A separate post covers why the same part gets sold twice. This one is about what you can actually do, starting today, on whatever you are running now.
There are three options. Only one of them ends the problem, and it is worth being clear about what the other two cost before you pick.
One: list on one channel
This works completely. If a part exists in exactly one place, it cannot be bought twice. It is also the option nobody wants, because the reason you are on a second channel at all - a storefront, Facebook Marketplace, the counter - is that it does not reach the same buyer as the first.
Worth naming anyway, because it is the baseline every other option is measured against. If a second channel is not paying for itself, the safest inventory system is not having one.
Two: sync faster
The usual answer. Push the stock count between channels every fifteen minutes, then every five, then every minute. Each step makes the overlap window smaller.
It never makes it zero. A five-minute sync means five minutes, all day, in which both channels believe they have the part. It also has a nastier property: the rarer the failure, the less likely you are to have a process for it. A yard that oversells weekly has a refund routine. A yard that oversells twice a year handles it from scratch, badly, on a busy afternoon.
Three: one record owns the count
No channel keeps its own number. Every channel reads from one record, and when a sale lands the other listings are ended as a consequence of that sale rather than at the next sync. That is the cascade, and it is the only version where the window is genuinely zero.
It is not magic and it is not free. It requires every channel to be integrated rather than exported to, and it means an integration that fails has to fail loudly instead of quietly falling behind.
It also only covers the channels that are actually built. Today that is eBay Motors, a Shopify storefront and your counter. Not Facebook Marketplace, not Car-Part.com, not Google Shopping. If Facebook is your second channel, option three does not cover you yet and the habits below are what you have - the full list of what it does not do is on its own page.
What to do today, with no tool
Most yards reading this are on option two by default. Until that changes, three habits do most of the work:
- End the other listing before you pack the box, not after. The gap between "sold" and "ended" is the entire risk, and packing takes long enough for a second buyer to arrive.
- Keep one sheet that owns the count, and treat every channel as a copy of it. Not two sheets. Not a sheet and your memory.
- Put your slowest-moving, highest-value parts on one channel only. The exposure is not evenly spread - it concentrates in the parts that sit long enough to be found by two people at once.
None of this is satisfying. All of it is labour, and labour fails on exactly the day you are busiest, which is also the day two buyers are most likely to collide.
Why used parts make the stakes different
A new-parts seller who oversells orders another unit. You cannot. The part came off one car, the count is one, and there is no second alternator from that Transit. The apology is the whole remedy, and the buyer knows it.
The refund is also not the real cost. On eBay a cancellation for stock you do not have is a transaction defect, and the threshold is tighter than most sellers expect - two percent, measured over a window that can be twelve months long.
If you want to see what your own rate of double-sales is costing, the calculator does the arithmetic.