Development and Termination


It may see obvious, but the first thing you should do when starting a product development project is to define clearly when you will be done!  Just what, specifically, is the project expected to deliver at the end?  If there is to be acceptance testing of some form, who will do it, when and where.  You might give serious consideration to beginning the project by writing the acceptance test!  That would give the project a clear and specific goal to accomplish and would help guard against creeping expectations and specifications.  As a practical matter this is not easy to do and circumstances may prevent it, but it is worth getting into that frame of mind and doing as much of this as is possible.  It can save a lot of misunderstanding and confusion.


When a development project starts, there is enthusiasm for its objectives -- or the project would not have been started.  What follows as the project progresses is an "enthusiasm curve" which typically looks like this:


As the project get organized and going, enthusiasm tends to build in the excitement of something new.  But as the real work begins and progresses, the natural problems will begin to surface and all the issues and hard work become visible, so enthusiasm begins to fall.  As succeeding problems are encountered there will continue to be falling enthusiasm until some point where enough problems are overcome and new ones appear less often, with the result that enthusiasm will begin to rise again, hopefully to a successful conclusion.  Of course, some projects will just keep going down and will collapse of their own accord, but most will have some sort of minimum.  Lucky (or well-conceived) projects will have a shallow minimum and not much concern will arise at the bottom.  But many projects will have a significant minimum or will be in danger of continuing to descend.  It is at this point -- the minimum or the continuing low point -- that you will find out whether you are a good manager of science and engineering.  Project evaluation and the actions that follow at these critical points is the hard part of management and a real skill and is where you will earn your salary and continuity.


First and foremost, you need to determine if the project is in trouble for the right or wrong reasons.  All projects will have problems of one sort or another.  Is this project having normal, natural problems of the kind to be expected in similar projects?  Or is this project having an abnormal problem such as violating its original assumptions or evaluation criteria?  In the former case, the project is likely to be healthy, but perhaps needing some additional resources or, perhaps, just a little patience.  In the latter case, serious consideration needs to be given to reconsidering the project at the most basic level, including considering cancellation.  In my experience, projects too often ignore or forget the assumptions made and the criteria established at their beginning.  They need to be reminded of both on a regular basis, even if the project is not in trouble.  But when trouble looms is the most important time to bring up these issues for examination and to determine if they are a surprise to the project personnel or if, indeed, they believe they are in compliance.


And, finally, it needs to be said that killing a project that is off the rails is not easy or fun to do, but should be regarded as a mercy killing.  In they end it will be best for both the personnel involved and for the organization on whose behalf the project was being carried out.  Needless to say, if a project is going to be killed, doing so as early as possible is better for all involved, so don't put it off or procrastinate once it is clear that something basic is wrong.


On the other side, however, is the need for forbearance and patience for those projects which are having "normal" troubles and which are not in violation of basic criteria.  They are fighting their way through the task and have not found anything which negates the value of the project, so continued effort is justified.