RN
A company that sells cam search published a study finding that cam search is bigger and more interesting than you thought. That is not a reason to throw the numbers out, but it is the frame you read them in, and it is missing from most coverage of them.
The dataset is genuinely unusual. Lemoncams indexes profiles across the major platforms rather than operating one, so it can see across a market where every operator sees only its own share. The supply-side figures — roughly 1.6 million profiles, with the United States at 26.8% and Colombia at 25.3% — appear in third-party coverage as well, so they are really published rather than invented for a press cycle.
What follows is not a debunking. It is a sorting exercise: which of these findings are load-bearing, and which are measurement artifacts wearing the costume of a discovery.
Two different instruments, one voice
The reports combine two things that are not the same kind of evidence, and the distinction gets lost the moment a figure is quoted.
The supply-side numbers come from a database of indexed profiles. That is a census of listings, and it is about as good as this kind of data gets — though what it records is the location field on a profile, which models set themselves and which platforms do not verify. Location is a marketing decision as much as a fact.
The demand-side numbers come from Google Analytics on the aggregator's own website. That measures people who visited lemoncams.com. It does not measure people who watch cams. Anyone who goes directly to a platform — which is most of the market — is invisible to it. So every sentence beginning "viewers of X are…" is really a sentence about the subset of viewers who arrive via one search intermediary, and that subset is selected in ways nobody has characterised.
Both instruments are legitimate. Reporting their outputs in the same breath, in the same units, is where the trouble starts.
Virginia is not a finding
The most quotable result in the US report is that Virginia leads the country in per-capita traffic, at roughly two and a half times the national rate, while California sits below average.
There is a well-known explanation for a small East Coast state topping a per-capita internet metric, and the report needs to rule it out before the number means anything. Loudoun County, Virginia is the densest concentration of data centres on the planet; estimates of how much global internet traffic moves through that corridor range from about a third to around seventy per cent depending on who is counting, and AWS's oldest and largest region sits in the same few square miles. IP-based geolocation resolves VPN exit nodes, cloud egress and proxied requests to where the hardware is, not to where the person is.
For a category of browsing where VPN use is unusually common, that is not a footnote. Northern Virginia appearing at the top of a per-capita chart is the expected signature of the measurement method. It might also be a real behavioural finding — but you cannot tell which from the outside, and the burden is on the publisher to show they excluded known data-centre ranges.
The same mechanism weakens the country-level rankings generally. Every VPN user is counted in the wrong bucket, and the direction of that error changes depending on where people route through.
The Iran number needs a denominator
The finding that gets the most attention — a country where same-sex relationships are criminalised placing third for share of its own traffic going to trans cam pages — is built on a ratio, and the instability lives in the bottom half of it.
Share-of-own-traffic is volatile wherever the total is small. A country sending a modest number of visits can swing several places on that metric from month to month without anything changing about the people in it. And in a jurisdiction with heavy access restriction, whoever reaches the site without a VPN is a peculiar, self-selected group; everyone using one has been counted somewhere else entirely.
The general observation the number is used to support — that official restriction and private behaviour diverge — is well established and does not need this statistic. The specific ranking should not be repeated as fact.
What "29% of viewers are women" actually reports
Age and gender in Google Analytics 4 are not observations. They require Google Signals to be switched on, they cover only visitors signed into a Google account with ad personalisation enabled, they are modelled and aggregated rather than counted, and they are aggressively suppressed by privacy thresholds when samples get thin.
So a gender split from GA4 describes Google's inferences about the identifiable minority of a website's visitors. On a site where the audience has an obvious motive to browse logged out or in a private window, the gap between that minority and everyone else is likely to be large and non-random.
The claim may still be directionally right — the assumption that this audience is uniformly male has been wrong for years. But the figure is not a measurement of who watches trans cams, and the decimal point gives it a precision it has not earned.
What survives
Quite a lot, and it is the structural material rather than the surprises:
- Latin America is a major supply centre, not a niche one. This shows up across independent slices of the data at magnitudes too large for sampling noise to produce.
- Platforms have distinct national characters. A roster that is overwhelmingly Brazilian, or overwhelmingly American, is a fact about how a platform was built and who it recruits. That is durable and it is genuinely useful when picking where to browse.
- Volume and attention live on different devices. Mobile carries the sessions; the longer sessions happen elsewhere. The shape of that split is consistent enough to trust even if the exact minutes are not.
- Concentration and raw count answer different questions. A large platform hosts many models of every kind because it is large. Whether a category is central to a platform's identity is a proportion, not a total — that distinction is the most useful analytical idea in the whole report.
The story the supply figure is standing on top of
One quarter of the world's indexed cam models listing a single country is presented in this genre as a fun fact about geography. Reporting on the same phenomenon from outside the industry treats it as a labour story — an investigation published by LatinAmerican Post examines Colombia's webcam sector in terms of working conditions, studio arrangements and the economics that drive people into it, which is a different article about the identical number.
Nothing here adjudicates that reporting; it is one outlet's account and it was not independently checked for this piece. But it is worth registering that "Colombia supplies a quarter of the market" is a sentence about several hundred thousand working people, and that a dataset built by counting profiles has no way to say anything about their conditions — which means its silence on the subject is a property of the instrument, not evidence of anything.
If you want to browse the market the report describes, the live cam index and the trans cam listings are where the platforms named in it are catalogued. Read the study for the shape of the industry. Do not read it for the decimal places.