Every video gets a trial before it gets a chance. The jury is a thumb, the courtroom is a phone screen, and the verdict lands in about three seconds. We have shot a few thousand short videos at Updraft, and the single biggest difference between the ones that travel and the ones that vanish is not the edit, the music or the budget. It is what happens before the viewer has decided whether to care.
So let us talk about sandwiches.
The sandwich test
Picture two videos about the same sandwich. The first opens on a wide shot of a deli, a slow push in, a logo, then a person saying "Today we are making our signature club." The second opens on a knife going through the middle of the sandwich, the crunch turned up slightly too loud, and a caption that reads "We have been told this is illegal in three countries."
Same sandwich. Same person. Same kitchen. The second one wins, every time, and not because it is louder. It wins because the first frame already contains the payoff and the caption opens a question the viewer needs closed. That is the whole trick, and it is surprisingly hard to do on purpose.
Three jobs, three seconds
A good hook does three jobs at once, and it has to do them before the viewer has consciously read anything.
- Stop the thumb. Movement, contrast or a face looking straight down the lens. Something that breaks the rhythm of the feed.
- Make a promise. The viewer should know, roughly, what they get if they stay. A result, a story, a laugh, a fix.
- Leave a gap. Something unresolved. A question, a strange detail, a before without its after.
Most weak hooks do one of these. Pretty footage stops the thumb but promises nothing. A clear title promises something but does not stop anyone. The rare video that does all three in the first frame tends to be the one your client forwards to you at eleven at night with six exclamation marks.
Start at the end
The fastest way to improve a hook is to steal from your own ending. Watch the edit back and find the most satisfying moment: the reveal, the result, the punchline. Then ask whether a version of it can sit in the first second. Not the whole payoff, just a slice. The knife going in. The finished room before the renovation. The score on the screen before the match.
The first frame should feel like the middle of something, never the start of something.
We call this starting at the end, and it works because people are more patient with a story once they know it is going somewhere. A slow intro asks for trust. A glimpse of the ending earns it.
Words on screen are a second hook
Plenty of people watch short videos with the sound off, at least some of the time. That means your on-screen text is not a subtitle, it is a second hook with its own job. Keep it to seven words or fewer, put it where thumbs do not cover it, and never repeat what the picture is already saying. If the frame shows a cake collapsing, the text should not say "my cake collapsed". It should say "this was the wedding cake".
A small rule we use in the studio: the text should change what the viewer thinks they are looking at. If it does not, cut it.
Test hooks, not videos
Here is something we learned in the Lab. When a video underperforms, teams usually reshoot the whole thing. That is expensive and slow, and it rarely tells you what went wrong. Instead, we cut three or four versions of the same video with different opening seconds and run them side by side. Same body, same ending, different first frame.
The results are often embarrassing. A hook we argued about for an hour loses to the one the intern suggested as a joke. Last spring, a version for a ceramics client that opened on a mug being dropped, and caught, kept 41 percent more viewers at the five second mark than the version that opened on the potter's hands. Same clay. Different first second.
Five hooks to borrow this week
- The reveal first. Show the finished thing, then rewind to how it happened.
- The wrong answer. Open on a common mistake, done with total confidence.
- The overheard line. Start mid-conversation, as if the viewer walked into the room.
- The impossible number. A specific, surprising figure on screen before anyone speaks.
- The direct dare. A face to camera saying something the viewer will want to argue with.
What not to do
Please do not open with your logo. Please do not open with a cheery welcome back. Please do not spend the first second on a transition that exists only because the editing app had one. And please do not promise something the video does not deliver. A hook that lies gets the view and loses the follower, and the apps are getting better at noticing the difference.
The short version
Put a slice of the ending in the first frame. Make the text change the meaning of the picture. Test the first second, not the whole video. And when in doubt, cut straight to the sandwich.
