Post-assessment process improvements need to be driven by senior management to succeed. There is no substitute for a managing director’s or a president’s authority, commitment, and resources.
Senior management, however, must be not only committed but also informed. Senior managers need to learn what questions to ask and how to ask better ones as the effort progresses. In addition, a management structure needs to be created that will extend the reach of senior management without diluting its effect.
On the other hand, the plan needs to provide for initiative and modification by the managers and engineers who are responsible for implementing it.
Senior Management: The Push from Above
No improvement effort will get beyond empty words unless senior management commits its authority, effort, and resources. This means more than establishing an SEPG or process improvement group to "do what they can." It involves taking responsibility for leading the effort, vesting real authority in subordinates who are directly responsible to senior management, and learning enough about the logic of the process (and its payoffs) to keep a personal thumb on the pulse of change.
Senior managers must not only lend their authority to the improvement effortthey must lead it. This requires learning about process improvement and asking the kind of questions that elicit useful answers.
Accepting a responsibility for process improvement amounts to increasing senior management’s own process discipline in an area, that is, software, in which few managers are expert. As Pat O’Toole has accurately observed, "If all your senior management does is sponsor the effort, it’s probably doomed anyway! I believe that senior management must lead the charge by exhibiting the same process-disciplined behavior that they expect the troops to adopt…Unless management is willing to align their own behavior with the direction they are trying to establish, it is simply another case of ‘do what I say, not what I do.’ The commitment of senior management to establish a process-disciplined organization is reflected much more clearly by their own behavior than it is in a policy statement lifted from the pages of the CMM/CMMI" .
Two Time-Honored Mistakes
Mistake 1: Appointing a Process Improvement Manager Who Has No Real Authority
All too often, senior management begins an improvement effort by appointing a process improvement manager and instructing him that he (or she) is responsible for moving the organization to the next maturity level by a certain date. The process improvement manager enthusiastically takes on the assignment but soon discovers that his decisions are likely to be overruled whenever there is a more "critical" goal to achieve. If the PI manager is not on the senior management staff or has not been given the authority of the senior manager, he cannot hope in an immature organization to be anything but the fall guy when the organization does not reach its desired quality maturity level. Conflicting business decisions always threaten to take priority over an improvement effort, and only senior management can ensure that improvement stays on track.
Mistake 2: Hiring a Consultant to Make Everything Better
A second common mistake involves farming out all or part of the improvement effort to a consultant entirely outside the organization. Here the distance between the improvement effort and the ordinary business of the organization is even more pronounced.
At its most basic, process improvement means articulating and refining existing processes. It often seems that these can be produced (for instance, by SEPG groups) without involving developers, but this is not the case. Developers are unlikely to accept them if they are not involved in their creation.
Organizations change only when they agree about which parts of their current procedures work and which do not. This must come from an internal perspective, or else it is seen as alien and either silly or threatening.
Hiring consultants produces shiny new processes that no one is ever likely to take off the shelf. Working out new processes within the projects themselves is more cumbersome, but it leads to a genuine basis for real improvement.
As Stan Rifkin has observed, unless an organization knows "why" it is adopting new processes, it never gets the "how" right.