How Shortzly's Viral Score Works
When Shortzly analyzes a long video, it does not just hand you a pile of clips. It gives every moment a Viral Score: a number from 0 to 100 that predicts how well that moment is likely to perform as a short. The score is what lets you render the two or three clips most likely to land instead of guessing, and it comes with plain-English reasons so you know why a moment scored the way it did. This guide explains what the Viral Score measures, how it is calculated, and how to use it to get more out of every upload to TikTok, YouTube Shorts and Instagram Reels.
What the Viral Score is
The Viral Score is a single 0-100 prediction attached to each moment Shortzly finds. A higher number means the moment has more of the qualities that short-form audiences reward: a strong opening, a reason to keep watching, a point that stands on its own, and a clear payoff. It is deliberately a prediction, not a verdict on your content. Two moments that both score in the 80s can perform very differently once they are posted, because the algorithm, the thumbnail, the caption and the audience all matter too. What the score does reliably is rank the moments in one video against each other, so the best candidates rise to the top.
What the Viral Score measures
The score is built from several sub-scores, each looking at a different quality of the moment. No single signal decides the outcome; the score is a weighted blend of all of them.
- Hook strength. How well the first couple of seconds grab attention. The opening line of a short does most of the work, and a weak hook caps everything that follows.
- Retention. Whether the moment keeps people watching to the end, based on pacing, pauses and how tightly the idea is delivered. Dead air and rambling pull this down.
- Context. Whether the clip makes sense on its own, without the rest of the video around it. A moment that needs ten minutes of setup is a poor short.
- Appeal. How broadly interesting or shareable the subject is. Niche can still score well, but mass-appeal topics score higher here.
- Emotion. Whether there is a genuine emotional beat, a laugh, a surprise, a strong opinion. Flat, purely informational stretches score lower.
- Audio energy. How the loudness and delivery move relative to the rest of the video. Energy change tends to mark the moments people remember.
- Visual change. How much the picture moves or cuts during the moment. Static talking-head stretches get a smaller bump here than dynamic ones.
Hook strength and retention carry the most weight, because they map most directly to whether a short survives its first few seconds and gets watched through. The softer signals, emotion, audio and visual change, refine the ranking rather than dominate it.
How the score is calculated
This is the part that makes the Viral Score trustworthy: the final number is computed in code, not improvised by a language model. Shortzly measures the real signals it can measure, pauses, pacing, loudness against the video's own average, cuts and the opening, directly from the transcript timing and the audio. The AI model is asked only for the judgments a model is actually good at, such as how self-contained a moment is and whether it carries an emotional beat. Those judgments and the measured signals are combined with fixed weights to produce the 0-100 score.
When a signal genuinely cannot be determined for a moment, its weight is redistributed across the signals that are present rather than filled in with a guess. That is why the score never pretends to know something it does not, and why the same moment scores the same way every time you analyze it.
Every number comes with a sentence
A score on its own is not very useful if you cannot see why. So each moment carries a short list of reasons and, where relevant, suggested improvements, written as fixed plain-English sentences built from the same measurements the score uses. You will see things like:
- "There is a 2.3 second pause at 0:14. Removing it may improve retention."
- "The opening line is strong and sets up a clear question."
- "This moment ends on a natural beat."
These are not model opinions or chain-of-thought. They are generated from the numbers, so they are consistent and actionable. If a moment loses points for a long pause, the suggestion tells you exactly where it is, and you can fix it before you render.
Moments that have very little speech
Not every great short is built on talking. For videos with very little spoken content, Shortzly does not force a transcript-based score. Instead it reads the audio and visual timeline and picks the peaks, the moments where the energy and the action are highest, and labels them as picked by activity rather than by a spoken hook. This means a highlight reel, a gameplay montage or a mostly musical clip still gets usable moments instead of failing for lack of a transcript.
Categories
Alongside the score, Shortzly tags each moment with a content category, so you can tell at a glance whether a high-scoring moment is a story beat, a reaction, a tip, and so on. Over time, paired with your real results, this is what tells you which kinds of moments actually work for your audience rather than which ones simply scored well.
How to use the Viral Score
The score is a workflow tool. Here is how to get the most out of it:
- Render top-down. After analysis, the moments are ranked by score. Start with the highest and work down. One upload rarely needs every moment; the top few are where your time is best spent.
- Read the reasons before you trim. If a strong moment is marked down for a pause or a soft opening, you can often fix it in the editor. Trimming a slow lead-in or removing filler words and long pauses can lift a moment's retention before you ever render it.
- Act on the improvements. A suggestion like "removing this pause may improve retention" is a concrete edit, not a vague note. Make the edit and the clip is tighter.
- Pair the hook with the score. The hook is the single biggest lever. If you want to push a borderline moment higher, a stronger opening line usually does more than anything else. Our guide to viral hook formulas covers the patterns that work.
- Compare the prediction with reality. Once a clip is posted, Shortzly's analytics put the real views, likes and comments next to the moment's predicted Viral Score, so you can see where the prediction was right and where your audience surprised it. That feedback is how the score becomes useful for your channel specifically, not just in general.
It is a prediction, not a promise
We are careful about this, and the product is too. The Viral Score predicts likely performance; it does not control it. A 90-scoring moment can underperform if the thumbnail is weak or the posting time is wrong, and a 60-scoring moment can take off because it caught a trend. Treat the score as a strong first filter that saves you from posting your weakest moments, not as a guarantee. Combined with good captions, the right aspect ratio and consistent posting, it tilts the odds in your favor across TikTok, Shorts and Reels.
Where to see it
The Viral Score shows up as soon as analysis finishes. Every moment in the review screen displays its score, its category and the reasons behind it, so you can make the render decision with the evidence in front of you. There is nothing to turn on and no extra step; it is simply how Shortzly presents the moments it finds. If you have never run a video through it, the fastest way to see the score in action is to point the AI clip generator at a talk-heavy video, a podcast, an interview or a webinar, and watch how it ranks the moments.
The short version
The Viral Score turns "which of these clips should I post?" into a number you can trust, backed by reasons you can act on. It measures the hook, retention, context, appeal, emotion, audio energy and visual change of each moment; it is computed in code so it is consistent; it explains itself in plain sentences; and it can be checked against your real results after you post. Use it to render fewer, better clips, and to learn what actually works for your audience on TikTok, YouTube Shorts and Instagram Reels.