Pareto analysis — finding the vital few defects worth fixing first
Most scrap comes from a handful of causes. A Pareto chart shows you which handful, so improvement effort lands where it pays off instead of being spread evenly across problems that aren't equal.
A Pareto chart ranks defect categories by frequency as descending bars, with a cumulative-percentage line overlaid on a second axis. It applies the Pareto principle (the 80/20 rule): roughly 80% of the defects come from about 20% of the causes. The categories to the left of where the cumulative line crosses ~80% are the vital few worth fixing first; the rest are the trivial many. It directs effort — it doesn't explain the causes, which still need root-cause analysis.
What the chart combines
A Pareto chart is two plots in one. The bars are defect counts per category, sorted tallest-to-shortest. The line is the running cumulative percentage of the total, read on a right-hand 0–100% axis. Together they answer one question fast: how few categories do I need to fix to cover most of the problem?
How to read it well
1 · Find the 80% crossing
Trace the cumulative line to where it passes ~80%. Everything to its left is the vital few — usually two or three categories. Those are your targets; fixing them removes most of the defects.
2 · Sort by the right unit
Ranking by count answers "most frequent." Ranking by cost or downtime can reorder the bars entirely — a rare defect that scraps an expensive assembly may matter more than a common cosmetic one. Choose the unit that reflects impact.
3 · Re-Pareto after you fix
Once you eliminate the top category, the chart reshapes and a new vital few emerges. Pareto analysis is iterative — it keeps pointing at the current biggest lever.
4 · It points, it doesn't explain
A Pareto tells you which defect to attack, not why it happens. Hand the top category to a fishbone or 5 Whys for the cause.
Common mistakes
- Counting when cost is what matters. The most frequent defect isn't always the most expensive — Pareto by impact, not just tally.
- A giant "other" bucket. If "miscellaneous" dominates, your categories are too coarse to be actionable.
- Comparing unstable periods. Mixing data from before and after a process change blends two different Paretos into a misleading one.
- Stopping at the chart. Identifying the vital few is step one; the causes still need analysis and the fix still needs verification.
Where this gets slow by hand
One Pareto from a clean defect log is quick. The reality is messy logs, inconsistent categories, and the need to re-cut the data by cost, by line, by shift, and by time window — and to rebuild the chart every time the top cause is fixed. Keeping a live Pareto across all of those slices is a standing data-wrangling job.
A Pareto that re-cuts itself and points to the cause
Niobia builds the Pareto from raw defect logs — normalizing inconsistent categories and letting you re-cut by count, cost, line, shift, or time window on demand — and keeps it live as the mix changes, so the current vital few are always in view. Then it carries the top category straight into root-cause analysis, correlating it against process and material data so you move from "which defect" to "why" without re-assembling the data by hand.
Frequently asked
What is a Pareto chart?
A Pareto chart ranks categories (such as defect types) by frequency as descending bars and overlays a cumulative-percentage line. It's used to identify which few categories account for most of the total, so improvement effort can be focused there.
What is the 80/20 rule in Pareto analysis?
The Pareto principle observes that roughly 80% of effects come from about 20% of causes — for example, most defects come from a small number of defect modes. The chart makes this visible so you can target the vital few rather than spreading effort evenly.
What are the vital few and the trivial many?
The vital few are the small number of categories (to the left of the ~80% cumulative crossing) that cause most of the problem and deserve priority. The trivial many are the larger number of categories that together contribute relatively little.