⚠ Each bucket is a separate comparability bucket — they are NEVER merged or ranked against each other. Solicited reviews, ambient social/forum voice, and a brand's own owned audience are fundamentally different speech genres with different error models. Read each on its own terms; a number in one bucket is not comparable to a number in another.
How to read this. Every headline is computed in SQL over that bucket's governed metric tables (the model never generates figures); open a report for the full evidence, with verbatim quotes behind each number. Buckets with no headline simply had no metric computed yet — never a fabricated figure.