Swimwear Specification Change Control: Engineering Change Orders, Re-Testing and Cost Allocation

Suggested Meta
– Description: How B2B swimwear buyers run engineering change control — ECO fields, tech pack versioning, mandatory re-testing, cost allocation and what to do when a change lands after cutting.
– Publish Config: category = b2b-wholesale-tips | focus kw = swimwear specification change control

Most swimwear programs do not fail at the design stage. They fail somewhere in weeks three to six, when somebody asks for a small change — a slightly deeper neckline, a different elastic, a cheaper lining — and the answer is “fine, we’ll just adjust it on the line”. Six weeks later the same program is paying for three re-tests, a half-stitched batch, a delayed container, and a quality report that does not match the tech pack that signed the PO.

That is the gap this guide fills. Almost every B2B buyer on this site has read our material on tech pack essentials and our purchase order essentials. Both stop at the moment of approval. Neither covers what happens when the spec moves after approval — which, in swimwear, is not an edge case. It is the normal state of a development program.

This article is the buyer-side operating manual for that situation: what counts as a change, what an engineering change order must contain, which changes force re-testing and who pays for it, and how to handle a change that lands after the fabric is already cut.

1. Why swimwear is unusually change-prone

Change is cheaper to control in a woven shirt than in a swimsuit, and understanding why changes your control strategy.

The fit is a function of stretch, not of length. A woven garment’s size is mostly set by pattern dimensions. A knitted swim fabric is under tension on the body; its recovered dimensions depend on yarn content, knitting structure, finishing, and how many hours it spent under chlorine or sun. Change any one of those and the same pattern produces a different garment. The pattern is not the specification — the pattern plus a stretch and recovery spec is, and that makes every material substitution a fit change.

The facings are load-bearing. The lining, the shelf bra, the power mesh, the edge binding and the elastic are not hidden trim. They carry the load. Swap a power-mesh for a lighter hand and the top cups migrate; swap a binding tape for a narrower one and the neckline rolls. Swimwear has more structural components per seam inch than almost any other apparel category, which means more places for a change to quietly degrade performance.

Compliance is attached to the composition, not the style number. If you change the fibre blend, the labelling requirement changes with it. If you change a component to one containing a restricted substance, the EU and US filing you already paid for may no longer describe the product you are shipping.

Several disciplines must agree at once. A change to fabric weight touches cost, hand feel, coverage, drying time, print behaviour, colour fastness and the measurement tolerances in your size spec. A change to binding touches seam construction and wear comfort. A change to lining and support construction touches fit and cost. In a woven garment you can often make the change and inspect the result. In swimwear you usually cannot see the failure until the garment is wet.

The practical rule: in swimwear, almost every change is a fit change, and every fit change is a re-test question. That single sentence should drive your whole change control structure.

2. The four change families

Not all changes are equal, and treating them as equal is why change control stalls. Sort every request into one of four families first.

Family	Typical triggers	Who usually raises it	Re-test required?	Default cost owner
A — Design / aesthetic	neckline depth, strap width, colourway, panel placement, print scale, trim colour	Brand, retailer, marketing	Usually no, but a fit-sample re-sign-off is required if the change alters any seam path	Buyer
B — Material / component substitution	fabric weight change, fibre blend, lining, elastic, binding tape, hardware, interlining	Factory (cost), buyer (supply), sometimes regulator	Almost always yes — mechanical performance, colour fastness, coverage	Split; see §5
C — Process / method	seam type, stitch density, bonding vs stitching, print method, finishing step	Factory (capability, cost)	Only if it changes the finished behaviour (stretch, seam strength, hand)	Factory if factory-initiated
D — Commercial / volume	quantity, mix, ship window, Incoterm, packaging configuration, delivery splits	Buyer, merchandising	No, unless packaging change affects transport or labelling	Buyer

Two things follow from this table. First, family B carries the re-test burden and should consume most of your governance — it is also the family factories raise most often, because it is where their margin sits. Second, family A is where most of the volume of change requests arrives, and it is the family most often handled over WeChat, which is exactly how undocumented changes get launched.

