Methodology

Score = deterministic public early traction.

score = age + uploads + views/day + first_5_views + shorts_ratio + freshness

What we measure

NicheBreakout combines public channel age, public upload count, total views, views/day, first-five upload traction, shorts ratio, max_video_views, median_video_views, scanner freshness, and same-age cohort context.

What we don't measure

No watch time, no retention, no RPM, no revenue, no traffic source, no impressions, no CTR, and no average view duration.

Cohort sizing

Percentile claims are computed only against same-age scanner cohorts, for example top 12% of n=140 25-day-old channels, when enough public rows exist.

Refresh cadence

Scanner jobs run daily. Last-scan timestamps are surfaced on channel detail facts and read models.

Data source

YouTube Data API v3, public metadata only.

Last updated: May 5, 2026