How to Tell If a Video Will Go Viral Before You Post It
You can't guarantee virality, but you can measure whether a clip is built to survive the scroll. Here's a pre-publish checklist — hook, hold, pacing, sound — and what each signal actually predicts.
Let's be honest about the question first, because the honest version is the useful one. You cannot know that a video will go viral before you post it — virality depends on your account, the algorithm, timing, and luck no tool can see. What you can know, before posting, is whether the clip is built to survive the scroll: whether it stops a thumb and holds attention long enough to get distributed at all. That part is in the file, and you can measure it.
Here's the pre-publish read that actually predicts something.
1. Does the first second interrupt a departure?
A scrolling viewer is already leaving. The opening frame's only job is to interrupt that motion before it completes. Watch your own first second with the sound off and ask: is there a pattern break, an open loop, or an instant stake in the first few frames — before any words land?
If the answer is "the payoff is at 0:04," you have a hook problem, not a content problem.
2. Does attention hold through the middle?
Most short videos don't die at the start — they die in the sag around the halfway mark, after the hook spent its energy. The fix is a re-hook: a new visual, a question, a cut, a beat change that resets attention before it drifts. Map where your clip gets visually quiet and assume that's where viewers leave.
3. Is the pacing doing any work?
Flat pacing reads as "nothing is happening," and nothing-is-happening reads as "skip." You don't need frantic cuts — you need change: a zoom, a new angle, a motion shift every couple of seconds so the eye always has a reason to stay.
4. Is the sound pulling, or is it dead air?
Silence in the first second is one of the most common silent killers. A strong beat or a confident voice in the opening moment is a retention signal; flat or absent audio quietly gives permission to scroll.
None of these four are about taste or production value. A polished video with a buried hook loses to a scrappy one that interrupts the scroll. Quality is judged after the interrupt — it can't save an open that never stopped the thumb.
Turning the checklist into a number
Eyeballing this works, but it's slow and biased — you made the clip, so you already know the payoff is coming, which is exactly the blind spot a fresh viewer doesn't share. That's the gap a pre-publish diagnostic fills.
Scrollproof reads the actual footage — visual saliency, motion, scene cuts, audio energy — and returns the same four signals as scores: a Hook Strength read on the first second, a predicted Hold Rate for the middle, a second-by-second attention curve, and a visual-attention heatmap of where the eye is pulled. It's a creative diagnostic, not a crystal ball: it predicts whether the clip is built to hold attention, and flags the weak seconds to fix — it does not promise views.
The pre-publish loop that compounds
The creators who improve fastest treat every clip as a test:
- Cut a version.
- Read the hook, hold, and pacing — by eye or by score.
- Fix the single weakest second.
- Re-cut and compare the two side by side.
Do that ten times and you stop guessing. You start to recognize a thumb-stopping open before you ever post it — which is the closest thing to "knowing" that honestly exists.
Want to see it on a real clip? Scan one free and read the curve.
Stop guessing. Scan the clip.
Drop a short video and get Hook Strength, Hold Rate, a second-by-second attention curve, and a real attention heatmap — in about a minute. First scans are free.