In any given Scrum project, the hardest decision that the team has to make is the order in which to build any given set of features on the backlog. The value of that order is of less importance than other considerations (such as quality, completeness, appropriateness of feature, etc.), but it is still a decision that has to get made. The main reason why it is so hard to make that decision is because it is almost always based on shoddy information! A good Scrum team thrives on data to make decisions, and using real data to make decisions takes a serious commitment. I think there is an approach that can enable a team to prioritize their backlog with real data, will address the above issues, and will help the team achieve some other ancillary goals.