I wish I had a better understanding of the bimodal distribution of customer satisfaction with Amazon. I know high volume customers who have 99.5% success with

I wish I had a better understanding of the bimodal distribution of customer satisfaction with Amazon. I know high volume customers who have 99.5% success with Amazon orders, and others who would say "pretty much everything I buy is used or fake". First assumption is that it boils down to differences in product types. Certainly some types attract more seller fraud. But I wonder if there's also a region or distribution center variability, like the East Cupcake Regional warehouse does a crappy job of inspecting returns. Or the delivery manager in West Cupcake hires crappy drivers. Amazon must do this analysis internally. I'd love to understand it better.