Each page on the Graveyard Atlas makes a specific claim: this idea doesn't survive scrutiny, for this reason. Here is exactly how those claims are checked before publishing.
Every autopsy is classified into one of eight blocking mechanisms — the actual structural reason an idea dies, not a vague "too competitive." Free floor, gold-rush graveyard, ToS/legal, credential gate, labor-in-disguise, distribution-unreachable, shelf-occupied, and no-demand. Naming the mechanism is what makes a page useful: it tells you whether the door could ever open, and what would have to happen for it to.
“Market-tested” means a real shipped product or experiment measured the failure directly (a live user count, a dead auction, a documented shutdown). “Predicted” means the verdict is an evidence-based forecast from the receipts below, not a direct measurement. A market-tested verdict is the strongest evidence we can publish — it isn't our opinion, it's what actually happened when someone tried.
Every factual claim on an autopsy page — a price, a user count, an entrant count, a shutdown date — is backed by a public source URL listed on that page. If a fact couldn't be independently verified, it isn't stated as fact. Where evidence was weak or a claim didn't survive re-checking, the page says so plainly rather than rounding up to a cleaner story.
We don't publish a verdict to be clever or to mock an idea — the reader is someone about to spend months of their life. Every autopsy ends with what would honestly have to change for the verdict to flip, even when that's a narrow or unlikely condition. And a "no" here is never permanent: markets change, incumbents die, and terms of service get rewritten — a page's verdict reflects the evidence at the time it was checked, not a law of nature.