Comments
Diagnosis: Your Data Is not 'Normal'
View Comments: Newest First | Oldest First | Threaded View
Re: Statistical Presentation
  • 1/26/2015 3:20:37 PM
NO RATINGS

I am a big fan of the approach Mack Hannan take to risk in his book "Competing on Value", where he lays out a three-tiered criteria of: How much, How soon, and How certain.  I wrote a series of blogs about this back in mid-2010.  To your point, you can boil "How much" down to one number - the invoice amount, and you can even boil "How soon" down to one number - either the NPV or the IRR.  But my best attempt at the time still required me to represent risk as: one 3x3 matrix, three 1x3 matrices, and two more related numbers - 20 numbers in total.  But I found that you need to have the narrative to go along with it, in order to focus attention on the imporant aspects of the risk assessment - management needs context in order to make a good decision - numbers without context are nearly worthless.

Statistical Presentation
  • 1/25/2015 9:49:58 PM
NO RATINGS

I must be some type of total nerd because your article had me laughing in agreement.  Hurricanes can be outliers and ignore them at your own peril. 

I believe much of the problem is that one can not communicate that which requires two parties to have a language to communicate it in.  And on top of that, managers like yes or no, right or left and done or not yet done answers.  In between answers are not what people are hired for. 



INFORMATION RESOURCES
ANALYTICS IN ACTION
CARTERTOONS
VIEW ALL +
QUICK POLL
VIEW ALL +