The trade's scale is manufactured, and the manufacturing is visible from outside. Four patterns account for most of it: a single catalogue sold under several factory domains; pages injected into compromised but authoritative websites; a storefront that shows one item while the checkout charges another; and reviews, ratings and prices with nothing behind them. Each of these can be measured by a buyer without an account and without trusting anyone, and each is described below with the numbers we found.
Do "different factory" websites sell different watches?
The most common way this trade looks bigger than it is: one catalogue sold under several "factory" domains. We measured two such domains in October 2026, each presenting itself as a different factory, and 278 of 300 product names were byte-identical, with 246 of those 278 also carrying an identical price. The two sites ran the same theme, the same page builder and the same payment plugin. Nothing in the product titles named a factory at all: the factory identity lived in the domain and the branding, not in the catalogue. A buyer comparing two "factories" on price is often comparing two shopfronts over one inventory.
Why do watch pages appear on government and charity websites?
Second pattern: pages hosted on sites that have nothing to do with watches. When we checked the front page of a Japanese search for replica terms in September 2026, the results included pages sitting on an art magazine, a national UNESCO commission domain, a government domain and several institutional sites, none of which sell watches. The mechanism is a compromised site: an attacker adds pages in a deep directory, the host domain's existing authority carries them into the results, and the injected copy matches the search phrase word for word. The pages disappear when the host notices. A result that appears authoritative but has nothing to do with watches is the clearest signal there is that the site itself was never checked by anyone.
Can the product page and the checkout show different items?
Third pattern: what you see is not what the checkout sells. We documented a shop whose storefront products and checkout products were different items — a public mapping file paired a "vitrine" product ID with a separate "checkout" variant ID, so the page that gets shared and screenshotted is not the item being charged. Buyers then report that the checkout shows a generic or unrelated name. That structure exists to survive review, and it is detectable: the mapping file sits on a public endpoint, and the checkout domain does not match the shop domain.
How common are fake reviews and fake prices?
Fourth pattern: reviews and prices that do not exist. Template text left in review widgets (one catalogue we checked carried three reviews for a different product category entirely), a "4.5 market rating" displayed with no rating data behind it, and strike-through "original" prices that were never charged. A related version of the same trick is availability theatre: in one 348-product catalogue we sampled, 24% of listings showed as out of stock, which produces urgency without any stock movement.
How can a buyer tell these apart?
None of this is hard to check from the outside, and none of it requires trusting the seller. Compare two "different" factory sites by product name and price before believing they are different businesses. Look at whether the product titles even mention a factory. Check whether the host domain has anything to do with watches. Compare the shop domain with the domain the checkout actually charges from. And where a rating or a review is displayed, look for the underlying data: if there is none, the number was decoration.
What these patterns have in common
Every one of them exists to survive a review rather than to serve a buyer, review by a search engine, by a payment processor, or by the buyer at the moment of paying. That is also why they leave traces: a structure built to fool an automated check usually looks arbitrary to a human, and a catalogue copied between domains keeps its fingerprints.
What we are not claiming
We report what pages show and what our measurements returned. We make no claim about who owns any domain, and we are not a party to any dispute between a buyer and a seller. No domain is named anywhere on this site; the counts and dates are published so the measurement can be repeated, not so that anyone can be pointed at.
Reviewed 2026-10-01. Sources on this page: our own measurements, dated — catalogue comparison of two factory domains (October 2026, 300 listings each), a search-results check on the Japanese market (September 2026), a shop whose storefront and checkout products were paired by a public mapping file, and product-level sampling of a 348-listing catalogue. This site sells nothing and does not link to shops.