# Separate the Vital Few From the Trivial Many

## A Pareto diagram can help you decide which improvement efforts to make first

by Melissa G. Hartman

This article was featured in January 2016’s Best Of Back to Basics edition.

Pareto diagrams graphically depict categorical data. A Pareto diagram is a bar graph, and each bar represents a category. The bars are rank ordered in descending order from left to right. The bar on the left represents the category with the greatest value, and the bar on the right the category with the least value. The tallest bar is always on the left and the shortest on the right.

Pareto diagrams are based on the principle of separating the vital few from the trivial many. This Pareto principle was developed by Joseph Juran, based on the work of Italian economist Vilfredo Pareto (1848-1923).1, 2 Pareto determined 85% of the wealth in Milan was owned by 15% of the citizens. Similarly, Juran observed that a majority of organizational effects resulted from just a few causes.

Today, the Pareto principle implies 80% of process problems can be accounted for by 20% of process factors. The premise behind the diagram says process improvement efforts will be more effective if the categories on the left (the vital few) are addressed first. The rank ordering of categories simplifies detection of the vital few.

Follow these simple steps to construct a Pareto diagram:3

1. Decide which categories will be used for grouping, such as the quantity of each type of lab test or reasons for patient falls.

2. Decide how you will measure the categories. Common measures include the number of occurrences in each category or total cost of the occurrences in each category.

3. Gather data and place them into one of the categories.

4. Enter the categories, counts and cumulative frequency into a table or spreadsheet. This is a summary step that prepares the data for the Pareto diagram. The data in Table 1 depict the occurrence and cost of customer complaints at a small hotel. The seven categories of complaints are in the first column, the number of complaints in the most recent month is in the second column, and the approximate average cost of resolving each complaint is in the third column.

5. Develop the diagram. Figure 1 shows the categories of complaints in descending order. The number of complaints is on the left axis. A cumulative percentage distribution--the line going upward from the first category--is scaled on the right axis. This shows the proportion of the total number of complaints accounted for as each successive category is added.

6. Interpret the diagram. The most frequent complaints shown in Figure 1 are that the room is not stocked or is dirty. These two categories account for about 59% of the total complaints. An appropriate action in response to this information might be to investigate housekeeping practices and identify opportunities for standardization and improvement.

However, if your goal is to reduce the total cost of complaints, a Pareto diagram can provide a different perspective. Calculate total cost by multiplying the cost of each complaint by the number of occurrences. According to Figure 2, the most costly complaints are the rooms are too noisy or are not ready when the guest arrives. This diagram requires a different type of response from the one indicated by Figure 1.

Pareto diagrams are useful in situations where there are categorical data. The diagrams are easy to develop and provide powerful insight into organizational problems.

### References

1. Joseph M. Juran, Juran on Leadership for Quality (New York: The Free Press, 1989).
2. James R. Evans and William M. Lindsay, The Management and Control of Quality (Cincinnati: South-Western College Publishing, 1999).
3. Nancy R. Tague, The Quality Toolbox (Milwaukee: ASQ Quality Press, 1995).

Melissa G. Hartman is assistant professor of management at Baker University School of Professional and Graduate Studies in Overland Park, KS. Hartman is an ASQ Fellow and certified quality manager, quality engineer, quality auditor and mechanical inspector.

Out of 0 Ratings