How to Write the First Three Seconds of a Short

The first three seconds of a Short are not an introduction. They are the entire argument compressed to the part that makes someone stop scrolling. Most scripts lose here because they warm up, and warming up is the exact behavior a vertical feed punishes. A viewer decides to stay before your setup has finished, so the setup has to go.

I write the opening last, after the rest of the clip is finished. Only then do I know what the video actually delivers, and the first line almost never survives the first draft. Knowing the payoff changes which words matter and which ones just delay the point.

Cut the runway before the claim

Every word before the point is runway, and runway is what a thumb skips over. If the clip is about a failed product launch, it opens on the failed launch, not on the year you incorporated the company. The viewer does not need your timeline. The viewer needs the crash.

I read the first sentence and ask one question. Does this tell a stranger what the clip is about? If not, the sentence is setup, and setup belongs in line two at the earliest. Honest answers here save a lot of wasted recording.

The instinct to explain feels polite, because context helps a reader who has already committed. On a vertical feed there is no commitment yet, and the next video opens faster than yours does. Politeness is not the metric. Attention is.

Context is a luxury you earn after someone decides to stay. The first three seconds are a loan, not a gift.

Say one thing, not a topic

A topic is not a hook. Social media growth is a topic. The post that got four million views had no hashtags is a claim, and claims stop thumbs while topics do not. If your opening could belong to a hundred channels, it belongs to none.

Find the single most surprising true sentence in the whole script and drag it to the front. Every other line exists to support that one, and a beat that supports nothing is padding you can hear. The strongest sentence is usually buried in the middle, right after you got comfortable.

Rewrite the open after you write the close

The test is simple. Read the first three seconds blind, then read the last line. If the two do not point at each other, the clip has two subjects and needs one. Loose endings and loose openings are usually the same problem.

This is why writing the ending first helps. A known destination turns a vague opening into an obvious one, and the rewrite takes two minutes instead of an afternoon. The payoff is a map, even when the viewer cannot see it yet.

Test the open on a dark screen

Record the first line alone and play it on a phone with the screen dark. If it fails to make you curious with nothing to look at, it will not make a scroller curious either. Audio has to carry more than we usually give it credit for.

I keep a note of cold opens, about forty first lines that survived this test. They name a concrete thing and then hide the reason behind it, and that note does more for my retention charts than any editing trick I have picked up. The formula is boring, which is exactly why it works.

Key takeaways

• Write the opening last, once you know what the clip delivers.
• Name a concrete thing inside the first eight spoken words.
• Cut every sentence that only sets up context.
• Test the line with the screen dark and no sound.