Pareto Analysis

Rank categories by measured contribution and focus improvement on the few that matter most for the chosen outcome.

In one minute

A Pareto chart displays categories as descending bars. The bar height represents a chosen measure such as frequency, cost, delay or harm. An optional cumulative line shows how quickly the categories add up to the total.

The familiar 80/20 split is a heuristic, not a requirement. Real data may show 70/30, 90/10 or no strong concentration. The method’s value is disciplined prioritisation, not forcing the data to match a slogan.

Best for: many categories with a common measurable outcome.
Avoid when: events are not comparable, data quality is poor or rare severe events would disappear in a frequency chart.

The problem it addresses

Teams can spread effort equally across many issues or prioritise the most visible complaint rather than the largest contributor. Pareto Analysis shows where the measured burden is concentrated.

The intended outcome is a defensible focus area and a baseline for comparing results after improvement.

When to use it

• to rank defect or complaint categories;
• to prioritise causes after data collection;
• to identify which delay types consume the most time;
• to compare cost, frequency or weighted impact;
• before a focused Root Cause Analysis;
• to compare distributions before and after an intervention.

When not to use it

Do not use a basic Pareto chart:

• when categories overlap or are inconsistently coded;
• when the denominator changes across groups without normalisation;
• when rare events have unacceptable severity;
• to claim causation from category frequency;
• when a time series or control chart is needed to understand variation;
• when “Other” contains a large, unexplored share.

Inputs required

• a decision question;
• mutually exclusive, operational category definitions;
• a meaningful measure: count, cost, time, harm or another unit;
• a consistent observation period and denominator;
• validated data with missingness documented;
• a rule for handling small categories and “Other.”

Step-by-step process

1. Define the decision and measure

“Which complaint types should we investigate first to reduce handling time?” may require minutes, not complaint count.

2. Define categories

Make categories mutually exclusive and useful for action. Train or check coders if classification is manual.

3. Select the period and denominator

Record whether values are totals or rates. Comparing 100 complaints from 1,000 orders with 80 from 400 orders requires normalisation.

4. Aggregate and validate

Subtotal the chosen measure by category. Review missing values, duplicates and a large “Other” bucket.

5. Sort descending

Place the largest contributor first.

6. Calculate shares and cumulative share

For each category:

category share = category value / total value

Add categories from left to right for the cumulative percentage.

7. Interpret the pattern

Identify the smallest useful set of categories that captures a material share of the outcome. Do not use an automatic 80% cutoff if a different boundary is more actionable.

8. Choose the next analysis

Pareto tells you where to look, not why it happens. Use RCA, 5 Whys or Ishikawa on the priority category.

9. Compare after action

Rebuild the chart with the same definitions and a comparable period. Track both the target category and the total outcome.

Visual model

Text alternative: wrong-time complaints consume the most handling time, followed by duplicate reminders, non-delivery, language issues and other complaints.

Interactive example

Northstar recorded reminder complaints:

Your move

Which category ranks first by frequency, and which ranks first by total handling time?

Worked answer

Wrong appointment time ranks first under both measures. In another dataset, the priority could change with the measure. The team should select the measure that matches the decision.

Facilitation notes

• Agree on the metric before revealing category totals.
• Review the coding guide with people who entered the data.
• Show both counts and rates when exposure differs.
• Keep rare high-severity items in a separate risk view.
• Drill into “Other” if it becomes material.

Expected output

• a clear question, metric, period and denominator;
• validated category totals;
• descending values and cumulative shares;
• a selected focus area with rationale;
• a linked causal-analysis task;
• a plan for the comparison chart.

Common mistakes

1. Forcing an 80/20 result.
2. Using frequency when cost or severity matters.
3. Treating the largest category as a root cause.
4. Combining overlapping categories.
5. Comparing totals with different exposure.
6. Hiding too much in “Other.”
7. Failing to rebuild the chart after action.

Quality checklist

• The measure matches the decision.
• Categories are mutually exclusive and consistently coded.
• Period and denominator are explicit.
• Missing and “Other” data were reviewed.
• Rare severe risks are not hidden by frequency.
• The chart identifies focus, not causation.
• A comparable follow-up measure is planned.

Template

Knowledge check

A Pareto chart shows that one complaint category represents 42% of total handling time. What can you conclude?

A. The category is the root cause of 42% of complaints.
B. Removing it will definitely reduce total handling time by 42%.
C. It is the largest measured contributor and deserves focused investigation.
D. The analysis is invalid because it does not equal 80%.

Answer: C. The chart identifies concentration, not causation or guaranteed savings.

Related tools

• Often follows: Ishikawa Diagram or data collection
• Supports: Root Cause Analysis
• Followed by: PDCA and a comparative Pareto chart
• Not to be confused with: ABC inventory classification or proof of causation

References

1. Juran, J. M. Quality Control Handbook. McGraw-Hill, 1951. Primary quality-management source that popularised applying the “vital few” idea to quality problems.
2. Juran Institute. “Pareto Principle (80/20 Rule) & Pareto Analysis Guide.” Juran Institute guide (opens in a new tab). Authoritative organisational explanation.
3. American Society for Quality. “What is a Pareto Chart?” ASQ guide (opens in a new tab). Authoritative procedure and usage guidance.
4. Juran, J. M. “The Non-Pareto Principle; Mea Culpa.” 1974. Juran Institute archival paper (opens in a new tab). Primary reflection on the principle and terminology.
5. Radson, D., & Boyd, A. H. “The Pareto Principle and Rate Analysis.” Quality Engineering, 10(2), 1997, pp. 223–229. ASQ record (opens in a new tab). Independent analysis showing that a basic Pareto ranking is not statistically grounded and that rate analysis can change management priorities.

Sources reviewed 27 July 2026.