3. What an engineering change order must contain

A change order is not a friendly email. It is a document with enough fields that two people six months apart read the same thing. If your ECO cannot survive an absence of context, it will not survive a dispute.

#	Field	Why it matters
1	Change ID (e.g. ECO-2026-014)	Sole reference in PO, tech pack, QC report and correspondence
2	Style / SKU + revision number currently superseded	Prevents “same style, different spec” ambiguity across two factories
3	Change category (A/B/C/D from §2)	Sets the re-test and approval path
4	Exact field(s) affected (fibre %, gsm, component, seam, measurement)	“Fabric changed” is not a specification; “back lining 190 gsm → 165 gsm” is
5	Reason (buyer request / factory cost / supply availability / regulatory)	Determines cost allocation and who needs to sign
6	Effective rule — applies to which POs / ship windows	The single most skipped field; without it, half an order runs old specs
7	New sample requirement (physical? pre-production? both?) and who pays	Defines what “approved” means after the change
8	Re-test list with standard references	Converts a subjective worry into a line item
9	Impact on unit cost, MOQ, lead time, and minimum dye lot	Makes the change a commercial decision rather than a technical one
10	Affected documents to be re-issued (tech pack rev, lab dip, print file, hangtag, care label)	Otherwise the factory “knows” the change and still builds the old sample
11	Approval block — roles and dates on both sides	Needed later for any claim or CAP conversation
12	Record of the old specification	Without a diff, you cannot prove a defect is caused by the change
13	Notification deadline relative to cutting date	A change raised after cutting is no longer a change; it is a write-off
14	Closure condition and archive location	Closes the loop and returns evidence to the program file

Field 6 and field 10 are the two that cause most grief. Field 6 stops a change from silently applying to the wrong shipment. Field 10 stops the factory from producing to an old tech pack with new fabric and calling it approved.

4. Version control: the three documents that must move together

In swimwear, change chews through three documents at different speeds, and buyers who version only one of them end up with a spec that no one can reproduce.

The tech pack is the source of truth. It needs a revision number, a revision date, a change summary line, and a clear statement of which revisions are superseded. Revise the tech pack before you write the ECO into the PO, not after. Our tech pack essentials guide sets the module list; the change module is simply the twelfth version of that same file.

The physical approval samples are the second moving part. A lab dip signed at revision 3 does not describe a fabric specified at revision 5. If the change alters colour, fibre, weight or component, the lab dip approval has to be re-issued with a new reference, and the old sample has to be marked superseded in the same room it is stored in — not in a spreadsheet.

The measurement sheet and grade rule are the third. A change in fabric stretch shifts every measurement in the size run, not just the one style. If the change touches the body fabric or the main lining, re-issue the pattern grading and spec consistency sheet with the new tolerance band, and re-run the fit validation at the two boundary sizes rather than the middle.

A workable naming convention that survives contact with a factory: style-SKU / fabric-subsystem / revision / date / author. SWD-4412-LIN-R4-20261008-JH. It looks pedantic in an email and it saves a container.

5. Re-testing: which changes force a new test

This is where change control either saves money or quietly destroys it. The default assumption among buyers is “it’s only a small change”. The correct default is: assume a material or process change invalidates the previous test report until a test proves otherwise.

Change	Tests that must be re-run	Who normally pays	Link to evidence
Fibre blend or polymer change	Mechanical performance (stretch, recovery, growth), dimensional stability, pilling	Buyer if buyer-requested; factory if substituting	fabric performance testing
Weight / gsm change	Coverage when wet, drying behaviour, seam behaviour, hand	Buyer	fabric weight and gsm guide
Any colour change or new dye house	Colour fastness to chlorine, light, water, rubbing; shade tolerance	Buyer	colour fastness testing
New elastic, binding, tape or mesh	Elastic recovery, retention after wash, roll / curl behaviour	Buyer	trims, hardware and elastic
Lining or support change	Support level, cup stability, chafe, moisture feel	Buyer	lining and support construction
Seam method or stitch density change	Seam strength, seam slippage, stretch in the seam	Factory if factory-initiated	seam construction methods
Print or dye method change	Fastness, hand, crocking, coverage, print registration	Buyer	print methods and artwork
Certification-linked component	Certificate scope check against the new component	Buyer	sustainability certifications
Labelling or component destination	Market labelling rules for the new composition	Buyer	EU 1007 / REACH labelling · US CPSIA / Prop 65

Three disciplines make this table usable in the real world.

