Pareto charts are a useful tool for prioritizing your efforts. If you can't fix everything, what should you fix first? Pareto charts are most commonly used to prioritized root causes by frequency. But they can also be used to prioritize problems. The idea was developed in the 19th century by an Italian economies Vilfredo Pareto. He develop, the 80/20 rule, he determined in Italy that 80% of the land was owned by 20% of the people. You may have heard of this rule. The idea has been widely adopted and modified. For example, 20% of your customers account for 80% of your business or 80% of your profit comes from 20% of your products. This concept was adapted for quality by Joseph Juran. He suggested that 80% of defects come from 20% of the causes. Juran recognizes that most problems have many root causes. He said that a Pareto chart will help us to separate the vital few from the trivial many. Here is an example from a student survey on the cafeteria. The bars indicate the frequency, or count, of poor ratings for each category. The corresponding scale for the bars is on the left of the chart. The line indicates a cumulative percent from left to right. The corresponding scale for the cumulative percent line is on the right of the chart. A Pareto chart is always arranged by frequency with the most frequent on the left. To create the chart, first collect the data. In this case, it is student dissatisfaction with various aspects of the cafeteria. When you've collected the counts, sort the data from largest to smallest. The largest bar is always on the left. You can tell at a glance that service efficiency is the biggest problem contributing to student dissatisfaction with this cafeteria. Using the 80/20 rules, the first two or three bars usually account for about 80% of the problem. If you can fix them, you can make significant improvement. Discreet data is a set of numbers where only certain values are possible. On a survey you may be asked to write something on a scale of 1 to 5. There are no other possible values. You can't rate it three and a half. Continuous data is a set of numbers with all of the values in between. This might include one through five, and all of the possible values between one and two, two and three, etc. For discreet data, we just need to count the number of occurrences for each condition. Continuous data can be converted to discrete data by creating groups, the same way we did for histogram. Then we count the values that fall in each group. Discrete data is most likely going to be things that are counted and will be whole numbers. Continuous data is infinitely divisible in a meaningful way. For example, an inch can be divided by two to get half an inch. That can be divided by two again, to get a quarter of an inch, and so on. There's no end to the divisions you can make. Continuous data is usually a result of direct measurement. This is an example of a Pareto chart of workplace injuries in a manufacturing facility. Data shown in the table, in the upper right, was collected by the type of injury. Hand injuries are easily the most common type of injury in manufacturing followed by foot injuries. So if we wanted to reduce the number of injuries, we would probably focus on those. But, what if the cost of injuries was the most important thing? If we collected data on the cost of injuries by type showing on this table, we get a very different picture. The most costly injuries tend to be back injuries because of lost time. The point here, is to think carefully about what you want to know and what you want to accomplish. We'll talk briefly about how to create Pareto charts in Excel. These are the major steps for creating a Pareto program in Excel. We'll go through each one of them. The specifics of the steps may vary, depending on what version of Excel you're running. This is the data from the injury by type example. First, determine your categories then collect the data. Notice that the injuries are in random order. You won't know how to order it until after you collect the data. The next step is to sort the data by count, from largest to smallest. You may need to use the Custom sort function in Excel. Select the column you want to sort on, which is count, and sort largest to smallest. Now the most frequently occurring injury will be first on the list. There are a couple of ways to calculate the cumulative percent. An easy way is to create a percent column first, to get a percentage for each type of injury. Divide the count by the total count. Click on the percent icon to make it a percent. To get cumulative percent, add each percentage to all the preceding ones. Now the data is ready for a graph. You may want to hide the percentage column. Select Type, Count, and Cumulative Percent cells, but do not include the Total. Insert a column graph, and it will look like this. All of the bars are in order from tallest to shortest, but it does not yet look like a Pareto chart. First we'll need to do some formatting. From the chart tools, turn off the legend, label the axes, select the bars and click on Format data series. Change the gap width to No gap. Select the secondary axis which is a cumulative percent, and change the type to line. Now we just need to set the scale. For the primary axis, which is the bars or count, set the maximum to 500, which is our total count. For the secondary or cumulative percent, set the maximum to 1.0 or 100%. If you want, you can add data labels to the cumulative percent line. Now, we have a proper Pareto chart. Notice that the first two bars account for 80% of the problem. Pareto diagrams are useful tool for getting a visual picture of what your priorities might be. They use discrete data that is counts or frequency. The data is arrange from most frequent to least frequent so that the tallest bars always on the left and the scale for the bars is on the left side of the chart. The line represents cumulative percent, and its scale is on the right side of the chart. Visually, it's easy to see where our efforts need to go for maximum effect.
