How we know what we know about our members.
Where Fairly’s audience data comes from, how it’s verified, what stays current, and what we never let advertisers see. If you’re evaluating us against a data broker or an inference model, start here.
1. Consent
Every Fairly member opts in to specific attribute categories — demographics, financial profile, lifestyle, life events, and so on. Consent is per-category and binary at the moment of the action: a category is either currently active or it isn’t.
The truth about a member’s consent lives in a per-member event ledger. Three event types matter — grant (initial opt-in), modify (changing the scope within a category), and revoke (full opt-out). The most recent event for any category is the one that’s true.
We don’t prompt members to re-consent on a schedule. A member who granted consent in January doesn’t lose it on a 30-, 60-, or 90-day clock. They lose it when they revoke it — which they can do at any time from one screen in their dashboard.
What happens when a member revokes
When a member revokes consent for a category, the change propagates within minutes. They cannot appear in any new survey delivery, sponsored offer, or research invitation that depends on that category — even from a segment you already saved. There is no copy of the audience in your hands to reconcile: because campaigns run inside Fairly, a revocation takes effect everywhere at once, and the member’s participation ends at the moment they revoke.
2. Sources of signal
A member’s profile in Fairly is assembled from three sources. Each one has its own collection method, its own verification mechanism, and its own set of queryable attributes. All four feed the same profile, but they’re labeled separately in the segment builder so you always know whether you’re filtering on self-reported data, purchases from real receipts, or AI estimates.
3. Freshness
Self-reported profile data — demographics, life events, preferences — is permanent until the user updates it. It doesn’t go stale on a schedule. Inferred behavioral attributes do: each one carries an expiry date.
4. Sensitive categories
A subset of categories sits behind a heightened consent requirement. Health conditions and religion are the most familiar examples; the broader set covers several other dimensions around identity, beliefs, and life circumstances.
For these categories, each one is turned on separately. The member must switch on that sensitive category by itself, and confirm it, before it becomes queryable. We don’t accept bundled consent, default-on toggles, or implied consent for any sensitive category.
What this means when you target one
Two things happen when a segment filters on a sensitive attribute:
The net effect: sensitive-category targeting costs more. You pay a premium for an audience that has explicitly chosen to be reachable on this dimension.
5. What buyers never see
A handful of things are never available to buyers — not on any terms, not for any premium. These aren’t features we’re holding back. They’re product constraints baked into the architecture.