[Casey]

Every experience I have had with AHT is that in its basic form, looking at it from a per call basis and not an average of averages it is un-normal data skewed to the right.  On this note if you see a shift in your variation you will also see a shift in your mean since the median is your central point and your mean is not.  
Where you choose to set your goal should depend on why you are even addressing AHT.  If it is from the financial standpoint then you should use the mean since it translates into dollars, but be careful to remember that changing variation here will change your mean.  If you are pursuing this for other reasons such at Customer Satisfaction with completion time of call resolution then you should pursue variation.

0

[Casey]

FF,
I am curious on how you would use a Chi Square test for a problem like this.  It seems that he is focusing on continuous data and I have always seen the Chi used with discrete data.  Are you recommending a change of the data before such a test?
It is an fascinating choice of test for this type of data and I am interested in finding out how it would work in this scenario.

0

[Casey]

Motorola (aka Mikel Harry) does not have data showing that processes shift 1.5 standard deviations. If anyone can produce that data, I will of course, retract my claim and apologize. Harry’s original book describing the shift was based on two references, both of which show merely a theoretical calculation, and a simulation. He does not state the assumptions of his simulation, which could have been made to produce any shift desired. In addition, his simulation also showed a 40% increase in variability. (And you thought that the nonsensical phrase Dynamic Fluctuation only made the mean shift!). Somehow this shift never entered the scale.
Those familiar with control charts, know that there are two sources of variability: common cause and special cause. Which is this Dynamic Fluctuation? Why neither! Harry’s discovery of another type of variation is muted. Perhaps because the first two are exhaustive and mutually exclusive, this discovery is not a real one.
Can you imagine what kind of study would have to be done to show that a change in a process was not due to special cause variation? I can not.

0

[Casey]

Jack Welch never won squat in the front office.  Championships are won on the field.  The cost of front office operations is not directly correlated to team performance.  Commit to buying good players and wins will follow.  It’s a quick hit baby!

0

[Author]

Posts