How will people use the new system? How will it support their business objectives? This is a critical part of the product owner's job - prioritizing the work that needs to be done. One approach is usage-based (how many users will be affected by the change?). Another approach is financial-based (how much money will we make/save/lose?). The most realpolitik approach is to think in terms of the political weight of the request. As you develop your stories/use cases/features, the team should also develop a common language for assessing the priority of stories. This can be a huge motivator for the team, and a great way to clarify focus - nothing like assigning a story a $100,000+ value to make the developers and testers really stress the edge cases.