How we source
Every fact on a record has to come from somewhere. We think that somewhere should always be visible to you, and to the machines reading it, not just accurate behind the scenes.
Where facts come from
Most of what’s on a record comes from three places, and every fact carries one of three labels: Sourced, Stated or Verified.
- A Creator’s own public presence. If it’s something the Creator said about themselves, like a bio line or a caption, it’s labelled Stated: their word, not yet independently confirmed. If it’s something the platform itself displays on their profile, like audience or media stats, it’s labelled Sourced.
- Public, licensed knowledge sources, like Wikipedia and other structured public knowledge bases, where a Creator is already documented independently. These are labelled Sourced.
- Verified platform data. Once a Creator claims their record and confirms their identity through the platform itself, we switch from an estimate to the platform’s own authoritative figures directly, and the label changes to Verified. A small number of facts, like Notable work, only ever appear this way: they come from the platform’s own data once a Creator has claimed their record, and a Creator can’t add them just by stating them.
- Curation for machine-readability. A small number of things, like a Creator’s content categories, come from us reviewing their public content and organising it so people and AI can find them by what they actually make. This is labelled Sourced too, since it’s built from public material, not invented. We don’t use a separate label for it, whether we’ve reviewed one post or twelve. Once a Creator claims their record, they can choose or correct their own categories directly. At that point it’s their own confirmed choice, not our analysis.
A note on third-party sources
We’re careful with third-party sources. When a public source is really just relaying something the Creator has said or done themselves, we’re comfortable treating it as reliable. But when it’s someone else’s independent reporting, commentary or estimate about a Creator, we treat it more cautiously, and it might not make it onto every record as a result.
What we never collect
There are things we simply won’t collect, no matter where we find them.
- Private information a Creator hasn’t chosen to make public. No home addresses. No private contact details.
- Sensitive personal categories: things like health, religion, or political and sexual identity. We’re careful not to collect or publish these ourselves, and we actively screen for them when we bring in new sources. The one exception is a Creator’s own words. If a Creator has already made something public themselves, in their own bio or their own post, for example that they’re a cancer survivor or a Pride supporter, we can mirror what they’ve already said. We don’t go looking for it, and we don’t take it from anyone else’s account of them.
Keeping things current
Unclaimed figures on a record carry an “as of” date. That’s because we’re bound by the same rules the platforms set for anyone using their public data. We can’t publish fresh, unauthorised statistics more often than their terms allow.
Once a record is claimed, that limit disappears. From that point, the figures come straight from the platform’s own verified data.
Adding a record
Most records get built the way this page describes: automatically, from what’s already public. If someone you know is missing, though, anyone can suggest a handle for a new record. We build it the same way, from public sources, and the Creator can claim it, correct it or remove it just like anyone else.
You might also want to read About this record and our Editorial policy.