Guides / Writing
Why viewers leave your video at 0:30, and how to fix it before you film
Your retention graph tells you where people left. By then the video is already out. Most drops have a cause you can see in the script before you film it.
This guide covers the six most common causes, how to find them in your own script, and how to fix each one.
The six reasons people leave
- A weak hook. The first two lines do not say what the viewer gets. See our guide to YouTube hooks for fixes.
- Throat-clearing. A greeting, a channel intro, a disclaimer or "before we start" in the first 15 seconds.
- A promise with no payoff. You tease something early and never come back to it, or it pays off with something smaller.
- Drag. One idea running for more than about 90 seconds (around 200 spoken words) with no turn, new question or new stake.
- A call to action at the wrong moment. "Subscribe" in the middle of the tension, before the payoff lands.
- A weak ending. The video ends on a summary instead of the thing the hook promised.
How to spot them in your script
Read your script out loud with a timer, or count words: most people speak about 150 words a minute. Then check:
- Read only the first two lines. Would you keep watching?
- Highlight every promise ("later I'll show you", "the surprising part"). Find where each one pays off.
- Mark every point where something changes: a new question, a turn, a result. Any gap longer than about 200 words is a drag risk.
- Find your call to action. Is it before or after the main payoff?
Check it in Cuely
Cuely's Retention check reads your finished script and predicts where viewers are most likely to stop. Each finding quotes the exact line, gives a rough timestamp, says why it loses people and how to fix it.
Fix the high-severity findings first, then run it again. If you use the Writing Agent, it runs the same check on its own draft and rewrites the weak parts before you see them.
Fixes that work
- Move the payoff up. If the best part is at 6:00, tease it in the hook and get there sooner.
- Cut the first 15 seconds of setup. Start where the interesting part starts.
- Add a turn every minute or so: "But that's not what happened." "Here's where it went wrong."
- Put the subscribe ask after the payoff, or near the end.
- End on the result you promised, not a recap.
Questions
Can I really predict retention before publishing?
Not exactly. Thumbnail, title and topic matter too. But the common causes of early drops are in the script, and those you can catch before you film.
What is a good audience retention on YouTube?
It depends on length and niche. Compare your videos against each other: find where your own graphs drop and look at what you said at that moment.