and that means that those in the best position to have hard data (Amazon, B&N, etc.) probably aren't inclined to share it. However, you can certainly find some research that has relevance. Here's a poll about Australians using reviews to decide whether to deal with a company, for example. Here's a paper on hotel reviews that finds a link between sales and reviews. Here's a page listing headline quotes from marketing research about customer reviews, pretty much all asserting that they affect decisions and sales.
Perhaps you are unimpressed with the above. You don't think they're relevant because they're not about books or you just don't trust marketing companies to be honest about marketing (I don't blame you). How about this then? Here's a paper from Yale's School of Management that specifically looks at book reviews on Amazon and B&N and finds that "an improvement in a book’s reviews leads to an increase in relative sales at that site". In terms of research data from an academically objective source and specifically about the topic I don't think you're going to get any better. I'm certainly not going to look any further.
But maybe you still aren't happy. You don't trust any of the above for whatever reason. Okay, how about a plain argument based on the evidence that's in front of your eyes every time you go to Amazon or B&N or pretty much any online retailer? Here goes.
Space on a web retail site is very much like window and shelf space in a bricks-and-mortar business -- you don't just throw it away on stuff that doesn't work for you. The very fact that user reviews are still present on the most mature and competitive commerce sites is evidence that those businesses see value in them. How much value? On Amazon, every product has the star rating right next to it and the ability to sort both by Top-Rated and Average Customer Review is provided for most listings. At almost every turn Amazon wants to show you a rating or a review or get you to give a rating or review. If you review a lot, Amazon rewards you with badges and free stuff. In short, Amazon puts cash-money behind maintaining, improving and promoting the review system. That would seem to suggest that it matters to some people.
Even if you don't believe Amazon's motives are anything to do with customers using reviews (odd, but I'll run with it), the reviews are there. People have written reviews (unless you believe every review is fake) and people respond to those reviews with comments and ratings on the reviews. Hell, there are reviews that mention how they were duped by reviews they now believe are fraudulent.
And there I rest my case. If none of this is enough for you to think that maybe a decent number of customers actually look at reviews to at least steer them towards or away from further investigation of a book then I have to congratulate you on the depth of your convictions. If you're now shifting your objection to "Yeah, but all this proves is that customers shouldn't trust reviews. They should read the samples." then don't worry, I'll get around to that shortly.