>> Tony Rizzo wrote:
>>
>> > The problem is that managers and executives have no practical
>> > understanding of variation or of methods for managing it.
>> > This is particularly devastating in the area of
>> > project management and enterprise management.
>> >
>> > Would anybody here care to sink his/her teeth into
>> > a chunk of reality?
>>
>> A couple of observations.
>>
>> I believe that the first PERT suystem (for the Polaris submarine
>> program?)
>> did use three estimates of task duration. With a few [questionable but
>> probably
>> OK] assumptions about distributions, they could give a pdf for project
>> completion date. I think that some other network planning systems have
>> followed up this idea over the years.
>>
>> For things like maintenance planning, probabalistic SNAFU charts have
>> been used (UK electricity supply I think). General idea: you open up a
>> noisy turbine and there's an x% chance of needing a new bearing which
>> takes N hrs (or opens a subSNAFU chart); there's a y% chance of
>> something even nastier having happened; etc.
>>
>> Peter
>
>
>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?
I suggest you consider looking at the quality literature. David Wheeler, has a
written a couple of excellent books on variation that go beyond the
traditional SPC approach. Also H. James Harrington has some books on Business
Process Improvement that address the subject.
>From my perspective, starting with an understanding of the difference between
special cause and common cause is a good point of departure. Naive managers of
require the identification of a special cause for any variation, however
slight.
=================================================================
Instructions for joining and leaving this list and remarks about
the problem of INAPPROPRIATE MESSAGES are available at
http://jse.stat.ncsu.edu/
=================================================================