Test at the right time point. Tests split into fabric confirmation (before bulk yarn is committed), batch verification (first production roll), and pre-shipment. A change approved at fabric-confirmation time may not need the full pre-shipment re-test if the batch verification report covers it — but that must be a written decision, not an assumption made at the loading dock.

Read the report, not the certificate. Most test reports say “passed” and contain nothing about specimen count, pre-treatment, number of wash cycles, or whether the pass threshold was 3 or 4. If the change was marginal, a report-only pass with n=1 at threshold 3 is weaker evidence than a failure with n=5 at threshold 4. Our QC and AQL inspection guide covers how to read these in the inspection context; the same reading skill applies to lab reports.

Batch the changes. If you are already funding one re-test cycle for a lining change, adding the elastic change and the binding change to the same cycle costs almost nothing extra and avoids three rounds of disruption. Buyers who sequence changes individually pay three times for the same factory visit.

6. Cost allocation: who pays, by default

Change costs are real: new lab dips, new sample rounds, re-test fees, re-cutting, re-work hours, and the factory’s lost line time. The default split below is what most factories expect, and stating it up front is cheaper than arguing later.

Scenario	Default allocation	Negotiation note
Buyer-requested aesthetic change before cutting	Buyer	Standard; ask for tiered pricing rather than a flat re-charge
Buyer-requested material or component change	Buyer for testing and samples; factory absorbs re-cutting only if the cut fabric was already bought to your spec	Push for the “one free change per style per season” clause
Factory-initiated substitution for cost saving	Factory — including re-test, re-sample and delay cost	Never accept a substitution in production without a written ECO
Factory substitution due to stock-out or force majeure	Shared; buyer may reject and take a delay	Buyers usually accept an equivalent-grade substitute only with a signed ECO
Joint improvement (both parties gain)	Split 50/50, or split against documented saving	Link this to your cost-down program so the saving is tracked, not taken once
Regulatory or market-compliance forced change	Split; compliance cost is usually shared by contract	Keep the compliance note in writing — it is the strongest cost-defence you have
Change after cutting	Buyer absorbs the full write-off unless the factory raised it late	See §7

The single most expensive pattern in swimwear sourcing is the silent substitution: the factory, facing a stock-out or a margin gap, quietly substitutes a component and ships to spec on paper. It is detectable — the fabric weight in the packing list differs from the tech pack, the test report’s measured value sits just inside the threshold, or the wear test complaints cluster around one component. The defence is procedural: require written ECO approval before any component substitution, and make it a PO term.

7. When the change lands after cutting

This is the decision most buyers handle badly, because by that point the price of the change has already been paid by somebody.

Stage when the change is raised	Option 1	Option 2	Usually right when
Before cutting / before marker making	Approve change, re-cut	Reject and re-develop	Any time — change is nearly free
After cutting, before sewing	Approve, re-cut the affected panels	Ship the current batch as “as-is” with a documented concession	Current fabric is unusual, or the change is cosmetic and low-risk
During sewing	Stop line, re-cut and re-stitch; balance line efficiency	Accept the in-process goods and apply change to the next order	The fault is systemic, not local
After sewing / before shipment	Rework at factory, or sort-and-segregate the affected units	Re-test and ship with a documented deviation note	Only the boundary units are affected
After shipment	Claim window — see §9	Write off	Never re-cut; you cannot send goods back into production

Two rules keep this table from becoming a negotiation.

