Method & caveats

• Publish-month attribution. An asset's lifetime views count in the month it was published, so recent content has had less time to accumulate. This understates the present rather than flattering it.
• Coverage grows too, not just audience. Backfill runs keep discovering already-published content, which raises a past year's total without anyone earning a new view. Early years also hold YouTube only — Facebook, X, TikTok and Rumble first appear in 2026 collection. Compare the audience figure against the asset count beside it, and treat reach per asset as the coverage-independent signal.
• Asset counts and length detail are collected-only. The report anchors are monthly totals with no per-asset rows, so they contribute to audience but not to the asset count or the length buckets.
• Report anchors fill gaps only. Platform back-end exports and Sprout figures cover months collection missed. Where both exist for a month, collected data wins. No anchored months in this view.
• Same window each year, complete months only. Every year is compared over January through the same month, and the current calendar month is excluded while it is still partial — 21 days of a month against full ones would read as a collapse.