Pareto’s Principle, the 80-20 Rule, and the “Vital Few and Trivial Many Rule.” Is this rule valid today? For example, do 20% of SDMB members contribute 80% of the value to SDMB?
I have been in Computer Science “quite a few years” and “80-20” rules pop up all over the place. In fact, you’re lucky if the split is “90-10”; it’s usually even more unbalanced. 10% of the users consume 90% of the resources, 10% of the files use up 90% of the space, and on and on. It’s a good “guesstimate” value to use in many situations to decide on a general approach to a solution, but to really fine tune you have to collect data.
2% of the records of one database account for 98% of the accesses is an even more unbalanced example I’ve come across.
I first heard about it in the early days of Ethernet. Back then, it was pretty close to 80-20 on bandwidth.
The “Pareto Principle” is widely misused in common speech and certainly beyond the bounds intended by the author who referred to an observation that “80% of the wealth is controlled by ‘about’ 20% of the population.”
Even within this limited scope, usage of the 80/20 rule is often off the mark. In some nations without a significant “middle class,” greater than 90% of the wealth is controlled by less than 2% of the population.
Other times, the “51-49” rule would hold or almost any combination summing to 100.
It seems that the usefulness of this rule is in identifying the truth that in many situations resources are not distributed uniformly, but to specifically ascribe fixed percentages to an untested population does not seem so useful.
I find the 80/20 rule is only applicable about 4/5 of the time
I never realized that this principle had a name.
It sounds like Pareto’s Principle holds for any case in which there is a large variation in the utilization of resources. I’ve never understood the indignation of someone pronouncing, “The 10% wealthiest people control 90% of the wealth.” Well, how much do you *think *they should control?
It strikes me as a simple principle, but if you put it down on paper first I guess you get to name it.