A good hook names the viewer's problem or promises a specific payoff within the first three seconds, working visually and verbally at once. The strongest hooks open a loop the brain needs closed. Get this right and average content performs; get it wrong and brilliant content dies unseen.
We script and shoot hooks for client ads every week, often capturing five openings for a single video body. Here is what consistently separates winners.
Because that is the decision window platforms give you. Feeds judge your video on early engagement, viewers judge it faster still. The hook effectively buys the rest of the video its audience. This is why two identical ads with different openings can show dramatically different results all the way down to cost per acquisition: the hook changes who keeps watching, which changes everything downstream.
Three layers firing simultaneously. The spoken line makes a claim or asks a question. The on-screen text reinforces or twists it. The visual shows something worth watching: motion, a result, a face reacting. Single-layer hooks (a static logo with a voiceover, say) waste two of the three channels. Specificity beats cleverness every time: "this took our edit time from 6 hours to 90 minutes" outperforms "work smarter, not harder" in almost any test.
Seven formats earn their place in nearly every account we touch: the named problem ("if your espresso never tastes right, this is why"), the bold claim with a number, the open question, the result shown first, the contrarian take ("hiring an agency? listen to this first"), the curiosity gap ("nobody talks about this part of padel") and the direct callout of the audience ("founders who hate being on camera"). Rotate formats rather than wearing one out.
All three at once where possible. Sound-off viewers need the text and the visual; sound-on viewers bond with the voice. If forced to choose, prioritise the visual plus text combination because most feed viewing starts muted. Whatever the mix, the meaning must land inside three seconds without any single channel doing all the work.
Keep the body constant, vary only the opening. Three to five hooks per body is the practical sweet spot, as we covered in how many ad creatives to test. Shooting them costs minutes when planned: the same speaker delivers five different first lines back to back, then editing does the rest. If your ads are underperforming, start the diagnosis at the hook with our guide to why video ads fail.
Write ten, shoot five, run three. Writing is free, so over-generate then cull. The discipline of producing ten forces you past the obvious first ideas, which are usually the ones every competitor also wrote.
We bank hooks for clients in every shoot, which is one reason a single Content Sprint feeds an ad account for months. Want a library of openings worth testing? Let's talk.
Hook rate is 3-second video plays divided by impressions, showing what share of people who saw your ad actually started watching. It is the cleanest signal of opening strength: strong performers sit around 25 to 30 percent on Meta placements.
Yes, when the question is specific and the viewer feels personally addressed. "Why does your coffee taste better at the cafe?" beats "want better coffee?" because it names a real moment. Generic questions invite a scroll; precise ones open a loop.
For direct response, usually not; lead with the viewer's problem rather than your logo. For brand awareness campaigns, early branding has value. The honest test is performance: run both and let hook rate plus conversion data decide.
Whenever frequency climbs and CTR slides, typically every four to six weeks on active accounts. Refresh the opening before replacing the entire creative; a new hook on a proven body is the cheapest performance lift in paid social.
.avif)
How much content will I get?
What content can you create?
Where can content be shot?
What if I need revisions?
Never done a Content Sprint?
Unhappy with the content?
Need models, locations or props?
.svg)
.svg)