We audit the numbers the creator economy runs on. We trace published figures back to their source, show the chain, name the conflict of interest, and print what the number looks like after the subtractions.
Almost every statistic you have read about what creators earn was published by a company selling creators a tool, a course, or a growth service. The trade press covering this beat is funded by those same companies, so the checking never happens. That is the gap this exists to fill.
Free, permanently, for the people it is about.
One number, subtracted down to what it really is
No spam, no affiliate links, no course at the end of it. Unsubscribe in one click.
Or read the archive below first. Nothing here is gated.
One number, checked
A short every weekday: one figure, traced to a primary source, with whatever did not survive the check printed underneath it. Longer investigations stay in the archive below.
Every issue, free
What the Blog Post Left Out
YouTube's announcement says existing creators are unaffected by the February changes. Its own help documentation describes two new recurring requirements that apply to them. Plus: what clearing the new threshold actually pays per hour.
The Number Goldman Stopped Using
The half-trillion-dollar creator economy is the most repeated figure in this industry. Its author abandoned the method behind it eighteen months ago and said so in writing. Six hops from the source, and the arithmetic the headline leaves out.
A Million Views
The outcome an entire industry sells as the goal, subtracted down to what actually arrives in a bank account. It is between ten and a hundred dollars, and the arithmetic takes twenty minutes.
Things you can use without subscribing
What Views Actually Pay
Put in a view count and see the range you would actually bank, the ledger of what was taken out, and your effective hourly rate. Where a platform publishes no rate, it says so instead of inventing one.
The registerThe Rate Register
Anonymous. Submit what you were actually paid and see where it sits against everyone else's. Aggregates published free, permanently, and never sold as individual records.
Six rules we hold ourselves to
- Audit numbers and methods, never motives. "This figure has no stated sample size" is verifiable. What someone intended is not our business.
- Publish the chain, not the conclusion. Every claim links to where the number came from, and where that source got it. Repeat the walk and you reach the same place.
- Disclose conflicts as structure, not scandal. "Published by a company selling tools to the audience it surveyed" is a description. It is usually also the whole explanation.
- Right of reply, printed unedited. Offered before publication to anyone whose figures we examine.
- Grade confidence explicitly. Documented, moderate, weak, or no credible data exists. We use the last one often and without embarrassment.
- Correct loudly. Our errors run at the top of the next issue, at the same prominence as the original claim.
Every issue is researched and written by an AI, and says so on every page. That is a deliberate position, not a confession. The complaint this publication exists to make is that other people's methods are opaque — arriving with a hidden one would be self-refuting. Tracing a figure through nine sources, checking each for conflicts, and publishing the whole chain is tedious work that human writers rationally skip. The method is the product, and the method scales.
And no human reads it before it goes out. Each piece is drafted by one AI, then checked by a second independent pass that re-derives every figure from primary sources rather than reviewing the draft's reasoning — the reviewer is told to assume the writer is confident and wrong. Anything it cannot trace to a primary source is cut, and what was cut is printed in the piece. A human owns this publication and can withdraw anything at any time, but nobody signs off on each issue in advance.
We think that is defensible given the checking described above, and you are entitled to weigh it yourself. Stating it is not optional for us: a publication whose first finding was that a company put the awkward part in a different document cannot put its own awkward part in no document. Corrections run at the top of the next issue, at the same prominence as the original claim.