Five defect patterns inspectors find during during production inspection (dpi)
The same five defect patterns appear in a large share of during production inspection reports: dimensions that drift, seams that fail, surfaces that mark, assemblies that do not function, and labels or cartons that do not match the packing list. None of them is exotic, and that is exactly why they are dangerous. This guide explains what each pattern looks like on the line and how experienced buyers respond before the defects get sealed into finished cartons.
Key takeawaysDPI samples live output at roughly 20 to 30 percent completion, when correction is still inexpensive compared with post-packing rework. · The five recurring families are dimensional drift, stitching and seam faults, surface and finish damage, assembly and function failures, and labeling and packing errors. · Each finding should be graded critical, major, or minor against the reference sample, because the grade drives the pass, fail, or hold decision [1]. · Most patterns trace back to process control gaps such as worn tooling, skipped first-article approval, or unclear work instructions rather than isolated worker mistakes [2]. · A useful DPI report names the cause, the containment for finished units, and a dated re-check so the buyer can verify the fix on the line.
When buyers hear about during production inspection, many picture an inspector walking the line with a clipboard. The real value is narrower and more practical: a trained third party pulls samples from live output, compares them against the approved reference, and flags the defect patterns that tend to repeat once a line is running at speed.
Those patterns matter because mid-production is the last moment when correction is still cheap. Once goods are packed and cartoned, fixing a stitching skew, a weak seal, or a color drift means unpacking, rework, and often a delayed vessel. Catching the same issue at 20 to 30 percent completion usually means adjusting one machine or one workstation.
This article explains the five defect families that show up most often during DPI visits across sewn goods, assembled products, and packaged consumer goods. For each family you will see what the inspector checks, why the pattern starts, and what a reasonable corrective action looks like. The aim is to help you read a DPI report with confidence and decide between continue, correct-and-continue, or hold.
Why mid-production defects repeat in patterns
Defects found during production rarely arrive as one-off surprises. Once a line settles into its rhythm, the same combination of material, machine setting, and operator habit produces the same result hundreds of times per shift. A needle tension set slightly tight, a sealing jaw running slightly cool, or a cutting die with a worn edge will mark nearly every unit the same way until someone intervenes.
That repetition is good news for buyers. A pattern can be traced, measured, and corrected with a single intervention, while random scattered faults are far harder to eliminate. The inspector's job during DPI is therefore not only to count defects but to describe the pattern clearly enough that the factory engineer can find the station responsible. Photos grouped by defect type, with notes on where on the line the affected units came from, are the most useful part of the report.
Quality management thinking supports the same approach: control the process and the product follows [2]. When work instructions are clear, first-article approval is documented, and in-process checks happen at defined intervals, drift gets caught within the hour. When those controls are missing, the line can run a full shift of defective output before anyone notices. Many buyers ask UTS Inspection to attend the line at the point when output first reaches a samplable quantity, so the pattern is identified before it fills a container.
Pattern 1 and 2: dimensional drift and stitching faults
Dimensional drift means finished units measure consistently over or under the specification, often by just a few millimeters. Knit panels relax after cutting, plastic housings shrink as they cool, and metal brackets vary when a press brake stop slips. A single unit slightly off tolerance may still assemble, but a systematic shift breaks mating parts, packaging fit, and ultimately the customer experience.
Inspectors check this pattern with calibrated tapes, calipers, and go or no-go gauges, measuring across sizes and cavities rather than a single sample. They record the range and the average, not just the worst piece, because the spread tells the factory whether the issue is a wrong setting or an unstable process. A tight spread offset from nominal points to a setting error; a wide spread points to unstable material or a loose fixture.
Stitching and seam faults form the second family, covering skipped stitches, puckering, open seams, and seam strength below the reference. Common causes include wrong thread tension, a needle past its replacement interval, or an operator hurrying a curve to meet rate. Inspectors pull seam samples and stress them by hand, then compare stitch density against the approved sample under a magnifier. As with dimensional checks, acceptance logic follows the same discipline used in formal acceptance testing: compare observed results against agreed criteria and record the evidence [1].
