What a Virality Score Can — and Can't — Tell You

Let's be blunt about the category we're in. "Virality predictor" is a phrase that invites overpromising, and a lot of tools take the invitation. They imply a number that tells you how many views you'll get, or your percent chance of going viral, and they're careful never to define what the number actually means — because if they defined it, you'd see it can't deliver that.

We'd rather define it. A virality score is a real, useful instrument and a thing that genuinely cannot tell you whether you'll go viral. Both halves are true. The honest product is the one that's clear about which is which.

What "virality" actually depends on

Whether a video goes viral is the product of several factors, and the creative is only one of them.

Look at that column on the right. A score computed from your clip, before you post, can only see the first row. It reads the creative — and it reads nothing else, because nothing else exists yet. The account, the timing, the algorithm, and the luck are all downstream of publishing, and a pre-publish tool is, by definition, upstream.

This isn't a flaw to apologize for. It's the definition of what a pre-publish score is. It measures the one thing you can change with a recut, and stays silent on the four things you can't.

So what is the score, honestly?

A virality score — the kind Scrollproof produces — is a predicted creative-strength signal: a transparent, weighted blend of measured channels (visual saliency, motion, scene cuts, audio energy, face presence) that estimates how well a clip is built to stop the scroll and hold attention.

Read that carefully. It estimates how the video is built, not how it will perform. A well-built clip with a strong open and a tight spine has better odds — but odds aren't outcomes, and a tool that can't see your distribution can't promise an outcome.

A high score means: this is well-constructed to win attention. It does not mean: this will get X views, or this will go viral.

The honest caveats, stated plainly

We label three things on purpose, and they're worth saying out loud.

The index is uncalibrated. Right now, the composite Virality Index is an informed heuristic, not a validated probability. We don't have a published accuracy number, so we don't claim one. As creators opt in to share real outcomes, we're building the dataset to calibrate it — and when we do, we'll publish the honest figure, including where it's weak. We'd rather earn a real number than borrow an impressive-sounding one.

The heatmap is illustrative, not medical. The attention heatmap is a visual-saliency model of where the eye is likely pulled. It is not an fMRI, an EEG, or any neurological reading, and we never imply one. "Brain-style" describes a metaphor, not a measurement.

We don't promise views, reach, or revenue. There is no honest path from a pre-publish creative read to a guaranteed audience outcome, so we don't pave one. Anyone who does is selling the gap between what they can measure and what you want to hear.

Why a bounded tool is the more useful tool

Here's the counterintuitive part: the score is more useful precisely because it's bounded.

A number that claimed to predict your views would be unfalsifiable and unactionable — when the prediction missed, you'd never know whether the creative, the timing, or the algorithm was at fault, and you'd have no specific edit to make. A score that only reads the creative gives you something you can act on today: it points at the open, the dead spot, the flat stretch. The relief of a bounded instrument is that its claims are small enough to be true and specific enough to be useful.

The right way to use a virality score:

• As a diagnostic, not a verdict. Low score? Don't ask "is this video bad?" Ask "which beat is the score reacting to?" — then recut that beat.
• As a comparator, not an oracle. Its most reliable use is relative: this cut reads stronger than that cut. A clean comparison between two of your own versions is worth more than any single absolute number.
• As one input, not the decision. It catches unforced errors. It doesn't replace your judgment about what's worth making in the first place.

A virality score can tell you whether your clip is built to win the part of the game that's in your hands. It can't tell you the final score of a game that hasn't started. A tool that's honest about that distinction is the only kind worth trusting — and, not coincidentally, the only kind that actually helps.