Rule one: freeze the effective date at the moment of the PO, not the moment of the change. A written clause that says “specification changes apply only to POs issued after the ECO effective date” is the one line that prevents the factory from claiming the change applies to goods already in production. It belongs in the purchase order next to the specification block.

Rule two: a change after cutting is a cost event, and the cost model should be known before it happens. Re-cutting a marker costs scrap plus labour; re-stitching a partially assembled garment costs line hours and reduces line efficiency for the whole shift; a post-sew rework can cost more than the garment. Agree the rates in writing — re-cut rate, rework hour rate, minimum charge — so the conversation in week five is arithmetic rather than blame.

If you are running short batches where a change can still be absorbed cheaply, our small-batch and low-MOQ guide covers how flexible quantities change the math. If the change is going to move the ship window, the downstream damage is production lead time and freight and consolidation planning — a two-week slip can push a container past a seasonal cut-off and cost more than the change itself.

8. Compliance and quality consequences of undocumented change

An undocumented change does not only risk the wear test. It risks the paperwork that keeps the goods sellable.

If a change alters fibre composition, the fibre content statement, the care label symbols and the country-of-origin logic may all need updating. If it introduces a component from a new supplier, the restricted substance file and any market filing must be revisited — see UK market compliance and the US compliance guide for what the labelling obligations attach to. If it changes the packaging, the shipping marks and the EPR-relevant packaging data change with it.

In the quality system the same pattern appears. The QC and AQL inspection must be re-levelled against the new spec, otherwise the inspector is judging goods against a superseded standard and the report is unusable in any dispute. And if a change conceals a defect — a lighter elastic that fails after eight weeks, a lining that goes translucent when wet — the eventual claim has to run through the returns, warranty and defect claims process, where the factory will argue the goods met the original spec. That argument is won or lost on records: the ECO, the rev number on the tech pack, and the date the new test report was issued.

Sustained change without governance also erodes the relationship itself. If your factories change your spec four times a season and never close the loop, their planning team stops holding material for you. Tracking that — change frequency, cycle time to approval, cost recovered — is one line of the vendor scorecard and one input to the supplier ESG and compliance audit, because an uncontrolled spec makes traceability impossible, and traceability is the precondition of a defensible ESG claim.

9. Record-keeping, traceability and the claim window

Assume that any change you cannot produce a record for will be treated, in a dispute, as if it never happened.

Keep per style, per factory: the ECO register (every request, decision, cost and closure date); the tech pack revision history; the sample approval chain with physical samples retained for the return window; the test reports mapped to revision numbers; the PO amendment showing the new spec and effective date; and the inspection reports flagged with the applicable revision. Store it all in one place with one naming convention, and keep at least the IP protection and NDA cover in place so the revision history is not itself the thing that leaks.

The claim window runs against all of this. A defect discovered nine months later is only recoverable if you can show that the goods were built to a spec, that the factory knew the spec, and that the deviation from it caused the failure. That is three documents and a diff.

10. Building change control in four sprints

None of this needs to be built in one sitting. Four sprints over a single season is enough to make a program defensible.

Sprint 1 — Classify (2 weeks). Write the four-family list into your email template. Nothing leaves your desk without a family tag. Measure how many changes you actually raise in the first month; most buyers are surprised that it is double what they thought.

Sprint 2 — Document (3 weeks). Issue the 14-field ECO template and the re-test matrix from §5 as a single attachment. Get the factory to confirm in writing that they will not substitute a component without an approved ECO. Add the effective-date clause to the PO template.

Sprint 3 — Bind the documents (3 weeks). Make the tech pack revision, the re-issued lab dip, the revised measurement sheet and the PO amendment a mandatory four-way package. Nothing ships on a partial set — even under time pressure.

Sprint 4 — Measure and close (ongoing). Track change count, average time to ECO approval, cost charged back, and re-test spend per style against the development budget. Feed the two lagging numbers into the next season’s assortment and range planning.

