Every textile plant has more quality problems than it has time to fix. Pareto analysis is how you choose which ones to fix first. Rank defect categories by their impact, add up the running share, and focus on the few categories that account for most of the loss. It sounds simple, and it is, but most textile Paretos are built on the wrong measure, the wrong categories or the wrong time window, and they point teams at the wrong problems. This article shows how to build a Pareto that holds up, why scrap and customer complaints need separate charts, and how to turn the top bars into root causes. To see a live Pareto built from inspection data, book a short walkthrough.
Textile Defect Pareto Analysis: Rank Scrap and Complaints by What They Actually Cost
Defects ranked by points, meters and cost rather than raw counts, split by loom, lot and shift, so improvement work starts where the losses are.
Why Most Textile Paretos Point the Wrong Way
The Pareto principle, popularized in quality management by Joseph Juran, says a small number of causes usually produce most of the effect. In a weaving or finishing plant, that often means three or four defect types are behind most of the second-quality fabric. Find them, fix them, and quality moves faster than any broad campaign could manage.
The trouble is that the chart only ranks what you feed it. Count every defect equally and hundreds of tiny slubs will outrank a handful of long shade bars that downgraded whole rolls. Mix symptom names with cause names and the same problem appears under three labels, each too small to rank. Use one day of data and the chart changes every morning. Each of these mistakes sends the improvement team after the wrong problem.
A good Pareto fixes all three: the right measure, clean categories and a stable window. We can review your current charts with you on a call.
How to Build a Textile Defect Pareto Step by Step
The method is the same whether you are ranking weaving defects, finishing faults or garment complaints. What changes is the data behind it.
Merge synonyms and name causes, so each defect sits in one bucket.
Choose points, scrap meters or cost to match the question.
Use a period long enough that small categories are stable.
Sort largest first and add the running percentage.
Mark the categories that reach roughly 80% of the total.
Split the top bars by loom, lot, shift or article.
Categories deserve the most care. Cotton Incorporated groups fabric defects into warp-wise, filling-wise, isolated, pattern, finishing and printing defects, which is a useful top level. Below that, keep a fixed code list, such as broken end, slub, float, oil stain, hole and shade bar, and do not let free-text names creep in.
When defects are recorded by AI vision, the category comes from the model’s defect library, so naming stays consistent across frames and shifts. Our engineers can map your current codes into that library.
Count Versus Cost: Why the Measure Changes the Answer
The same week of data can produce two very different Paretos depending on what is measured. Here is an illustrative example from a finishing plant.
Illustrative data. By count, slubs dominate and shade variation barely registers.
Illustrative data. Measured by scrap, shade variation leads because each case downgrades long lengths, while slubs fall to the bottom.
Neither chart is wrong. They answer different questions. The count chart shows where inspectors spend their time; the scrap chart shows where money is lost. For improvement priorities, a measure tied to value, such as four-point points, meters downgraded or cost, is almost always the better choice.
Where fabric prices are known, converting meters to cost makes the chart speak the language of management. That conversion is set up once in the software.
Scrap and Complaints Need Separate Charts
Internal scrap and customer complaints describe different failures. Mixing them hides both.
- Built from inspection records
- Shows defects that were caught
- Measured in points, meters or cost
- Points at process problems in the plant
- Owned by production and QC
- Reviewed daily or weekly
- Built from customer claims and returns
- Shows defects that escaped inspection
- Measured in claims, units or claim value
- Points at gaps in detection as well as process
- Owned by QC and commercial teams
- Reviewed weekly or monthly
Comparing the two is revealing. A defect that ranks high in complaints but low in scrap is escaping inspection, so the fix is detection, not only the process. A defect that ranks high in scrap but never reaches customers is being caught; the fix is upstream, to stop making it.
Both charts should draw on the same defect codes, so a complaint can be matched to the inspection record of the roll behind it. That link is shown in a short demo.
Split the Top Bars to Find Where the Problem Lives
A top-ranked category tells you what, not where. Stratifying the top bars by source turns a category into a location.
If one or two looms carry most of the broken ends, the fix is maintenance, not a plant-wide campaign.
Slubs and barre that follow a yarn lot point to the spinner or the yarn store.
A defect concentrated on one shift points to set-up, handover or training.
