Skip to content

Practical guide · Root cause analysis and problem solving

Pareto chart: how to build and read one, with a worked example by count and by cost

Short answer

A Pareto chart is a bar chart of problem categories sorted from largest to smallest, with a line showing the cumulative percentage. It shows which few categories make up most of the total, so you work on those first. The bars can count occurrences, minutes or cost.

A Pareto chart ranks categories of one problem, such as defect types or downtime reasons, from the biggest to the smallest, with a cumulative percentage line over the bars. To build one, count each category, sort them, work out each share of the total and the running total, and draw the bars and the line. Read it from the left: the first few bars are the ones to work on. Build it by count and by cost, because the order often changes.

Last updated:

What a Pareto chart is

A Pareto chart is a bar chart with its bars sorted in descending order and a line on top showing the running total as a percentage. The bars use the left axis (a count, minutes or money) and the line uses the right axis (0 to 100%). It rests on the 80/20 idea: when several things contribute to a result, a few of them usually account for most of it. The chart shows which few, with your own data. It is one of the seven basic quality tools, and it is the usual first step before a fishbone or 5 Whys, because it tells the team which problem deserves the investigation.

The 80/20 split is a rule of thumb, not a law. In one set of data the top two categories may make up 85% and in another it takes five categories to reach 60%. Whatever your data says is the answer; a flat chart is a finding too, because it says there is no single big lever.

How to build a Pareto chart in six steps

  1. 1

    Choose one problem and one period

    Scrap on one line, downtime on one machine, complaints in one quarter. Mixing problems gives a chart nobody can act on.

  2. 2

    Choose the categories and the measure

    Use reason codes the floor already uses, such as a defect check sheet or the downtime reason codes. The measure is a count, minutes or cost. Keep it to 6 to 10 categories plus an Other.

  3. 3

    Collect the data

    Tally for long enough that the order of the bars is stable. A few events are not enough to rank anything.

  4. 4

    Sort and total

    Sort the categories largest first, with Other always last, however big it is. Add them up for the grand total.

  5. 5

    Work out the share and the cumulative share

    Share = category value ÷ grand total × 100. Cumulative = this share plus all the shares above it. The last one must be 100%.

  6. 6

    Draw it and mark the cut

    Bars in sorted order on the left axis, the cumulative points joined on the right axis, and a line at 80% if you use that cut. The Pareto chart template does the sorting and drawing in Excel.

Worked example: scrap on one line, by count and by cost (illustrative)

This example is made up. A machining cell scrapped 500 parts in a month across eight causes. The first table is the Pareto by count.

Pareto by count: 500 scrapped parts (illustrative)
CausePartsShareCumulative
Burrs18036.0%36.0%
Wrong dimension11022.0%58.0%
Scratches7515.0%73.0%
Porosity5010.0%83.0%
Dents357.0%90.0%
Paint run255.0%95.0%
Wrong label153.0%98.0%
Other102.0%100.0%

Check: 180 + 110 + 75 + 50 + 35 + 25 + 15 + 10 = 500. Four causes reach 83%, the first bar to pass 80%.

By count, burrs are the biggest problem and the first project. Now put a cost on each scrapped part, which is what it cost the plant to make up to the point it was scrapped. The costs per part, again illustrative, are: burrs $2, wrong dimension $8, scratches $3, porosity $12, dents $5, paint run $4, wrong label $20 and other $5.

Pareto by cost: the same 500 parts (illustrative)
CauseParts × cost per partCostShareCumulative
Wrong dimension110 × $8$88032.7%32.7%
Porosity50 × $12$60022.3%55.0%
Burrs180 × $2$36013.4%68.4%
Wrong label15 × $20$30011.2%79.6%
Scratches75 × $3$2258.4%87.9%
Dents35 × $5$1756.5%94.4%
Paint run25 × $4$1003.7%98.1%
Other10 × $5$501.9%100.0%

Check: 880 + 600 + 360 + 300 + 225 + 175 + 100 + 50 = $2,690.

Burrs fall from first to third: they are 36% of the parts and 13.4% of the money. Wrong dimension and porosity, 32% of the parts between them, are 55% of the cost. Wrong label is only 3% of the parts and 11.2% of the cost. Which chart to act on depends on the goal: to cut scrap cost, start with wrong dimension; to cut the number of rejects the operators handle, burrs are still worth a quick fix. Showing both charts side by side is what makes the choice an informed one.

