Managing Change Control and Traceability with a Contract Window Manufacturer

Auditing a factory tells you what the system looks like on the day you visit. It does not tell you what that system will look like in eighteen months, when a profile supplier has been swapped, a sealant has been reformulated, or a machine has been rebuilt.

Change control and traceability are how you manage that gap. They are also the two areas where a supplier can damage a European brand most easily without any intention of doing so.

Why changes are the real risk

A declaration of performance describes a product type. The technical documentation behind it is built on the elements that relate to the system of assessment and verification of constancy of performance — that is the structure set out in Article 11, paragraph 1 of Regulation (EU) No 305/2011, with factory production control defined in Article 2, point 26 as the documented, permanent and internal control of production in a factory.

The practical consequence is simple. Your documentation describes the product as it was when the evidence was created. If the product quietly drifts — a different profile alloy, a different thermal barrier, a different sealant, a different glazing build-up — then at some point you are selling something your file no longer describes. Nothing dramatic happens on the day of the change. It becomes visible years later, under a question you cannot answer.

That is why the useful work is not a longer audit. It is a shorter list of triggers, agreed in writing, plus a traceability structure that lets you verify any single order.

The change triggers to agree in writing

Categories to cover, with examples. This list should be in the supply agreement, not in a presentation.

A supplier who accepts this list without argument has understood the question. A supplier who proposes to notify only "major" changes, without defining major, has not.

The minimum traceability fields

Traceability does not need to be elaborate. It needs to be sufficient to reconstruct one order. These nine fields are the practical minimum:

Two tests tell you whether this is real. First, ask for the fields for a shipped order — not a template. Second, ask which operative performed the final inspection on that order. A traceability system that cannot answer the second question is a filing system, not a traceability system.

Deciding when a change needs re-testing

This is where most brands either over-react or under-react. The judgement rule we would propose:

• If the change cannot affect a declared characteristic — a change of packaging supplier, for example — record it, notify it, and no further action.
• If the change might affect a declared characteristic, treat it as potentially affecting the product type. Notify. Document. Then take the decision to the party that owns the declaration of performance, together with the notified body where the applicable system requires their involvement.
• If the change removes or replaces the evidence base — a different glass build-up, a different thermal barrier material — assume the existing product-type determination no longer covers what you are selling until someone qualified says otherwise.

The third rule is the one that protects you. It is also the one that a supplier is most likely to resist, because it costs time. That is precisely why it should be written down before the first order rather than argued about in the middle of one.

A note on scope: the decision belongs to the entity that owns the declaration of performance — the manufacturer in the sense of Article 2, point 19 of the regulation. If that is your brand, the decision is yours, taken with advice. A supplier can inform it; a supplier cannot make it.

Running it without bureaucracy

Two habits carry most of the value:

1. Keep a change log — one running record of every change to product or process, with the date it took effect, its scope, who approved it and which customers were notified. Ask to read it at every visit. Its absence is the finding.
2. Sample one shipped order per quarter. Run the traceability exercise on it: order, run, batches, inspection record, shipment. Then verify that any change logged in that period was notified before it reached production.

That is roughly an hour of work per quarter. It will not catch every problem, but it reliably catches the class of problem that damages documentation — drift that nobody announced.

Frequently Asked Questions

What counts as a change that should be notified to me?

Any change that could affect a declared characteristic: profile alloy or wall thickness, the type or source of thermal barrier, gasket or sealant materials, the profile supplier or extrusion plant, hardware family, model or surface finish, glass build-up, glass supplier, spacer type or gas fill, and process parameters such as cutting tolerances, corner assembly method or glazing method.

What is the minimum set of traceability fields I should require?

The order number, the production run or batch identifier, the profile batch, the thermal barrier batch where separate, the glass batch or unit identifier, the hardware batch, the inspection record identifier with its date and inspector, and a reference to any change log entry that applied to that run.

When does a change require re-testing?

The working rule is this: if the change could alter a characteristic you declare, treat it as potentially affecting the product type and take it to the party that owns the declaration of performance together with the notified body. Neither extreme is safe, because re-testing everything is unaffordable and re-testing nothing means your documentation describes a product you no longer sell.

How often should I verify traceability?

A practical rhythm is a random sample of one shipped order each quarter, plus verification after every notified change. The point is not to audit continuously but to notice drift while it is still cheap to correct.

Can a supplier change a material without telling me?

They can, which is exactly why notification has to be a written, contractual obligation with a defined trigger list rather than a matter of goodwill. The audit question that tests it is whether an unnotified change has occurred in the last twelve months, and what happened as a result.

What is a change log?

A single running record, held by the manufacturer, that lists every change made to the product or the process, the date it took effect, its scope, who approved it, and which customers were notified. It is the document that turns change control from an intention into something you can inspect.

Next step

Change control only works if the supplier is willing to be measured against it — which is a question you can settle before the first order, in an hour, with the right questions. We manufacture to customer drawings and specifications in Yantai and expect to be asked for exactly this. Our contract manufacturing page explains how we handle notification, traceability and documentation, and what we provide at shipment. If you are already sourcing and want to test a current supplier, start with the factory audit checklist and the manufacturer responsibility guide.

> This article is practical guidance for buyers. It is not legal advice, and it does not state that any approval is held by us.