A factory that produced a beautiful sample has demonstrated one thing: it can make one good unit, eventually. Whether it can make ten thousand good units on a schedule is a different question — and it is the question a capability assessment answers before you commit a purchase order. This article provides the assessment framework: what to look at, how to verify without a site visit, and how to design a pilot order that measures capability. It is written for buyers selecting factories for first orders or scaling an existing relationship.
The sample-to-capability gap
Samples and mass production are different manufacturing events. A sample may be built by the factory's best hands — engineers and sample-room specialists — with parts selected from the best lots on the shelf. Mass production is the average line worker, on cycle time, with material lots purchased at volume pricing, repeated thousands of times. The gap between those two events is where first-order disasters live: defect rates that only appear at volume, lead times that only stretch at volume, materials that get substituted when the good lot runs out.
Assessment exists to price that gap before the purchase order. Not to predict perfection — to establish, with evidence, whether the factory's system can hold the standard your sample defined, and what it will take from both sides to keep it there.
What to assess
| Dimension | What good looks like | How to verify |
|---|---|---|
| Machine and process fit | Equipment and processes suited to your product's materials and tolerances — not generically present | Line photos or video of your product type in production; process list mapped against your BOM |
| In-house versus outsourced | Core quality-determining processes in house; outsourcing disclosed and bounded | Ask which steps are subcontracted, and verify the subcontractor question directly |
| QC infrastructure | Inspection stations, calibrated instruments, defect boundary samples, written standards | Photos of stations and records; a redacted QC report from another order |
| Materials depth | Established sources for your key materials; change-control discipline on substitutions | Material list with sources; a description of how substitutions get approved |
| Capacity headroom | Quoted lead time is believable against current load — not heroic | Ask what else is running on the relevant lines, and when your order would slot in |
| Engineering support | Someone technical who can discuss tolerances, materials and failure modes | A technical call — not the salesperson — on your product's two hardest specifications |
| Communication depth | Questions flow both ways; the factory asks about your standard, not just your quantity | Review the sampling correspondence: who asked about what, and when |
Assessing without a site visit
Most first assessments happen remotely, and the sequence matters:
- Structured questionnaire. Capacity, processes, in-house versus outsourced steps, QC staffing, current load — scored answers, not essays.
- Live video walk-through. A video call walked through the workshop that will make your product, requested during production hours. Empty lines tell you nothing, and the factory's comfort with the request is itself a signal.
- Document verification. Business credentials, export records and references — the lighter cousin of the full process in our supplier verification guide.
- Reference checks. One or two customers of similar product type and volume. Ask what broke and how the factory handled it — recovery behavior is the real data.
- Onsite audit for larger commitments. Before meaningful capital, an audit remains the standard — see our factory audit checklist.
Design the pilot order to measure capability
A pilot order is an instrument, not a discount. Size it large enough to cross from sample-room logic into production logic — large enough that the factory must actually use the line, materials at volume and normal staffing — but small enough that total failure is survivable. Then measure what the pilot exists to reveal, against criteria written before it starts:
- Defect rate at pre-shipment inspection, against the agreed AQL
- On-time completion against the confirmed schedule — including whether early warnings came from the factory or from you
- Communication: latency, precision of answers, whether problems were reported or discovered
- Rework behavior when defects were found: root cause and correction, or quiet sorting
- Golden sample fidelity: does the mass-produced unit match what you approved
Reading the signals
Capability leaks through small behaviors. A factory that asks about tolerances, edge cases and packaging details is planning to meet them; a factory that says yes to everything is planning to try. Realistic lead times that include buffer signal production experience; instant promises usually signal sales targets. An early warning about a material delay is the behavior of a factory that expects a long relationship — problems surfaced early are nearly always cheaper than problems discovered late. No single signal is conclusive, but a pattern of them, consistent from sampling through the pilot, is about as close to a prediction as sourcing offers.
Walk away, or dual-source
Assessment pays off most when it produces a "no." Walk-away signals: capacity answers that change every time you ask; refusal to show the relevant lines live; core processes outsourced to an unverifiable subcontractor; quality control that exists only as a final gate; a golden sample the factory defends as a one-off rather than a reference. Where the product matters but the factory is imperfect, dual-sourcing caps the damage: split the first volume order between two assessed factories, compare their pilot data, and let performance — not sentiment — allocate the second order. The discipline costs coordination; it removes single-supplier hostage risk from your launch calendar.
Where this sits in a managed program
In FULVERA's gate system this is Gate 01: supplier and product verification — business credentials, export records, capability interviews and references before the first deposit, with factory audits added for material programs. The assessment scales with the stakes: questionnaire and video for a trial order, a full audit for a line you plan to build a brand on. The framework lives on our quality and compliance page, with the buying-side context in our sourcing program.
The pre-PO checklist
- Questionnaire scored, with answers consistent across two separate asks
- Live video walk-through completed during production hours
- Credentials, export records and two references verified
- Technical call completed on your two hardest specifications
- Pilot order sized, with acceptance criteria written before it starts
- Pilot measured against AQL, schedule, communication and rework behavior
- Decision recorded — proceed, dual-source or walk — with the evidence filed
