Inspection Method

Full Inspection vs Random Sampling: Which Quality Check Fits Your Order?

Published May 2026 · By SSTI Inspection Team

When planning a quality check for an order from China, one of the first decisions is whether to inspect every unit or a statistically representative sample. Both approaches are legitimate tools used in pre-shipment inspection. The choice between them depends on the order characteristics, the product's risk profile, the supplier's track record, and the buyer's specific quality control objective — whether that is confirming the batch quality level or sorting acceptable from unacceptable units.

Understanding what each method actually does — and does not — accomplish is the starting point for making the right choice. This guide explains both approaches, describes the situations where each provides the most value, and addresses the cost and time tradeoffs that buyers need to consider.

What Full Inspection (100% Inspection) Actually Means

In a full inspection, every unit in the shipment is physically checked against a defined set of criteria. The result is a direct count of how many units pass and how many fail — and typically a sorted output: acceptable goods separated from defective goods.

Full inspection is primarily a sorting exercise rather than a statistical assessment. Its purpose is to identify and remove defective units from the shipment, not to estimate the defect rate across the batch. A buyer who uses full inspection before shipment approval receives the highest possible unit-level coverage available from an on-site check. The practical output is that the shipped goods represent only the units that passed inspection.

However, full inspection does not mean zero defects in the shipped goods. Human inspection carries inherent limitations — fatigue, attention span over a long inspection session, and the practical difficulty of detecting certain internal or latent defects through visual or manual checking. For products where defects are subtle, involve components not visible in external inspection, or require functional testing under specific conditions, 100% inspection of externally visible characteristics may still leave some defective units undetected.

Full inspection is also more time-consuming and more expensive than AQL sampling for the same batch size, since every unit must be individually handled and assessed.

What Random Sampling (AQL Inspection) Actually Means

AQL sampling inspects a statistically determined subset of the batch — typically between 80 and 315 units depending on batch size and inspection level — and uses the results to draw a conclusion about the quality level of the full quantity. If the defects found in the sample fall below the acceptance number defined by the chosen AQL level, the batch passes. If they exceed it, the batch fails.

AQL sampling is a batch-level quality assessment, not a sorting exercise. Its output is a judgment about whether the batch as a whole is within an acceptable quality threshold — not a direct count of defective units across the full quantity. A passed AQL result means that the defect rate in the sample supports accepting the batch; it does not mean that no defective units exist in the shipment.

Random sampling is significantly faster and more cost-effective than full inspection for the same batch size. It is the standard method used in pre-shipment inspection across most product categories and most buyer-supplier relationships because it provides a reliable quality assessment at a practical cost and within a practical time frame.

When Full Inspection Is the Right Approach

Full inspection provides value that sampling cannot replicate in specific situations where the buyer needs to act on unit-level results rather than a batch-level judgment.

The most common case is post-rework verification. If a batch previously failed an AQL inspection and the factory has reworked the identified defective units, the buyer may want 100% inspection of the reworked goods to confirm that all affected units have been corrected before approving shipment. A sampling result on a post-rework batch can produce a misleading picture if the reworked units are not evenly distributed through the batch; full inspection eliminates this risk.

Small-to-medium batches where the quantity makes full inspection practically feasible within the available inspection time are natural candidates. For a batch of 200 or 300 units, the time difference between checking every unit and checking an AQL sample may be marginal, while the certainty provided by full inspection is substantially higher.

High-value products where the financial consequence of a single defective unit reaching the end customer is significant may justify the additional cost and time of 100% inspection. Luxury goods, precision instruments, and products subject to specific safety regulations that demand unit-level verification fall into this category. So do products where the buyer has received customer complaints on previous shipments and needs to establish the actual defect rate across current stock before further sales.

Products with defects that are both clearly visible and easily assessed during a brief visual check lend themselves to full inspection because the per-unit time is low and the sorting benefit is direct. Goods where defect detection requires more extensive individual handling or testing are less practical candidates for full inspection at scale.

When Random Sampling Is the Right Approach

AQL sampling is appropriate for the majority of standard pre-shipment inspection situations — and is the approach most buyers use as their default quality control method in ongoing supplier relationships.

Large batch orders where full inspection is not feasible within a standard inspection visit make AQL sampling the only practical option. A batch of 5,000 units cannot be individually inspected within a single working day; a statistically sized sample of that batch can be, with a result that provides a reliable indication of the batch quality level.

Repeat orders of the same product from an established supplier, where previous inspection results provide a baseline for expected quality levels, are well-suited to AQL sampling. The sampling result confirms whether the current batch is consistent with the supplier's established quality performance. Deviation from the expected result — an unexpected fail, or a passed result with a higher defect count than prior shipments — is itself informative.

Product categories where sampling is the recognised industry standard for pre-shipment quality assessment, and where the buyer's objective is batch-level quality confirmation rather than unit-level sorting, fit the AQL framework well. Most consumer goods, industrial components supplied in bulk, and textile or apparel orders fall into this category.

The Cost, Time, and Risk Tradeoff

Choosing between full inspection and random sampling involves a practical balance between three factors: cost, time, and the level of quality assurance the buyer requires.

Full inspection costs more and takes longer for any given batch size. The cost difference increases with batch size: for a 500-unit batch, the difference may be modest; for a 3,000-unit batch, the cost of full inspection may be several times that of AQL sampling. The time implication also matters if the inspection is on the critical path between production completion and the vessel loading date.

AQL sampling costs less and completes faster, but provides a probabilistic rather than exhaustive result. The confidence it provides increases with sample size and decreases with batch size relative to sample size. A buyer who understands AQL statistics can make an informed judgment about whether the standard inspection level provides sufficient confidence for their specific order value and risk tolerance.

A practical approach that some buyers use is to apply AQL sampling as the primary inspection method, but to apply 100% checking to the highest-risk individual criteria within the sample — for example, scanning every inspected unit's barcode, or checking the label content on every unit, while applying AQL criteria to workmanship and appearance. This combines the efficiency of sampling with targeted completeness on the criteria where errors have the highest practical consequence.

Making the Decision for a Specific Order

The right inspection approach for any given order should be determined by the combination of product type, order volume, supplier history, and the buyer's specific quality concern rather than by a fixed rule.

If a batch previously failed inspection and has been reworked, full inspection on the reworked quantity provides the most reliable confirmation. If a supplier has a strong track record across multiple completed orders and the product is a standard repeat item, AQL sampling at the appropriate level is the proportionate approach. If a product is high-value, the batch is small enough to be feasible, and the buyer's customer has specific quality expectations, full inspection may be justified even on a first order.

Discussing the inspection approach with the inspection partner before booking — rather than accepting a default method — tends to produce more appropriate results. An experienced inspection team can advise on the right sample size, inspection level, and defect classification for a specific product category and order situation. More information on how SSTI approaches inspection method selection is on the Services page. Buyers who want to discuss the right approach for a specific order can contact SSTI through the Contact page. For context on how inspection fits into the broader order process, see the Process page.

Summary

Full inspection and AQL random sampling are complementary tools, not competing alternatives. Full inspection provides unit-level sorting and is most appropriate after rework, for small batches, and for high-value or compliance-driven products. AQL sampling provides a batch-level quality assessment and is the standard approach for most volume pre-shipment inspections. Choosing between them — or combining elements of both — based on the specific order characteristics produces better quality control outcomes than applying a single method uniformly regardless of context.

Back to Insights