Dealing with a bad batch

Stop the stock moving. Nothing goes out to customers, nothing gets mixed with good stock, and nothing gets repacked. Once a defective batch has been partly shipped, partly returned and partly restacked, the evidence is gone and so is the ability to say how many units were affected.

Count it properly. Draw a real sample across the whole batch — different cartons, different pallets — and count the defect rate rather than describing it. "About a third of them are bad" is a feeling; a counted rate from a documented sample is a fact, and it is the number every subsequent conversation will be about.

Photograph in a way that survives being forwarded. The fault itself, close up; the same unit next to a good one; the batch or date code; the carton and its label with the goods visible; and a wide shot showing the quantity. Photographs with no context are the commonest reason a supplier can honestly say "that is not from our batch".

Keep the packaging. Cartons, inner boxes, labels and inserts are evidence of what was shipped and when — and if the defect turns out to be transit damage rather than manufacture, the packaging is the only thing that can tell you.

And check the obvious alternative explanations before you accuse anybody: is it transit damage, is it the wrong version of a correct product, is it a specification that was ambiguous, is it a batch from a different production run mixed in? Being right about the cause is worth more than being fast about the complaint, and a buyer who has already checked those is credible in a way that one who has not is not.

Still at the factory, balance unpaid. This is the strongest position there is and it is the one an inspection before the balance is designed to create. Everything is available: sort at the supplier's cost, rework, replace, re-inspect, or refuse the shipment. The cost of putting it right is also at its lowest here, which matters — a supplier can afford to fix a batch that has not shipped much more easily than one that has.

Shipped, balance unpaid. You hold money and they hold goods you have not received. It is a real negotiating position and a temporary one, because the goods will arrive and the balance will fall due; the useful move is to reach an agreement in writing BEFORE arrival rather than to simply withhold and see what happens.

Landed and paid. Now you have the goods and no money, and the realistic outcomes narrow to a credit against a future order, a discount on the next one, replacement stock sent with the next shipment, or the supplier bearing some of the cost of sorting locally. Notice that every one of those depends on there BEING a next order, which is why the "never speak to them again" instinct is usually expensive.

Already sold to customers. This is a different problem and it stops being primarily commercial. Returns, refunds and reputation are yours to deal with immediately and regardless of who caused it, and if the defect touches safety or compliance there may be obligations on you as the party that placed the goods on the market. Deal with that first and pursue the supplier second; the order matters.

Underneath all four sits a question people forget to ask: where did risk pass? If the goods were damaged after the point at which risk transferred to you, the loss may be yours and the conversation is with an insurer rather than a factory. That is decided by the Incoterm and it is worth checking before the first message, not after it.

Money back, replacement goods, a discount, and a supplier who still wants your business are four different outcomes, and in practice they compete. The first message you send chooses one whether you intended it to or not, because it sets the register of everything that follows.

If the relationship is over, say what you want and evidence it. Numbers, photographs, the specification clause that was breached, and a specific ask. Do not threaten anything you are not going to do — a threat you do not follow through on is the cheapest possible way to lose the leverage you had.

If the relationship is worth keeping — and with a factory that has performed before, it usually is — then the most effective opening is a factual one that gives them a way to fix it rather than a verdict. Suppliers who are told "here is the counted defect rate, here are the photographs, what do you propose?" respond far more usefully than suppliers who are told they are dishonest, and you can always escalate later. You cannot de-escalate later.

Ask for a cause, not just a remedy. A supplier who can say what went wrong — a material lot, a new operator, a tool that needed maintenance, a substituted component — is a supplier who can prevent it. One who offers only a discount and no explanation is telling you it will happen again.

And write down whatever is agreed, however friendly the call was. What is being done, by when, at whose cost, and what happens if it does not work. A remedy agreed in a voice call and not recorded is a remedy that quietly becomes a misunderstanding.

Find the gate it passed through. Every bad batch that reaches a buyer got past something: the specification was ambiguous, the defect was not on the defect list, no inspection was booked, the inspection was booked too late, or the sample it was compared against was not the production reference. "The factory got worse" is almost never the whole answer, and it is the answer that prevents nothing.

Fix that one gate, not all four. The instinct after a bad batch is to add every check at once, which is expensive and does not survive contact with the next busy month. One change that closes the specific hole — a defect added to the list with a photograph, an inspection moved earlier, a first-article check added — is worth more than a policy nobody keeps.

And be honest about the commercial reality: most bad batches are settled commercially rather than legally. Cross-border litigation over a single consignment is slow, expensive and rarely proportionate, which is precisely why the leverage that works is future orders and an unpaid balance rather than a contract clause. That is not a reason to have no contract — a clear specification and a written consequence are what make the commercial conversation short — but it is a reason to keep the commercial route open.

No typical settlement figure is published here. There is no honest one, and a percentage printed in a guide would become an anchor in a negotiation it knows nothing about — including yours.

Finally, keep the file. The counted rate, the photographs, the correspondence and whatever was agreed. It is what makes the second conversation with the same supplier shorter, and it is the only thing that makes a pattern visible if there is one.