Nobody outside TikTok knows the exact ranking formula, and anyone selling certainty is guessing. But creators and researchers widely observe the same mechanic: new videos get tested with small audiences first, and early response decides whether the audience grows. Understanding that early window is the highest-leverage thing a creator can learn.
Test audiences: the rounds model
The widely observed pattern: a new video is shown to a small initial group. If watch time, likes, comments, and shares clear some bar, the next round is bigger. Each round is a gate, and the gates get harder. This is why "the algorithm" feels random — a video that fails round one never gets the chance round three would have given it.
Which signals matter in the window
Watch time (especially completion rate) is consistently reported as the heaviest signal — a finished video beats a liked-but-skipped one. Shares and comments indicate deeper engagement than likes. Rewatches are gold. None of this is officially confirmed in detail, but it matches what large-scale creator experiments show again and again.
What you can control
- The hook: the first 3 seconds decide whether round one even starts well.
- Posting time: publish when your audience is actually online so the first rounds have real viewers.
- The first comment: pin a question to seed comments early.
- Replies as new videos: they recycle engagement back to the original.
- Series formats: "part 2" mechanics pull viewers back and lift completion on the next one.
Where early boosts fit
This is the honest case for early-window growth services: concentrating likes and views into the first hours feeds the exact signals the rounds model reads. It does not rescue a video nobody finishes — watch time still has to come from real viewers. Think of it as making sure a good video survives round one, not as a substitute for being good.