| Pareto (pa-RAY-toe) analysis is named after Vilfredo | | | | the left vertical axis in increments from zero to the |
| Pareto, an Italian economist who lived in the late 19th | | | | grand total of all the problem classifications. Mark the |
| and early 20th centuries. In 1897, he presented a | | | | right vertical axis in increments from zero to 100%. |
| formula that showed that income was distributed | | | | 9. Construct the vertical bar diagram beginning on the |
| unevenly, with about 80% of the wealth in the hands | | | | left with the highest percentage classification and |
| of about 20% of the people. | | | | progressing to the lowest and ending with |
| In a similar way, a disproportionately large percentage | | | | “other.” The height of each bar should |
| of errors or defects in any process are usually caused | | | | correspond with the value or number of occurrences |
| by relatively few problems. Pareto analysis helps | | | | on the left axis and the percentage of the total on the |
| identify those significant few problems so people can | | | | right axis. The width of the bars should be the same |
| target them for action. It is particularly helpful in the | | | | and they should be touching. |
| measure and control phases of Six Sigma | | | | 10. Label the bars under the horizontal axis. |
| methodology. | | | | 11. Beginning at the left zero point, plot a line showing |
| What can it do for you? | | | | the cumulative percentage total reached with the |
| There are so many aspects of work that can be | | | | addition of each problem classification. The line should |
| improved, knowing where to begin is often difficult. | | | | end at the 100% mark on the right axis. |
| Pareto analysis will help you: | | | | 12. Title the chart. It is also a good idea to write a brief |
| • Categorize and stratify such things as errors, | | | | summary telling how and when you collected the data |
| defects, delays, customer complaints or any other | | | | you used to produce your Pareto chart. |
| measures of the resulting quality of your process so | | | | NOTE: The first pass at a Pareto chart may identify a |
| that you can identify different classes or types of | | | | significant problem that is still too big to work on. A |
| problems. | | | | second Pareto analysis may be necessary to break |
| • Graphically display your results so that the | | | | this most significant problem into workable pieces. |
| significant few problems emerge from the general | | | | Now what? |
| background. | | | | If everyone tried to make improvements individually |
| How do you do it? | | | | with no definite basis for what they were doing, much |
| 1. Decide how to classify your problems. This may | | | | energy would produce disappointing results. A Pareto |
| involve looking at the reasons cited for returned or | | | | chart is a useful tool to draw the attention and |
| rejected material, talking to your customer, or | | | | cooperation of all concerned to target the most |
| examining the rework aspects of your process. | | | | important problems affecting quality. |
| 2. Create a preliminary list of problem categories. Try | | | | • Use the Pareto chart as a basis to “divide and |
| to keep this list to no more than six or seven. (You | | | | conquer” problems. It tells you what to work on first. |
| may find yourself modifying this list as you continue | | | | • It is normally easier to reduce a tall bar by half |
| your analysis.) | | | | than to reduce a short bar to zero. Use the Pareto |
| 3. Decide on a time frame or other scope limitations on | | | | chart like a map to “island hop.” Significantly |
| the material you will consider. The time frame or scope | | | | reduce one big problem, then hop to the next. Leave |
| you choose should be representative of the whole | | | | the smaller problems for “mopping up” later. |
| process you are examining. | | | | • Of course, if you find a small problem that is easy |
| 4. Tally the occurrences in each problem classification. | | | | to fix, don’t ignore it. But don’t get tangled in |
| (If cost or time is an issue, you may wish to tally the | | | | chasing small problems or trying to fix everything at |
| cost or time involved in each problem classification.) If a | | | | once. |
| problem does not fit in any of your classifications, tally | | | | • After you have worked on the problems for a |
| it as “other.” | | | | while, use the same techniques to collect data again |
| 5. Determine the total occurrences (or total cost or | | | | and create a new Pareto chart. |
| total time) in each classification. Add these totals to | | | | The new chart will tell you how you are doing. Use the |
| produce a grand total. | | | | data on the new chart to identify and target the new |
| 6. Divide each classification total by the grand total to | | | | most significant problems. |
| determine the percentage that each individual problem | | | | Pareto charts can be used to identify problems to |
| classification represents of all the problems. | | | | work on. They can help you produce greater |
| 7. Arrange the problem classifications in order from | | | | efficiency, conserve materials, reduce costs or |
| highest value or most frequent to lowest value or least | | | | increase safety. They are most meaningful, however, |
| frequent. The “other” category should always | | | | if your customer–the person or organization that |
| be put last even if it is larger than some of the others. | | | | receives your work–helps define the problem |
| 8. Draw a horizontal axis and two vertical axes. Mark | | | | categories. |