What Does the Fifth-Reorder Test Consist Of?
By the fifth run a programme has been through four quiet changes it never recorded. The fifth-reorder test is the moment to stop comparing with the previous run and compare with run one. This article sets out the three comparisons, which properties deserve a laboratory at that point, and how the result changes run six.
['MOQ 500 pcs per style and colourway; 300 pcs open to discussion on a repeat.', 'Sampling 6-10 working days from a signed-off tech pack.', 'Bulk production 35-50 days from written sample approval.', 'Inspection at AQL 2.5, ISO 2859-1, General Inspection Level II.', 'T/T 30/70, FOB Xiamen; CIF and DDP available.', 'Run five is compared with run one, with run four, and against the bands.', 'A slow drift is visible only when two runs that are not neighbours are compared.']
Most reorder programmes compare each run with the one before it. That is the right check for a sudden change and the wrong check for a slow one. A difference of a small amount per run is inside every band on every comparison, and by the fifth run the total movement is larger than any band would have allowed.
The fifth-reorder test exists to catch that. It compares run five with run one, with run four, and against the agreed bands, and it asks a different question from the routine inspection: not does this lot pass, but has the product moved since it was approved.
Why Is Run Five the Right Moment for a Full Test?
Three reasons. By run five a year has usually passed, which is the period over which materials, machines and people change. Four comparisons have been filed, which is enough to build a trend line. And the cost of the extra testing is spread over a programme that has already proved it is going to continue.
It is also the moment when the retained references are most likely to have aged. The gold sample sealed at approval has been out of the drawer several times, and the archive sample is the only thing that can say whether the working one still describes run one.
There is a fourth reason that is purely commercial. By the fifth run the buyer has usually made a claim about the product in the market - a capacity, an hours figure, a material statement - and the fifth test is the natural moment to confirm that the claim still describes what is being shipped.
What Is Compared Against Run One Rather Than Run Four?
Everything with a band, plus the properties that were never given one. Shade, dimensions, foam thickness, stitch density, hardware codes and artwork position are compared against run one, because the tolerance was written against the approved run and not against the previous one.
| Comparison | Run four comparison | Run one comparison | What the gap means |
|---|---|---|---|
| Body shade | Shows whether the last lot moved | Shows whether the programme has moved | A drift that every single comparison passed |
| Dimensions | Catches a pattern or die change | Catches slow panel movement | Cutting drift accumulated over a year |
| Foam thickness | Catches a purchase change | Catches supplier tolerance creep | A material supply drifting inside its tolerance |
| Stitch density | Catches a setting change | Catches a thread or needle substitution | A component that moved without a note |
| Hardware code | Catches a substitution | Catches an accepted one-run exception that returned | A temporary fix that became permanent |
| Artwork position | Catches a print setup change | Catches accumulated fixture wear | Tooling ageing rather than a decision |
How Do You Detect a Slow Drift Nobody Noticed?
By laying the retained units from each run side by side in order and reading them as a series rather than as pairs. A drift of a small amount per run is invisible in any pair and obvious across five, which is the entire argument for retaining one unit per run and for dating them.
The second method is the trend line on the comparison sheets. Where the readings for one property move in one direction across four runs, the property is drifting whether or not each reading passed. That is the signal, and it is only visible on a sheet that records numbers rather than pass or fail.
A third method is simply to ask the people who make the bag what has changed. Asked plainly, and not as an accusation, a line supervisor will usually name the material lot that behaved differently or the fixture that was replaced. That answer is not evidence, but it tells you where to look, which is the expensive part of the exercise.
Lay the units out under the same light, in the same orientation, with the run numbers visible, and photograph the series once. The photograph is supporting evidence rather than a standard, but it is the fastest way to show a drift to someone who was not in the room, which is usually the person who has to approve the correction.
Which Properties Deserve a Laboratory on Run Five?
The short answer is fewer than most buyers expect and more than most programmes commission. Sending everything to a laboratory each year costs a great deal and produces a thick file that nobody reads; sending nothing means the only properties ever verified are the ones visible on a bench.
Three, typically: any property covered by a test report at approval, any property that supports a claim made to the market, and any property where the trend line has moved even though every reading passed. The third is the one programmes forget, and it is the one that finds the problem.
Use a laboratory working to a recognised competence standard so the report names its method. A number produced by an unnamed method cannot be compared with the number taken at approval, which makes the whole exercise decorative.
