Last Updated on 9 September 2023
Pareto Charts, also known as Pareto analysis, are a fantastic way of visualizing data (especially for defects), and prioritizing how you tackle them.
Pareto Charts are named after the Pareto Principle, where (approximately):
80% of effects are due to 20% of the inputs
The idea is to find the 20% of issues, which are causing you 80% of the problems; the Pareto charts are bar charts which let you quickly see major sources of issues – this enables you to focus your efforts on the biggest problems.
The Principle
The Pareto Principle (or Pareto Effect) is named after economist Vilfredo Pareto, and says that there is an unequal relationship between causes and effects, specifically 20% of causes will usually cause 80% of the effects. It is supposed to apply to all sorts of situations in all areas; Pareto himself used it for showing that around 20% of Italians at the time owned 80% of the land, but it applies to e.g. income distribution, effort in business and causes of business issues. I really enjoyed Richard Koch’s book ‘The 80/20 Principle – The Secret of Achieving More with Less’ on the subject, which helps you to apply it to all areas of your life.
Using Pareto Analysis in Lean Six Sigma
You can use the Pareto Principle to greatly magnify the advantages of Six Sigma:
| 80% of: | Will be 20% of: | Advantage |
|---|---|---|
| Benefits of Lean Six Sigma | Projects | Prioritizing will bring the benefits a lot faster |
| Defects | Root causes | Tackling them in order will have cause huge improvement with relatively low cost and effort |
| Stock usage | Stock items | Can get huge efficiencies from 5S organization |
| Stock value | Stock types | Makes it a lot quicker to cut down on inventory waste |
| Stock volume | Stock types | Lets you quickly cut down on space waste |
| Issues | Locations | You won’t waste resources improving processes in strongly performing sites |
| Customer complaints | Issues | Pareto Analysis will help you to quickly analyze voice of the customer |
The time when you will most frequently actively use the effect is usually the last of those – eliminating defects. If you create a chart of all your defects and root causes (a ‘Pareto Chart’), you can quickly find which causes need to be tackled in order to significantly improve your Sigma level.
Creating a Pareto Chart
Pareto charts are essentially bar charts, except instead of a scale on the horizontal axis, they represent categories. Creating a Pareto chart is relatively simple; the process is:
Data collection
Collect the details of a large number of defects, such as by using check sheets. The larger the number the better but this can be updated as you get more details. you could have months of data on the graph.
Categorize the data
Categorize the causes into core categories, making them broad enough that you have a sensible number of categories (preferably 10 – 20)
Rank your categories
Sum over for each category the number of times that cause has occurred in the data. Rank the causes of issue in descending order of frequency, so your ‘worst’ category is at the start.
The last small categories can be grouped into an ‘other’ category to save space; this will make the chart a lot easier to interpret. If you manage to resolve the larger issues so that they go away, you can break these out later.
CARE – not all issues are created equally!
A defect that leads to a customer cutting their hand off is a lot more serious and costly than one that leads to the product being the wrong color.
If the defects being considered are of very different nature to each other, you might get more use adding them up in a different way, such as:
- Cost to the company
- Time taken to fix the error
- A risk factor showing the potential effect if left
If you rate each defect with these, you can then sum over that instead by category, in order to make your Pareto chart.
Alternatively, only categorize issues that the customer cares about, using CTQ (critical to quality) issues only.
Create a bar chart
The pareto chart is a bar chart where the x (bottom) axis is the issues categories; either defects or causes of defects/errors. The a y axis of number of defects, so the totals from step 3.
Cumulative %
Add a cumulative line for % of errors, so that you can see when you reach 80% (or whatever you have as your significance line)
Interpret the results
There is a Pareto effect if the 80% threshold is hit with only a few categories (generally, less than 30% of the categories), and there is no Pareto effect if the 80% involves a large proportion of the categories.
With a Pareto effect
If there is a Pareto effect, you can focus your efforts on this small number of issues, and with relatively little effort (compared to tackling all issues), you can remove 80% of your issues. This will make your progress a lot quicker than if you just start working through your issues in a random order.
No clear Pareto effect
This could mean either that your process improvement efforts will have to be more far-reaching, or simply that you need to refine the categories that you’ve used; you may need to group your data better.
It could mean that you are measuring the wrong thing. If instead of instances, you change to cost of issue, or time taken to fix issue, you’ll see that a different method of measuring the issue will show a Pareto effect on how the issues are affecting the business.
Leave a Reply