What We Check Before Signing a Customer NDA

Most drawings arrive after a confidentiality agreement rather than before one, which is a reasonable order. The agreement itself, however, tends to get read far less carefully than the drawings that follow it — and it decides what both sides may do with those drawings for years afterwards. This is what we look at when a customer sends us an NDA, which clause types cause the most back-and-forth, and the two commercial points that usually sit in the same conversation: who pays for design work, and who pays for samples. It describes how we work, not legal advice; what goes into the document is for you and your counsel to decide.

Why the Agreement Lands Before the Drawings Do

Confidentiality first is a sensible habit, and we do not push back on it. What we do ask for is a few minutes spent on the text, because the agreement does not only protect the drawings — it also defines what we are allowed to have learned by the time the project ends.

Our starting position is narrower than most people expect. Independently of anything signed, we treat customer drawings, models and specifications as confidential, and the handling rules we apply by default are set out on our confidentiality page. A signed agreement therefore does not switch our behaviour on; it gives you something you can enforce. That distinction shapes how we read it. We are not trying to promise less — we are trying to avoid signing something we cannot actually operate, because a clause we cannot honour is worth nothing to you either.

Mutual or One-Way: The Question We Answer First

A one-way agreement protects the party disclosing information. In a typical enquiry that is the customer, and on the face of it that seems fair — you are the one sending files.

The complication is that we disclose things too. A manufacturability review produces suggestions: a bend that could be formed instead of bought, a fastener pattern that would remove a fixture, a panel split that would cut welding time. Those suggestions originate with us, and under a one-way agreement drafted broadly they can end up captured as your confidential information. We would then need your permission to use our own engineering judgement elsewhere, which is not a position any workshop can accept.

So we ask for mutual, or for the reciprocal carve-out to be written in, whenever the project will involve design feedback rather than plain build-to-print. For straightforward build-to-print work a one-way agreement is usually fine. Our engineering review page describes what that feedback actually contains, which is why the distinction matters.

Independent Development: The Clause That Gets Rewritten Most

Independent development clauses exist for a good reason. You do not want a supplier to become unable to serve its own other customers simply because it once worked on your product, and we do not want to lose the right to run our own workshop. When the clause is drafted carefully it protects both sides.

The difficulty is that these clauses are frequently drafted wide enough to swallow the general knowledge of the trade. Folding processes, standard fastener practices, corner treatments, fixture approaches, the ordinary ways a cabinet is stiffened or a door is hung — these are not your design. They are how sheet metal is made. A clause that leaves their status ambiguous creates a risk nobody intended and no court would enjoy untangling.

What we ask for is two clean boundaries rather than one sweeping sentence. What you provide — your drawings, models, specifications and the product-specific detail in them — is yours and stays confidential. What we already know, and what we develop independently, is ours; and we remain free to fabricate enclosures and parts for other customers, including ones serving a similar purpose. Framed that way, the clause protects your product without requiring a workshop to stop practising its trade.

This is the clause we most often return with comments on, and it is almost always settled with one sentence rather than a negotiation.

Scope: Describing What Is Actually Confidential

The broadest scope wording covers information in any form, including things mentioned orally. We understand the intent and we do not argue with it in principle. We do ask whether it can be operated, because a scope with no boundary cannot be complied with — nobody can later demonstrate where an idea came from.

The test we apply is simple. Looking at a specific piece of information, can both sides tell whether it is inside or outside the agreement? If not, the definition needs work. In practice that means confidential information should be identifiable: written and marked, or if spoken, confirmed in writing within a reasonable period.

It also helps to write down what is explicitly not confidential — information already public, information the receiving party can show it held beforehand, information lawfully obtained from a third party. Those exclusions are standard, and their presence usually signals a document someone has thought about rather than downloaded.

Term, Return and the Long Tail After the Project Ends

Three practical points sit near the end of most agreements and get the least attention.

Duration. Terms of three, five and ten years all appear, and perpetual obligations appear too. A perpetual obligation is difficult for a manufacturer to honour credibly: staff change, archives move, backups rotate. We usually propose a fixed term — commonly three to five years, or a period running from the end of the project — because a commitment we can actually administer is worth more than one we cannot.

