We analyze the flat rate and bundle pricing nature of the Amazon Prime subscription. Based on two customer surveys (n=2062, n=906) and using a best–worst scaling (BWS) discrete choice experiment, we determine how important the different Prime benefits are to the subscribers. We perform a correlation and cluster analysis. We find mostly very weak correlations between the importance scores of the different Prime benefits. This is beneficial for a bundle pricing model such as Amazon Prime. We also find that the importance scores of the individual benefits are only weakly correlated with the actual usage rates of the benefits. This is beneficial for a flat rate pricing model such as Amazon Prime. Furthermore, we see indications customers value having various benefits included in Prime, even if they do not use them. We also discuss how Amazon Prime fits into the business strategy of Amazon.