This post is is inspired by Tim Rahschulte and his LinkedIn discussion in the Program Management Academy group.

Some of the practices that have worked best for me are:

1)  Educate people early to appreciate (if not understand) that a big project and a program are two different things. There should be at least some delta in the benefits metrics you're measuring because of that. 

2)  Do it, whatever it is, only if the benefit can be described in tangible, non-political terms. If "because it sounds good (ego)" is the only reason to do it, it probably should not be done.

3)  Recognize the mole hills that will become mountains and deal with (perhaps even laud) them as successful learning opportunities.

4)  Appreciate the current speed at which you are traveling. There is an optimal point between managing to the "dumbest common denominator" and the soldier's creed of 'leave no man behind'. It's important to regularly assess whether it is still optimal.

These four things won't guarantee success but ignoring them will almost certainly refute the opportunity to achieve it. I (like many a reader here I'm sure) have seen way too many really good strategic initiatives become question-marks and dogs from stars and cash cows by sponsors and other clouted stakeholders who make it about personal egos rather than corporate benefits.