A third-party laboratory sits outside the commercial relationship, which is exactly why its result carries weight. That weight has limits. Knowing which questions a third party can answer, and which ones it cannot, saves you from spending money on a test that changes nothing.
There is a moment in most intermediate supply relationships where a buyer starts wondering whether the numbers are right. The certificate of analysis arrives. Your own incoming test confirms it. Then three months later a quality issue appears downstream, and the question is whether the material was ever what the certificate said.
Independent lab testing is what you reach for in that moment. Send a retained sample to a laboratory with no commercial stake in your supplier, and let the result stand on its own. It is a strong instrument, and it is also frequently used in ways that produce less than buyers expect.
This guide covers what a third party can establish, what it cannot, when a sample is worth sending, and what to agree with your supplier before it leaves your building. The last matters more than most buyers realise — a surprising share of inconclusive results trace back to a conversation that should have happened first.
Four Testing Relationships And What Each One Proves
Most buyers think about testing as one activity with one result. It is four different activities answering different questions. Confusing them is the most common reason a third-party test fails to settle the question it was sent to settle.
| | Supplier’s own QC | Your internal lab | Independent third party | Arbitration lab |
|---|
| Who runs it | The manufacturer | You | A lab you and your supplier both accept | A lab appointed by agreement |
| Commercial interest | Direct | None | None, by construction | None |
| Proves | The batch met the specification on the day it was tested | Material arriving at your dock meets your limits | A second, unbiased measurement agrees or disagrees | Which of two conflicting results is correct |
| Does not prove | The process is stable across batches | The sample represents the batch | The whole batch is uniform | Which supplier is at fault |
| Best used when | Every consignment, as baseline | Every consignment, risk-based | There is a specific doubt worth resolving | Two results conflict, relationship still live |
Read the third row carefully, because it is the one people skip. An independent result proves that a measurement was taken, by someone with no interest in the outcome, and came out a particular way. It does not prove the material was made the way you think, that the batch is homogeneous, or that the process can repeat that result. For those you need the batch documentation — see our guide to reading a batch manufacturing record.
When You Should Not Send A Sample
Third-party testing has a cost that rarely appears in the conversation: your staff time, the sample, the delay while the result is pending, and the signal a test sends to the supplier. A buyer who commissions a test without a specific question usually gets a number and no decision.
When the question is already answered by your own data. If your incoming testing shows a clean trend across several batches with no drift in the attribute in question, a third-party result on one more batch will not change your view of the supplier.
When the doubt is really about the specification. If a result is marginal against your internal limit, check the limit first. A test cannot prove a material conforms to a specification that was set incorrectly in the first place. That conversation belongs with your supplier and your own quality team, and it starts with where the acceptance criteria come from.
When the sample will not represent the batch. A composite sample drawn from bulk material, or a grab sample from a drum that sat at the dock through a wet afternoon, tells you about that sample. If you suspect non-uniformity, the sampling plan needs to address uniformity first — our article on incoming QC checks for intermediates covers how to draw samples that carry weight later.
One constraint is worth stating plainly. In most of our target markets, many buyers have no laboratory of their own. If that describes you, the internal-lab column above does not exist, and your choice is between the supplier’s QC and a third party.
When The Result Is Worth Paying For
A first commercial batch from a new source. Before you commit to a supply relationship, you want a measurement that is not the supplier’s own. This is the cleanest use of independent testing, because there is no prior data to contradict. It is also where our first article approval checklist asks for it as a matter of course — an early signal costs far less than a late one.
A specific quality question with a defined answer. “Is the residual solvent profile within the agreed limit?” is testable. “Is this material good?” is not. The difference is whether you can write down what result would change your decision. If you cannot, you are not ready to commission it.
A conflict that will otherwise stall. When your incoming result and the supplier’s certificate disagree, both sides can be technically correct and the argument can run for months without resolving. A laboratory both parties have accepted in writing breaks the deadlock, because its result is neither of your claims. Where such a conflict becomes a rejection, see handling a rejected intermediate batch.
Two smaller situations are worth recognising: a periodic re-verification — testing a fraction of batches from an approved supplier — and an escalated attribute needing a method your supplier does not routinely run, such as elemental impurities in intermediates.
Four Things To Agree Before The Sample Leaves
Here is where a lot of independent testing quietly fails. The sample goes out, the report comes back, and the number cannot be used to decide anything — because the basis of the measurement was never defined. Each item below takes one email to settle first.
The method and its version. A laboratory will follow a method you specify. If you do not specify one, they will use their own internal method, which may differ from the supplier’s in ways that produce a disagreement that has nothing to do with the material. This is the most common reason two results diverge. A published method validated for a similar matrix — see our article on analytical method validation under ICH Q2 — removes the variable. A revised method version can also shift the number for identical material.