Return and destruction. Requiring every copy to be returned is easy to write and impossible to satisfy once files have entered backup systems. The workable version is a commitment to delete or isolate on request, with confirmation, and it is better for both sides because it describes something that can be done.

Subcontracted processes. Some operations — coating in particular — are commonly carried out by specialist partners. If the agreement does not permit disclosure to a subcontractor under equivalent confidentiality, we cannot use them, and that constrains the finish we can offer. Our finishing work runs on exactly that basis, so the clause is worth a line rather than an omission.

None of these three is contentious on its own. Left vague, they are the ones that surface later, usually at the point where somebody asks for files back.

Where Confidentiality Meets Money: Design and Sample Charges

Two commercial points regularly arrive in the same exchange as the agreement, and it is cheaper to settle them together than to discover them one at a time.

Design work is charged. Producing 2D and 3D geometry, working through manufacturability, and iterating a prototype is engineering time, and we do not give it away on the expectation of an order that may never come. There is a second reason that has nothing to do with effort: ownership is clearer when work is paid for. Paid design delivers its output to you, while the general process knowledge behind it stays ours — which is the same boundary the independent development clause is trying to draw.

Samples are paid in advance. A sample exists for one project and has no resale value once built; if the programme stops, the parts have nowhere to go. Charging the sample up front keeps that risk with the party who decides whether the programme continues, and it is a reasonable signal in both directions about how serious a project is. What comes out of that stage is inspected the way a production part would be — the checks are the ones described in our first article inspection article.

Both points are ordinary commercial terms rather than expressions of distrust, and both are easier to agree before drawings change hands. The staged structure we use for this is described in our staged quotation article, where the sample stage and the production stage are quoted as separate things.

What We Ask to Change, and What It Costs in Time

In most cases we return one to three comments, not a redraft. The independent development clause accounts for the majority of them; duration and return obligations account for most of the rest. Where the document is a standard template from your side, the comments are usually accepted without discussion.

Review takes a few days in our experience, and it does not hold up the technical side. We can assess general feasibility from a description before any drawings are released, so the conversation about whether something is buildable can run while the paperwork finishes.

What stops us is narrower than people expect. We decline obligations we cannot bound — an undefined perpetual duty, or a requirement to give up process knowledge we held before the project. When that happens we say so plainly and explain which sentence is the problem, rather than letting an agreement sit unsigned while both sides assume the other is stalling.

For ongoing programmes, agreeing the terms once and reusing them is worth the effort. Repeat projects move faster when the paperwork is already settled, which is one of the quieter advantages of working with the same supplier over several product generations.

Questions We Get Asked About NDAs

Do you need an NDA before we send drawings?

Not before we talk. Customer files are treated as confidential by default, and the handling rules are set out on our confidentiality page. A signed agreement adds something you can enforce rather than changing how we behave.

Will you sign a one-way NDA?

Usually yes for straightforward build-to-print work. Where the project involves design feedback from us, we ask for mutual cover or a reciprocal carve-out, because our own engineering suggestions would otherwise be captured as your confidential information.

Which clauses do you comment on most often?

Independent development, first and foremost — usually to make sure general process knowledge stays outside the agreement. Then duration, and return or destruction obligations that cannot be satisfied once files are in backups.

Can you use subcontractors such as coaters under the NDA?

Yes, provided the agreement permits disclosure to a subcontractor under equivalent confidentiality. Without that line, some finishes cannot be offered at all.

Do you charge for design work?

Yes. Design and prototyping are engineering time and are quoted as such. Paying for design also makes ownership clear: the output is delivered to you while the general process knowledge behind it is not transferred.

Why is the sample paid in advance?

Because a sample is built for one project and has no resale value if that project stops. Advance payment puts that risk with whoever decides whether the programme continues.

How long does the review take?

Typically a few days. It does not block feasibility discussion, which can proceed from a description before any drawings are released.