How do we find a reliable manufacturing partner for a hardware product we already sell?

You recognise a reliable manufacturing partner by what they can evidence on paper and by how they behave on the first small order, not by what their website says. Before you commit to anything, ask for three things: proof that the company exists and is insured, a quality certificate you can verify yourself, and a quotation that states tool price, piece price, cycle time and guaranteed shot count all four. Then put tool ownership in writing, and test the supplier with a paid pilot run before you move any live production. The fact that your product already sells makes that order more important rather than less: your own delivery promise keeps running while you do it.

What is different about this choice when your product already sells?

You have a deadline that is not yours to move, and you have evidence a startup does not have. Both change how you run the process.

Take the timing first. A company with nothing in the market can trial a supplier calmly and push a quarter if things slip. You cannot. Every week without stock is revenue you do not get back and a buyer who starts looking for an alternative. So you qualify alongside your existing supply, on stock you build before the switch. If you are adding a supplier because your current factory cannot keep up with demand, that order of work counts double. That stock build is almost always the largest and least budgeted cost of the whole exercise, because it lands in the same quarter as the tooling invoice.

Against that sits a real advantage: you have numbers. You know what you sold last year, in which months, how many units came back and on which complaint. Any supplier writing a serious quote needs exactly those numbers, because annual volume drives cavitation, steel grade and minimum call-off. Hand them over in full and ask for something back. A supplier who proposes a different tooling layout than the one you run today, based on your real volumes, is doing the arithmetic with you. One who takes the drawing and puts a price under it is not.

Which documents do you ask for before committing?

The ones that prove something a sales meeting cannot: that the company exists, that it is insured, that the certificate is real, and that they have made something resembling your product before. Ask for all of them in one request. How a supplier responds to that request is information in itself.

| What you ask for | What it evidences | How you verify it | | --- | --- | --- | | Company registration extract and latest accounts | The company exists and can carry the series financially | Public register of the country of registration, or a credit report | | Proof of general and product liability insurance | There is cover if a batch goes wrong | Policy schedule with territory and term, not just a statement | | Quality certificate with number, body and scope | A controlled process sits behind the work | IAF CertSearch, by company name or certificate number | | Material declarations for your application | The resin is allowed to do what your product does | Declaration per batch, with the standard and the supplier named | | Dimensional report for a comparable part | They measure, they do not only make | Format of the report, instruments used, drawing revision | | List of processes done in house and subcontracted | Where the chain actually runs | Ask per step who carries responsibility |

Two of these lines get ticked off too easily. The certificate is the first. Accredited certificates for standards such as ISO 9001 are listed in IAF CertSearch, the International Accreditation Forum database, searchable by company name or certificate number, showing validity, certified sites and the issuing certification body. What usually goes wrong is not whether the certificate exists, but what it covers. A certificate covering trade and assembly says nothing about the moulding shop doing the actual work, and a certificate for one site does not cover another.

Insurance is the second. The new EU product liability directive has to be transposed into national law by 9 December 2026, and it applies to products placed on the market from that date. You put the product on the market under your own brand, so you are the one your customer comes to. Whether you can pass that claim on to the company that made it depends on their cover and on your contract, not on who was at fault.

How do you test whether they can really make your kind of product?

By asking about products that resemble yours in tolerance, material and batch size, not about the list of processes on their website. Almost anyone can injection mould. The question is whether they have run your wall thickness, your surface finish and your annual volume before.

Three questions are hard to talk around. First, ask what scrap rate is built into your piece price, and who carries the cost when actual scrap runs higher. Second, ask who pays for corrections after the first trial shots, and which drawing revision parts are measured against. Third, ask what they do themselves and what they subcontract. Printing, surface treatment, assembly and packaging often run through third parties, which is fine, as long as it is clear who stays responsible when something goes wrong and who your single point of contact is.

The strongest signal comes from having your design reviewed before anyone puts a price on it. A manufacturability review across the whole chain shows you how a supplier thinks: do concrete proposals come back on wall thickness, draft, part count and packaging, or just a quote with a surcharge for things they never mentioned. Higo Sense ran to more than sixty parts across thirteen moulds, manufactured across Europe and Asia with the electronics on a separate track. With a build like that, the question is not who can make it, but who keeps the overview.

Who ends up owning the tool, and what do you agree about it?

Not automatically you, even if you pay for the whole tool. Ownership follows what the contract says, and that is exactly what decides whether you can switch manufacturer later without stopping production.

So agree five things before the tool is built: who legally owns it, where it is physically kept, who pays for maintenance and refurbishment, which documentation you receive, and on what terms you can collect it. That documentation matters more than it sounds. Without the tool drawings, the steel grade, the runner type and the maintenance log, a tool that arrives at a new supplier is a puzzle you pay for in machine time.

Watch the structure behind an attractive tool price. If the tool is fully or partly amortised into the piece price, ownership usually stays with the manufacturer and you buy it out when you leave. That can be a sound choice when cash is tight, but it is a different deal from "we paid for the tool". In our process, a signed mould property contract is in place before production starts, rather than at the moment you want to move.

How do you test a new supplier with a small first order?

With a paid pilot run on the production tool, the real resin and the real packaging, and with the acceptance criteria agreed up front. Free samples from a prototype tool prove nothing about a supplier who will soon ship thirty thousand units per call-off.

Five things to judge in that pilot run:

• The dimensional report, against the drawing revision you supplied, on the dimensions that are functional rather than on every dimension.
• The agreed sampling plan. ISO 2859-1 is the common standard here. At general inspection level II, a lot between 3,201 and 10,000 units gives a sample of 200; consumer goods commonly run AQL 2.5 for major defects and 4.0 for minor, cosmetic ones.
• Delivery on the agreed date, because the first late shipment predicts the second.
• The response to the first deviation. This is the real test. Does the supplier flag it, or do you only hear about it once you find it?
• Reply speed on an ordinary email, outside the sales phase.

A worked example on invented assumptions, to show the proportions. You sell 120,000 units a year and have them made in call-offs of 30,000. A pilot run of 3,000 units costs you the machine set-up, the measuring work and roughly three weeks of lead time. If it goes wrong there, you are holding 3,000 bad units. If the same thing goes wrong in the first full call-off, you are holding ten times that, plus a delivery you have to explain to your own customers and a recovery round with a tool correction and fresh sampling.

Name the honest downside too. A pilot run costs money and six to ten weeks you would rather have spent shipping, and a successful pilot guarantees nothing about batch twenty. What you buy is that the first mistake stays small, and that you see how a supplier behaves when something is not right. No certificate tells you that.

Production Heroes takes products from design to series production, from feasibility and cost price to sourcing and supply chain, and supplies moulds and tooling to manufacturers across Europe, from the Netherlands and its own presence in China.

Sources: accredited certificates for standards such as ISO 9001 are verified through IAF CertSearch, the International Accreditation Forum database (2026); ISO points to the same route in its own guidance on certification. The sample size of 200 at general inspection level II for a lot of 3,201 to 10,000 units and the AQL levels come from ISO 2859-1, the standard for sampling inspection by attributes. The 9 December 2026 transposition date and the application to products placed on the market after it come from Directive (EU) 2024/2853 on liability for defective products (2024). Lead times for tooling, first article inspection and the overlap period come from our own process. The worked example uses invented assumptions and is not a quotation.