Raw material quality control is the first line of defence of any production process. A defect that comes in with the raw material is amplified throughout the entire chain, so that every euro invested in a rigorous incoming inspection prevents many euros of rework, returns and complaints. This article develops the inspection and supplier acceptance procedures in the supply chain, with standardised acceptance sampling, supplier qualification and the traceability that regulated sectors require.
Why input quality determines output quality
Genichi Taguchi's "loss function" principle explains it clearly: any deviation of the raw material from its target value generates a loss that grows with the square of the deviation, even if the part remains within tolerance. That is why incoming control is not just about separating the conforming from the non-conforming, but also about measuring how much variability each supplier introduces into the process. Crosby's practical "factor of 10" rule adds the economic argument: a defect detected at receiving costs ten times less than the same defect detected in production, and a hundred times less than if it reaches the customer.
Clause 8.4 of the ISO 9001:2015 standard regulates precisely the "control of externally provided processes, products and services". It requires evaluating and selecting suppliers according to objective criteria, defining the controls to be applied to what is received and keeping evidence of conformity. Raw material control is not optional for a certified organisation: it is an explicit normative requirement.
Purchasing specifications: the basis of all control
You cannot inspect what is not specified. Every raw material must have a specification sheet that defines the critical characteristics, their nominal values, the admissible tolerances and the measurement method. The characteristics are usually classified into three levels: critical (affecting safety or legal compliance), major (affecting function) and minor (affecting appearance). This classification determines the rigour of the sampling and the acceptance criterion of each characteristic.
The specification sheet must include, alongside each characteristic, the standardised test method by which it will be measured and the certificate that will be required from the supplier. This is where the certificate of analysis (CoA) that accompanies each batch becomes valuable: it documents the results of the tests carried out by the supplier on that specific batch. A CoA, however, does not replace incoming inspection unless the supplier is fully validated; it is wise to periodically cross-check the values declared in the CoA against your own measurements to confirm that the certificate is reliable. Repeated discrepancy between the CoA and the receiving measurement is, in fact, one of the earliest indicators of a supplier whose process is degrading.
For the receiving measurements to be defensible, the measurement system must itself be under control. This is verified with a measurement system analysis (MSA), usually a repeatability and reproducibility study (Gauge R&R) that quantifies how much of the observed variation comes from the instrument and the operator rather than from the product. Measuring equipment without calibration traceable to national standards, compliant with the ISO/IEC 17025 standard when a laboratory is involved, can cause good raw material to be rejected or defective material to be accepted, invalidating the entire control effort.
Acceptance sampling under ISO 2859 and ISO 3951
Inspecting 100% of a batch is rarely feasible or reliable, because repetitive inspection induces fatigue and human error. Acceptance sampling makes it possible to decide on a complete batch from a statistically representative sample. For attribute characteristics (pass/fail) the reference is the ISO 2859-1 standard; for variable characteristics (continuous measurements) it is ISO 3951.
The key parameter is the AQL (Acceptable Quality Limit): the maximum percentage of defectives that is considered tolerable on a systematic basis. The standard defines single, double and multiple sampling plans, and three general inspection levels. An essential mechanism is switched inspection: if a supplier demonstrates sustained quality, you move to reduced inspection (lower cost); if it fails repeatedly, you switch to tightened inspection (more samples) until it recovers the level. In this way the inspection effort automatically adapts to the supplier's actual performance.
Understanding the operating characteristic curve (OC curve) of the chosen plan is decisive in order not to fool yourself with a false sense of security. The OC curve relates the real quality of the batch to the probability of accepting it, and makes clear that no acceptance sampling plan discriminates perfectly: there is a producer's risk (rejecting a good batch, normally set at around 5%) and a consumer's risk (accepting a bad batch, normally bounded at 10% for the limiting quality level or LTPD). Choosing the sample size and the acceptance number is equivalent to deciding how those two risks are shared. Reducing the sample size cuts the cost of inspection but flattens the OC curve and increases the probability of accepting defective batches, a trade-off that the quality manager must accept consciously and not by oversight.
It is also worth recalling a fundamental limitation: acceptance sampling detects bad batches, but it does not improve quality by itself. Deming harshly criticised mass inspection as a substitute for process control, because inspecting to find defects is always more expensive than producing without generating them. That is why receiving sampling should be understood as a temporary safety net while you work with the supplier to make its own process capable, at which point incoming inspection can be reduced or eliminated through certified-quality agreements.
