Eighty-Twenty

Give me a fruitful error any time, full of seeds, bursting with its own corrections. You can keep your sterile truth for yourself.

The quote above comes from Vilfredo Pareto, an Italian economist and sociologist from the early 1900s. He is perhaps best remembered for the "80-20" rule. He studied the distribution of real wealth in early twentieth century Italy and discovered that eighty percent of the land was controlled by twenty percent of the population.

A generalized rule followed from this observation: eighty percent of an effect is usually accounted for by twenty percent of the set of causes. For example, 80% of the revenue of a business is generated by 20% of its products. Or 80% of the time to manufacture a widget is taken up by 20% of the set of processes required to manufacture it. A company can use a Pareto distribution in each case to identify its most profitable products or improve its most inefficient processes.

 
This general law of statistical distribution, which seems abundantly obvious to us now in hindsight, has become a widely used tool for things like root-cause analysis and studies of economic and social phenomena. In manufacturing root-cause analysis, for example, it is found that eighty percent of production errors are caused by twenty percent of the process. A Pareto distribution can be used in this case as a prioritization tool in order to focus on and amend the errors with the highest severity.

Pareto also contributed to the study of the (micro-)economics of consumer choice by forwarding theories on utility maximization and indifference curves.

The 80-20 rule, by the way, is another example of hindsight bias. The rule seems so patently plain to us today because we have assimilated its wisdom over time. As a matter of fact, it almost seems like an insult to the intelligence of the readers of this blog to describe it in any detail. "I knew that, annoying cretin," is the average reaction I suppose, "As a matter of fact, I could have discovered it myself if this Pareto bloke hadn't beaten me to it."

Perhaps. But then again, perhaps not.