We expected competitors undercutting each other. In 227 AliExpress price gaps, sometimes both listings are the same shop.

I run a small tool that records AliExpress prices once a day and checks whether the discount badges hold up. This week I went looking for a different thing: the same product listed by two shops at very different prices. Everyone on Reddit has a theory about why that happens. Nobody has a list.

As of 8 September 2026 I am holding 227 product titles carried by two shops where one charges at least double, both sides with real recorded sales. Total spread across all 227: $2,002.19.

What I counted, and what I did not. A pair enters the list when two shops carry the same product title, the cheaper listing is at least $1, and each side has at least ten recorded sales. A shared title is a hint, not proof the two listings are identical — pack size, variant, size and colour can all differ, and neither price includes shipping. So I split the claim: of the 227 pairs, 52 use the identical image file. The other 175 share a title and nothing more, and I only label the rows that earned it.

The part I did not expect. In several pairs the two listings belong to the same shop. One A5 leather notebook: same shop, same title, same photograph, 2,760 recorded sales on each listing, $3.46 on one and $10.35 on the other.

I am not going to explain why, because I do not know. A shop testing price points, a leftover listing from an old pack size, a badly titled variant — all plausible. What I can say is the pattern is measurable and "different sellers charge different prices" does not describe it.

The build lesson, since this is IH. My product pages get almost no search traffic — the average one earned 0.06 impressions in 28 days. One informational page earned 271 in the same window. So I stopped trying to rank 65,000 product pages and started publishing the things only my own database can say, with the method and its limits stated plainly. Saying "not enough data, no verdict" turns out to read as more trustworthy than a confident number.

Happy to answer anything about the data or the method. Full 227 pairs with shop names and sales counts are on the site in my profile.