The first step includes developing a list of causes or error types.
The second step is to count the number of defects that can be attributed to each cause.
The third step is to rank the causes in descending order, according to the frequency of errors they result in.
The fourth step is to find the cumulative percentage of the total number of errors for each cause.
To do this, you
- add the frequency numbers to determine the total errors
- find the percentage for each cause by dividing the percentage of errors for which it’s responsible by the total number of errors, then multiplying the result by 100
- for each cause, add the percentage of errors it’s responsible for to the sum of the percentages for causes ranked above it
As the fifth step, you create a histogram that displays the data.
Frequency rates must be on the left vertical axis and bars representing the number of errors associated with each cause are plotted, in descending order, on the horizontal axis.
As the final step, you draw a line graph that plots the cumulative percentages for each cause.
The first point on the line graph should line up with the top of the first bar.