Most calls to action fail for one reason: they ask for everything at once, at the moment the viewer is already leaving. The CTA Generator writes five scripts — one each for subscribe, like, comment, share and playlist — with the exact second to deliver each, the persuasion principle behind the wording, a combined verbal-plus-on-screen version, and an A/B variant so you can test.
This guide explains why timing matters more than wording, how to pick which single ask each video gets, and how to write comment prompts that actually generate comments.
Timing beats wording
The end of a video is the worst place for a CTA: retention graphs show that the outro is the least-watched twenty seconds of any video. The generator places each CTA where its target behaviour is most likely — the subscribe ask right after the first major payoff (usually 25–40% through), the comment prompt immediately after posing a question the audience has opinions about, the like ask at the moment of highest value delivery, and the playlist or next-video ask in the final ten seconds where an end screen can support it.
One ask per video
Each video gets one primary CTA. Subscribe for videos likely to reach new viewers (browse, Shorts feed traffic). Comment for videos where discussion helps the algorithm and gives you ideas. Playlist for tutorials in a series. The generator gives you all five so you can pick; using more than one primary ask roughly halves the response to each.
Comment prompts that work
"Let me know in the comments" produces nothing. A prompt that asks for a specific, low-effort, opinion-based answer ("Which grinder are you using — drop the model") produces dozens. The generator's comment CTA always includes the specific question, and the A/B variant tests a different framing of the same question.
Tips
- Tell the generator the conversion goal and the video's structure (where the payoff lands). The timestamps are only accurate if it knows where the peaks are.
- Deliver the subscribe CTA with a reason and a timeframe: "if this fixed your shot, subscribe — the milk video is Thursday".
- Put the on-screen version of the CTA in the lower third, not the centre, and keep it up for four seconds.
Common mistakes
- CTA in the first 15 seconds — Asking before delivering value has the highest drop-off of any CTA placement.
- "Like, comment, subscribe, share, hit the bell" — Five asks is zero asks.
- Same CTA every video — Audiences tune out repeated phrasing. Rotate the A/B variants.
Worked example
Input: "12-minute espresso tutorial, payoff at 4:10, goal is subscribers."
What the report returned: Primary: subscribe CTA at 4:25 — "If that shot just made sense, subscribe; the milk steaming one is Thursday" (reciprocity + specificity). Secondary comment prompt at 9:40. Playlist ask at 11:45 with end-screen cue. A/B variant uses a social-proof framing.
Frequently asked questions
Where should the subscribe CTA go?
Right after the first major payoff, typically a third of the way in. The generator calculates the second from your video's structure.
How many CTAs per video?
One primary, optionally one secondary. More than that dilutes all of them.
Do verbal or on-screen CTAs work better?
Together. The combined script has the verbal ask with an on-screen text or graphic reinforcing it for four seconds.
Can it write end-screen copy?
The playlist and next-video CTA includes end-screen cues. The End Screen Optimizer in the Creator tier handles full end-screen layout.