
To make a video trailer, decide on the single promise you are making to the viewer, find the moments in your footage that tease that promise without resolving it, arrange them into a short rising structure, and export a timeline to your editor for music and finishing. A trailer is a sales pitch with a runtime. Its only job is to make someone want the full thing.
The classic mistake is treating a trailer like a tiny version of the project. A trailer is not a summary. A summary tells you what happens. A trailer makes you need to find out. Those are opposite goals, and confusing them is why so many trailers feel flat even when the footage is great.
Before you cut anything, finish this sentence: "This trailer makes you want to watch because ___." If you cannot answer in one line, you are not ready to cut. The promise might be a question (will they pull it off?), a feeling (this will move you), or a hook (you have never seen this done before). Everything in the trailer serves that one promise. Everything else is cut, no matter how good it looks.
Once you have the promise, the trailer becomes a selection job. You are pulling a handful of moments out of hours of material, and the fastest way to find them is to read, not watch. Reading runs around 238 words per minute for an adult on non-fiction text per Brysbaert's 2019 meta-analysis, while speech crawls along near 150. So transcribe the footage and read it looking for tension, not information.
You are hunting for lines and beats that open a loop without closing it:
When you find a moment that explains or resolves, flag it as a do-not-use. Resolution is the enemy of a trailer. The viewer should leave with the loop still open.
Most trailers that work, from films to course launches to documentary teasers, follow a three-beat rise. Reduced to its bones:
Pacing accelerates as it goes. Longer holds early, faster cuts as the tension builds, then a beat of silence right before the button. That silence is doing real work. It tells the viewer the important thing is coming.
Say you have a 90-minute documentary cut about a small-town team and the client wants a 60-second teaser. Here is how I would build it.
I read the full transcript first and tag every line that creates tension or sounds like a promise. From 90 minutes I might tag 25 candidates. Most get cut, because 60 seconds holds maybe 8 to 12 beats.
Then I arrange against the three acts. The setup is two quick lines that establish the team and what is at stake. The tension act is five or six escalating moments, intercut. For the button I look for the line that recontextualizes, the one that, on a second viewing, you realize was the whole point. I duplicate that line to the front as a cold open if it works as a hook out of context.
Inside each clip I trim hard. A trailer has no room for ramble. I cut the lead-ins and land on the clean line. Then I export the arranged sequence to Resolve or Premiere and do the music, sound design, and titles there. The music is chosen to fit the cut, not the other way around. If you cut to the beat first, the words stop landing.
You spoiled the ending. If the trailer answers the question it raised, there is no reason to watch. Hold the resolution back.
It is a list, not a build. Eight equally weighted moments in a row feel like a slideshow. The structure has to rise. Each beat should feel a little bigger than the last.
No silence. Wall-to-wall music and talking gives the viewer nowhere to feel the tension. A held beat before the button is often the most powerful second in the whole piece.
You cut to the music. On a footage-led trailer, the words and faces carry it. Find the cut in the transcript first, then lay music under it. A J-cut that brings the next line in early can stitch two beats into one rising motion.
A trailer lives or dies on the moments you choose and the order you put them in. That is selection and arrangement, and both are far faster in a transcript than on a timeline. You read, mark the tension, arrange the rise, trim the ramble, and bring a structured sequence into your editor for the part NLEs are actually good at: music, sound, grade, and titles.
ScriptCut handles the pre-edit: transcribe, mark your moments, arrange them into a structure, and export an XML or EDL to Resolve, Premiere, Final Cut, or Avid. You get to spend your time deciding what makes someone want the full thing, instead of scrubbing for it.
Name the promise, find the tension, build a rising three-act structure, and never resolve the question you raised. Do the selection and arrangement in the transcript, then finish in your editor. A great trailer is not a small version of the project. It is a question the viewer has to answer by watching.
Most trailers run 30 to 120 seconds, with 60 to 90 seconds being a common sweet spot. A teaser can be as short as 15 to 30 seconds. The right length is the shortest one that delivers your promise and leaves the viewer wanting more.
A highlight reel shows off the best moments and tends to resolve them. A trailer raises a question and deliberately withholds the answer to make people watch the full thing. A trailer sells, a highlight reel showcases.
Lock the structure and the moments first, then add music. On a footage-led trailer the words and faces carry the story, so cutting to the dialogue and laying music underneath keeps the lines landing. Cutting to the beat first usually buries the meaning.
Transcribe the footage and read it, looking for tension rather than information: stakes lines, reveals that raise questions, emotional peaks, and lines that sound like taglines. Reading is much faster than scrubbing, and it surfaces the moments that open a loop without closing it.