You sell an ECU. Three hundred dollars, tested, pulled from a low-mileage donor.
Nineteen days later it comes back. "Not as described." The casing has a fresh scuff. The connector pins look like something was on and off them more than once. It still works, probably, and you now cannot honestly describe it the way you described it before.
What happened is not a return. A mechanic had an intermittent fault, needed a known-good module to diagnose against, and used yours to do it.
Why electronics specifically?
Because a module is the only diagnostic tool that is also the repair. If you are chasing an electrical fault, the fastest way to find it is to swap in a part you know is good and see whether the symptom moves.
A workshop without a spare on the shelf has two options: buy a diagnostic-grade tool, or buy the part with a return window. The second is cheaper and the window makes it free. That is not a loophole somebody is exploiting maliciously - it is the rational move given how returns work, which is precisely why it keeps happening. You do not fix rational behaviour by being upset about it.
Two situations that look identical and are not
| What happened | Your description was | Whose problem |
|---|---|---|
| Genuinely not as described | Wrong, or overstated | Yours. Take it. |
| Used and returned | Accurate | Yours anyway, financially. |
The distinction that matters to you - whether your description was true - is not a field either of you fills in. Which means the whole game is evidence, gathered before dispatch, because there is no mechanism afterwards.
What record actually helps?
- Casting numbers, part numbers, any stamped identifier, in focus. This is the difference between a unit and YOUR unit. If what comes back has a different casting number, you have a fact rather than a suspicion.
- Connector pins and mating surfaces, close. Pin condition is where "installed and removed" shows.
- The condition notes exactly as published, not your recollection. A dispute eleven weeks later is argued over the listing as it stood when somebody bought it.
- Weight and packed dimensions. Boring, and occasionally decisive: a returned box materially lighter than the one you sent is a question worth asking.
What to do when it happens
- Open the box on camera. One continuous take, unopened box in frame first. Ninety seconds, and it is the only version of events that exists independently of both parties.
- Compare identifiers first, condition second. A different part is a different argument entirely - that is not a return and should not be treated as one.
- If it is your unit and it has been used, decide fast. Fighting a remorse-coded return you will probably lose costs you the shipping and risks a case closed without resolution, which is the more expensive outcome.
That last point is the uncomfortable one, so plainly: sometimes accepting a return you consider unjust is the cheaper decision. Not because it is fair, but because the defect maths punishes the fight more than the loss.
The arithmetic that makes this worse than it looks
2
Cases closed without seller resolution is all it takes to go Below Standard. Not two percent - two cases.
So the borrowed-ECU problem is not a $300 problem. It is a $300 problem that also spends a sixth of your annual margin for error, on one transaction, on behalf of somebody who never intended to keep the part. eBay's seller standards policy has the current thresholds.
What actually reduces it
Honestly, not much, and I would rather say so than sell you a trick.
- Sell electronics on one channel, not several. Fewer simultaneous exposures on your highest-risk units.
- Describe testing precisely. "Tested" invites a broad reading. "Powers up, communicates on CAN, not tested under load" does not.
- Keep the evidence. Every point above comes back to this.
There is no mechanism after dispatch. There is only the record you kept before it.