If you've shipped short-form for any length of time, you already know the math: more than 60% of viewers swipe within the first 2 seconds. Whatever you do in those 90 frames decides whether the algorithm pushes you to 10,000 more people or buries you.

The good news: hooks are not creativity. They're a pattern library. After cutting clips for podcasters, founders, and creators across every vertical you can imagine, we've found 5 patterns that consistently outperform whatever your gut tells you to do first.

Here's what works in May 2026.

1. The "wait, what" hook

This is the most reliable pattern we've ever tested. You take the most counterintuitive sentence from your clip — the one a listener would do a double-take on — and you lead with it.

Example, real client: a podcast guest casually said, "I fired him on day 90." On its own that's a normal sentence. On the Reel, we cut it as the first frame, paired with a captioned overlay: "Wait, you fired him on day 90?" The clip then cut to the explanation.

2.4M views in 28 days. The same clip without the hook framing did 14k.

The principle: set up a question the viewer has to stay for the answer to. Pure curiosity, no payoff in the hook itself.

2. The "this is wrong" hook

Make a statement that contradicts conventional wisdom. The brain can't help but stay to see if you can defend it.

"Your podcast doesn't need better audio. It needs better hooks." "Hiring an in-house editor is a mistake." "MrBeast thumbnails are over."

The hook itself is the value — the rest of the clip just has to deliver on the promise of "and here's why." This works especially well for founder-led content where the host has earned the right to be contrarian.

3. The visual-first hook

If your raw footage has anything visually unusual in it — a chart, a sticker on someone's laptop, an unboxing, a graphic — lead with the visual and let the audio catch up.

Most podcasts don't have great B-roll, so this hook is less common in our work, but when it lands it lands hard. A chart that says "Your podcast loses 73% of viewers at 4 minutes" as the opening frame, no audio, no captions, just the number — viewers will sit through 30 seconds of context to know why.

4. The "I made a mistake" hook

Self-deprecation is a cheat code on short-form. "I almost killed my podcast last year." "I spent $40k on the wrong editor." "Don't do what I did."

It works because the brain pattern-matches vulnerability to authenticity. The viewer assumes: this person is going to tell me something true. They stay.

One caveat: this can't be performative. If you're not genuinely about to share a mistake, don't use this hook. Audiences detect fake humility instantly.

5. The "number + claim" hook

Numbers ground the brain. "I edited 200 podcasts last year and one mistake kept coming up." "We tested 47 hook variations and the worst one taught us the most."

The specificity of the number signals competence. A round number ("hundreds") feels like marketing. A weird specific number ("47") feels like data. The brain leans in.

Use this hook when you have actual proof to deliver in the next 20 seconds. If you don't, the audience will feel the bait-and-switch.

What we got wrong before this

For a while we used "POV:" and "Day in the life:" framings because they were viral in 2023. They've stopped working. The platforms are saturated with them and the algorithm penalizes repetitive openers now.

Also dropping in volume: any hook that requires the viewer to read text for more than 1 second before the action starts. Word-by-word captions on the hook itself are fine — but a paragraph of setup is a swipe.

"If the viewer has to work to understand the first 2 seconds, they won't."

How we apply this to client work

Every clip we ship through our short-form service (or as part of podcast clip cutdowns) goes through a hook review. Our lead editors flag the top 3 candidate hook patterns from the raw footage, then we ship at least one A/B variation so you can test which lands better on your audience.

That A/B step is where most creators lose. They commit to the first hook that "feels right" and never see how a different framing would have performed. We've seen 4x performance gaps between the same clip with two different openings.

If hooks are eating your retention and you don't have time to test 5 variations of every clip, that's literally what we do for a living. Book a free call and we'll cut your next clip with 3 hook variations so you can see what we're talking about.