The most frequent problem may not be the most expensive. A Pareto chart sorted by frequency answers “Which categories occur most often?” A chart sorted by cost answers “Which categories account for the most recorded cost?”
Both can be useful. Choose the measure that fits your decision, and compare the two when the cost of an occurrence differs substantially between categories.
ASQ describes Pareto bars as representing frequency or cost, including time or money, arranged from largest to smallest. The measure does not have to be a count. ASQ: Pareto Chart
1. Define what you are counting
All data in this article are fictional and provided for learning. They are not customer results or industry benchmarks.
Imagine a packing team reviewing 100 final-inspection exceptions over four weeks. Each affected item receives one primary exception category, so no item appears twice.
The team wants to examine both the number of affected items and the estimated direct correction cost. For this simplified example, every occurrence within a category has the same estimated cost.
| Exception category | Occurrences | Estimated cost per occurrence | Estimated category cost |
|---|---|---|---|
| Smudged label | 50 | $2 | $100 |
| Missing insert | 30 | $5 | $150 |
| Wrong accessory | 15 | $30 | $450 |
| Damaged product | 5 | $260 | $1,300 |
| Total | 100 | $2,000 |
Here, category cost equals occurrences multiplied by estimated cost per occurrence. For example, five damaged products at $260 each produce an estimated $1,300 cost.
In a real dataset, sum the recorded costs when occurrences have different costs. Define what those costs include, apply the same rules to every category, and avoid counting the same loss twice. Label estimates as estimates.
2. Build the frequency Pareto
Sort categories by occurrence count. Calculate each share as the category count divided by the total count, then add the shares in order to obtain the cumulative percentage.
For smudged labels, 50 ÷ 100 × 100 = 50%.
| Frequency rank | Category | Occurrences | Share of exceptions | Cumulative share |
|---|---|---|---|---|
| 1 | Smudged label | 50 | 50% | 50% |
| 2 | Missing insert | 30 | 30% | 80% |
| 3 | Wrong accessory | 15 | 15% | 95% |
| 4 | Damaged product | 5 | 5% | 100% |
Draw the bars in that order, with occurrences on the left axis. If you add a cumulative line, use a right axis from 0% to 100% and align 100% with the total of 100 occurrences on the left axis. ASQ explains the descending bars and optional cumulative-percentage line in its Pareto chart procedure.
What this shows: Smudged labels are the largest category by count. Labels and inserts together account for 80 of the 100 exceptions.
These percentages describe the exceptions in this dataset. They are not defect rates across all items packed, because the total production volume has not been supplied.
3. Build the cost Pareto
Now sort the same categories by total estimated cost, rather than by count or cost per occurrence. Recalculate the shares using $2,000 as the denominator.
For damaged products, $1,300 ÷ $2,000 × 100 = 65%.
| Cost rank | Category | Estimated cost | Share of cost | Cumulative share |
|---|---|---|---|---|
| 1 | Damaged product | $1,300 | 65% | 65% |
| 2 | Wrong accessory | $450 | 22.5% | 87.5% |
| 3 | Missing insert | $150 | 7.5% | 95% |
| 4 | Smudged label | $100 | 5% | 100% |
Draw a separate chart with dollars on the left axis. For its cumulative line, align $2,000 with 100% on the right axis.
What changes: Damaged products move from last place by frequency to first place by estimated cost. Smudged labels move from first to last. Damaged products and wrong accessories account for only 20 exceptions but $1,750, or 87.5%, of the estimated correction cost.
4. Use the ranking to choose an investigation
For this fictional team:
- If the immediate aim is to reduce the number of final-inspection exceptions, smudged labels are a reasonable starting point for investigation.
- If the aim is to reduce the estimated correction cost, damaged products deserve closer examination.
- If either category creates a safety, compliance, or serious customer concern, assess that concern separately. Frequency and cost rankings cannot represent every consequence.
The $1,300 damage total is not a savings forecast. The team has not yet established what causes the damage, what proportion it can prevent, or what a change would cost.
Likewise, a category such as “damaged product” describes an observed problem. It does not establish whether packing, transport, handling, or another condition caused it. Inspect the affected cases and check evidence before choosing a countermeasure.
Questions to check before using either chart
Do the charts have to show an exact 80/20 split? No. Report the distribution in your data. In this example, two of four categories account for 80% of occurrences; a different pair accounts for 87.5% of cost.
Can you keep the frequency order in the cost chart? Re-sort by cost to make it a cost Pareto. Otherwise, the cost bars will not be in descending order.
What if cost estimates are uncertain? Show the assumptions and test plausible alternatives. Here, damage remains above the $450 accessory category only while its average estimated cost exceeds $90 per occurrence, assuming the other figures stay fixed.
Try it with your own data
Start with a defined period, consistent categories, and a clear question. Keep the frequency and cost views separate so readers can see what each ranking means.
The $9 Seven Quality Tools Worksheets pack includes a printable Pareto worksheet and fictional worked example. Use the frequency worksheet to organize your counts, and record any additional cost calculations separately with their assumptions.