A one-year warranty is a commercial commitment with a predictable cost, and on a first OEM order that cost is almost never modelled. It shows up later as a percentage of units that come back, the freight to move them, the time spent handling them, and the credit notes issued to keep a channel customer happy. The suppliers who handle this well are not the ones promising the longest warranty; they are the ones who can tell you the expected failure rate and what happens to a failed unit.
Where the money actually goes
Four cost centres, in rough order of size:
- Replacement units. Whether they come out of the next order, out of a spare pool, or out of a separate shipment.
- Logistics. Cross-border return freight on a low-value item frequently costs more than the item. This is the part that surprises people.
- Handling. Someone has to receive, log, test and dispose.
- Channel credits. Retail and marketplace customers often want their own customer made whole immediately, before any analysis happens.
Because return freight dominates, the standard commercial answer for low-value goods is not "return and repair". It is: verify the failure remotely, ship a replacement from a local or next-order pool, and either scrap the failed unit in the destination country or ask for a photo and a destructive disposal. A policy that requires physical return of every failed unit is usually a policy that will not be followed.
The spare parts convention
The usual arrangement is a spare allowance shipped with or alongside the order, commonly around 0.5% to 1% of quantity, covering the field-replaceable parts and sometimes complete units. The percentage is negotiable and it should be negotiated: it depends on the product, the channel, and how replaceable the failure modes are.
Two details that matter more than the percentage:
- Which parts. A spare allowance of the wrong parts is worthless. Agree the list — typically the parts that actually fail or get damaged in transit, not the cheapest components.
- When it ships. With the order, or held at the factory for call-off? Shipped spares consume duty and freight now; held spares are slow when you need them.
DOA and the first thirty days
Dead-on-arrival is treated separately from warranty failure, and it should be, because it points at a different cause — packaging, transit, or a production escape rather than field life.
A workable DOA clause: units failed within a defined short window (commonly 30 days from receipt) are replaced on the next shipment or credited, with evidence defined in advance — a photo, a serial number, or a short test result. Defining the evidence matters. Without it, every DOA claim becomes a negotiation.
What counts as a defect on a lithium product
This is where warranty arguments start, so it belongs in the agreement before the order:
- Capacity fade is normal. A cell losing capacity over cycles is physics, not a defect. A common, defensible line is capacity retention: a battery that has dropped below roughly 80% of rated capacity within the warranty period is a failure; one at 92% is a used battery behaving normally.
- Physical abuse is not a defect. Impact damage, water ingress beyond the stated rating, and use outside the specified temperature range are exclusions, and they should be written as exclusions.
- Cell swelling is always a defect. It is a safety issue and it is not a judgement call.
Write the capacity threshold into the specification. Without a number, "battery does not last as long as it used to" is a dispute you cannot settle.
Failure data is worth more than the parts
The most valuable output of a warranty programme is not the replacement units. It is knowing which failures happen and how often, so the next order can be different. That requires failed units or failure descriptions to come back to engineering in a structured way.
The minimum viable version: a monthly log with serial number, symptom, date of sale and date of failure. From that you can compute a failure rate by production batch, which is the number that tells you whether a problem is a design issue, a component lot issue, or normal field attrition. Without it, every warranty conversation is anecdotal.
Three questions to settle before the deposit
- What is the spare parts percentage, which parts does it cover, and when do they ship?
- What is the DOA window, what evidence is required, and who pays the freight?
- Is the failure log part of the agreement, in what format, and how often?
A supplier who answers all three concretely has run warranty programmes before. That is worth more than an extra six months on the warranty card.
Related reading: manufacturing capabilities, MOQ and pricing, about ODMKit and the ODMKit OEM/ODM programme.
Top comments (0)