Frame-level signals, tied to seconds
The analysis starts from the file itself, not a summary of it. The model steps through the finished cut frame by frame and measures the things a retention graph reacts to: how long the hook takes to show motion or a face, how often the visual state changes, where near-identical shots repeat, how the audio sits, and when on-screen text arrives. Those raw signals then run through the ten frameworks, from Hook Power to Payoff and Outro, weighted for your niche, platform, and country.
Every finding is anchored to a timestamp. The report never says the pacing is weak somewhere in the middle. It says the energy flattens at a specific second because three similar angles run back to back, and it names the cut that fixes it. The score itself is a probability signal, not a guarantee. It tells you where viewers are likely to slip, and the creative decision stays yours.
What a 32 looks like
The walkthrough above shows a strong cut at 74. A weak cut reads very differently. Picture a 52 second short that comes back at 32. The cold open spends five seconds on a logo sting and a spoken greeting, so Hook Power lands in the twenties and most viewers are gone before the content starts. The premise the title promised finally arrives at 0:19, buried behind setup nobody asked for. The mid-section holds one static angle for eleven seconds, so Tension and Pacing flags a long flat stretch with no visual change. And the video simply stops at 0:52 with no payoff beat, nothing that rewards the watch or invites a loop.
The playbook for that cut is the mirror image of the demo above. The first step is not to keep the cold open but to rebuild the opening three seconds around the strongest moment in the footage. Then move the premise into the first spoken line, break the static stretch with two angle changes or a cutaway, and end on an actual payoff, even if it is a single reaction shot. A 32 rarely means the idea is dead. It usually means the order is wrong, and order is the cheapest thing to fix before you post.
Analysis or Score first?
The two tools work at different moments. Score rates an idea while it is still a sentence: describe the concept, pick your platform and niche, and decide whether the idea deserves hours of production at all. Analysis belongs at the end, when the cut exists and the question is what to change before you upload.
In a normal week the loop looks like this: score your ideas, film the winner, analyze the finished cut, apply the playbook, then post. If you are starting from zero, begin with Score to test ideas before filming, and see how the full workflow fits together from first idea to upload.
