Early on, quality control for most sellers is informal: check a sample from the first shipment, keep an eye on customer complaints, and trust the relationship as it develops. That approach scales poorly — as order volume, SKU complexity, and the financial stakes of a bad shipment all grow, an ad hoc approach means quality problems are caught later (often from customer complaints or return rates, well after a defective batch has already reached buyers) and cost more to fix than a systematic inspection process would have.

The difference between a factory audit and a shipment inspection

These are related but distinct, and a mature QC program uses both:

  • A factory audit evaluates the factory itself — its production capacity, equipment, working conditions, quality-management systems, and general capability — typically done before committing to a new factory relationship or at a scaled cadence (annually, or when volume increases significantly) for an existing one. It answers "can this factory reliably produce what we need, consistently, at the volume we're planning."
  • A pre-shipment inspection evaluates a specific production run before it ships, checking a statistically-meaningful sample against your specifications and an agreed quality standard. It answers "is this specific batch acceptable to ship," and is the inspection type most directly tied to catching a defective run before it reaches your inventory or customers.

Building a formal QC system

  1. Define written quality specifications and an acceptance standard for each product — not just "it should look right," but explicit dimensions, materials, functional tests, and cosmetic tolerances. A common framework here is AQL (Acceptable Quality Limit) sampling, an internationally recognized statistical standard for how many units to sample from a given batch size and how many defects are acceptable before the batch is rejected — using a recognized standard rather than an ad hoc sample size gives you and the factory a shared, defensible basis for accept/reject decisions.
  2. Decide who performs inspections: an in-house QC hire (justified at high enough volume and value at risk), a dedicated third-party inspection service (common for sellers not yet ready to hire in-house QC staff, and available in most major manufacturing regions), or, less rigorously, the factory's own QC team (useful as a first layer, but represents a conflict of interest as your sole inspection source for anything beyond low-stakes orders).
  3. Set a formal audit and inspection cadence tied to risk and volume — for example, a full factory audit before a new relationship begins and annually (or after any major change, like a new production line or ownership change) thereafter, plus a pre-shipment inspection on every production run above a defined order value or unit count, with a lighter sampling approach for smaller or lower-risk orders.
  4. Document a corrective-action process for what happens when an inspection fails: does the factory rework the batch, is a partial shipment accepted with a price adjustment, or is the order rejected entirely? Agreeing on this process in advance, ideally in the supplier contract (see Supplier and Manufacturing Contracts: What to Include), avoids an ad hoc, adversarial negotiation happening in the moment a failed inspection is discovered.
  5. Track defect data over time, by factory and by product, not just per shipment in isolation — a factory whose defect rate is trending up over several orders is a different, more strategic problem than a single bad batch, and trend data is what surfaces that distinction.

What to actually put in an inspection

A thorough pre-shipment inspection generally checks: quantity (does the shipment match the PO), visual/cosmetic quality against your specification, functional testing appropriate to the product (does it work as intended, tested on a representative sample), packaging and labeling accuracy (including any compliance labeling required for your market), and basic safety/compliance spot-checks relevant to your product category — see Product Compliance Basics for the category-specific compliance layer this should tie into.

Common mistakes

  • Relying solely on the factory's own QC reporting for anything beyond low-stakes, low-volume orders, without independent verification.
  • No written, agreed quality specification, leaving "acceptable quality" as a subjective judgment call that differs between you and the factory when a dispute arises.
  • Inspecting only the first shipment from a new factory and then discontinuing inspections once trust is established, missing gradual quality drift over time (a common pattern as a factory takes on more customers or substitutes materials/subcontractors without notice).
  • No documented corrective-action process, turning every failed inspection into an ad hoc, relationship-straining negotiation.
  • Treating a single failed inspection as a factory-ending event without investigating root cause — sometimes a systemic future risk, sometimes a genuinely isolated incident worth working through rather than immediately re-sourcing.

Best practices

  • Use a recognized sampling standard (like AQL) rather than an arbitrary sample size, so accept/reject decisions are consistent and defensible to both sides.
  • Maintain inspection and defect-rate records by factory over time, not just per shipment, to catch gradual quality drift before it becomes a serious problem.
  • Agree on the corrective-action process in the supplier contract before you need it, not while negotiating a specific failed batch.
  • Scale inspection rigor to risk and order value — a full formal inspection on every small reorder of a low-risk, well-established product may be unnecessary overhead, while a first production run of a new, higher-risk product warrants the most thorough inspection you can arrange.

FAQ

Do we need a full factory audit for every supplier, even smaller ones? Not necessarily at the same depth — scale the rigor of both audits and inspections to order volume, product risk (safety-critical or compliance-heavy categories warrant more), and how much is financially at stake if a batch fails. A lighter-touch approach for smaller, lower-risk orders is reasonable.

Should we use the factory's in-house QC team or an independent inspector? An independent third-party inspector (or in-house QC staff of your own) gives more reliable, unbiased results than relying solely on the factory's own reporting, especially as order value and risk increase. The factory's own QC can be a useful first layer but shouldn't be your only check for higher-stakes orders.

What's a reasonable AQL level to use? This varies by product category and how critical defects would be (a cosmetic flaw is judged differently than a functional or safety defect) — many third-party inspection services can advise on an appropriate AQL level for your specific product rather than defaulting to a generic standard that may be too loose or too strict for your situation.