Pareto Chart Explained: How to Make, Read and Report It
A Pareto chart answers a practical question: which few causes account for most of the problem? It combines a bar chart sorted from largest to smallest with a cumulative-percentage line, so you can see at a glance where improvement effort will pay off most.
What is a Pareto chart?
It is a descending bar chart of category frequencies (or costs, or time lost) with a second line showing the running cumulative percentage. It applies to categorical data counted or weighted by a common unit. The idea is linked to Joseph Juran's "vital few and useful many" and the broader Pareto principle, but the 80/20 split is an empirical tendency, not a law; your data may show 60/40 or 95/5.
When to use it
Defect analysis, complaint categories, reasons for protocol deviations, lab error sources, equipment downtime, and any prioritisation task where categories are mutually exclusive. Do not use it for continuous measurements or when categories overlap.
Calculation
For categories sorted so that n1 ≥ n2 ≥ … ≥ nk with total N = Σni:
Percenti = 100 · ni / N Cumulative%j = 100 · (n1 + … + nj) / N
Original example (illustrative)
A packaging line rejects 100 sample units in a week. Defects were classified into five categories.
| Defect | Count | % | Cumulative % |
|---|---|---|---|
| Label misprint | 48 | 48 | 48 |
| Seal failure | 31 | 31 | 79 |
| Underfill | 12 | 12 | 91 |
| Cracked cap | 6 | 6 | 97 |
| Other | 3 | 3 | 100 |
The counts sum to 100, and the cumulative column ends at 100%. Two of five categories (label misprint and seal failure) explain 79% of the defects, so these two are the first targets. Always place "Other" last, even when it is not the smallest, because it is a residual group.
How to make it in StatFAS
Open the Pareto Chart Generator. Prepare two columns: category labels and numeric counts (or costs).
Reading the chart
Look for a steep early rise in the cumulative line and a clear drop in bar height. If bars are nearly equal and the line is almost straight, there is no "vital few", and prioritisation by Pareto will not help. Judge importance by the right unit: frequency can mislead if a rare defect is very costly, so consider a cost-weighted Pareto as well.
Reporting
Of 100 rejected units, label misprints (n = 48, 48%) and seal failures (n = 31, 31%) together accounted for 79% of defects (Figure 1).
Report N, the unit of measurement, the classification scheme and the period covered.
Common mistakes and limits
Avoid too many categories (merge small ones), inconsistent classification, using a Pareto for before-and-after comparison without the same category order, and treating the chart as proof of causation. It describes frequency; it does not test differences. To test whether category distributions differ between groups, use an appropriate categorical test. For general charts and summaries, see descriptive statistics and charts.
FAQ
Does a Pareto chart need to show 80%? No. The 80% line is a visual aid only.
Can I use percentages as input? Yes in principle, but raw counts preserve the sample size.
Conclusion
A Pareto chart is a prioritisation tool: sort, accumulate, and focus on categories that carry most of the burden, while remembering that frequency, cost and causation are separate questions.
1.
statfas-pareto-chart-interface.png, alt: "StatFAS Pareto Chart Generator interface"2.
statfas-pareto-chart-example-data.png, alt: "Example data entered in the StatFAS Pareto Chart Generator"3.
statfas-pareto-chart-settings.png, alt: "Pareto Chart Generator settings in StatFAS"4.
statfas-pareto-chart-result.png, alt: "Pareto Chart Generator result generated in StatFAS"References
- Juran, J. M. (Ed.). Juran's Quality Control Handbook (McGraw-Hill). Check edition and year before citing.
- Montgomery, D. C. Introduction to Statistical Quality Control (Wiley). Check edition and year before citing.