OEM production: running a factory on the customer's drawings

What the drawing hands over, and what it withholds

The customer's pack fixes form, fit, function, materials and often the approved vendor list. The factory decides how to make it: fixturing, sequence, cycle, gauging, operator instructions. Process know-how developed that way is genuinely yours, though it is worth little outside the programme unless the same techniques transfer. Engineering changes arrive as instructions rather than proposals, and each one carries a cost-recovery negotiation over scrapped stock, reworked tooling and re-qualification. Plants that fail to log change impact at the moment it lands rarely recover the money afterwards, because the buyer's own change record closes long before the invoice is raised.

Demand you can see and cannot shape

Releases come as forecasts with a firm window and a flexible tail. The firm portion authorises build; the tail authorises material commitment, and the distance between those two horizons is where a plant's exposure lives. Component stock bought against a forecast that softens becomes the factory's problem unless liability terms say otherwise. Finished units are staged to call-offs, so warehouse space is consumed by someone else's shipping calendar. A factory serving a handful of programmes should treat forecast accuracy history as a supplier metric applied in reverse — measured, recorded and raised in the quarterly review.

Tooling, test rigs and the amortisation argument

Programme-specific tooling, fixtures, end-of-arm attachments and functional test rigs represent capital with no resale value once the model retires. Who funds it, who owns it and who holds it decides a great deal: buyer-owned tooling can be recalled to another plant at the end of a contract, while factory-funded tooling has to be recovered through piece price and dies with the programme. Amortisation schedules should match volumes the plant believes rather than the buyer's launch curve, and the agreement needs a clause covering what happens when actual offtake falls short.

Proving conformance in the buyer's language

Approval regimes run on submission packages: dimensional reports, capability studies, material certificates, process flow, failure-mode analysis and control plans. The burden is front-loaded before first shipment and then re-triggered by any change to process, location or sub-supplier. Buyer audits examine both the system and the specific line. Defect liability usually extends past delivery into warranty recovery, sorting costs and line-stoppage claims, which can dwarf the value of the parts involved. Traceability depth — batch, shift, machine or serial — is a design decision taken early and expensive to retrofit once production is running.

Where a build-to-specification plant runs out of room

Scaling means winning more programmes, and each brings its own machinery: buyer portals, electronic data interchange formats, labelling rules and reporting cadence. Two or three large customers can fill a factory and then own it; margin erodes at every re-quote because the buyer knows the tooling is paid for and the alternative is a transfer they control. Systems have to hold buyer revision levels against internal part masters, translate call-offs into schedules, and keep change orders auditable. Buying happens inside the customer's approved sources, so substitution is restricted and cost reduction becomes a joint negotiation rather than a unilateral move.

Frequently asked questions

Who should own the tooling on an OEM programme?

It depends on who is prepared to carry the risk of the programme ending early. Buyer-funded tooling keeps the factory's capital free but gives the customer a portable asset and a credible transfer threat. Factory-funded tooling supports a higher piece price and some protection, provided the contract sets a minimum offtake or a termination payment. Either way, physical custody, maintenance responsibility, refurbishment triggers and the condition tooling must be returned in should be written down before the first shot.

Can we sell process improvements we develop on a customer's product?

Usually yes for the method, no for the product. The design, drawings and specification belong to the buyer, and most agreements assign any product-related improvement to them as well. General process know-how — a fixturing approach, a welding parameter set, a gauging method — normally stays with the factory unless the contract sweeps it up in a broad intellectual property clause. Read that clause carefully, because a wide assignment can quietly transfer capability you intended to reuse elsewhere.

How much customer concentration is too much for a plant that builds only to customers' designs?

There is no single threshold, but the practical test is what happens to fixed-cost recovery if the largest programme stops. Should losing it leave the plant unable to cover overhead through the notice period plus a realistic replacement sales cycle, concentration has already become a solvency question rather than a commercial one. Mitigations include staggered contract end dates, a deliberate mix of industries with different cycles, and tooling terms that make an abrupt transfer expensive for the buyer.