How rankings work
Every ranking prints its inclusion rule, its primary metric and its data date, so the order can be checked against the number.
What a ranking is
A ranking is a filter, a sort and the metric printed beside each broker. The filter says who is in, the sort says why they are in that order, and the metric is the number the sort used. A ranking is published only when its inclusion rule and minimum data coverage are met, and nothing is invented to fill a table: a value that is not available is shown as not available where the template permits it, and otherwise the row is left out.
Sponsored positions
A sponsored broker appears in a dedicated slot above or between the organic results. It carries no organic rank number, and payment cannot alter a measured metric, an editorial score or a regulatory status. When no eligible sponsor exists for the reader's country the slot collapses.
The terms the ranking pages use
- Compatible data
- Data that shares the required definition, unit, instrument, account basis and reporting period.
- Coverage
- The proportion or count of eligible records represented in the published result.
- Normalisation
- Conversion into a common unit using the method stated for that ranking.
- Outlier treatment
- Clearly invalid, duplicate or extreme records are excluded only under the published quality rules.
- Tie rule
- The category's stated secondary factor, then the latest verified dataset timestamp; if still equal, a shared rank.
- Rank movement
- The change in a broker's organic position against the immediately preceding equivalent period, shown once a previous period exists.
- Missing broker
- A broker may be absent because it does not meet the eligibility, country, verification or data-coverage requirements.
Corrections
A material correction updates the affected result and its last-updated date. If a figure on a ranking looks wrong, use the ranking data issue form and name the ranking, the broker, the data point and your source.