We found that approval only for high-risk changes was not correlated with software delivery performance. Teams that reported no approval process or used peer review achieved higher software delivery performance. Finally, teams that required approval by an external body achieved lower performance.
