The Wheel Was Covered in Year Two and Billable in Year Three

One unit, one wheel. In year two it fails and the replacement costs nothing. In year three the same wheel fails and a line appears on a quotation. The product has not changed and the warranty has not expired. What changed is the category the wheel was filed under. The periods here are illustrations.
Why does the same wheel become billable in year three?
Because it was handled as a failed component for two years and as a wear item in the third. The line between those two words is not drawn in the purchase contract. It is drawn in the product structure, and nobody signs that drawing at the point of sale, so both sides usually discover the classification the first time a claim is raised.
A wheel is an assembly: tyre, hub, bearing, axle. The tyre wears against the floor, so its life follows distance covered and the surface travelled. The bearing's life follows load and contamination, and a single wet season of grit does more damage than a year of clean indoor use. The hub and the axle are not supposed to fail at all. One failed wheel can therefore be three entirely different events.
The workable approach is to cut the warranty by failure mode rather than by part name. Names drift with suppliers and with language; the same part is a caster in one quotation and a wheel assembly in the next. Failure modes do not drift, because they follow physics rather than vocabulary.
Following physics has a second benefit: the criterion can be written as a sentence both parties can check, instead of a sentence that needs a third party to interpret it after the goods have already been rejected.
Written that way, it reads like this: the tyre is excluded; the bearing is covered in year one and judged after that by axial play measured at the hub; the frame is covered for the full term. Three sentences carry three different treatments, and an inspector can apply them without asking who drafted the clause.
How should the scope be cut so it can be checked?
The table below is the only one in this article: part category on the left, the criterion that places a failure in that category in the middle, and the treatment worth writing into the clause on the right.
| Part category | Criterion | Treatment to write into the clause |
|---|---|---|
| Structural (frame, welds, hinges) | Should not fail within design life; failure shows as cracking or permanent deformation | Covered for the full term, traced by serial number |
| Moving parts (bearings, hubs, steering) | Wear is an outcome of use; failure follows load and contamination | Fully covered in year one, then judged on measurable wear |
| Consumables (tyres, pads, tips, grips) | Designed to be replaced; life shorter than the unit | Explicitly excluded, or given a separate short term |
| Electrical (battery, controller, motor) | Failure tracks charge cycles and can be read out | By years or by cycles, whichever comes first |
| Surface finish (paint, plating) | Failure is mostly adhesion or corrosion, and follows environment | State the start point and the exclusions |
| Fasteners and adjustment (bolts, cables) | These describe assembly state and a maintenance action | State who adjusts, and how often fasteners are re-torqued |
Every line on the right answers three questions: what is covered, on what basis it is judged, and what happens once it has been judged. Remove the middle column and the other two collapse in any dispute, because the factory and the customer rarely mean the same thing by the wheel failed. One calls it play in the bearing; the other calls it an adjustment.
The opposite error is just as common: a clause that lists every part in the bill of materials and gives no criterion for any of them. It looks thorough, and it leaves every judgement to be settled after the fact, by whoever happens to be holding the goods when the argument starts.
Which day does the term start?
This is the item most often skipped and the one with the largest consequence. Three start points are common in practice: date of manufacture, date on the bill of lading, and the date the end user first commissions the unit.
Date of manufacture is easiest for the seller and worst for the buyer. Goods sit in a warehouse for six months, a quarter of the term is gone, and the machine has never carried anyone. The bill of lading is a document both sides hold and can verify, but it says nothing about use either, because a distributor may hold stock for a season before the first unit moves.
First-use registration comes closest to reality, provided a mechanism exists: serial number, commissioning date, user or facility. Most B2B transactions have no such mechanism, so the clause can be written and still cannot be executed when the claim arrives. A clause that cannot be executed is worse than no clause, because it invites a dispute that it cannot settle.
What can be executed is a start event paired with a form of evidence. Write one without the other and the expiry date becomes two dates calculated by two parties. Write the evidence without the event and the evidence quietly ends up defining the time, which is how a warehouse receipt becomes a warranty document.
A form worth copying: the term runs from the bill of lading date, evidenced by the bill itself; where both parties agree to run it from commissioning, the commissioning record governs, and that record must carry the serial number, the date and the facility name. That sentence gives an event, its evidence and a fallback order.
One distinction belongs beside it: the warranty end date and the claim window are two different clocks. The first says when cover stops; the second says how long after a failure the buyer may still raise it. Merging them into one date is a common way for a valid claim to become a late one.
Whole units or serial numbers?
Batch-level handling is fast. One claim covers a shipment, nobody checks units one by one, and the paperwork stays short. The price is that a single unit's history becomes invisible: a machine that comes back three times leaves its record scattered across three batch settlements, and on the third failure nobody can see it is the same machine.
Serial-level tracing needs three things at once. A durable carrier for the number in the design. A numbering rule that resolves back to the production batch and the material lots. A record that stays searchable for the whole term. The third is the one usually missing, because it needs a system that survives staff turnover rather than a drawing that survives a file transfer.
One engineering detail decides the whole question: label durability. A paper label on the underside of a frame, after two years of sun, cleaning agents and washing, is routinely unreadable. An unreadable serial is no serial, and traceability ends at that point. That is a materials choice made at design, not a management failure discovered later.
Why does spare parts supply matter more than the warranty term?
The warranty term decides who pays when something breaks. The supply period decides whether it can be repaired at all. The second matters more to buyers, especially institutional ones — care homes, rehabilitation centres, rental fleets — whose asset is a fleet rather than a single machine, and whose downtime cost sits entirely outside the warranty.
Four things belong in the clause: the number of years of supply, the channel (original manufacturer or authorised), the fallback after discontinuation, and ownership of tooling and dedicated gauges. The fourth is most often missed. Plastic parts usually come from a dedicated mould, and who holds that mould decides whether anyone can still make the part after the line stops.
The basis of the supply period also needs stating: counted from discontinuation, or from the last delivery. The difference is a few words on the page and a whole product generation on a ten-year asset book.
The channel matters as much as the years. Supply from the original manufacturer keeps the part identical; supply through an authorised channel keeps it traceable; a local equivalent keeps the fleet running but changes the part. All three are legitimate, and the clause should say which one the buyer is promised.
FAQ
Is there a legally fixed warranty period?
This article gives no legal conclusion. Scope, term and start point are best agreed in the contract and confirmed in writing by both parties; where the place of sale imposes mandatory rules, local legal or compliance staff must check them.
The customer asks for a lifetime warranty. Can we accept?
Break the phrase apart first. Structural parts can be discussed over a long term; consumables cannot. Accepting without splitting it means promising something physically impossible, and the promise will be tested within a few years.
Battery: by years or by cycles?
Whichever comes first is the safer wording. It only works if cycles can be read out; if the controller keeps no cycle count, the clause cannot be executed either.
What happens when a spare is discontinued?
State the supply period counted from the last delivery, and agree how a substitute part is confirmed. Without that line the supply period is a wish, and the buyer discovers it at the worst possible moment.
Every period in this article is an illustration rather than a value taken from a standard or a regulation; the terms themselves are best agreed in the contract.