Many buyers begin supplier diligence by trying to place a company into one of two boxes: manufacturer or trading company. The classification can be useful, but it often creates false confidence.
A licensed manufacturer may outsource the process that determines your product’s performance. A trading company may manage a narrow network with stronger technical control than a nominal factory. A systems integrator may own almost no component production but be the only party capable of delivering the working system. The label does not tell you where control lives.
The practical question is not “does this supplier own machines?” It is “can this supplier reliably control the decisions and processes that determine our result?”
Four operating models appear repeatedly
1. Process manufacturer
This company owns the critical production process: machining, molding, casting, electronics assembly, cell manufacturing, coating, furniture construction, or another defined capability. It can be an excellent choice when that process dominates product risk and the buyer can coordinate everything around it.
2. Final assembler with managed subcontractors
This company controls final configuration, assembly, inspection, packaging, and shipment while buying components and external processes. Its strength depends on incoming controls, engineering change discipline, supplier management, and whether it can trace a field problem back through the chain.
3. Systems integrator
This company combines hardware, software, controls, and external equipment into an operating system. For automation, robotics, production lines, and inspection systems, integration may be more valuable than component ownership. The risk is that the integrator’s commercial authority over its vendors may be weaker than its technical responsibility to the buyer.
4. Commercial orchestrator
This company coordinates factories, communication, purchasing, export, and logistics. A strong orchestrator can reduce friction across many small suppliers. A weak one adds distance without adding control. The difference becomes visible when engineering changes or defects cross company boundaries.
Map ownership across the order
Instead of asking for a yes-or-no factory identity, build an ownership map. For every critical part or process, identify five forms of control.
| Control | Question to answer | Evidence |
|---|---|---|
| Technical | Who approves specifications, materials, tolerances, software, and changes? | Controlled drawings, change records, named engineer |
| Process | Who sets parameters and decides whether the process is capable? | Work instructions, parameter limits, capability or trial records |
| Quality | Who defines inspection and owns nonconforming output? | Control plan, inspection records, reaction plan |
| Commercial | Who can require a subcontractor to act, remake, or prioritize? | Purchase structure, agreements, escalation path |
| Recovery | Who contains a failure, finds cause, and verifies correction? | Recent corrective-action example and closure evidence |
If the same party owns all five, the structure is simple. If ownership is distributed, the structure can still work—but the interfaces require diligence. “Our partner factory handles that” is not an ownership answer.
What to trace during a factory visit
Choose one critical characteristic and follow it backward. If coating thickness matters, start with final inspection, then trace the specification, sampling method, incoming or subcontracted process, acceptance records, and reaction to failure. If firmware revision matters, trace the loaded version from the finished unit to the programming station, release record, source owner, and update approval.
This method is stronger than asking for an equipment list. It tests whether the supplier can connect product requirements to operating control.
We also ask to speak with the person who handles exceptions. Salespeople usually know the standard flow. The quality engineer, production supervisor, application engineer, or planner knows what happens when material is late, a process drifts, a customer changes the design, or a subcontractor rejects responsibility.
Subcontracting is not the problem
Modern manufacturing is networked. Few products are made entirely inside one building, and forcing vertical integration can remove access to specialist capability. The risk is uncontrolled subcontracting: an external process is selected mainly on price, changes without approval, cannot be visited, or operates without a clear acceptance standard.
A supplier that manages subcontracting well should be able to explain:
- which processes leave the site and why;
- how external providers are approved and monitored;
- which specifications and records travel with the work;
- how incoming output is accepted before it enters the next step;
- whether the buyer is notified before the source changes;
- who bears the cost and schedule impact when the process fails.
The commercial structure must match the control structure
If a supplier is contractually responsible for the finished outcome but has no authority over a critical subcontractor, the buyer may own more risk than the contract suggests. If the buyer nominates a component source, the assembler should not automatically carry unlimited responsibility for that component’s performance.
This is why the ownership map should appear in the quotation, quality agreement, statement of work, and escalation process. Technical reality and commercial responsibility need to describe the same system.
A better conclusion than “real factory”
At the end of diligence, the useful conclusion is not a label. It is a statement of control: which work is internal, which is external, who owns each decision, what evidence will be available, and where the buyer must add its own oversight.
A trading company with transparent boundaries can be manageable. A manufacturer hiding a critical external process can be dangerous. A systems integrator with strong interface ownership can be more valuable than any single component factory.
The job is to make the operating model visible before the deposit turns every unclear boundary into a negotiation.