Pareto Chart: What It Is and How to Make One (with Example)

Most problems have many causes, and they are rarely equal. A support team might log twenty complaint reasons a month, but three or four of them often fill most of the queue. A Pareto chart makes that visible in one picture, and it takes minutes to build in a spreadsheet.

This guide covers what a Pareto chart is and how to read it, how to make one step by step in Excel or Google Sheets, a worked example on customer complaints, the mistakes that make the chart misleading, how it fits with other root cause tools, and when to re-cut the data.

What is a Pareto chart?

A Pareto chart combines two charts on one set of axes:

• Bars, one per cause or category, sorted from tallest to shortest. Their height is the count (or the cost) on the left axis.
• A line showing the cumulative percentage: the first bar's share, then the first two together, and so on until 100%. It is read on the right axis.

Many charts also draw a dashed horizontal line at 80%. The bars to the left of the point where the cumulative line crosses it are the causes to tackle first.

Because the bars are sorted, the cumulative line always rises steeply at first and flattens out. How steep that first part is tells you how concentrated the problem is.

Where the 80/20 idea comes from

The chart is named after Vilfredo Pareto, an Italian economist who studied how unevenly income and wealth were distributed. The observation usually quoted from his work is that about 80% of the land in Italy was owned by about 20% of the people.

Joseph M. Juran, one of the pioneers of quality management, applied the idea to quality problems: a small number of causes produce most of the defects. He called it the Pareto principle and described the causes as "the vital few and the trivial many", later preferring "the vital few and the useful many" so that the rest would not be dismissed as worthless. The Pareto chart became one of the basic tools of quality improvement.

That is where "the 80/20 rule" comes from. It is a pattern, not a law of nature: in real data the split might be 70/30, 90/10 or anything else, and the two numbers do not have to add up to 100.

How to read a Pareto chart

1. Look at the left side first. The tallest bars are the most frequent causes.
2. Find where the cumulative line crosses 80%. Count the bars up to that point. Those are your "vital few".
3. Check the shape. A steep line that passes 80% after two or three bars means the problem is concentrated, and fixing a few causes will move the total. A line that climbs slowly and evenly means the causes are spread out, or that the categories need rethinking.
4. Check the "Other" bar. By convention it sits at the far right whatever its size, and it should be small. If it is large, the chart is hiding something.
5. Read the axis. Make sure you know whether the bars count tickets, defects, minutes or money. The top cause by count is not always the top cause by cost.

When to use Pareto analysis

• To decide where to start when a problem has many causes: complaints, returns, defects, ticket reasons, churn reasons in exit surveys.
• To check whether a fix worked, by comparing the chart before and after. The bar you worked on should shrink.
• To explain a priority to your team or leadership with data instead of opinion.
• To narrow down a root cause analysis: run 5 Whys or a fishbone diagram on the top bars, not on everything.

It is less useful when you have only a handful of records, when the categories overlap heavily, or when one rare event matters more than all the frequent ones put together (a safety incident, for example).

How to make a Pareto chart, step by step

The method is the same in any tool:

1. Define the problem and the period. For example, "all customer complaints received in October". Make the period long enough to smooth out one-off spikes.
2. Choose what to measure. Usually a count of records. Use cost, time or revenue instead if the causes differ a lot in impact, but use one unit for the whole chart.
3. Define the categories. Each record should fit exactly one category. Aim for roughly 5 to 10 categories plus "Other".
4. Count the records in each category.
5. Sort the categories from largest to smallest, with "Other" always last.
6. Calculate each category's share and the cumulative percentage. Share is the category's count divided by the total; cumulative is the running total divided by the total.
7. Draw the chart: bars for the counts on the left axis, a line for the cumulative percentage on a right axis from 0% to 100%, and a dashed line at 80%.

In Excel 2016 or later (built-in Pareto chart)

1. Put the categories in column A and the counts in column B, with a header row.
2. Select both columns.
3. Go to Insert → Insert Statistic Chart → Pareto (it sits under the histogram charts).

Excel sorts the bars and adds the cumulative line for you. It does not draw an 80% line or mark the top causes.

In any version of Excel (combo chart)

1. Sort your table by count, largest first, and move "Other" to the bottom.
2. In column C, calculate the cumulative percentage with the formula below, fill it down and format the column as a percentage.
3. Select the categories, counts and cumulative column, then go to Insert → Insert Combo Chart → Create Custom Combo Chart.
4. Set the counts to clustered columns and the cumulative percentage to a line on the secondary axis.
5. Fix the secondary axis to run from 0% to 100%. To add the 80% line, add a column with 0.8 in every row and plot it as a second line on the same axis.

The cumulative formula for cell C2, with data in rows 2 to 10:

=SUM($B$2:B2)/SUM($B$2:$B$10)

In Google Sheets

Google Sheets has no Pareto chart type, so you build the same combo chart. Sort the data (Data → Sort range), add the cumulative column with the same formula, then choose Insert → Chart → Combo chart. In the chart editor, under Customize → Series, put the cumulative series on the right axis and set that axis from 0 to 1.

Or start from a template

The free Pareto chart template does all of this for you: type causes and counts in any order, and it sorts them, calculates the shares and cumulative percentage, marks the top 80% and draws the chart. The formulas work in Excel, Google Sheets and LibreOffice, with no macros.

