Seven QC Tools
Tool #4: Pareto Chart
Lesson Structure
Pareto Chart
STEP 1: Identify the categories of problem or causes to be compared
STEP 2: Select a standard unit of measurement and the time period to be studied
STEP 3: Collect and summarize data
STEP 4: Draw the horizontal and vertical axes
STEP 5: Plot bars on the Pareto Chart
Follow-up: Decide On The next Steps
Summary
Pareto Chart
The Pareto Chart is a special type of bar graph you can use as an interpretation tool in:
Determining the relative frequency or importance of different problems or causes
Focusing on vital issues by ranking them in terms of significance
Note: Before you can construct a Pareto Chart, you need to know how to use Check
Sheets (since the data you collect with the Check Sheet will be used to construct the
Pareto Chart) and other basic data-gathering tools.
Creating a Pareto Chart consists of five major steps:
STEP 1: Identify the categories of problem or causes to be compared
STEP 2: Select a standard unit of measurement and the time period to be
studied
STEP 3: Collect and summarize data
STEP 4: Draw the horizontal and vertical axes
STEP 5: Plot bars on the Pareto Chart
The example that follows shows how the Pareto Chart is used to help distinguish opinion
from fact.
Mahesh, Correspondence Center Supervisor,………..
had been hearing from the center’s internal customers about frequent typos and
misspellings in documents created by the Center, and he wanted the “problems”
corrected. Mahesh and Yogesh, an Administrative Assistant, decided that a Pareto Chart
would help them pinpoint the specific problems and their causes…
Step 1: Identify the Categories of Problems or Causes to be Compared
Begin by organizing the problems or causes into a handful of categories. Narrow down a
long list to manageable number of eight categories or less.
For example, if you are measuring reasons why you were late to work over a period of
time, possible categories could consist of: too busy, traffic, oversleeping, arguing with
your spouse or children, etc.
Note: this information can be obtained from brainstorming, a Cause and Effect Diagram, Cheek Sheet,
existing reports, data etc.
Mahesh called together a small group of people………
who had been on the receiving end of some of the center’s errors? He asked them to
brainstorm a list of errors they had noticed in the last few months. Mahesh then asked
the group to identify five error categories. The categories chosen were: punctuation
errors, keystroke errors, spelling errors, late delivery, and wrong page numbering…
Step 2: Select a Standard Unit of Measurement and the Time Period to
be Studied
The measurement you select will depend on the type of situation you are working with. It
could be a measure of how often something occurs (such as defects, errors, overflows,
cost overruns, etc.), how often reasons are cited in surveys as the cause of a certain
problem, or a specific measurement of volume or size.
Time
Defects
Frequency
Size
Note: when selecting a sample time period, allow a long enough span to gather the required data. For
example, if you were measuring reasons why people get to work late, you would not use data gathered from
just one day of the week. On Monday, traffic might be heavier than Tuesday or Wednesday. By spanning a
week or more, any special cases, such as Monday morning traffic, will balance out against the other reasons
why people are late to work.
Mahesh gave Yogesh the go-ahead…
to collect data and build the Pareto Chart. Yogesh had a training session the month
before and learned that measuring the right variable on chart gave him valid insight. He
began tracking the errors in a 30 day span - a long enough time period to get a clear
picture of what was going on. Yogesh realized the customers were not reporting all of the
errors to the center. He personally contacted each customer requesting they report all
errors. Yogesh even mailed each customer special card to record this information. When
he explained to them that the information gathered would reduce errors, they were only
too happy to help!
Date:___________
Error:_______________________
____________________________
Error:_______________________
____________________________
Step 3: Collect and Summarize Data
Begin by creating a three-column table, for which the headings should be “error
category,” “frequency,” and “percent of total.” The items in the “error category” column
should include the different types of errors (or causes of errors) that occur. This
information can be taken directly from your check sheets.
Under the frequency column, write in the totals for each of the categories.
Now divide each number in the frequency column by the total number of measurements.
This will give you the percentage of the total. For example, if the frequency for a given
category is 30, and the total measurement add up to 80, the percentage (30/80) equals
37.5%. Write your calculated percentage under the heading titled “Percent of total” for
each category.