Shade and stain problems grouped by dye lot direct attention to recipes and machines.
Some constructions are simply harder to make; they need tighter process windows.
Defects clustered at roll starts or ends often come from handling, doffing or joins.
This is where traceability pays off. Stratifying is only possible when every defect carries its roll, and every roll carries its loom, lot and shift. Without that link, the Pareto stops at the category.
Normalize before comparing. A loom that ran twice as many meters will naturally show more defects, so compare points or scrap per thousand meters produced rather than raw totals.
A second-level Pareto, such as broken ends by loom, is usually the chart that gets action. Ask our team for examples from similar plants.
From the Top Bar to a Fix That Holds
A Pareto identifies the problem to work on. It does not solve it. The steps below turn a top-ranked category into a verified fix.
Name the defect, where it occurs and how big it is, using the stratified Pareto.
Use a fishbone across machine, material, method, people, measurement and environment.
Check each likely cause against the records: does the defect follow the loom, the lot or the shift?
Act on the confirmed cause, not the symptom, and record what was done.
Watch the same Pareto bar over the following weeks to confirm it falls and stays down.
The verify step is the one most often skipped. A bar that falls for a week and then returns means the fix treated a symptom. Keeping the before-and-after view on the same chart makes that visible.
Linking each action to the Pareto bar it targets is what turns a chart into an improvement system. That link is built into the action log.
Pareto Mistakes to Avoid
These are the errors that most often make a textile Pareto misleading. Each has a simple remedy.
Most of these are prevented by fixed codes, a value-based measure and a standard review rhythm. We can help you set those up during the pilot.
How iFactory Delivers Defect Pareto Analysis
Defect categories from one library across all frames.
Rank by points, meters downgraded or cost, not only counts.
Two linked Paretos from inspection and customer claims.
Split any bar by loom, yarn lot, shift, dye lot or article.
Each fix linked to the bar it targets and tracked over time.
Daily and weekly Paretos sent to the right owners.
It uses your existing ERP and complaint records alongside inspection data. See a Pareto from data like yours in a session.
See Which Defects Really Cost You Most
Share a month of inspection and complaint records, or run AI vision on one frame. We build count, scrap and complaint Paretos and split the top bars by source.
Oil stains moved from fifth to second place by scrap meters this week, all on one stenter.
A Weekly Pareto Review With iFactory
This is how a finishing plant manager might run the weekly quality review with iFactory.
iFactory ships as a pre-configured NVIDIA AI server, racked and ready with the defect classification and Pareto analytics models loaded. Rack it, plug in power and Ethernet, and the AI is live on your network. Our scope covers cameras and lighting on inspection frames, PLC/SCADA and ERP integration, cabling and network setup, operator and QC team training, and 24×7 remote monitoring.
Server installed, cameras and lighting mounted, historical inspection and defect records loaded.
Models trained on your own fabrics and styles, then piloted on one line with your QC team reviewing every call.
Rollout to the agreed lines, inspector and supervisor training, ERP hand-off and 24×7 remote monitoring in place.
Hardware, software and integration come as one package. For a quote scoped to your plant, contact our sales team.
Frequently Asked Questions
It ranks defect categories by their impact and shows the running share, so teams can focus on the few categories that account for most of the loss. In textiles it is used for inspection scrap, customer complaints and process faults.
For improvement priorities, use a measure tied to value, such as four-point points, meters downgraded or cost. Counts treat a tiny slub and a long shade bar as equal, which can rank the wrong problems first.
Scrap shows defects that were caught; complaints show defects that escaped. Comparing them tells you whether the fix is in the process, in detection or both.
Enough that the ranking is stable from one period to the next. Many plants use weekly or monthly windows normalized by meters produced, rather than a single day.
Stratify the top bars by loom, lot or shift, run a root cause analysis on the confirmed source, fix it and verify that the bar falls and stays down.
Most plants have Paretos running within the first weeks of a 6–12 week rollout, with AI inspection and complaint links added during the pilot. Plan it with our team.
Fix the Few Defects That Cost the Most
iFactory ranks defects by what they cost, splits the top bars by loom, lot and shift, and tracks every fix until the bar comes down.
Ranked by scrap meters, not counts, so the costly causes rise to the top.