11. Eight red flags

1. A change discussed but never written. The most common pattern, and the one that produces the unrecoverable claim.
2. A material substitution raised by the factory with no ECO and no price impact shown. Usually a margin move.
3. The change is approved verbally but the tech pack revision is never issued. Production then runs on the old drawing.
4. No effective date. Half the order ships to the new spec and half to the old, and nobody can see it until the fit complaints arrive.
5. Re-test requested but never scheduled. The test report arrives after shipment, so the information is worthless.
6. A “we already tested that” answer to a change that alters the fabric’s weight or blend. Previous reports describe the previous material.
7. A change requested after the cut date. By definition unrecoverable, and the price of knowing it late should be visible in the cost line.
8. The same change raised three times by three people. Without an ID field, you approve it three times and pay for it three times.

Frequently asked questions

1. Is a change to print placement an engineering change?
If it changes a seam path, a panel join or a coverage zone, yes — it needs an ECO and a re-fit check. If it only moves print within an existing panel and does not touch any measurement, it is a design revision that needs the artwork re-issued and signed, not a full re-test. Draw the line at “does any measurement or load path change”.

2. Who owns the re-test cost when the factory wants a cheaper lining?
The factory, in practice, and it is worth holding that line hard. They are making the margin improvement; they should carry the qualification cost. Letting this slide once sets the precedent for every substitution that follows.

3. Do I need a new lab dip if only the bag and the care label change?
No. Lab dips govern the garment colourway. A packaging or labelling change needs the label artwork re-issued and the market compliance check re-run, but not the fabric colour approval.

4. What is the latest point at which a change is still cheap?
Before the fabric is cut. After cutting, cost jumps by the scrap rate plus re-labour; during sewing it jumps again by line-efficiency loss; after sewing, rework may cost more than the garment. Push every change you can into the pre-cut window by keeping development sample rounds closing early — see our sampling and MOQ guide and 3D virtual sampling for how to shorten that window without cutting corners.

5. How do I stop factories from substituting components silently?
Contractually first: no substitution without an approved ECO, and the factory bears the cost of any unapproved substitution plus the delay. Operationally second: spec the components with a supplier name or a part number wherever the market rules allow, so a substitution breaks the spec visibly rather than quietly.

6. If a change improves the product but raises unit cost, can I reject it?
Yes, if it was factory-initiated and not required by a regulation or a reported safety issue. Make the rule: factory-initiated changes that raise cost need the buyer’s written approval before they are applied; factory-initiated changes that reduce cost need the same approval but with the saving documented against the next cost-down cycle.

7. How long should a change record be kept?
At least the length of your returns and warranty window, and typically two seasons — swimwear sells in short seasons and the claims arrive in the following one. Store the ECO register, the test reports and the inspection reports together; a claim without the revision history is an argument you lose on procedure rather than on facts.

8. Does digitising development remove the need for change control?
No — it relocates it. 3D virtual sampling catches a large share of aesthetic and fit changes before physical sampling, which genuinely reduces change volume. But the moment a real fabric, a real elastic or a real compression test enters the loop, the same governance applies. Digital tooling shortens the change cycle; it does not make changes free.

Conclusion

Specification change control is unglamorous, and that is precisely why it is where margin leaks. Every buyer on this site knows how to write a tech pack. Far fewer can state, on request, what happens when the spec moves three weeks into production — who approves it, what has to be re-tested, who pays, and which POs it applies to.

The structure is not elaborate: sort changes into four families, put fourteen fields on every one of them, bind the tech pack revision to the lab dip to the measurement sheet to the PO amendment, and assume a material change invalidates the last test report until one proves otherwise. Do that and the change stops being a crisis and becomes a line item you chose to spend.

Start with the two fields that cause most grief — the effective date and the affected-document list — and add them to your ECO template this week. Then send the four-sprint plan to your factories before the next season opens, not during it.

Put change control into practice on your next program. Send us the style you are about to revise and we will return a change-order form pre-filled with your spec, the re-test list that applies to it, and the cost impact of each option. Request a change-control template and spec review · See our custom swimwear specification capabilities · Walk the production and QC process