The battlefield your ad enters
Scroll. Scroll. Skip.
That's three people who never saw your ad. The feed is brutal because leaving costs nothing. A hook is the one thing that makes someone stop, and hook analysis is the practice of figuring out why some openings work and most don't.
What is a hook?
The first 1-5 seconds of a video. The first frame of an image. The first line of copy.
Its job is to interrupt the scroll before the person even decides they want to stop. Good hooks work before the audience is paying attention. That's the part most ads miss entirely.
They show up as a striking visual, a statement that doesn't immediately add up, a problem that hits close to home, something that breaks the visual rhythm of everything else in the feed, or a question that doesn't resolve fast enough to ignore.
Why the hook is actually a distribution mechanic
Facebook and TikTok reward fast engagement. If people react in the first two seconds, the ad gets pushed to more people. If they don't, it quietly dies.
That means your hook isn't just a creative decision, it's a distribution decision. A stronger hook means lower CPMs, more reach, and more conversion attempts per dollar. Teams that study their hooks seriously tend to outspend competitors on the same budget without realizing that's what they're doing.
The five hook types
1. Problem-first
"If you're struggling with X, this is for you."
Works because people are primed to notice their own unsolved problems. The mistake most brands make here is keeping the problem vague. "Tired of low energy?" hits nobody. "Your 3pm crash isn't about sleep, it's about what you ate at lunch" hits a specific person.
Check: is the pain visible within 2 seconds? Does it name a specific scenario, or is it broad enough to apply to anyone, which means it applies to no one?
Watch: thumb-stop rate, 2-second view rate.
2. Pattern interrupt
Unexpected visuals, strange framing, silence when everything else is loud, visual chaos when everything else is clean. The feed has a rhythm and these break it.
Mute test: if the first frame doesn't stop you with the sound off, nothing is saving it. Also worth asking whether the interruption reads as intentional or just incoherent, because incoherent looks like bad production rather than a hook.
Watch: impressions to views ratio.
3. Curiosity
"Nobody talks about this..." / "This mistake cost me $50,000."
Opens a loop people want to close. The risk is that vague curiosity reads as clickbait and trained audiences skip it reflexively. Specific curiosity ("the one landing page element costing you 30% of conversions") holds better than generic mystery.
Check: is the payoff teased early enough to earn the wait? Does it feel like something real is being withheld, or does it feel like a trick?
Watch: 3-5 second retention, hold rate.
4. Proof-first
Show the result before the explanation. Lead with the screenshot, the before/after, the number.
The problem with most proof hooks is that the proof isn't immediately readable. Viewers scroll before they register what they're looking at. Make the number or result the first thing the eye lands on, not the brand name or a setup sentence.
Check: is the proof legible at a glance? Does it feel documented or does it feel staged?
Watch: watch time, saves, shares.
5. Relatability
Raw, messy, human moments. The opposite of a brand ad.
An imperfect video of someone talking to their phone often outperforms polished production because it looks like content rather than advertising. If it would make sense as an organic post, it's probably doing this right.
Watch: comments, profile clicks.
How to actually run hook analysis
Start with the first 3 seconds, muted. Loop it. If nothing is happening visually, the hook is doing no work. Most people never mute-test their own ads.
Label every creative. Pick one type from the list above and assign it. Do this across 20-30 ads and patterns appear fast. You might find you've been running 80% problem-first hooks with no read on whether curiosity even works for your audience.
Run hook variations in isolation. Same offer, same CTA, different openings. This is the only clean way to separate hook performance from everything else. If you change the hook and the offer at the same time, you learn nothing.
Then map to metrics:
- Low CTR: the hook isn't earning the click
- High views, low retention: the hook promised something the ad didn't deliver
- High engagement, low conversion: you're attracting the wrong audience
Mistakes worth knowing
Opening with your logo or brand name. Nobody stops scrolling for a name they don't recognize.
Explaining too much in the first five seconds. The hook's job is to create a reason to keep watching, not to summarize the whole ad.
Generic openers. "This will change your life" gets filtered automatically. The brain has seen it too many times.
Designing for desktop. If the first frame doesn't work at 9:16 on a phone, most of your impressions aren't seeing it right.
Hooks as a library, not a lottery
The teams that consistently outperform aren't necessarily more creative. They're more systematic. They tag every hook they run, log what worked, and build swipe files they can pull from and recombine. A strong hook from six months ago often becomes a new winner when paired with a different offer or a fresh audience.
"Shock," "confession," "demo-first," "social proof cold open" -- these become shared vocabulary. Once your team can name what they're making, creative testing stops being random.
One last thing
Most ad post-mortems look at targeting, bidding, or the offer. The hook gets treated as an aesthetic choice, not something you can systematically test and improve.
It's probably the highest-leverage variable in the whole ad. Worth treating it that way.