How to read the chart

  • Start at the left. The first bars are the ones to work on, and the cumulative line tells you how much of the total they cover.
  • Look at where the line bends. A steep start and a flat tail means a few causes dominate. A line close to straight means the causes are about equal and a different cut of the data is needed.
  • Check Other. If it is bigger than the largest named bar, the categories are too coarse. Split it and count again.
  • Do not stop at one chart. Break the top bar down by machine, shift or product (a second Pareto inside the first) to find where it is made, as the reduce scrap and rework guide does for one month of reason codes.
  • Re-draw it after the fix with the same categories and the same period length. If the top bar did not come down, the fix did not work.

Mistakes that mislead

Common mistakes and what to do instead
MistakeWhat goes wrongDo this instead
Only counting frequencyThe cheap, frequent problem wins over the rare, costly oneBuild it by cost or minutes as well
Not sorting, or putting Other in the middleThe chart no longer shows the vital fewDescending order, Other last
A big Other barThe largest problem may be hiding in itSplit categories until Other is small
Too few eventsThe order of the bars changes next weekCount for long enough that the top bars hold
Treating 80% as exactWork stops at a line that means nothingUse it as a guide to where the bars level off
Stopping at the chartThe chart picks a problem but nothing is fixedHand the top bar to a root cause method and an owner

What to do after the chart

The Pareto names the problem. It does not explain it. Take the top bar to a fishbone if it could come from several areas, or to the 5 Whys if it looks like one chain, and use the problem solving methods guide to choose between them. For downtime, the reduce machine downtime guide works a month of stops through the same sort-and-cut.

To keep a Pareto alive instead of redrawing it by hand each month, LeanSuite's loss and cost management page describes putting a cost on each loss so the biggest can be fixed first, and its micro-stops page covers logging short stops so they show up in the ranking.

Definitions and notes

  • The Pareto chart is one of the seven basic quality tools: bars in descending order with a cumulative percentage line (ASQ, "Pareto chart"). It applies the 80/20 idea attributed to Vilfredo Pareto and brought into quality management by Joseph Juran.
  • The scrap counts, costs per part and totals are made up for illustration. They are not from a real plant. Every share is the cause's value divided by its table's total, rounded to one decimal.
  • Cost per scrapped part here is the cost put into the part up to the step where it was scrapped. How you set it is a decision for your own finance team.

Go deeper

This page is the short version. These guides cover each part in full.

All our articles on this subject are in the root cause analysis and problem solving topic.

Free templates and tools

Key terms

Lean glossary

Doing this in LeanSuite

See it on your own lines

Bring the problem you are working on to a 45-minute demo. Or look around with sample data first, or pilot LeanSuite on your own lines for 90 days, with the setup done by our team.

Book a 45-minute demoClick through the interactive demoStart a 90-day pilot

FAQ

Questions about pareto chart

  1. It is used to decide what to work on first. It ranks the categories of one problem, such as defect types or downtime reasons, from largest to smallest, with a cumulative line, so the few categories that make up most of the total stand out. It is usually drawn before a root cause investigation.

  2. Read from the left. The tallest bars are the biggest contributors, and the cumulative line shows how much of the total they cover together. Look for where the line starts to flatten: the bars before that point are the ones to work on. A near-straight line means no single cause dominates.

  3. It is the idea that a few causes, often about 20% of them, account for most of a result, often about 80%. It is a rule of thumb, not a law. Your own data may show 90% from two causes or 60% from five, and the chart shows which it is.

  4. Build both when you can. A count shows what happens most often and a cost shows what hurts most, and the order often differs. In the worked example on this page, burrs are first by count at 36% of parts but third by cost at 13.4% of the money, so the choice of measure changes the first project.

  5. A Pareto chart ranks separate categories, such as defect types, in descending order. A histogram groups measured values of one characteristic, such as a diameter, into ranges in their natural order to show how the values are spread. Use a Pareto to choose a problem and a histogram to understand variation.

Related guides and tools

More method guides

Or browse the topics, the full library and the lean glossary.

Rather see it on your own problem?

Bring the problem you are stuck on, and we will show you on a call how LeanSuite handles it.

Schedule Demo

Pass it on

We want LeanSuite to be the best place on the internet for lean help. If this was useful, send it to someone on your team or in your network who needs it.