How do I write Instagram Reels hooks that keep people watching?

When I ran Reels for Painex Clinic in Pune, one account hit 742,000+ views in a stretch where 94% of viewers were non-followers. That reach didn't come from luck or a big budget. It came from rewriting the hook on every Reel until the first 3 seconds carried their own weight. Instagram will push a Reel with a strong hook to strangers, and it will kill one that opens with "Hey guys, today we're going to talk about…". Here's how I write hooks now.

What a hook is (and what it actually decides)

A hook is the first 3 seconds of a Reel: the opening line spoken, the caption on screen, and the visual you open on. It answers one silent question the viewer is asking, "Why should I stay?" Instagram measures retention in that window brutally. If people swipe in the first two seconds, the algorithm reads it as "not for this audience" and stops distributing. The hook doesn't have to be clever. It has to promise something specific.

The 3-second rule, and why 85% watch on sound-on but text still wins

Instagram's own data has said for years that around 85% of Reels are watched with sound on. Great. But most of that audio-on viewing starts with the phone unmuting mid-scroll, so the first frame is often silent. That's why on-screen text carries the hook. I write the text hook and the spoken hook to say the same thing. Audio-off viewers get the promise from the caption. Audio-on viewers get it reinforced by the voice. Skip either half and half your audience walks.

• Text on screen: the promise, in a single readable line, 6 to 10 words max.
• Spoken audio: the same promise, said naturally, no throat-clearing.
• Opening visual: movement or a face or a pattern break, not a static logo.

Six hook formulas I actually use

I don't invent hooks from scratch every time. I keep a short list of formulas that work and fit the topic into whichever one suits it. These are the six I've used most across Painex, local brands and my own account. They're what I hand to clients when we're planning a batch of Reels.

• Curiosity gap: "Nobody told you this about back pain." Names the topic, hints there's a missing piece, doesn't reveal it.
• Sharp contradiction: "Stop stretching before workouts. Do this instead." Kills a default belief and promises a replacement.
• Named number: "3 mistakes I made getting a Reel to 742K views." Concrete number, concrete outcome, obvious payoff.
• Direct callout: "If you own a clinic in Pune, watch this." Filters the audience so the right person can't scroll past.
• Confession: "I've been running ads wrong for two years." Vulnerability earns attention because it promises a real lesson.
• Question the viewer already has: "Why did your Reels reach drop this month?" Answers a question they were googling that day.

Pick one before you shoot. Don't cycle through three in the first sentence. If you're not sure why your reach fell off recently, I wrote a fuller breakdown in why your Instagram reach dropped that pairs well with hook rewrites.

The mistakes I've stopped making

Most weak hooks aren't unlucky. They're recognisable. These are the ones I catch myself and clients writing, and cut before we shoot. If any slip in, the first three seconds go soft and the algorithm reads it instantly.

• "Hey guys, today we'll talk about…": promises nothing, wastes the whole window, trains the viewer to swipe.
• Introducing yourself first. The viewer doesn't care who you are yet. Deliver value; introduce yourself in second 8.
• Being vague to sound clever. "This changed everything" isn't a hook; it's a horoscope. Say what changed.
• Burying the promise under a joke. Humour works after the hook, not before it.
• Mismatched text and voice. If your caption says one thing and you say another, the viewer feels the friction and leaves.
• Static opening frame. A logo or title card in second one is dead space. Open on motion or a face.

The hook isn't a warm-up. It's the whole ad for the rest of your Reel. Write it like a headline you'd pay to run.

How I write and test a hook before shooting

The hook isn't the last thing I write. It's the first. Before I shoot a single frame, I write 5 to 8 hook options for the same idea on a notes app, read them aloud, and pick the one that would stop me. This costs 10 minutes and it's the highest-leverage step in the whole workflow. If the hook is weak, the best editing on earth won't save the Reel. If it's sharp, even a rough shoot can travel.

1. Write 5–8 hooks for one idea. Force yourself past the obvious first one.
2. Read them out loud. Anything you stumble on won't survive on camera either.
3. Cut anything vague. If it could apply to any topic, it isn't a hook. It's filler.
4. Pick the one with a specific promise. Name a number, a person, a place, a mistake.
5. Shoot two versions. Same body, two different hooks. Post one, keep the other for a rewrite in 30 days.
6. Check retention at second 3 in Insights. That single number tells you if the hook worked, independent of views.

The same discipline applies to writing Meta ad copy. The first line is 90% of the job, and the same tests (read aloud, cut the vague, name a number) apply. If you're a local brand thinking about how this fits your wider content, I broke down the strategy in Instagram marketing for local brands in 2026.

What Painex Clinic's 742K views actually taught me

The Painex account didn't hit 742,000+ views because we spent more or shot fancier. It hit those numbers because we treated the hook as its own deliverable, separate from the rest of the Reel. Every script started with the hook on line one and the rest of the video was written to earn it out. 94% of viewers were non-followers, which only happens when the first 3 seconds are strong enough for Instagram to push the Reel to cold audiences and for those cold audiences to actually stay.

The pattern I saw repeat: Reels with named-number hooks and confession hooks travelled furthest with strangers. Curiosity-gap hooks retained best with existing followers. Neither is universally "the best". It depends on whether you want reach or depth this week. But both beat "hey guys" every single time.

Your one-line rewrite for this week

If you only take one thing from this: open your last three Reels, look at the very first line of on-screen text, and ask honestly, does it promise something specific? If not, rewrite it into one of the six formulas above and repost, or use it on the next Reel. That single change moved the needle for Painex, and it's the change I make first for every client whose Reels aren't travelling. If you'd like help auditing your last month of Reels and rewriting the hooks that are killing distribution, get in touch and I'll go through them with you.

Frequently asked questions

How long should an Instagram Reels hook be?

The hook is the first 3 seconds, visually and verbally. In text, that's a single line, usually 6 to 10 words. In audio, that's one short spoken sentence. Anything longer and the viewer swipes before your promise lands. I write the hook first, in one line, then build the rest of the Reel to deliver on it. If it doesn't fit that window, it isn't a hook yet.

What makes a Reels hook actually stop the scroll?

A hook stops the scroll when it promises something specific the viewer wants and hints they don't already know it. Curiosity gaps ("Nobody told you this about…"), sharp contradictions ("Stop doing X, do Y instead") and named-number promises ("3 mistakes I made at 742K views") work because they're concrete. Vague openers like "Hey guys, today we'll talk about…" fail because they promise nothing.

Should the hook be text on screen, spoken, or both?

Both, and they should say the same thing. About 85% of Reels are watched with sound on, but the text-on-screen hook is what stops thumbs on mute. Match your on-screen caption to the first spoken line so audio-off viewers get the promise and audio-on viewers get reinforcement. For Painex Clinic's Reels, this single change (matching text and spoken hooks) pushed retention noticeably in the first week.

Related guides

• Why your Instagram reach dropped, and how the hook fits into fixing it.
• Instagram marketing for local brands in 2026: the full content system Reels sit inside.