| Pattern | What the inspector sees | Likely process cause | Sensible correction |
|---|---|---|---|
| Dimensional drift | Consistent over or under size across units | Slipped machine stop, wrong pattern, material relaxation | Reset and lock the stop, re-approve first article |
| Stitching and seam faults | Skipped stitches, puckering, weak seams | Thread tension, worn needle, rushed curves | Retension, replace needle, retrain that station |
| Surface and finish damage | Scratches, dents, coating misses | Unprotected handling, dirty jigs, thin spray pass | Add station protection, clean jigs, recoat |
| Assembly and function failures | Loose fits, dead buttons, weak bonds | Wrong torque, missed step, under-cured adhesive | Calibrate tools, add checklist step, extend cure |
| Labeling and packing errors | Wrong labels, mixed sizes, weak cartons | Outdated packing sheet, mixed materials | Update packing sheet, segregate materials |
Pattern 3 and 4: surface damage and function failures
Surface, coating, and cosmetic damage
Surface damage is the most photographed DPI finding because it is visible and easy to dispute. Scratches from unprotected stacking, dents from hard handling, dust trapped under paint, and thin electroplating at edges all belong here. The root cause is usually handling and workplace organization rather than the core manufacturing technology, which means the fix is often simple: sleeves between units, lined bins, cleaned fixtures, and a defined recoat procedure.
Assembly fit and functional failures
Function failures carry higher stakes because the customer discovers them at first use. Buttons that stick, hinges that bind, connectors that seat loosely, and adhesive joints that peel under a light pull all signal that assembly or curing parameters drifted. Inspectors test a defined subset of functions on every sampled unit and record which station the failed units came from. When failures cluster by operator or by hour, the cause is usually a skipped step or an uncalibrated tool rather than a design flaw.
Pattern 5: labeling, barcodes, and packing errors
The fifth family surprises first-time importers because the product itself can be fine while the shipment still fails. Wrong size marks, outdated care labels, unscannable barcodes, mixed assortments in one carton, and cartons below the agreed bursting strength all create chargebacks and warehouse rejections. These errors usually start at the packing table, where an outdated packing sheet or mixed label rolls meet time pressure.
Inspectors verify this family by scanning barcodes with a phone app, checking label content line by line against the purchase order, and confirming carton counts and markings. Drop and stacking checks on a sample carton reveal whether the agreed packaging actually protects the goods. Because labeling and packing are the last operations before the container door closes, errors here are the most expensive to fix after the fact and the simplest to prevent with an updated packing sheet and segregated materials.
A DPI report that only lists defect counts leaves the buyer guessing. Ask the inspector and the factory to close three questions before the line restarts at full speed: what caused the pattern, which finished units are contained, and when will a re-check confirm the fix. Write the answers into the inspection follow-up so the next visit can verify them directly on the line.
Sources
- Wikipedia: Factory acceptance test —— Overview of factory acceptance testing performed before shipment or installation.
- Wikipedia: ISO 9000 —— Overview of the ISO 9000 family of quality management standards.
Frequently asked questions
How do inspectors decide whether a defect is critical, major, or minor during DPI?
Inspectors typically pull units from current output and recently completed cartons, then classify each finding as critical, major, or minor against the approved reference sample and specification sheet. The classification matters because a single critical defect can fail a lot, while major and minor findings are judged against agreed AQL limits. Photos of each defect, with a ruler or coin for scale where useful, go into the report so the factory and the buyer see the same evidence. This discipline is what separates a casual walkthrough from a real in-process control.
How long does a during production inspection visit take, and when should I schedule it?
A DPI visit usually lasts one working day for one product line, covering opening meeting, line walk, sample pull, testing, and a closing meeting with the factory. The written report with photos normally follows within one working day after the visit. Buyers should plan the visit when at least 20 percent of the order is finished so there is enough output to sample, and leave time for a re-check if corrections are needed.
The DPI report shows repeating stitching or sealing defects. What should I ask the factory?
Ask for the suspected cause, the containment action for goods already produced, the correction for goods still on the line, and the date of the re-check. A credible reply names a specific process change, such as recalibrating a heat sealer or replacing a worn cutting die, rather than promising that workers will simply be more careful. If the factory cannot explain the cause, keep the shipment on hold and arrange a follow-up visit focused on that defect family.
Should I order a follow-up inspection if the first DPI finds systematic defects?
Yes, when the same pattern appears across two visits or the factory cannot show that the root cause is fixed. Re-checks can be limited in scope to the failed checkpoints, which keeps cost down while confirming that the correction actually worked on the line. Many buyers write this rule into the purchase order so the supplier expects it. The goal is to verify, not to punish, and a short targeted visit usually resolves the question.