How to build a video landing page that converts

Put one video above the fold with a thumbnail that states the outcome, keep it under two minutes for a cold audience, repeat the promise in text beside it for the majority who never press play, and place the CTA at the moment the retention curve says the video has proved its point — usually right after the demo, not at the end.

Updated 2026-08-18 · Markdown version

The structure

1. Headline — the outcome, in words. Not "Watch our video".
2. One video — the demo or the proof, not the brand film.
3. Text beside or under it — the same promise, readable without playing anything. Most visitors will not press play; the page must still work for them.
4. A single CTA, repeated: once beside the video, once at the end of the copy, once inside the player.
5. Proof — numbers, logos you are actually entitled to use, or a named customer story.

Length

For a cold audience, under two minutes. Every extra minute of a marketing video is paid for in drop-off, and the retention curve makes that visible: look at the share still watching at 50% and multiply by your conversion rate before defending a five-minute cut.

For an audience already in a trial, longer is fine — they came for detail. Split it with chapters so they can skip.

The thumbnail does more work than the video

Play rate is plays ÷ impressions, and it moves more with the poster frame than with anything else on the page. A thumbnail should show the product doing the thing, with three or four words of text. A random mid-video frame of a slide is the default and it is always beaten. A/B two thumbnails and let the numbers decide.

Where the CTA goes

Read the retention curve, then place the in-player CTA immediately after the payoff moment and before the ordinary drop-off. Practically:

• Overlay at the payoff — "Start free" the second after the result is shown.
• End screen — for the minority who finish, the strongest offer.
• Email gate — only on a video people are already finishing, and only if the rest is worth an address.

The mistakes worth naming

• Autoplaying with sound. Browsers block it; visitors who do hear it leave.
• A YouTube embed on the money page: it ends by offering the visitor a competitor's video.
• Two videos on one fold.
• No captions, so the sound-off majority sees a mime.
• A video that opens with a fifteen-second logo animation. The first five seconds decide the rest.

Measuring it properly

Track the chain, not just the page: impressions → play rate → retention at 50% → CTA click → lead. A page that "converts badly" is usually failing at exactly one of those steps, and each has a different fix. Video analytics walks through reading it.