One of the problems with feedback systems is that they truly fail to present a balanced view of the user or customer base. The people with the motivation and drive to post reviews and feedback in ecosystems such as the App Store and Amazon are those with problems. And so the rating that is produced isn't a true reflection of how well an app is received or how good a product might be.
We happen to have both a physical product business operating on Amazon and an app business on the Apple App Store. Based on this experience I'll say most apps and products receive feedback from 1 out of 500 to 1,000 people who purchase the product. Put another way, the rating figure is based on the opinions of approximately 0.1% of users, if that. And it is also far more likely that the review-posting population is highly biased in favor of people experiencing difficulties.
In other words, based on my experience, I think these rating systems that consist of a linear vote count equation are deeply flawed as they fail to capture reality in any measurable way.