The reporting basis. Assay can be reported as an area percentage against the main peak, weight-by-weight against a dried sample, or with a correction for water and residual solvent. These are different measurements that land in different places. If your supplier’s certificate reports one basis and the laboratory reports another, the gap you see is entirely arithmetic. Decide the basis in advance and write it into the request.
Which standard the laboratory uses. The reference standard is the calibration anchor for the whole measurement. If your supplier uses a working standard traced to a different reference standard than the laboratory does, a small systematic difference is expected. Establishing that the reference material is properly qualified — see reference standard qualification — lets you recognise a systematic offset rather than chase it as a quality problem.
Who pays, and what happens if the result disagrees. This point decides whether a dispute can be resolved at all. If the result confirms the material was non-conforming, the cost of the test and the material’s fate should follow that outcome. If it confirms the material was conforming, the cost of your own test should not disappear. Agreeing this in advance turns a negotiation under pressure into a procedure — see what to look for in a quality agreement.
Regulatory frameworks treat this as a written obligation rather than a commercial nicety. The FDA’s guidance on quality agreements for contract manufacturing describes how the parties define their testing responsibilities, and 21 CFR 211.84 sets out the requirement for component testing before release. EU GMP Chapter 7 covers the same ground for outsourced activities. All make clear that testing responsibilities belong in writing, in advance.
When Two Results Disagree, Work Through This Order
You have a certificate saying the material passed, and a laboratory saying it did not, or the reverse. The instinct is to escalate immediately. That is almost always the wrong first move, because a meaningful share of apparent conflicts are differences of method or of basis rather than of material — and escalation resolves none of those.
First: check the laboratory’s own system suitability. Every chromatographic or spectroscopic method has performance criteria the laboratory must meet before a result is meaningful — resolution between the peaks of interest, tailing factors, sensitivity. A result produced on a run that failed those checks should not be compared with anything. Ask for the system suitability data. This is a documentation request, and it costs nothing.
Second: compare the reporting basis. This is the most frequent cause of a gap and the easiest to check. A certificate reporting 98.0% area and a laboratory reporting 95.1% on a weight-by-weight dried basis describe the same material, not two different ones. Put both on the same basis before deciding anything.
Third: examine sample handling. A sample that was held, shipped across a border, or stored at a temperature the material does not tolerate can shift measurably. See our article on retest dates versus expiry dates — a sample past its retest date no longer describes the batch it came from.
Fourth: check the standard and the calibration. If both labs used qualified standards traced to the same source and correction factors, a real difference is more likely. If they did not, the difference may be a calibration artefact.
Fifth: test the batch itself. Take a fresh sample from the retained material and run it, ideally in a laboratory you have already qualified, alongside your incoming test as a parallel determination. A single discrepancy on one sample is a weak signal. Two fresh samples agreeing with each other and disagreeing with the certificate is a strong one.
Sixth, and only then: escalate. If the conflict survives all five, it is real. The useful move is then a joint session with the supplier and the laboratory together, batch documentation open in front of everyone. The question shifts from whether the material is good to whether the process that produced it is understood — which is where the pattern across batches beats any single result.
Writing The Arbitration Clause Before You Need It
Where a relationship has broken down and the material is contested, the way forward is an arbitration laboratory appointed by agreement, whose result binds both parties. Independent testing answers a scientific question; arbitration settles a commercial dispute. An independent result favouring your supplier does not oblige you to accept the material; an arbitration result does.
The mechanism is straightforward. Agree the arbitration laboratory, or the procedure for appointing one, and the method it will use, and who bears the cost depending on the outcome. Put it in the quality agreement rather than in an email during a dispute. A clause written in advance, when both parties are still working, is honoured more readily than one improvised later.
Where material has already been rejected on quality grounds, the sequence begins with evidence rather than a test result — see the questions to ask when a deviation report arrives.
What Independent Testing Is Actually For
Independent lab testing is one of the better instruments in a buyer’s toolkit, and it is used far less carefully than it deserves. Its value is not that it produces a number. Its value is that the number comes from someone with no reason to reach a particular conclusion — and that holds only if the question, method, basis and standard are fixed before the sample leaves your building.
Used properly, it answers a narrow question well: that a particular measurement, taken to an agreed method, produced a particular result. Everything else — whether the process is stable, whether the batch was representative, whether the material will behave as expected in your plant — comes from other evidence, and most of that is already in your supplier’s documentation. The certificate is the part everyone requests; the record behind the certificate is where the harder answers live.
