How much? When?
Too often software development is driven by costs and costs drive schedules. But it is difficult to win when everyone thinks it should cost less and someone, somewhere will offer to do it for less – even if they are not capable.
Given that we know that iterative software development can work much better than ‘big bang’ or waterfall, should we apply the same process to organisational change? How about a backlog + iterations for shifting to a team structure which takes Conway’s Law into account, so that the organisation has a better chance of delivering the software systems it needs?