Quality Accountability When a Supplier Subcontracts Part of an Order
Buyers remain liable for subcontracted work they never see or audit.

A buyer signs one purchase order. One facility is named on the quote, one facility passed the audit, and one facility's name sits on the paperwork. But by the time the parts ship, two or three shops may have touched them: one for machining, another for heat treatment, a third for plating or finish. The buyer never sees most of this. The contract, though, never splits along with the work.
Subcontracting is standard practice across aerospace, medical devices, defense, and precision manufacturing generally. Heat treatment, plating, and specialty finishing routinely leave a primary facility and come back finished days or weeks later. No engineer should be alarmed to learn that a contracted shop doesn't do every operation in-house. The trouble starts with a quieter assumption: that the facility a buyer toured, quoted, and signed with is the only facility with hands on the part.
In regulated markets, the position on this is settled. The OEM or brand owner stays accountable for product conformity and compliance no matter how much of the production gets pushed downstream to other shops. A primary supplier cannot hand off accountability by routing work to a sub-tier vendor, even if that vendor made the defect. The contract the buyer signed covers the full scope of work, regardless of how many hands the work passes through before it comes back.
That creates a structural mismatch most engineers never plan around. A buyer's legal and commercial recourse runs to the prime contractor and nowhere else. But the actual quality risk, the place where a defect gets introduced, may sit two tiers downstream at a shop the buyer has never visited, never audited, and doesn't know exists. The accountability stays put. The risk doesn't.
How undisclosed subcontracting hides quality risk
Subcontracting itself doesn't create the danger. The danger comes from invisibility. An audit checks the controls inside one building. If a second shop performs part of the process and the buyer never learns about it, that second shop's controls were never checked at all, because the buyer didn't know to look for them.
Quality certifications validate a management system; they say nothing about whether the buyer can see the full chain of facilities that touched the part. Without tight oversight, a portion of an order can get quietly pushed to lower-tier workshops, and that rerouting can stay undetected until finished products start failing out in the field.
The processes most often subcontracted without disclosure are also the hardest to inspect once assembly is done. Heat treatment changes a material's internal structure in ways you can't see. Plating conceals the metallurgical history that produced it. Specialty finishing hides seams and transitions. Once a part is built into an assembly, there's no way to look backward and confirm which shop actually ran that process step.
That invisibility produces a specific failure pattern known as quality fade. A supplier shows the buyer a first batch, built and inspected under close attention. Later batches get quietly routed to a second shop, and that shop substitutes materials or skips inspection steps that the original batch would never have gotten past. Without batch-level traceability data tying each shipment back to its specific production run, there is no way to catch any of this before parts are already in a customer's hands.
Lead time erodes the same way, more quietly. Every undisclosed handoff between shops adds freight transit, a receiving backlog, and a redundant incoming inspection step at each facility in the chain, none of which the buyer's schedule accounted for, because the schedule assumed a single shop handled everything.
The ceiling case for where this goes is Aventura Technologies. The company sold security equipment, including networked cameras marketed as made domestically, to U.S. government agencies. In large part, the products were actually imported from China. "Made in the USA" labels were applied by the overseas manufacturer before the equipment ever shipped, and Chinese markings were scrubbed off the circuit boards to complete the cover. What started as an undisclosed sourcing decision ended as a criminal case, not a quality dispute.
Supplier flow-down requirements in regulated industries
Regulated industries didn't solve this problem through vigilance. They solved it through clause structure, written down and made contractually binding at every tier of a supply chain. Commercial buyers outside those regulatory regimes can borrow the same architecture, even though they aren't subject to the rules that produced it.
GMW's supplier quality requirements offer a clean illustration. Under that framework, a supplier must flow down the substance of each applicable quality clause, specifically the ones individually marked with a flow-down instruction, to every sub-tier shop involved. The obligation doesn't stop at the first supplier a buyer contracts with. It travels with the work.
A few specific flows from that same document show how the architecture holds together. Suppliers must give advance notice of any change: to the product or process, to a sub-tier supplier, to the manufacturing facility's location, or to the management representative responsible for quality. So a supplier cannot quietly move heat treatment to a different shop without tripping a formal notification requirement first.