Send the laboratory units from the production run rather than from a fresh sample. The purpose of the fifth test is to find out what the programme has been shipping, not what it could make if the best operator made it carefully from the best material on the shelf.
Ask for the report to state which revision of the spec file the units were made to. A laboratory report that names the revision can be filed against that revision and read correctly a year later; one that does not has to be re-interpreted every time it is opened.
How Many Units Should Be Pulled for the Fifth Test?
More than a routine inspection and fewer than a full qualification. Two samples: the standard attribute sample for the lot under ISO 2859-1 at General Inspection Level II and AQL 2.5, plus a fixed panel of five units taken for measurement and destructive checks across the run.
Take the five from the first, middle and last pallets rather than from one carton. A drift that entered partway through the run is invisible in a single carton, and the fifth-reorder test is specifically looking for changes that entered partway through a year.
Keep the five units afterwards rather than returning them to stock. The fifth test is the first of a series, and next year comparison starts from these units rather than from nothing. A programme with three years of retained units can answer questions about its own history that no document can.
What Does a Trend Line Across Five Runs Show?
Four shapes are worth naming. A flat line means the property is controlled. A step means a change entered at a known run, and the change log should have an entry there. A steady slope means a slow drift, usually a material or a tooling one. And a scatter means the property is not under control at all.
- Flat
- Controlled; leave the band alone and keep recording
- Step
- A change entered at a named run; check the log for an entry there
- Slope
- Slow drift; look for material supply or tooling wear
- Scatter
- Not controlled; the band is not being held run to run
- Reversal
- A correction was applied; confirm it was recorded
- Missing
- The property was not recorded on that run; fix the sheet
The missing reading is more common than any of the others and it is the most damaging, because a trend line with a gap cannot distinguish a step from a slope. Recording the same properties on every run is what makes the fifth test possible at all.
How Is the Comparison Sheet Different on Run Five?
It carries three columns rather than one: the reading, the run one value, and the run four value, each with a pass or fail line against the band. That is the whole difference, and it is what turns a routine inspection into a programme review.
It also carries a conclusion line: whether the product has moved since approval, and if so on which property and by how much. A sheet without a conclusion is a sheet that gets filed and never acted on, which is the fate of most inspection paperwork.
Where the conclusion is that nothing has moved, the sheet should still name the properties that came closest to their band. Those are the ones to watch on run six, and naming them turns a clean result into a useful one rather than a formality that tells the reader only that the exercise happened.
What Happens If Run Five Fails a Property Run One Passed?
Treat it as a change that was never proposed. Establish which run introduced the movement by reading the trend line, find whether the change log has an entry there, and then decide: accept with a revised reference, re-qualify the property, or correct the process and return to the original band.
| Finding | What it usually means | What to do on run six |
|---|---|---|
| A step with a log entry | An accepted change whose reference was never updated | Update the reference and the register |
| A step with no log entry | A silent change that entered at that run | Record it, correct it, and tighten the loose register line |
| A steady slope | Material supply or tooling drifting inside tolerance | Ask the supplier for the batch record and shorten the review period |
| Scatter across runs | The property is not controlled | Widen the band honestly or fix the process first |
| One unit far outside | A single-run event rather than a drift | Check that run for a setting or a material lot |
| A missing reading | The sheet was not completed on that run | Fix the sheet before drawing any conclusion |
How Is the Result Written Into the Spec File?
Three lines: the conclusion, the properties that moved, and the decision taken about each. The file moves to a new revision where a reference was updated, and the revision log records the fifth-reorder test as the reason. That is what lets the next test compare against a known starting point rather than against a memory.
Where nothing moved, say so on the record as well. A fifth-reorder test that found no movement is evidence that the controls work, and it is the strongest argument against the proposal - made at some point in every programme - to stop doing the paperwork.
What Changes About Run Six Afterwards?
Three things, in the useful case: the bands are confirmed or tightened, the properties that drifted get added to the routine sheet rather than only to the annual test, and the review date is set. In the less useful case, the same three happen plus a correction to the register.
Six things the fifth-reorder test should leave behind
- One retained unit per run, dated, stored with the gold samples.
- A comparison sheet with the reading, the run one value and the run four value.
- A conclusion line stating whether the product has moved since approval.
- A trend chart for each property that has a band.
