The Video Got 40,000 Views. The Product Page Got Three Clicks.

Forty thousand views. Three clicks on the link in bio. The creator scrolled through the comments looking for the reason and found nothing wrong there either. People watched. They just didn't do anything after.

That gap between a video that performs and a video that works is where most creators get stuck. View count only tells you people stopped scrolling. It doesn't tell you where they mentally checked out before the point landed, or whether the ending gave them any reason to act once it did.

Use Diagnose Why Your Short-Form Video Underperformed to find the exact section costing you the result, not just a list of things that could theoretically be better.

A worked example

Take a 45-second script for a skincare brand. The hook: "I stopped using this after a dermatologist told me why." Strong, curiosity-driven, does its job. The video pulled 40,000 views and a 68% average watch time, numbers most creators would call a win.

Feed the script and those numbers into the prompt and the diagnosis narrows fast. The hook promises a reveal about a specific ingredient. The video does deliver that reveal, but not until the 38-second mark, and the 25 seconds before it are backstory the viewer doesn't need to follow the point. That's the drag. Then the video ends right on the reveal itself, no line asking the viewer to do anything with what they just learned.

The fix isn't "post more" or "try a different hook next time." It's cutting 25 seconds of setup and giving the last 3 seconds an actual job again.

Why this beats guessing

Most underperformance reviews turn into a list of ten minor tweaks that all sound reasonable, and none of which move the number that actually matters. This prompt is built to resist that. It forces one specific verdict, and it treats a strong hook with a weak finish as a completely different problem from a weak hook with a strong finish, because the fix for each is nothing alike.

Get that distinction wrong and you reshoot the hook on a video whose real problem was always the ending. Get it right and the next video only needs one thing changed, not a full rewrite.

How to use it

1. Pull the script or transcript from the video, plus its real numbers: views, average watch time or completion rate, comments, shares.
2. Paste both into the prompt along with the platform it ran on.
3. Apply the single biggest fix it identifies before touching anything else. Ship the next version, then run the same check once its numbers come in.

A video with 40,000 views and three clicks isn't a failed idea. It's a script with one specific section still costing you the result, and now there's a way to find it before you spend another week reshooting the wrong part.