Should You Inspect Every Order? A Risk-Layered QC Frequency Matrix

Two failure modes cost importers the most money, and they sit at opposite ends of the same dial. Inspect everything and you burn budget on lots that were never at risk. Inspect only when you get nervous and the one lot you skipped in the middle of peak season is the one that arrives with the wrong motor. The fix is a coverage plan rather than a reflex: risk decides how often a lot gets eyes on it. A visit runs from $169 per man-day, reports land typically within 24 hours, and after 2,000+ lots we can say the plan is priceable arithmetic rather than a philosophy.

⏳ The two mistakes

Inspecting everything feels rigorous and is mostly theatre. If a factory has shipped six clean lots with no change of tooling, material or line, a seventh identical inspection buys you very little new information — it buys reassurance, which is a fine thing to buy but not at full coverage cost on every order.

Deciding by feel is the expensive one. The skip usually happens exactly when it should not: Peak season, when the factory is outsourcing, hiring temporary staff and running the lines flat out. That is the lot where a 30% during-production check pays for itself many times over, because the defect is still fixable inside the factory instead of inside an Amazon receiving queue.

🧭 Four risk factors that decide coverage

🗺 The matrix: three coverage tiers

Note what never disappears: an unannounced visit somewhere in the quarter. Predictability is what a factory optimises against, so keep at least one visit on a date the production manager cannot guess.

💵 The break-even arithmetic

A visit is worth it when the expected escape cost exceeds its price. Expected escape cost = units × defect rate × cost per escaping unit, where cost per escaping unit is rarely just the refund. It is the refund, the return freight, the removal or disposal fee, the air-freight replacement, and the sales you lose when a four-and-a-half-star listing drops a tenth.

The arithmetic is deliberately rough, and the point is the shape of it: cheap stable lots tolerate a document-only release, expensive or drifting lots do not. Sampling also has a blind spot worth knowing — at AQL 2.5 with a 200-piece sample you accept up to 10 major defects, so a two percent defect rate is expected to produce four bad units and still pass the lot. Run your own numbers through the AQL calculator before you assume a pass means clean.

📅 One quarter, twelve lots, seven visits

A client shipping four categories from four factories had been buying twelve pre-shipment inspections a quarter — one per lot — and still got surprised twice a year. We rebuilt the calendar around risk instead of around lots. Two full-tier visits went to the new molded-power-supply SKU and to the first lot after a factory swapped a subcontractor. Five standard-tier pre-shipment inspections covered the FBA-bound lots. The remaining five lots, all from a supplier with nine clean lots and no change events, went to watched release with line photos and carton marks reviewed by our team.

Seven visits instead of twelve, roughly $1,183 instead of $2,028, and more coverage where it mattered: the during-production visit at the 30% mark caught a material swap on the biggest lot of the quarter, 2,600 units, with three days of rework instead of a container of returns. Escapes across the quarter measured 0.6 majors per 100 — better than the twelve-visit quarter that preceded it.

🚫 When skipping is the expensive option

Small orders are not automatically exempt. If a $3,600 lot is going to FBA with a safety gate on it, or it is the first lot from a new tool, an inspection is still cheap insurance. Where the volume really is tiny, the better answer is consolidation: group two or three small lots from the same factory or the same consolidator into one visit rather than buying three separate ones. CloudSpects schedules that way often enough that it shows up in the booking confirmation as a shared man-day, and you still get a report per lot.

Want the coverage plan written out for your next quarter? Send us your PO list and we will mark which lots need a visit and which ones do not, before you book anything through our inspection services team.

❓ FAQs

Should I inspect every order from China?

Not automatically. Inspect every order when the supplier is new, the product is safety-gated or high value, or something changed in the tooling, material or line. For an established supplier with a clean record and no change events, alternate-lot coverage plus one unannounced visit per quarter usually buys more information per dollar.

How often should repeat orders be inspected?

Every lot until you have six clean ones. After that, keep pre-shipment inspection on alternate lots and review line photos and carton marks on the others. Any change event — new mold, new material, new subcontractor, new packaging line — resets that supplier to full coverage for the next lot.

What does one inspection visit cover?

A standard man-day covers one factory visit on one production lot: sampling per AQL 2.5 at level II, functional and dimensional checks, packaging and carton marks, FNSKU verification, photos of every defect found, and a report typically within 24 hours. Larger lots, multiple manufacturing lines or multi-SKU containers may need more than one man-day.

Is inspection worth it on small orders?

Sometimes, and the arithmetic tells you quickly. If a small lot carries a safety gate or a new tool, inspect it. If the value is genuinely tiny and the product is stable, consolidate two or three lots from the same factory into a single visit instead of buying one inspection per lot.

What events force an inspection even on a stable supplier?

Six things: a new mold or tooling, a material or supplier substitution, a new or relocated production line, a packaging change, peak-season outsourcing, and any lot that will be stored or shipped long enough for humidity and handling to matter. Each one resets the risk score regardless of past performance.