The states every fact on this site carries, and how they turn into — or block — a number. This cluster is the site's own vocabulary about itself.
Verified
Also known as: checked, sourced
A value we hold, confirmed against a named source on a named date, and still inside its re-check cadence — 90 days for fees and rates, 180 for documents, a year for features.
- Why it matters
- It is the only state in which a number on this site carries our full weight. Everything else on the page is calibrated against how many of its facts reached it.
- How we treat it
- Verified facts are what a score is built from, and each one shows its source and check date on the profile.
- Do not confuse with
- Verified means we checked the source, not that we opened an account to test it — that is a separate, higher source tier.Unverified
Unverified
Also known as: imported, unchecked
A value we hold, from an import or a provider's own material, that no editor has yet confirmed against a source. The value exists; the check does not.
- Why it matters
- It tells you how much of a profile is our claim versus the provider's. A page of unverified facts is a starting point, not an answer.
- How we treat it
- Unverified facts are labelled as such, may be displayed and filtered on, and cannot lift an entity to score eligibility on their own.
- Do not confuse with
- Unverified means unchecked, not missing — a missing value is a different state.Unknown (not disclosed)
Stale
Also known as: overdue for re-check, expired verification
A value that was verified against a source but has passed its re-check cadence. It is not known to be wrong; it is known to be old.
- Why it matters
- Fees and rates change quietly. A rate verified fourteen months ago is a historical record, and presenting it as current would be the most ordinary way to mislead.
- How we treat it
- Stale facts are shown with their age, feed the re-verification queue, and hold an entity back from the top of the evidence ladder.
- Do not confuse with
- Stale is not unverified — a check happened, it simply expired.Unverified
Unknown (not disclosed)
Also known as: not disclosed, no value on file
We hold no value for the field: either the provider does not publish it, or we have not collected it yet. A first-class state with its own reason, not an empty cell.
- Why it matters
- Missing data is how comparison sites flatter empty entries. If a blank counted as "no fee", the least transparent provider would win every table.
- How we treat it
- Three rules, enforced in one place: an unknown can never improve a score, can never satisfy a filter, and always renders as "Not disclosed" rather than blank.
- Do not confuse with
- Unknown means we have no value at all; unverified means we have one that is unchecked.Unverified
Conflicting
Also known as: sources disagree, disputed value
Two or more credible sources give different values for the same field, and no editor has resolved which is right.
- Why it matters
- Picking the nicer number silently is the most damaging thing a comparison site can do. Publishing the disagreement is the honest alternative.
- How we treat it
- An open conflict blocks the editorial score for that entity entirely — the profile shows "score on hold" and lists what disagrees — until an editor resolves it against the more authoritative source.
Source tier
Also known as: source authority, evidence level
The authority ranking of where a fact came from: official sources first (provider terms, fee schedules, regulators), then structured data feeds, then our own account-opening tests, then corroborated community reports.
- Why it matters
- Not all evidence is equal. When two sources disagree, the tier is what decides the outcome instead of whoever was read most recently.
- How we treat it
- Every fact stores its tier; score eligibility requires at least one official source; community reports are used only with a citation and corroboration.
- Do not confuse with
- A source tier rates the evidence, not the service.User score
Editorial score
Also known as: our score, the 0–100 number
A weighted mean of six rated dimensions — foreigner onboarding, reliability, fee transparency, support and language, payout speed, app and UX — computed from verified facts with published weights, then capped by how much of the model is actually rated.
- Why it matters
- It is a stated model rather than an opinion: the weights are published, the inputs are sourced, and both can be argued with.
- How we treat it
- It is never sold, never adjusted for a partnership, and never published for a service that has not reached score eligibility.
- Do not confuse with
- It is not the user score, and the two are never merged into one number.User score
Coverage cap
Also known as: the honesty cap, rated-on-N-of-6
The ceiling an entity earns by being documented: if only a quarter of the weighted model is rated, the score cannot exceed 25 out of 100, however good that quarter is.
- Why it matters
- Without it, one glowing rating on an otherwise empty profile scores a perfect 100 and outranks a fully documented, honestly average service.
- How we treat it
- Score equals the lower of the weighted mean and the coverage cap, in both the site and the database, with "rated on N of 6" shown next to the number.
- Do not confuse with
- The cap is not a penalty for a bad rating — it is a limit on rating too little.Editorial score
User score
Also known as: review score, reader rating
The aggregate of moderated reviews from readers who used the service, kept as its own number with its own review count.
- Why it matters
- Lived experience catches what a fee schedule cannot: support that stops answering, an app that locks accounts, a transfer that took nine days.
- How we treat it
- Reviews are moderated before publication, aggregate ratings appear only above a minimum review count, and the user score is never blended into the editorial one.
- Do not confuse with
- It is not our assessment, and a high user score does not lift an editorial score.Editorial score
Maturity level
Also known as: evidence ladder, publication level
How far a service has earned the right to travel on this site, in four steps: candidate (a critical fact is missing), listed (comparable, not scoreable), score-eligible (two or more sources, one official, no conflicts), and fully verified (rated on the whole model and checked within the year).
- Why it matters
- It answers the question behind every score — not "what is the number" but "why should this number be trusted at all".
- How we treat it
- The level drives eight separate permissions: listing, scoring, ranking, recommending, comparing, indexing, the public profile and the referral link. Below score-eligible we publish no number, and below full verification a service cannot appear in ranked picks.
- Do not confuse with
- A high level means well-evidenced, not good — a fully verified service can still be a bad deal.Editorial score