More than 2400 people played the Pareto Game last week and we got some interesting results. This post provides my responses to each of the six questions that were posed. If you have not played the game yet, please stop reading this post & play the game at https://benchmarksixsigma.com/blog/let-us-play-pareto/. Once you play it, reading this post shall be more enjoyable and hopefully, a better learning experience.
Question 1 – Is it necessary that the categories used in the Pareto Chart are mutually exclusive?
A) YES, categories must be mutually exclusive.
B) NO, I do not worry about exclusivity as I just need some meaningful direction.
For this question, one can argue in favor of either side. My answer is option A) Yes, they must be mutually exclusive, especially if severity of the problem is high. You may go for option B) while using it as a preliminary analysis tool, as you may not know initially if categories are really mutually exclusive. It is understandable that we just want to begin with some numbers and direction but, while using option B) we should be aware of the risk which we are exposed to. In any case, we must avoid making the kind of logical mistake highlighted in the image above. In this example, the four reasons are not exclusive and actually mean just one.
Question 2 – While solving a business problem, the best way to work with a Pareto Chart is to plot number of occurrences (frequency) and corresponding cumulative percentage on the Y axis. Do you agree?
A) YES, the number of occurrences should be used on Y axis.
B) NO, instead of that, I prefer plotting a business impacting output like cost on Y axis.
My choice is B) – I prefer plotting something like risk, cost or revenue on the Y axis, if that is possible. 80% of occurrences of causes may not have 80% of the business impact. In other words, majority of occurrences do not always create majority of business impact. In the example shown in the image, it is easy to imagine that key card malfunction or delay in newspaper delivery are issues that guests have learned to live with, at several hotels. For such issues at most places, people may quietly lodge a personal complain without making much fuss. While these high recurrence problems need to be addressed, they do not classify as highest priority, especially when something like presence of insects is waiting for attention.
Question 3 – Pareto principle is universally applicable to every problem. Do you agree?
A) YES, Pareto Principle applies to each business problem without exception.
B) NO, it may or may not apply. It is not a universal principle.
My response is B) – The principle may or may not apply. For the shooter shown above, 80% of the mis-hits are due to 80% of guns. While categorizing in this way does not help him find a solution, it should tell him that type of gun does not seem to be a factor in accuracy of the outcome. So, when the 80-20 principle does not apply, it might tell us something about low likelihood of influence of the considered factor (here, gun type) on the outcome.
Question 4 – Repeated application of the principle – For a given problem, Pareto is not a principle meant for one time use. One may like to identify the most dominant categories from a high level pareto analysis and then IN MOST CASES, it is possible to carry out a next level Pareto analysis on each of the dominant categories. Do you agree?
A) YES, I agree with this and I find “more than one level use” practical in MOST cases.
B) NO, I find “more than one level usage” practical in very few cases.
My preferred choice is A, I find “more than one level use” practical in MOST cases. but even option B cannot be classified as wrong here. This question relies on your experience with Pareto Principle and depends on several things. If you are detail oriented and wish to dig deep, you are likely to see more opportunities. If you are a member of senior management or manage MIS, with access to data at various levels, you have more possibilities of doing Pareto Analyses at multiple levels. In the Lean Six Sigma world, many of the drill downs from big Y to small Y are done this way. The usage also depends on type of data. If your data is such that can be split in mutually exclusive categories repeatedly, you are likely to have opportunities for using Pareto multiple times.
Question 5 – The categories that end up as low priority as a result of Pareto analysis can be safely ignored. Do you agree?
A) YES, low priority categories can be safely ignored while solving the problem.
B) NO, low priority categories cannot be ignored in all cases. It really depends.
My answer is B) It really depends. Let us consider staple goods. These are daily need items providing low margins to retailers. If a retailer checks the contribution to profits, he may feel the urge to remove staple goods from his home store. However, if attractive staple goods prices bring bulk of customers to the store, their presence is vital to business success. They also need high priority and focus all the time. In this example, it may be affordable to do away with some slow moving items which provide the highest margins. In other words, what is low priority in one Pareto Analysis may be high priority in another Pareto chart. Multi-perspective analyses are needed sometimes.
Question 6 – The 80-20 rule is not an absolute rule. It can be 70:30 or even 90:10 – the total must remain 100. Do you agree?
A) YES, the total must be 100.
B) NO, I do not agree. The total may be a number other than 100.
The answer is B). The total does not have to be 100 as these percentages are calculated from different sets of data. In the above example, 98% complaints (49 out of 50) came from 10% (1 out of 10 outlets) of the outlet categories.
The data and Pareto Chart for this situation is here –
If you have liked this post, do share it with others using one of the share links on the left side. Feel free to comment and share your views. To comment please click here.