Report Offensive Message

The single biggest factor in the success of projects that I've been on is an obsessive desire to limit requirements. The absolute first thing you need to determine is what business need needs to be met, and for every single "Hey, can the system do this?" question that comes up, force the asker and the decision makers to explain why that "feature" is absolutely necessary to meet the final goal, or how the time/risk involved in adding the feature is outweighed by the benefit of the feature. This approach weeds out a lot of dumb ideas, and results in leaner, smoother processes and products that "just work" the first time. Every additional feature is just one more thing that can go wrong, so be vigilant in ensuring that whatever system/product/process you're working on is no more complex than it needs to be.