- A revised reference where an accepted change had never been written in.
- A date for the next full test, and the properties added to the routine sheet.
Every run can pass and the product can still have moved. Only a comparison with run one, made once a year, shows whether that has happened.
Send us four comparison sheets
Send the comparison sheets you hold from the last few runs, or say that there are none. We coordinate production with one dated retained unit per run and a trend chart per property, bulk in 35-50 days from written sample approval.
Frequently Asked Questions
What is a fifth-reorder test?
An annual review that compares the current run with run one rather than only with the previous run, across every property that has a band, plus a laboratory check on the properties that support a claim or show a trend. It asks whether the product has moved since approval.
Why compare with run one instead of the previous run?
Because a small movement per run is inside every band on every comparison, and only the total movement is visible against the approved run. Comparing with run one is what makes a slow drift detectable instead of perpetually passing.
When should the fifth-reorder test happen?
Once a year on a running programme, or at the fifth run, whichever comes first. By then a year of material, machine and staff changes has usually passed, four comparison sheets exist, and the extra cost is spread over a programme that is clearly continuing.
How many units should be pulled?
Two samples: the standard attribute sample for the lot under ISO 2859-1 at General Inspection Level II and AQL 2.5, plus a fixed panel of five units for measurement and destructive checks, taken from the first, middle and last pallets.
Which properties need a laboratory at that point?
Any property covered by a test report at approval, any property supporting a claim made to the market, and any property whose trend line has moved even though every reading passed. Use a laboratory working to a recognised competence standard so the method is named.
What does a trend line across five runs show?
Four shapes: flat means controlled, a step means a change entered at a known run, a slope means a slow drift in material supply or tooling, and a scatter means the property is not under control at all. A missing reading is the most common and most damaging finding.
What happens when run five fails something run one passed?
Treat it as a change that was never proposed. Read the trend line to find which run introduced the movement, check whether the change log has an entry there, then accept with a revised reference, re-qualify the property, or correct the process and return to the original band.
How is the result written into the spec file?
Three lines: the conclusion, the properties that moved, and the decision about each. Where a reference was updated the file moves to a new revision, and the revision log records the fifth-reorder test as the reason.
Should a retained unit be kept from every run?
Yes. One dated unit per run, stored with the gold samples, is what makes the series readable. Laying five units out in run order shows a drift that is invisible in any pair, and it costs one unit per run to maintain.
What if the fifth test finds nothing wrong?
Record that as the conclusion. A test that found no movement is evidence that the controls work, and it is the strongest argument against the proposal, made in every programme eventually, to stop keeping the paperwork.
Does the annual test replace the routine inspection?
No. The routine inspection asks whether this lot passes; the annual test asks whether the product has moved. Both are needed, and the annual one adds the run one column and a conclusion rather than replacing anything.
What changes about the next run afterwards?
The bands are confirmed or tightened, the properties that drifted are added to the routine sheet rather than only to the annual test, and the next review date is set. Where a silent change was found, the loose register line is corrected as well.
People Also Ask
How do you verify quality after many repeat orders?
Compare the current run with the approved run rather than only with the previous one, keep one dated retained unit per run, chart each bounded property as a trend, and add a laboratory check on anything that supports a claim or shows movement.
Why does quality drift slowly over repeat orders?
Because each run moves a small amount inside its band and every comparison passes. Only the accumulated movement is visible, and only when two runs that are not neighbours are compared against the same retained reference.
How often should you re-test a repeat product?
Once a year on a running programme, or at the fifth run, whichever comes first. Re-test sooner whenever a class three or class four change is proposed, or when a trend line starts moving even though readings pass.
What is a trend line in quality control?
The recorded reading for one property across successive runs, drawn against its band. A flat line is controlled, a step marks a change, a slope marks drift, and a scatter means the property is not under control.
How do you catch drift that passes every inspection?
Record numbers rather than pass or fail on every run, keep the same properties each time, retain a dated unit per run, and review the series annually against the approved run instead of only against the previous one.
Sources and further reading
Every figure on this page is either our own programme data or attributed below. Open the originals before you write them into a specification.
- ISO 2859-1:1999 — sampling procedures for inspection by attributesInternational Organization for Standardization
- AATCC test methods and evaluation proceduresAmerican Association of Textile Chemists and Colorists
- ISO 13934-1 — tensile properties of fabricsInternational Organization for Standardization