Methodology
Every score on this site is produced by the process described here. If a number appears without a path back to this page, treat it as an error and tell us.
What we score
Platforms are assessed on four dimensions, each scored out of 100 and then weighted:
- Settlement reliability (40%). Observed settlement times across the corridors the platform advertises, and the proportion of those corridors where the published cut-off was met.
- Cost transparency (25%). Whether the full fee — spread, transfer fee, intermediary charge — can be determined before a payment is sent.
- Coverage (20%). The number of currencies and corridors reachable directly, without a correspondent hop.
- Operational depth (15%). API surface, webhook reliability, reconciliation tooling and the quality of the published documentation.
How the weighted score is formed
The four dimension scores are multiplied by their weights and summed to a single number out of 100. We publish the dimension scores alongside the total so a reader can re-weight them. A platform that scores well on settlement but poorly on transparency is visibly that, not averaged into obscurity.
What the score is not
It is not a price comparison. A cheaper platform does not automatically rank higher; an expensive one that settles reliably can outrank a cheap one that does not. It is not a prediction of fitness for your use case — a platform that suits a marketplace may not suit a payroll provider.
Evidence and verification
Fee schedules and terms are collected from the platform's own published documents and dated. Settlement observations are recorded with the corridor, the amount band and the date. Each article states its last verification date; a platform that changes its terms after that date is re-assessed before the article is updated.
Scoring and structured data
Our editorial scores are published as visible text and tables. They are deliberately not emitted as review ratings in structured data, because a rating out of 100 on our four-dimension scale is not a customer review of the product and should not be read as one.