Error category Frequency Percent of Total
Punctuation 20 44%
Keystroke 12 27%
Spelling 7 16%
Late 5 11%
Page#’s 1 2%
Total 45 100%
Diagram-4.1 Create a data Table
Yogesh started by creating a table………..
for the data he collected during the past month (see diagram-4.1). He listed the various
errors made under the category column. Yogesh then recorded the frequency of each
error in the second column (this information was obtained directly from Check Sheet).
Finally he calculated the percentage of total errors for each of the categories and
recorded those in the appropriate spaces. Yogesh was now ready to draw his Pareto
Chart.
Step 4: Draw the Horizontal and Vertical Axes
Begin by drawing the horizontal axis. Draw a line from left to right on a piece of paper.
Make the line long enough so all the categories can be written under it. Write the
categories in the descending order with most frequently occurring category on the far left
(or beginning of the horizontal line). Label the axis. The label should tell the readers what
they are looking at (e.g., cause for being late to work, types of customer complaints, etc.).
Next draw a vertical line up from the far left point of the horizontal axis. This line will
indicate the frequency for each of the categories. Scale it so the value at the top of the
axis is slightly higher than the highest frequency number. Label this axis also. Again, the
label should tell the readers what they are looking at (i.e., frequency of occurrence).
Diagram-4.2- organize the Pareto Chart
Draw another vertical axis, this time from the far right side of the horizontal axis. This
line will represent the percentage scale and should be scaled so that the point for a
number of occurrences on the left matches with the corresponding percentage to the right.
Yogesh drew the horizontal and vertical axes…
He labeled the horizontal axis “Error Category” and the vertical axes “Frequency of
occurrences” and “Percentage of total”. Since punctuation errors occurred most
frequently, Yogesh made “punctuation” the first category on the far left. He wrote in
“keystroke” as the next category on the axis and so on (see diagram-4.2)………..
Step 5: Plot the Bars on the Pareto Chart
The final step is to plot the data by drawing a series of bars in decreasing height from left
to right, using the frequency scale on the left vertical axis.
Note: Categories with very few items can be combined into an “other” or “Miscellaneous” category, which
is placed on the extreme right of the last bar.
To plot the cumulative percentage line, place a dot above each bar at a height corresponding to the scale on
the right vertical axis. Starting with the first column on the left, draw a line to connect these dots from left
to right, ending with the 100% point at the top of the right vertical axis.
Diagram –4.3-plot the data
While Yogesh plotted the chart,……….
the biggest issue becomes clear. Two sources, punctuation and wrong keystroke, together
accounted for 71% of the problem. Mahesh and Yogesh decided to create a game plan
before presenting the chart to the rest of the correspondence Center (see Diagram-4.3)
………..
Follow-up: Decide On The next Steps
You now have an easy-to-read and understand chart that should help you decide what to
tackle first. While working on your problem (e.g., customer complaints, etc.), however,
continue to investigate. For example, just because a certain problem occurs most often
doesn’t necessarily mean it demands your greatest attention. Also consider the following:
What makes the biggest difference
to the customer?
What will it cost to correct this
problem?
What will it cost if we don’t correct
this problem?
Investigating all the angles will help you to decide which problem should be worked on
first.
Remember, while your data will lead you in a certain direction, the customer should have
the final say on what should be corrected first. Continue to collect and plot data to verify
your original findings, and also to evaluate any changes (improvements) you make.
Priorities as per Pareto Chart Priorities as per Customer
1. Punctuation (44%) 1. Late deliveries
2. Keystroke (27%)
3. Spellings (16%)
4. Late deliveries (11%)
5. Wrong page’s (2%)
Having completed his Pareto Chart,………..
Yogesh reviewed the data with Mahesh and agree to share it with a cross section of the
Correspondence Center’s internal customers. This meeting provided new insight. The
number one complaint for several of the center’s key customers was material not being
completed on time. This error did not happen often, but when it did, customers
experienced severe problems (e.g., missed deadlines, rescheduled meetings, additional
work hours, etc.).
Yogesh discovered firsthand that although the data highlighted one finding (i.e.,
punctuation error are the biggest problem), the customer’s perspective was completely
different (i.e. “get the material to us on time”). Based on this information, Mahesh and
Yogesh created a new improvement team to focus on eliminating work being delivered
late from the Correspondence Center.
Summary
In summary, use Pareto Chart when:
You and your team need to select a problem or process to improve.
You and your team need to evaluate improvement efforts that you have already
made.
You or your team wants to identify the distribution of the causes of a problem.