The certificate of conformity has to list every special process performed, by the supplier and by any sub-tier, and name the specific processor along with the process designation: specification, revision, type, class, grade, and material condition, as applicable. The buyer, the buyer's customer, or a government representative also retains the right to enter and survey facilities, reviewing every process, subcontractor, part, procedure, and record involved, and that right travels down to every tier in the chain, not just the one the buyer contracted with directly.
One more piece closes the loop: verification of product and processes by the buyer, the buyer's customer, or a regulatory agency never absolves the supplier of responsibility for delivering acceptable product. Passing an audit doesn't release anyone from accountability for what ships afterward.
The principle sitting under all four of these requirements is consistent. A prime contractor can delegate the work of execution to a sub-tier shop. What it cannot delegate is accountability for the outcome, and even delegating execution has to be documented, approved in advance, and traceable after the fact. Commercial buyers don't need to comply with a regulatory standard or a government contract clause to use this same logic. The documentation architecture already exists and has been field-tested for decades; it just needs adapting to a commercial purchase order.
The four documents that keep the chain of custody intact when work leaves the primary facility
Chain of custody in a subcontracted order is only as strong as the paper trail behind it. Four documents, used together, close most of the gaps where quality accountability tends to blur between tiers.
The first is an advance notification requirement. Before any portion of the work gets routed to a sub-tier shop, the prime supplier has to disclose, in writing and in advance, who will perform the work, what process they'll run, and where. A notification buried in a certificate after the fact doesn't count. Depending on the relationship, written buyer approval before the change goes into effect may be warranted on a case-by-case basis. This clause gives the buyer an actual decision point in the process, not just a record of a decision someone else already made.
The second is a sub-tier certificate of conformance. The certificate riding with each shipment needs to name every special process performed at every tier, name the specific processor, and spell out the full process designation (specification, revision, type, class, grade, and material condition, where applicable). A certificate that only lists the prime supplier's own work is incomplete the moment any sub-tier process was involved, and buyers should reject shipments that arrive with incomplete paperwork rather than treat it as a formality to chase down later.
The third is batch-level traceability. This is the mechanism that actually makes quality fade detectable in practice. A material substitution or a skipped inspection step buried in a later production run leaves no paper trail to find it by unless traceability is tied to specific lots or batches. Limited traceability of raw material sources and grades raises the odds of unauthorized substitution going unnoticed, and when finished goods can't be linked back to specific material batches, a root-cause investigation turns slow, expensive, and often inconclusive.
The fourth is a right-of-entry clause that explicitly extends to every sub-tier supplier, not just the primary facility. An agreement that grants audit rights only at the top of the chain is incomplete by design. The value of this clause isn't in how often a buyer actually exercises it. The value is in preserving the option, so a subcontractor the buyer never knew to inspect can't put its best face forward at a scheduled audit while the real work happens somewhere else.
These four are a minimum viable set, not a full compliance program. An engineer or procurement lead reviewing a supplier agreement can put all four on a purchase order today without needing a dedicated quality department to administer them.
Converting requirements into enforceable obligations through quality agreements and purchase order language
None of the four documents above protect a buyer unless they're written into the commercial relationship before any work starts. A verbal understanding that a supplier will disclose sub-tier work isn't recourse. Asking for a certificate after a shipment has already failed isn't recourse either. Both are just requests, and requests carry no contractual weight when a dispute actually happens.
Defect classification is usually the first place quality agreements break down. Contracts that refer broadly to "defects" without defining categories make accountability nearly impossible to enforce later. A shipment passes AQL sampling at receiving, the buyer rejects it anyway over cosmetic inconsistency, and neither side has a contractual basis for deciding who covers reverse logistics, rework, or replacement. The agreement needs to sort non-conformances explicitly: critical defects covering safety hazards or regulatory violations carry zero tolerance, major defects covering functional failures prevent the part from doing its job, and minor defects covering cosmetic issues affect brand perception without affecting function. Each category should carry its own remedy, agreed in advance.
