How short-form distribution actually works
Forget "hacking the algorithm." Short-form feeds all run on the same simple loop — and once you see it, most growth advice explains itself.
Click to enlarge
Every short-form feed — TikTok, Reels, Shorts, and the rest — answers one question, over and over: should we show this video to more people?
They answer it the same way: show the video to a small batch of viewers, watch what those viewers do, and let the result decide the next, bigger batch. That's the whole machine. Everything else is detail.
The loop
- You post. The platform shows your video to an initial audience — some followers, some strangers whose interests match your content.
- The platform measures. Not your follower count, not your hashtags — your viewers' behavior: Did they watch it? How much of it? Did they watch it twice? Like, comment, share, follow, tap your profile?
- The result decides the next batch. Strong signals → a bigger batch, and the measuring repeats. Weak signals → distribution slows and stops.
Two consequences fall straight out of this:
- Every video gets a chance. Distribution is per-video, not per-account. A small account can have a big video — the loop doesn't care who you are, it cares how the batch reacted.
- Nothing is owed to you. A big account can have a dead video, for the same reason.
The signals, ranked roughly by weight
Platforms don't publish exact formulas, and the details shift constantly — but across all of them, the same behaviors dominate.
| Signal | What it tells the platform | Weight |
|---|---|---|
| Watch time / completion | People chose to stay | Heaviest — a 20 s video watched to the end beats a 60 s one abandoned at second 8 |
| Re-watches | The video is worth seeing twice | Very high; loops and "wait, what?" endings get counted again |
| Shares and saves | Someone put their own name on it | The strongest active signals |
| Comments | Engagement and time on video in one | High |
| Follows and profile visits | The video created a relationship | High, and it compounds |
| Likes | Someone tapped | Weakest — cheap to give, so worth less |
Notice what's not in that table: posting time perfection, hashtag rituals, follower/following ratios. Those can matter at the margins; retention decides the game.
Why the first seconds dominate
The completion math is brutal at the start. Lose half your viewers in the first two seconds and even a perfect rest-of-video can only be watched by the half that stayed.
That's why openers get their own craft — see Hooks: the first two seconds.
What this means for how you work
- Volume is how you learn. Since distribution is per-video, every post is a test. Ten posts teach you ten things; one post teaches you almost nothing. This is the real argument for consistent output — not "the algorithm rewards activity," but you can't read signals you never generate.
- Retention beats production value. A phone-shot video people finish outperforms a cinematic one they skip.
- The platform is measuring the content, so fix the content. When a video dies, the useful question is "where did people leave?" — not "did I post at the wrong time?"
- Every platform runs its own loop. The same video gets independently tested on TikTok, Reels and Shorts — the same ticket in several draws. That's the mechanical case for cross-posting (and why tools that render one variant per platform, like VidVertex, exist).
Click to enlarge
One honest caveat
None of this is a formula for a specific video going viral — the loop has too much randomness in the initial batch for that. It is a formula for a body of work growing: enough videos, watched well, on enough surfaces. Play the averages, not the jackpot.
START HERE · 3 MINCross-posting right: what actually differs between platforms"Just post it everywhere" is correct — and still done wrong most of the time. The differences that matter aren't the ones people obsess over.