Last reviewed: June 19, 2026
A Pareto analysis for manufacturing scrap or downtime is a ranked chart that shows which defect categories or failure modes are causing the most loss. In plants we have assessed, the top one or two categories almost always account for 60 to 70 percent of total scrap cost. Building the Pareto takes one to two hours. Acting on what it shows is what produces results.
Key takeaways:
In plants we have assessed, the top one or two defect categories typically account for 60 to 70 percent of total scrap cost. Fix those two before working on anything else.
Use at least 30 days of reason-coded data. Below two weeks, the Pareto reflects the events of those specific days, not the underlying process.
The most common mistake is sorting by count when cost is the right variable. Five defects at $4,000 each outrank 200 defects at $5 each, every time.
A Pareto by category is a prioritization tool. A Pareto of individual events is a sorted list.
WHAT A PARETO ANALYSIS IS AND WHY IT WORKS IN MANUFACTURING
A Pareto chart applies the 80/20 principle: a small number of causes produce most of the loss. For manufacturing scrap, this means a small number of defect categories produce most of the scrap cost. For downtime, a small number of failure modes produce most of the lost time.
The Pareto works as a prioritization tool because it forces a choice. Without a Pareto, teams work on the most recent problem, the loudest problem, or the problem the plant manager noticed this morning. All three generate improvement activity that does not reliably target the biggest loss. The Pareto shows the biggest loss directly, sorted by magnitude, so the team works on the top category rather than the most recent one.
The Pareto does not tell you why the top category is high. That is what a 5-Why analysis is for. The Pareto tells you where to apply the 5-Why. The combination is the most reliable scrap reduction sequence we have used across plants of different sizes and process types.
HOW TO BUILD A PARETO CHART STEP BY STEP
Step 1: Collect reason-coded scrap or downtime data for at least 30 days.
Every scrap or downtime event needs a category label from a fixed list of reason codes. Eight to twelve codes cover most operations. The data should be collected at the time of the event, at the work center, by the operator or supervisor who was present. Reconstructed end-of-day logs are less reliable and often push events into "other." Below 30 days of data, random variation will distort the Pareto enough to send you after the wrong category.
Step 2: Assign a cost to each event.
Use standard cost per part for scrap. Use a fully loaded hourly rate for downtime: idle labor, overhead that runs regardless of whether the machine runs, and restart scrap. If you do not have an established hourly downtime cost, a starting estimate for a mid-market operation is $500 to $1,500 per hour depending on the work center and labor rate.
Cost is the correct sorting variable. Count tells you which event happened most often; cost tells you which event matters most to fix. These are rarely the same ranking.
Step 3: Aggregate by category, not by individual event.
Sum the total cost within each reason code category. The unit of analysis is the category, not the individual occurrence. If you have 47 separate burr events, they sum to one row in the Pareto table under "Tooling/Burr." A Pareto built from individual events is a sorted list, not a Pareto.
Step 4: Rank categories from highest cost to lowest.
Sort the aggregated table so the most costly category appears first and the least costly appears last. Calculate each category as a percentage of the total scrap or downtime cost.
Step 5: Build the cumulative percentage column.
Add a cumulative percentage column by summing category percentages from the top down. When the cumulative percentage reaches 70 to 80 percent, you have identified the categories that account for most of the problem. In most operations this is the top two or three categories. These are the ones that need structured investigation first.
Step 6: Assign a 5-Why investigation to the top category.
Pick the two or three largest individual events within the top-ranked category and run a 5-Why on each. Run it in front of the machine with the operator and supervisor who were present, within one week of the event. The Pareto tells you where to look. The 5-Why tells you what to change.
Use the blank Pareto analysis template to build your own. It opens in Excel or Google Sheets and includes the five columns (reason code, event count, cost, percent of total, cumulative percent) and the sorting and thresholds notes.
A WORKED EXAMPLE WITH SCRAP DATA
A precision stamping plant logged all scrap events for 30 days across five work centers. Eight reason codes were in use. Here is the resulting Pareto:
| DEFECT CATEGORY | EVENT COUNT | COST ($) | % OF TOTAL | CUMULATIVE % |
|---|---|---|---|---|
| Tooling failure / burr | 22 | 9,680 | 38.2% | 38.2% |
| Setup / first-off reject | 41 | 7,380 | 29.1% | 67.3% |
| Feed / transfer error | 18 | 3,960 | 15.6% | 83.0% |
| Incoming material | 9 | 2,430 | 9.6% | 92.6% |
| Handling damage | 27 | 1,350 | 5.3% | 97.9% |
| Other | 6 | 540 | 2.1% | 100.0% |
The top two categories (tooling failure and setup rejects) account for 67 percent of total scrap cost from 63 of 123 total events. The 5-Why investigation would run on the three largest tooling failure events and the three largest setup reject events.
Notice that handling damage produced 27 events, the second-highest count in the table. A count-ranked Pareto would have placed it much higher and sent the team after a $1,350 problem while a $9,680 problem waited. The cost ranking corrects this automatically.
Also notice that "Other" is 2.1 percent of total cost. The rule of thumb is that "Other" should not exceed 10 percent of total events or cost over a 30-day period. If it does, the reason code list is missing a specific category and needs one more code.
THE MOST COMMON PARETO MISTAKE
The most common mistake is sorting by count instead of by cost. Count tells you which event happened most often. Cost tells you which event matters most. These are rarely the same ranking. Using count alone can direct significant improvement effort at high-frequency, low-cost events while ignoring low-frequency, high-cost events.
The second most common mistake is building the Pareto from less than two weeks of data. A bad material lot, an unusual setup, or an atypical shift schedule can make a minor category look like the biggest problem when the window is too short. Thirty days filters most of this noise.
The third mistake is too many categories. A Pareto with 20 or 25 bars has been built at the wrong level of abstraction. The tail categories are fragments, not meaningful groups. Merge categories with fewer than five events or less than one percent of total cost into a single "Other" bucket. The Pareto should have no more than 12 to 15 categories.
WHAT TO DO ONCE YOU HAVE THE CHART
The Pareto is a starting point, not a finish line. Once you have the top category, the next step is a structured root cause investigation on the largest individual events in that category.
The 5-Why analysis for manufacturing defects covers the full investigation process: how to run the analysis in front of the machine, a fully worked example from event to countermeasure, and how to avoid the "operator error" conclusion that leaves the underlying system unchanged.
The Pareto and the 5-Why together form the core of an effective scrap or downtime reduction program. The Pareto tells you where to look. The 5-Why tells you what to change. Together they give you a countermeasure that addresses the biggest driver of loss with a root cause traceable to a specific system gap rather than a general call for more effort.
For a broader look at the full scrap reduction sequence including reason code design, first-pass yield tracking, and verification, the guide on how to reduce scrap rate in manufacturing covers the program from data collection through verified closure.