Incoming QC: 3 Checks That Stop Bad Batches on Arrival

The shipment has reached your warehouse. The supplier’s COA says 99.2% — inside specification. Your QC lab tests the same material and gets 98.1%, outside your assay limit. Whose number is right? Before accusing anyone, ask a harder question: did your incoming QC process give you a defensible answer?

This guide is written from the manufacturer’s side. We ship intermediates to buyers across India, Bangladesh, and Pakistan — and the buyers who argue with suppliers the most rarely have bad suppliers; they have no clear incoming inspection procedure. The ones who resolve disputes in days run the same seven checks.

Incoming QC Is Where Sourcing Ends and Quality Begins

Your sourcing process does not end when the truck arrives. It ends when a batch is examined against the agreed specification and released or rejected. Everything before this — the RFQ, the negotiation, the shipping — created an expectation. Incoming QC is where it meets a physical drum.

The checks are deliberately unglamorous. No single one proves a batch good; together they answer three questions: is it the batch you ordered, did it arrive intact, does the supplier’s data hold up in your lab?

The First Five Minutes: What to Check Before Unloading

Start before material is stored. The first five minutes catch more problems than any lab test — and cost nothing.

• Three-way batch identity match. The batch number on the COA must match the drums and the packing list. Any mismatch — stop, photograph, and notify the supplier before unloading. This check prevents the most common mix-up in receiving.
• Label against your order. Product, grade, and quantity should match your PO. If you wrote a tight specification into your RFQ, confirm the COA cites it — not the supplier’s looser in-house one.
• Container condition. Look for dents, punctures, moisture stains, or broken seals. A damaged container does not fail the batch, but it changes your sampling plan: damaged containers are sampled separately.
• Quantity and transport evidence. Count drums and check gross weight against the packing list. Read any temperature or humidity indicators before anything moves; material exposed outside its declared storage conditions in transit needs a documented decision.

Why a COA Is Not Proof of Compliance

The most expensive assumption in buying intermediates is that a supplier’s certificate of analysis is a guarantee. It is not. A COA is a declaration — the batch met their specification, tested with their methods on their instruments. It is evidence, not proof — and the difference matters when your lab and theirs disagree.

Regulators agree. Under US 21 CFR 211.84, a receiving operation must examine containers and labels and perform at least one identity test on each batch before use — because documentation alone is not enough. Three things a COA never tells you:

• Whether the batch meets your specification. Supplier specs and buyer specs differ. Reading a COA against your own acceptance criteria is a skill — see how to read a COA properly.
• Whether the document is authentic. COAs can be copied, edited, or recycled for another batch. The signs of a fraudulent COA are visible if you know where to look.
• Whether the methods behind the numbers match yours. An HPLC area-percent assay and a titration can disagree on the same sample, so test against the specification you actually set — the discipline of specification setting — not a number you assume means the same in both labs.

The Seven Incoming QC Checks

Checks escalate from paperwork to laboratory work: run all seven on a first shipment; for an established supplier, checks 1–5 stay mandatory and 6–7 follow the reduced-testing logic below.

Check 1 — Identity and Documentation Cross-Match

Verify the COA’s batch number, dates, and retest window are internally consistent — and that the batch is still within its retest period on arrival. A batch that lapsed in transit fails on arrival.

Check 2 — Container and Seal Integrity

Seals intact, drums unopened, desiccant indicators in range. One broken seal can void a hygroscopic batch’s moisture claim — the cheapest quality check you run.

Check 3 — Weight, Quantity, and Label Accuracy

Net weight per drum against the packing list. A persistent underweight pattern suggests an out-of-control filling process — document it even if you accept.

Check 4 — Visual Screen

Open a representative drum and look: color versus the supplier’s typical description, free-flowing versus caked, no foreign matter. A trained eye catches most gross deviations before any instrument.

Check 5 — Retest and Resampling Discipline

Define in advance how many samples, from which drums, with what tolerance. This discipline — behind a supplier’s batch-to-batch consistency — protects you from your own “test into compliance.”

Check 6 — Identity Test in Your Own Lab

Run at least one identity test on a representative sample: FTIR against a reference spectrum, melting point, or another identity reaction. This is the regulatory minimum in 21 CFR 211.84(c), and the one test that catches mislabeled material. It only has power if you hold a suitable reference standard — qualify one before you need it.

