Thursday, June 25, 2015

The Pareto Analysis Fallacies

This is in continuation to my previous post on the the dilemma of multiple objectives where we had talked about pareto analysis as one of the possible solutions......

Tuesday, June 2, 2015

The dilemma of the multiple objective functions

"To be or not to be?"

Through my various heated discussions with my colleagues / ex-colleagues (who I love to death - pun intended) and at times even at a personal level with my family members, what I have learnt is questions that lead to the highest amount of deliberations are the ones that have multiple possibilities and hence in all probability do not have cookie cutter answers.
Here is an example to illustrate what I mean further:
Assume that you want to answer the question -
"Should I quit my job?" - Perhaps not the best example -but hey if it were useful to you then what the heck.
Typically, here are 4-5 things that one would consider (this is not an exhaustive list, I am just using this to illustrate an example). They could be
1. Drag yourself to office
2. Do you feel like you are banging your head against a brick wall explaining things to your manager?
3. Instances when you have walked out of 1on1s totally confused about the next steps.
4. Instances where even after repeated requests have not got any proper feedback
5. Do you often feel hungry right after lunch because the office food  sucks?
6. Instances where you have found yourself having to explain yourself more than you would have liked. In the same meeting.
7. Instances where you have caught either yourself or your colleagues cribbing about the way "things are being run these days"
8. Kept wondering about your growth prospect
9. Generally tired without having done any work at all

etc etc

You get the idea. 
Which among the above points mentioned do you think can actually be dropped? Like is crappy food really that big a deal? 
To remove ambiguities such as these, people use various different kinds of techniques. One among them is the Pareto Analysis. It basically hinges on the fact that most things that matter are driven by the 80/20 rule. 
So  you end up providing weights (lets say on a scale of 10 with 10 being the highest  weight) for the points above in accordance of their perceived importance.  Arrange them in order of how many times they occur in each of the job offers / new opportunities.
  • Arrange the bar chart in descending order of cause importance, that is, the cause with the highest count first.
  • Calculate the cumulative count for each cause in descending order.
  • Calculate the cumulative count percentage for each cause in descending order. (Percentage calculation: {Individual Cause Count} / {Total Causes Count} *100)
  • Create a second y-axis with percentages descending in increments of 10 from 100% to 0%.
  • Plot the cumulative count percentage of each cause on the x-axis.
  • Join the points to form a curve.
  • Draw a line at 80% on the y-axis running parallel to the x-axis. Then drop the line at the point of intersection with the curve on the x-axis. This point on the x-axis separates the important causes on the left (vital few) from the less important causes on the right (trivial many).
Basis this, if you get lets say points 2,3 and 6 above. You could then decide whether you could change things there or if it is hopeless. If it is hopeless you have your answer :)

BTW there is a fallacy in the pareto analysis. More on that later, for now let me go figure if I really want to buy that 52 inches 3D TV ;)