Tony Rizzo wrote:
> Your observations are correct. But the solution that you mention does
> not
> apply to the problem that I describe. The technology is not lacking.
> It is the understanding that is lacking, in management.
>
> So, my question remains. If the problem is that most managers and
> nearly all executives know nothing about variation and effective
> methods for managing it, then what is a workable solution?
As always, the question is workable to whom, acceptable to whom.
If the managers are in an organization that does not `recognise'
variation -- and they don't worry about it themselves -- then
there seems like there is little pressure to change. What's in it for
them?
In other words, you need to spell out the problems and politics in
more detail before anyone else can usefully comment.
If you are in charge, a top down solution might start by demanding
that planning is done in terms of best/average/worst case estimates
(i.e. using the sorts of tool I mentioned). If the managers are forced
to estimate variability and talk about it, then they may start to become
more interested in it.
A bottom-up solution that is often applicable, at least in the s/w world,
is incremental delivery. Given three external constraints (time, quality
and content) at least one has to give if there are problems. It is often
better to [at least internally] plan progressive additions to
functionality.
So when an [arbitrary?] deadline arrives you meet time & quality, but not
necessarily full content. [Indeed was the specification detailed enough
for anyone to argue that it hasn't been met?]
Putting another spin on it. You commit to delivering a core functionality,
and strive to deliver most of a prioritized wish-list. [If you were unable
to commit to a core functionality that would be minimally acceptable,
then you shouldn't have committed at all.]
Peter
=================================================================
Instructions for joining and leaving this list and remarks about
the problem of INAPPROPRIATE MESSAGES are available at
http://jse.stat.ncsu.edu/
=================================================================