Check 7 — Key-Specification Spot Check

Verify the two specification points that matter most to your next step — typically assay and water, plus any critical impurity — and compare with the COA. The question is not whether the numbers are identical but whether they sit within the method’s inter-laboratory variation. A few tenths of a percent on HPLC is usually noise; beyond that range is a genuine discrepancy that triggers rejection.

When You Can Test Less: The Reduced-Testing Premise

Full testing of every batch is expensive, and most buyers cannot justify it once a supplier has a record. Reduced testing is a privilege the supplier earns — and it rests on three conditions. If any is missing, test fully.

• Supplier qualification is current — you have recently audited the manufacturing site and found its QC sound.
• Historical data is trustworthy — past COAs survived your verification, traceable to proper reference standards.
• Batch history is stable — few excursions across recent batch-to-batch consistency data.

Even then, keep checks 1–4 on every batch. Reduced testing relaxes laboratory work — nothing else.

When a Batch Fails Incoming QC: Six Steps

A failing batch is not a disaster; a disorganized response is. Six steps keep a rejection under control:

1. Quarantine physically. Move the batch to a marked hold area and label every drum HOLD — NOT RELEASED. If sensitive, quarantine under its required storage conditions so the investigation creates no second deviation.
2. Notify the supplier in writing within 24 hours, attaching test data, method summary, COA, and photos. Written notification with evidence turns blame into a technical discussion.
3. Ask for a root-cause investigation, not a new COA. Request their deviation report: what their lab sees, whether check standards were in range, what explains the difference. A supplier who replies with a fresh certificate instead of an investigation has told you something.
4. Handle retest rights carefully. Agree to a retest only under fixed conditions: same samples, same method, an agreed independent laboratory if needed. Never allow “send new samples” — that turns verification into a lottery. Retesting is for investigation, not for cycling methods until a result passes; that is test-into-compliance — a red line, not a tactic.
5. Decide: reject, replace, or accept with deviation. Rejection and replacement are the clean options. Accepting with a documented deviation fits only non-critical deviations, needs written quality approval, and records downstream impact. Never let a discount substitute for a quality decision — and remember that payment terms like TT versus LC decide your leverage when a rejected batch needs a refund or replacement.
6. Record, score, and act. File the investigation and score the supplier: one rejection with a credible root cause is an incident; two in a year is a pattern. Patterns feed your sourcing plan — more audits, or the moment when the hidden costs of switching suppliers look acceptable.

Six Supplier Red Flags to Recognize Early

One pattern deserves attention: the supplier whose impurity story is thin. If a rejected batch cannot be explained by the data they share, the gap matters downstream — impurities that survive your process reappear later. Read about impurity fate mapping before accepting a “harmless impurity” argument.

Frequently Asked Questions

If the COA says compliant, do I still need to test it myself?

On a first shipment, yes — at minimum an identity test plus a spot check of the points critical to your process, mirroring 21 CFR 211.84. Once a supplier’s history and batch consistency are proven, move to reduced testing, keeping documentation and container checks on every batch.

My result and the COA differ. How big a difference is a failure?

Judge against inter-laboratory method variation, not zero. A few tenths of a percent on HPLC is usually method noise; a difference beyond the method’s expected range — or any failure on a point critical to your next step — is genuine. When in doubt, test a joint split sample at an independent laboratory before calling it a rejection.

The supplier wants to retest. Should I agree?

Yes — but under a fixed protocol: same samples, same method, an agreed laboratory, never fresh samples chosen by the supplier. If the process continues until a result passes, that is test-into-compliance — a red flag, not a resolution.

What does accepting a batch with a documented deviation cost?

A concession releases the batch, not the risk. The deviation moves downstream and can reappear in your output, so concessions fit only non-critical deviations, need written quality approval, and should be rare. Conceding regularly means the specification or the supplier is wrong — change one.

Build Incoming QC That You Can Defend

Incoming QC is not about distrusting suppliers. It makes both sides’ numbers mean the same thing — and gives you a documented answer when they do not. The seven checks are deliberately ordinary: match the paperwork, inspect the containers, run an identity test, spot-check key specifications, and treat failure as an investigation, not a fight.

Buyers who do this get better supplier behavior, because suppliers respond to those who hold material to a standard. When a batch genuinely fails, the same record that protects you shows the supplier what to correct — the only outcome that makes the next shipment better. A second source gives you leverage before a rejection forces it: see when dual sourcing makes sense.