Iterative change is good, but, not if it is a change to a common pattern that your Users expect, and the change that you want to make is a compromise or work in progress. Don’t experiment with what your Users do most to help you do the homework that you should have done in the first place. if you are not completely sure that the change you are making is right, and you are changing something that the User does consistently and has huge impact on them, don’t make that change. Don’t tweak common patterns to see what happens and get usage data to help you figure out what you should have done in the first place. Your Users are your Customers, not lab rats to experiment on. Actually, I’m not sure we should even be experimenting on lab rats, but that’s another thought. Make sure you are 100% with what you are going to do, before you do it when experimenting with what your Users do most.