Supplier qualification and approval
Incoming control is the last net, not the first. Input quality is built earlier, in supplier approval. The process includes the second-party audit (the customer audits the supplier), the assessment of its management system — ideally certified to ISO 9001 or to the relevant sectoral standard, such as IATF 16949 in the automotive sector — and the approval of initial samples through the PPAP (Production Part Approval Process) procedure in industrial sectors.
Once approved, the supplier is monitored with a dashboard: defect PPM (rejected parts per million), on-time delivery percentage, number of open non-conformities and response time to incidents. This monitoring makes it possible to segment the supplier portfolio and concentrate audits and tightened control where the risk justifies it.
Traceability: from the batch received to the product delivered
Traceability links each batch of raw material with the final products into which it was incorporated. It is mandatory in sectors such as food (EU Regulation 178/2002), pharmaceuticals (GMP standards) and medical devices. Its value becomes evident in a product recall: without traceability you have to recall all the production of a period; with batch-level traceability the recall is narrowed to the units actually affected, drastically reducing the cost and the reputational damage. Traceability requires unique batch identification, recording of the incoming inspection and a documentary link with the manufacturing orders.
Comparison table: incoming inspection levels
| Level | When to apply it | Inspection cost | Reference |
|---|---|---|---|
| 100% inspection | Critical safety characteristics | Very high | Control plan |
| Tightened sampling | Supplier with a history of failures | High | ISO 2859-1 |
| Normal sampling | Standard approved supplier | Medium | ISO 2859-1 |
| Reduced sampling | Supplier with sustained quality | Low | ISO 2859-1 |
| Skip-lot / no inspection | Certified and validated supplier | Minimal | Certified-quality agreement |
Managing non-conformities at receiving
When a batch fails inspection, the non-conformity procedure is triggered. The batch is identified and physically segregated to prevent its accidental use, the deviation is documented and its disposition is decided: rejection and return, rework, use under a documented concession or scrapping. The root-cause analysis (5 Whys, Ishikawa) determines why it failed, and a corrective action is opened with the supplier. A well-managed non-conformity does not end with the return of the batch, but with proof that the cause that originated it can no longer recur.
Common mistakes in raw material control
- Inspecting without a clear specification: with no defined tolerances, the decision to accept or reject is arbitrary.
- Applying the same AQL to all characteristics: a critical safety characteristic cannot have the same criterion as a cosmetic one.
- Not switching the inspection level: continuing to inspect 100% of an impeccable supplier, or relaxing control with one that fails.
- Broken traceability at receiving: accepting batches without recording their identification prevents narrowing a recall.
- Closing the non-conformity without root cause: the defect reappears in the next batch.
Frequently asked questions
What is the AQL and how is it chosen?
The AQL is the maximum percentage of defective units that is systematically accepted in a batch. It is chosen according to the criticality of the characteristic: very low values (for example 0.065%) for critical characteristics and higher values for minor defects.
Is it mandatory to inspect all raw materials?
Not necessarily 100%. The standard requires controlling what is externally provided, but the level of control can be modulated according to the risk and the supplier's track record, even reaching certified quality with no incoming inspection for fully validated suppliers.
What is the difference between ISO 2859 and ISO 3951?
ISO 2859-1 applies to inspection by attributes (the unit passes or fails); ISO 3951 to inspection by variables (a continuous magnitude is measured). Sampling by variables usually requires smaller samples for the same protection, but it requires the characteristic to follow a known distribution.
How does traceability relate to a product recall?
Traceability makes it possible to identify exactly which final products contain a defective raw material batch, so that the recall is limited to the affected units rather than to all the production of the period.
Conclusion
Raw material quality control is where quality costs the least and pays the most: stopping a defect at the receiving dock is always cheaper than chasing it downstream. The combination of clear specifications, acceptance sampling standardised under ISO 2859 and ISO 3951, prior supplier approval and batch-level traceability turns incoming control into a self-regulating system, devoting inspection effort only where the risk demands it. At Summum Quality we help organisations set up this complete system — specification sheet, sampling plan, supplier dashboard and documentary traceability — so that the first barrier of the chain is also the most reliable.