Worked example: 1,240 customer complaints

An online retailer grouped one month of complaint tickets by reason. This is an illustrative example, not customer data.

How to read it. The three causes in bold (late delivery, damaged items and late refunds) make up 80.3% of all complaints: a third of the causes drive four fifths of the volume. Late delivery alone is 39.2%. The other six reasons together are 19.7%.

What the team did next. Each of the top three got its own root cause analysis. Late delivery had one clear chain of events, so the team used 5 Whys. Damaged items had possible causes in packaging, the warehouse, the carrier and the product, so a fishbone diagram fitted better. Late refunds turned out to be a process question, so they reviewed the refund process step by step. They planned to repeat the Pareto chart a month after the fixes.

What the chart does not say. It does not say why deliveries are late, and it does not say which cause costs the most. A billing error that charges a customer twice is rare here (23 complaints), but it may cost more trust per case than a late parcel. If that matters, make a second Pareto chart weighted by cost or by customer impact.

Common mistakes

• Too many categories. Twenty-five thin bars hide the pattern. Merge rare, similar reasons, or move them into "Other".
• Too few, too broad categories. "Delivery problems" as one bar tells you nothing you can act on. Split it until each category points to something a team could fix.
• Mixing units. Do not put ticket counts next to refund amounts or minutes of downtime on the same chart. One chart, one unit.
• An "Other" bucket that is too big. If "Other" is more than about 5 to 10% of the total, real causes are hiding in it. Open those records and reclassify them.
• Treating 80/20 as a law. If your top 20% of causes add up to 65%, the data is not wrong. Look for the point where the line flattens, not for a magic number.
• Recording symptoms as causes. "Customer unhappy" or "bad experience" is not a category. Use observable reasons such as "parcel arrived after the promised date".
• Counting each record more than once. If one complaint mentions two problems, decide on a rule (primary reason only, or count both and say so) and apply it every time.
• Using too short a period. One week of data can make a random spike look like the biggest cause.
• Treating the chart as the answer. A Pareto chart tells you where to look, not why it happens.

Pareto chart vs other root cause tools

A Pareto chart is a prioritisation tool. It works best as the first step, before the methods that explain causes.

A typical sequence: a Pareto chart picks the top causes, 5 Whys or a fishbone diagram finds the root cause of each, a corrective action plan fixes it, and a second Pareto chart confirms the bar has shrunk. For the full picture, see the root cause analysis guide for customer feedback.

When to re-cut the data

A single Pareto chart is a starting point. Make a new one when:

• One bar is too broad to act on. Break the top bar into its own Pareto chart, a second-level cut (example below).
• The chart is flat. If no cause stands out, the categories may be too narrow, or the problem may differ by segment. Try grouping the categories differently, or split the data by channel, region, product or customer type.
• "Other" keeps growing. New problems often show up in "Other" first. Read those records and add a category.
• Frequency and impact disagree. Make a second chart by cost, refund value or customer lifetime value and compare the two rankings.
• After a fix. Repeat the chart for the same length of period. If the bar you worked on did not shrink, the root cause you found was not the real one.
• On a regular rhythm. A monthly or quarterly Pareto chart shows whether the vital few are changing.

Example: a second-level cut

In the worked example, late delivery was the biggest bar but too broad to fix as one problem. The team split the 486 late deliveries by where the delay happened (illustrative):

Here two causes make up 79.2%, so the first three together pass 80%: the split is not a neat 80/20, and it does not need to be. Parcels that left the warehouse after the carrier had already collected that day are the single biggest cause, which gives the 5 Whys a precise starting point.

The hard part: clean categories from free-text feedback

A Pareto chart is only as good as its categories. Tickets that already carry a reason code are easy to count. Much customer feedback does not arrive that way: survey comments, app reviews, chats and call transcripts are free text, the same problem is described in many different words, and one comment often mentions two issues. Tagging a few hundred comments by hand works. Tagging tens of thousands every month does not, and two people often tag the same comment differently.

That is the step Pivony's root cause analysis is built for:

• Themes without a coding scheme. Pivony's NLP models discover themes from your feedback without a pre-defined list of categories, and emerging issues surface automatically, including ones your team has not anticipated.
• All your feedback in one place. Tickets, NPS surveys, call centre transcripts and app reviews are analysed together.
• Segments and operational data. Feedback is blended with data such as carrier, sales channel, customer tier and region, so a bar like "late delivery" becomes which carrier, which region, which customers and since when.
• Key driver analysis. It shows what drives satisfaction most, not just what is complained about most, which is exactly the gap a count-based Pareto chart can miss.

You can still chart the result as a Pareto and discuss it the same way. The difference is that the categories come from all of your feedback, not a hand-tagged sample.

Related reading

• Root cause analysis in customer feedback: the complete guide
• 5 Whys root cause analysis: method and examples
• Fishbone analysis: complete guide for CX teams
• Sentiment analysis: the complete guide for customer feedback

Next step

Make your own chart with the free Pareto chart template: an Excel workbook that sorts your causes and draws the chart, plus a Word worksheet and a printable PDF. If your causes are buried in thousands of comments, request a free root cause analysis on an export of your tickets, reviews or survey comments.