Chasing the Wrong 80%: Common Pareto Analysis Mistakes
The Pareto principle — the idea that a small number of causes drive most of an effect — is one of the most useful tools in the Lean Six Sigma kit. Built into the Analyze phase of DMAIC, a Pareto chart sorts categories of a problem from largest to smallest so a team can focus on the "vital few" rather than scattering effort across the "trivial many." It is simple to draw and easy to misuse. After a year of disrupted operations producing piles of messy data, the temptation to chart first and think later is high. Here are the mistakes that lead teams astray.
Mistakes in how you set the chart up
Charting by frequency when impact is what matters. A defect that happens often but costs little can dwarf a rare defect that costs a fortune. If cost, time, or risk is the real driver, weight the bars by impact, not just count. Counting occurrences answers "what happens most," not "what hurts most."
Categories that overlap or are too broad. If "shipping issues" and "late delivery" are separate bars that mean the same thing, your tallest bar is an artifact of how you sliced the data. Define mutually exclusive categories before you count.
A giant "Other" bucket. When "Other" is one of your biggest bars, the analysis has failed — the real causes are hidden inside it. Break it open before you draw conclusions.
Too little data, or the wrong window. A chart built on two weeks of pandemic-distorted data can point you at a problem that won't exist next month. Make sure the period is representative and the sample is large enough to be stable.
Mistakes in how you read it
Even a well-built chart gets misread. The common traps:
Treating the tallest bar as the root cause. A Pareto chart shows where a problem concentrates, not why it happens. The biggest category is where to point your root-cause analysis next — it is not the answer itself.
Forcing the 80/20 split. The principle is a tendency, not a law. Sometimes three causes drive 60%, sometimes one drives 90%. Read the actual breakpoint in your data instead of insisting on a tidy ratio.
Charting once and never again. After you fix the top cause, the order changes — what was the second bar may now be first. Re-run the analysis to confirm the fix worked and to find the next target.
Ignoring the cost of action. The tallest bar may be the most expensive to fix. Weigh the size of the prize against the effort before you commit the team.
A disciplined way to run it
Use the tool the way the Analyze phase intends. Decide first what you are optimizing — count, cost, time, or risk — and measure that. Define clean, non-overlapping categories. Pull a representative, sufficient sample. Sort largest to smallest, add the cumulative line, and read where the curve flattens; that breakpoint, not a memorized percentage, tells you where the vital few end. Then take the top category into a real root-cause method — five whys, a fishbone — rather than treating the bar as the diagnosis.
Done well, Pareto analysis is how a stretched team spends its limited improvement capacity where it actually moves the needle. Done carelessly, it lends a chart's authority to a guess. The difference is entirely in the setup and the reading.
If your team has plenty of data but keeps fixing the wrong problems, XNM's strategic advisory can help you focus improvement effort where the evidence says it belongs.