Riffkit Research
Original data, honestly reported.
We run a structural analysis engine over winning short videos — and publish what it finds, aggregates only, unflattering numbers included. Every chart is CC BY 4.0: use it, credit riffkit.ai, link the study.
August 2026 · n=1,504
We tracked 1,504 AI-generated TikToks. Here’s what actually happened.
The honest distribution — median 418 views, 20% cleared 1,000, top video 1.5M+ and climbing with 18% engagement — and the structural choice that doubled the odds.
August 2026 · n=1,111
We analyzed 1,111 winning TikToks. The hook isn’t where you think it is.
The median hook runs 4.1 seconds, not the 1–2 the folklore claims. 65% of winners have no spoken words, 83% still show a human, and 39% are a bare two-beat structure.
The engine behind the numbers.
The same analysis that produces these studies runs on any winning video you point it at — and rebuilds that video’s formula as a new one for your product. Here’s how that works.