GuerrillaMaths 2#
Piero Giacomelli
CIO | Group IT Manager at Fidia Farmaceutici SpA | Packtpub Book Author | Manning Publication Reviewer | Passionate about math
Today I am going to talk about another example of #GuerrillaMaths, on a subject that on the internet has often found rather curious drifts, also due to the fact that it is often presented as a mathematical principle when in fact it is an empirical observation.
The empirical observation first formulated by the sociologist Pareto in the early 1900s can be formulated in modern terms by saying that when observing a physical phenomenon of any kind, one notices that 80% of the effects are due to 20% of the causes. In purely numerical terms, this means that taken any set of numbers and ordered in descending order, the top 20% contributes to 80% of the result. This phenomenon has been observed in so many areas and in so many fields that it has in fact become a standard for evaluation. For example, if we want to tie this example to an organisational domain, we can see that 80% of sales are contributed by the top 20% of customers, the top 20% of suppliers contribute 80% of procurement expenditure. More disturbing is the fact that if we measure the performance of a work group numerically, 20% of the top performers contribute 80% of the performance of the group as a whole. Or if we move into the sociological domain we have the following examples:
In process optimization it has been observed that reducing the top 20% of inefficiencies improve by the 80% the efficiency of the whole process. Moving from this observation one can easily understand that for example in sales process it is better to focus on the 20% top customers to increase the sales instead of focusing on the total number of customers and, vice versa, focusing on the least bottom 20% of the customers have little to zero improvement on the sales.
Are you a PM? If you archive the 20% most difficult tasks (defined numerically with a certain weight) you have a probability to improve the achievement of the whole process by 80%.
Pareto's law has also a fractals behaviour, meaning that for example, a market in which 20% of the listed companies generate 80% of the stock market value, and in turn 20%(of that 20 per cent) generate 80% of the capitalisation and profits.
So, next time you have to evaluate a numerical dataset you could order it and find the 20% highest number and you will have a general clue on the 80% of the dataset. For the willing reader I leave some references that show how to apply in practical situation this law:
领英推荐