How to calculate a Pareto chart
A Pareto chart sorts the causes from largest to smallest and tracks how much of the problem is explained as each cause is added.
Cause share = cause count ÷ total
Cumulative % = sum of the counts down to the cause ÷ total
| Class | Rule used here |
|---|---|
| A (vital few) | causes up to crossing 80% of the total, including the one that crosses the line |
| B | the next ones, up to crossing 95% |
| C (trivial many) | the rest |
The cause that pushes the total past 80% belongs to class A because it is the one that brings the cumulative line up to the mark. Without it, the vital few would explain less than 80% of the problem.
Worked example
The values already in the tool are the parts an injection molding machine scrapped in a month, by defect, in the order someone logged them: scratch 58, flash 152, stain 18, voids 96, dimension 31, warpage 9, short shot 12 and dirt 6. The total is 382 parts.
Sorted from largest to smallest:
| Defect | Parts | % of total | Cumulative % | Class |
|---|---|---|---|---|
| Flash | 152 | 39.8% | 39.8% | A |
| Voids | 96 | 25.1% | 64.9% | A |
| Scratch | 58 | 15.2% | 80.1% | A |
| Dimension | 31 | 8.1% | 88.2% | B |
| Stain | 18 | 4.7% | 92.9% | B |
| Short shot | 12 | 3.1% | 96.1% | B |
| Warpage | 9 | 2.4% | 98.4% | C |
| Dirt | 6 | 1.6% | 100% | C |
Flash, voids and scratches are 3 of the 8 defects and add up to 306 parts, 80.1% of the scrap. Fixing flash alone removes almost 40% of the scrapped parts.
How to read the result
- The dark bars are the vital few. That is where effort pays off most. Start with the first bar, not with the easiest cause.
- The red line is the cumulative share, read on the right axis. Where it crosses the dashed 80% line is the class A cutoff.
- A curve that climbs slowly, with no bar standing out, means there is no clear vital few. That usually points to causes that are too broad (“machine problem”) or to count not being the right measure. Try the Pareto by cost or by downtime minutes.
- After acting, redo the Pareto with the next period’s data. If the bar dropped, the fix worked, and the next cause becomes the new first bar.
For machine breakdowns, a Pareto by asset pairs well with the MTBF and MTTR calculator, and the lost time shows up in the OEE calculator. When the cause you picked is a process risk, rate it in the FMEA RPN and Action Priority calculator.
Common mistakes
- Letting “Other” grow too big. If “Other” is one of the tallest bars, it hides causes that should be visible. Break it down.
- Mixing measures. Scrapped parts, downtime minutes and money lost do not add up. Build one Pareto per measure.
- Counting only occurrences when costs differ a lot. A rare defect that scraps an expensive part can matter more than a frequent, cheap one.
- Using too short a period. With little data the order of the bars changes from week to week. Collect at least a few weeks.
- Stopping at the Pareto. It shows where to act, not why it happens. The root cause comes from an analysis such as a fishbone with 5 whys.