Friday, October 28, 2011

Fixing Amazon with 6th Grade Math

I've read a number of posts over the last month or so where people complain about how pretty much everything on Amazon has a 4 or 5 star rating despite how good a book actually is. A lot of people are giving up on the rating system, and actually wading through all the reviews to find out if a book is good or not. Without an effective means for people to sort out what books are good, the reading public will abandon mid-listers and especially those Indy/Self-pub writers for whatever the New York Times or Oprah recommends.

But I say I still have faith in crowd sourcing...even if people are far too generous with their ratings. And I say lets bring some simple 6th grade math to the rescue.

First, we need to understand that the problem is people who give a 5 star rating to pretty much everything they read and only give a 4 star if they hated it...you know, don't want to hurt the author's feelings. I don't suggest trying to change the people who are giving the review; however, what I do suggest though is that we start to weight the reviews from each reviewer based on their history.

This is what I'm talking about: Say Mary Sue rates literally everything she reads 5 stars because she is in awe of anybody who puts pen to paper. Right now, she would tend to skew books up towards 5 stars, when in reality, 5 stars is essentially "average" for her. What I propose Amazon should do is take that info and weight Mary Sue's ratings so that when she gives say my book (whenever I get one published) a 5 star rating, Amazon actually computes it as a 3 star review.

The same thing would go for Dick Cheney when he is on Amazon giving everybody a 1 star review, then goes and reviews my book (Again giving it 1 star even though it may have reached in and touched him where his heart may have been.) Amazon treats that review as a 3 star as well.

Now if Joe Smith is running around giving 1, 2, 3, 4, and 5 star reviews pretty evenly, well then when he gives my book a 5 (or 1) star rating, it actually registers as that.

This is not a new idea either. Hotornot.com has used this same basic idea for the last probably 10 years. And when you go there (yes, admit it, you visit that site from time to time) you'll notice that there is a pretty good range of ratings that everybody falls under, and never will you see somebody with a 9.0 or above rating that makes  you want to gag. Amazon on the other hand generally has a range of 4 stars to 5 stars with plenty of 5 star (average) train wreck books that would give you a migraine just trying to get past the first chapter.

Amazon, it's time you broke out your 6th grade math books and apply some of those principles of weighted averages to your ratings system.



What do you think? Is the ratings skew on Amazon a problem? Why or why not? What do you think can be done?

3 comments:

B.T. Hoskins October 28, 2011 at 11:59 PM  

I spent so much time selecting only certain reviewers for just that reason. That's a very smart idea, but I wonder how many people would get annoyed that their opinion didn't register the way they meant for it to? Only problem I could see coming out of that, but I'm sure that can be remedied too.

Ciara Ballintyne October 30, 2011 at 7:01 PM  

Brilliant! Absolutely brilliant! Such a simple solution to the problem. I agree something needs to be done.

Andy November 6, 2011 at 5:51 PM  

Maybe I'm just a cynic, but I can't help thinking that people will find a way to game that system too. Like giving all trad books 1-star reviews and all indie books 5-star. Or vice versa. Still, maybe the fact that it would take more effort to do that would at least limit it.

Regardless, I think the ultimate test is the sample pages. Those can't be faked or manipulated. Although I guess having a typo-free first chapter doesn't guarantee that a book will be good.

  © Blogger template Brooklyn by Ourblogtemplates.com 2008

Back to TOP