3. Don't fix imaginary impediments
It can be tempting, if you "know" you are going to experience a given impediment, to try to come up with a solution before the problem itself ever arises. However, to do so is unnecessary and runs the risk of sending you in the wrong direction. Scrum is an empirical method and, therefore, has frequent inspect-and-adapt cycles built in to deal with just such issues as they arise.
A good example of this is documentation. Many organizations feel their documentation processes are too bulky for Scrum and want to "create a new system for documents" before starting their first Scrum project. Instead, I recommend they start Scrum and continue until the point in the project where they would create a certain document (say, a business requirements document). At that point they can decide how that document should be revised. They can then test their approach in the next sprint and get immediate feedback on its effectiveness. In essence, they are doing "just in time" process change and, because of that, are likely to make better choices.
4. Avoid creating world peace
When I visit organizations to provide coaching on Scrum adoption, they often present me with a laundry list of problems and issues that need improvement. They want quick and instant answers to tough, complex problems. I caution such organizations to avoid "creating world peace," meaning, trying to wrap every problem in their organization into a big tangled ball-which they then try to "solve." To do so overwhelms everyone and the answer quickly becomes, "You can't get there from here."
A better approach in the early days of Scrum adoption is to focus on incremental improvement. The advantage of an empirical approach is that there is no need to solve every problem up-front. Rather, Scrum allows companies to make use of the inspect-and-adapt cycle to try different solutions to problems, allowing the best choice to emerge.