The purchase order itself should name the quality requirements document and its revision number. That way, when clause language like advance sub-tier notification or right of entry applies, it applies because the PO names it specifically, not because someone assumes it was implied.
Recourse triggers belong in the agreement up front, too: what counts as a non-conformance event, what inspection right that event activates, what remedy the buyer can demand (a remake, a refund, a corrective action plan), and the timeframe for each. Leaving these undefined turns a quality dispute into an open-ended negotiation with no clear end point.
The same language regulated industries use to close the "passing the audit" trap belongs in commercial quality agreements too. Verification by the buyer or by a regulatory agency doesn't relieve the supplier of responsibility for the product. A quality agreement should state this directly, so a supplier can't later argue that a passed audit closes out a quality claim filed afterward. None of this requires specialized legal drafting. Standard commercial contract language can carry the full accountability structure, if you actually write the structure down.
Specialty processes that most often break the chain (and what to require at each handoff)
Some fabrication processes are far more likely than others to leave the primary facility, and far more likely to produce defects that stay invisible once the part is assembled. Those two facts together are why these specific processes demand the tightest documentation at every handoff.
Heat treatment and surface plating sit at the top of that list. Both routinely leave the primary facility and come back for final inspection. The prime supplier's incoming inspection is the only checkpoint standing between the sub-tier shop's work and the buyer's finished parts. Special process certifications need independent documentation at each tier; a blanket certificate from the prime supplier that doesn't name the specific heat treater or plater is structurally inadequate on its face. Buyers should require the sub-tier processor's name and the full process specification, revision, type, class, and grade printed on the certificate itself, not left as a detachable attachment that can get lost between the shop floor and the shipping dock.
Tube bending and formed assemblies carry a different kind of risk. Each bend depends on tooling matched precisely to radius, angle, and the material's own properties, and a part can measure within tolerance on its own while still failing once it's fitted into a full assembly, because a different shop used different tooling than the one that originally qualified the part. Under AS9102, a change in the manufacturing source for tube bending triggers a full or partial first article inspection re-accomplishment covering the affected characteristics, as long as the change could plausibly affect fit, form, or function. Buyers should write that trigger explicitly into their own quality agreement rather than relying on the standard's scope to cover it by implication.
Powder coating and other dimensional finishes introduce a quieter failure mode. Coating adds real thickness to a finished surface, enough to turn a designed clearance fit into an interference fit if nobody accounted for it at the drawing stage. Finish thickness affects hole sizing and thread engagement directly. A shop that controls finishing in-house can catch these problems before an assembly ever reaches the line; once finishing moves to a sub-tier shop, that internal catch disappears. Design-for-manufacturing review that accounts for finish thickness needs to happen before drawings are released.
Custom springs and gaskets round out the list. These are specialty items frequently sourced from sub-tier specialists, and their functional properties, spring rate, sealing force, fatigue life, can't be confirmed by looking at the part. They need documented process controls and incoming test data from the sub-tier shop itself, not just a dimensional check once the parts arrive at the prime facility.
Vendor consolidation as the structural fix
Every document and clause covered so far manages risk that already exists the moment work splits across multiple shops. Vendor consolidation removes the risk at its source by reducing the number of handoffs where accountability can blur.
A single qualified facility that performs machining, heat treatment, plating, and finishing under one roof removes the problem of invisible sub-tier work by keeping every process under one roof. There's no second shop to disclose, no sub-tier certificate to chase down, no handoff where a batch can quietly get rerouted to a cheaper, unaudited vendor. The four documents outlined earlier still matter wherever subcontracting is unavoidable, but the number of places they need to be enforced shrinks dramatically when one facility owns more of the process.
This is a structural decision about how a supply chain gets built, not a preference for fewer vendors on a spreadsheet. Engineers and procurement teams evaluating a new supplier should ask directly which processes happen in-house and which leave the building, then weigh that answer against how much documentation discipline the relationship can realistically sustain over time. A supplier who controls more of the process under one roof carries less invisible risk by design.


