Skip to content

Every change,
on the record.

The formula is versioned like software. Each version lists exactly what changed and when it took effect.

Scores are never retroactively rewritten without a changelog entry. Corrections to evidence are applied at the Monday 06:00 UTC recalculation.

  1. v1.3

    13 August 2026

    Current formula

    • Published board shortened from Global 100 to Global 50 — a tighter numbered list with the same Index Rating (0–100 points)
    • Places that previously sat at 51–100 remain scored and move to unnumbered Watch; scores were not silently rewritten
    • Watch now covers both Rising-board curation and established names just outside the Global 50 cutoff
    • Agency pages split numbered roster from Watch so #0 is never published
  2. v1.2

    26 July 2026

    • Documented score production mode: editorial pillar totals + automatic weights (not yet an automated line-item ledger)
    • Profiles labelled “Editorial pillars” so provenance is visible on every score
    • Clarified three lenses: Index (scored), Movers (snapshot deltas only), Watch (Rising curation)
    • Movers never substitute momentum or traffic for missing week-over-week deltas
    • Added Appendix B: regional house / title equivalence for non–Big Four markets
  3. v1.1

    20 July 2026

    • Published the full sub-metric point values for all four pillars
    • Added an end-to-end worked scoring example
    • Published the tie-break rule for equal Index Ratings
    • Quantified the evidence-confidence thresholds (High / Medium / Baseline)
    • Added Appendix A: the house tier list used to weight runway shows
  4. v1.0

    13 July 2026

    • Initial public formula: pillar weights (runway 40%, campaigns 30%, editorial 20%, cultural 10%)
    • Published the five-year eligibility window
    • Published the time-decay rule (100% / 50% / 10%)
    • Introduced the public confidence labels
← Read the current methodology