Pareto Chart

One of the first things new engineers are asked to do is to learn to prioritize tasks.  They are made aware early on to distinguish between the 'vital few' and the 'trivial many', which means that one has to focus on the few things that really matter, and not spend resources on the many others that have little or no impact.

There was even a mathematical expression for it, known as the 80/20 rule, which states that 80% of all problem occurrences are due to only 20% of the types of problems encountered. Another variant of this rule states that for any problem, 80% of its occurrences are due to only 20% of all the causes.

Thus, if one were to address a problem, the 20% of the causes that results in 80% of the problem (the 'vital few', so to speak) must first be identified and eliminated, before the rest are acted upon, if still necessary. The Pareto Chart is a very simple but effective tool for prioritizing problem causes, which is why it is widely used for problem-solving in the manufacturing industry.

The Pareto Chart is basically a descending bar graph that shows the frequencies of occurrences or relative sizes of either: 1)  the various categories of all problems encountered, in order to determine which of the existing problems occur most frequently; or 2) the various causes of a particular problem, in order to determine which of the causes of a particular problem arise most frequently.  The problem categories or causes are shown on the x-axis of the bar graph.

Aside from its main bar graph, the Pareto Chart may also include a line graph that indicates the cumulative percentage of occurrences at each bar of the bar graph.  This line graph, referred to as the 'cumulative percentage line', is used to determine which of the bars belong to the 'vital few' and which ones are relegated to the 'trivial many.'   Figure 1 shows an example of a Pareto Chart.

Bars that belong to the former group are those that account for bulk of the problems or problem causes encountered.  The last point of the line graph corresponds to the last bar of the bar graph (usually the 'Others' bar), and should correspond to 100% of the cumulative occurrences.

A pareto chart offers the following benefits: 1) it helps the team focus on the problems or causes of problems that have the greatest impact; 2) it displays the relative significance of problems or problem causes in a simple, quick-to-interpret, visual format; and 3) it can be used repeatedly in cycles to produce continuous improvements systematically (for each succeeding cycle, the major pareto bars are actually minor bars in the previous cycle).

To construct a pareto chart, the following steps are recommended:

1)  choose a problem that needs to be addressed;

2)  identify the causes of the problem based on existing data and through brainstorming;

3)  decide on how these problem causes will be monitored for the data collection for the pareto chart, e.g., frequency of occurrence?, cost?, etc.;

4)  define the duration and time frame of the data collection - it should be long enough to provide meaningful information about the real situation;

5)  conduct the data gathering according to the defined time frame, e.g., monitor the frequency of occurrence or cost impact of each problem cause encountered, ensuring that causes not identified earlier for monitoring must still be counted in a catch-all bin ('Others' category);

6)  tabulate the problem causes in order of decreasing frequency, and assign a column each for: a) the frequency of occurrence (or cost impact); b) the percentage share of the cause; and c) the cumulative percentage corresponding to each problem cause (see Table 1);

7)  construct the pareto chart using the tabulated data and following the pareto chart format discussed earlier (see Figure 1);

8)  interpret the pareto chart and select the 'vital few' that need to be addressed immediately.

Table 1. Example of a Tabulation of Causes of Ball Bond Lifting

for use in a Pareto Chart

 Ball Lifting Cause Frequency Percent (%) Cum Pct (%) Bonder Set-up Issues 19 38% 38% Unetched Glass on Bond Pad 11 22% 60% Foreign Contam on Bond Pad 9 18% 78% Excessive Probe Damage 3 6% 84% Silicon Dust on Bond Pad 2 4% 88% Corrosion 1 2% 90% Bond Pad Peel-off 1 2% 92% Cratering 1 2% 94% Resin Bleed-out 1 2% 96% Others 2 4% 100% Total 50 100% -

Figure 1. The Pareto Chart Corresponding to Table 1

The Pareto Chart in Figure 1 clearly shows that the greatest contributors to the ball bond lifting problems of the engineer who undertook the data collection are bonder set-up issues, unetched glass on bond pads, and foreign contaminants on bond pads. Thus, the engineer has to focus on these causes to reduce their ball lifting occurrences by as much as 78%.