CBCreatorBenchmarks
Changelog · Versioned in public

What changed,
and when.

Methodology updates, product releases, and fixes. The score on a creator's report on any given date can be reproduced from the methodology in effect on that date.

  1. April 24, 2026
    v1.4.0
    Methodology

    Authenticity now reads comment depth, not just ratio

    Authenticity sub-score now incorporates average comment length and emoji-only ratio alongside the existing like-to-comment check. Most accounts moved by ±2 points; weak signals dropped slightly to make room for the new dimension.

  2. April 12, 2026
    v1.3.0
    Product

    Niche correction is now inline, not a modal

    If the auto-detected niche is wrong, you can now switch it from the report header without leaving the page. The cohort recalculates immediately. Modal removed.

  3. March 28, 2026
    v1.2.1
    Fix

    Pinned posts excluded from engagement averages

    Engagement rate now excludes pinned posts, which were skewing the average for creators who pin a viral hit. Affected scores were recomputed once on March 28; no historical adjustments after that date.

  4. March 14, 2026
    v1.2.0
    Methodology

    Cohort minimum raised to 200 creators

    Niche/tier intersections smaller than 200 creators now widen the tier band (e.g. micro + mid combined) instead of falling back to platform averages. The widened cohort is labeled explicitly on every affected report.

  5. February 22, 2026
    v1.1.0
    Product

    Share card is now 1080×1920 first

    Share cards now render Instagram-Stories-first, with the 1200×675 OG-image variant generated from the same template. Stories share rate jumped 3.4× in the rollout cohort.

  6. January 30, 2026
    v1.0.0
    Product

    Public launch

    First public version. Engagement, Consistency, Growth, Authenticity. 350K creators in the cold-start dataset.