Stop writing hooks as one line. Build all four layers

Your team spends an hour on the opening line of an ad. Then an editor picks the first frame, someone adds a trending sound, and someone else types the text overlay at 6 p.m. on a Friday. The line was strong, and the viewer scrolled past before the creator said it.

A hook has four layers: the first visual, the text overlay, the spoken line and the audio. They reach the viewer in about the same second, and the viewer reads them as one signal. If you write the spoken line and leave the rest, you leave three of your four layers to luck.

The four layers, in the order a viewer meets them

On Meta and TikTok, a viewer decides to stay or swipe within the first one to three seconds [VERIFY]. Many of them watch that opening with the sound off [VERIFY]. For those viewers, the line most scripts treat as the hook arrives third, behind the picture and the words on screen.

1. First visual. The frame and the movement in it. It lands first and works with or without sound.

2. Text overlay. The words on screen. For a sound-off viewer, this layer does the work of the script.

3. Spoken line. The first sentence out of the creator’s or founder’s mouth.

4. Audio. Music, sound effects, room tone and the tone of the voice itself.

Your script covers layer three, but your results depend on how all four fit together, whether they agree or clash on purpose.

A sound-off viewer meets your hook as a picture and a line of text. The spoken line you spent an hour on reaches them third.

Layer one: the first frame stops the thumb

Pause your ad on frame one and look at it as a still image. If it shows a creator’s face in a bathroom with soft daylight, it looks like thousands of other ads in the same feed.

Strong first frames show something mid-action or out of place: a hand pouring a drink that’s the wrong color, or a gym bag dumped upside down on a kitchen counter. You put a small question in front of the viewer, and they stay long enough to get the answer.

Brief the frame with the same care you give the line. Write down what’s in it, what moves in the first second and how close the camera sits. If you write “Creator holds product,” your editor has nothing to work with. Write “Tight shot, creator drops the jar into a full sink, splash on frame one” and they have a shot list.

Layers two and three: split the work between text and voice

Teams often write a caption that repeats the spoken line word for word, which means you paid for two layers and used one of them. Viewers read faster than creators talk, so they finish the overlay first and then hear the voice repeat what they already know.

Give the overlay and the voice different jobs. The text sets the scene, the voice brings the tension.

Give each layer its own job. One split that works well:

• The overlay sets the context. The audience or the situation. “Your 3 p.m. crash, explained.”

• The voice delivers the tension. A confession or a claim that sounds wrong at first. “I cut coffee after lunch and my afternoons got worse.”

The sound-off viewer now has a reason to stay, and the sound-on viewer gets two pieces of information in the same second. Aim for a setup and a payoff, or a question and a hint at the answer.

Keep the overlay to five to eight words, about what a viewer reads in a second [VERIFY]. On TikTok and Reels, place it above the bottom fifth of the frame so the platform interface doesn’t cover it [VERIFY].

Layer four: audio sets the mood before the words land

Teams pick audio last and often swap it at the last minute. You can pull a trending sound from the library, but most of those sounds say nothing about your product. Treat audio as part of the hook and make two decisions on purpose.

• The first sound. A pop, a pour, a zipper or a sharp breath. Put a sound effect on frame one to pair with the visual and give sound-on viewers a second reason to stay.

• The delivery. Whisper, rant or deadpan. Your creator can make the same line land as a secret or a complaint depending on tone, so direct them on it and ask for two takes in different registers.

If music runs under the voice, cut it for the first line or keep it low enough for the words to come through. When you’re briefing creators for UGC content, put the first sound in the brief next to the line.

Brief all four, then test hooks that differ

Once your team writes all four layers, change how you test. If you swap one spoken line and keep the same frame, overlay and sound, Meta sees two ads that look almost identical. Platforms group near-duplicate creative more and more [VERIFY], so you learn little from that kind of swap.

Test hooks that differ across several layers, like these two for a sleep supplement, the kind of test we run with wellness brands:

• Hook A. Frame: alarm clock reading 3:12 a.m., phone light on a tired face. Overlay: “If you wake up at 3 a.m., read this.” Voice: “I tried everything for six months before I changed one thing at dinner.” Audio: a phone buzz, then silence.

• Hook B. Frame: close-up of powder stirred into warm milk, steam rising. Overlay: “The nightly drink I switched to.” Voice: “My partner asked why I stopped scrolling in bed.” Audio: a spoon clinking on ceramic.

A real hook test changes the frame, the overlay, the voice and the audio together, so the platform treats it as a new ad.

Each angle changes all four layers. Launch both, keep the winner and build on it by putting the same idea into new formats and handing it to new creators before you retire it. Teams often drop a winning angle too soon, so keep it running until results fall and change the layers around it in the meantime.

Choose the faces in those frames with care too. Meta has signaled that it expects creators to do more of the selling in paid media [VERIFY], so protect real faces and real hands on frame one in your briefs. Our creator ads team builds the four layers into the brief before anyone films.

Run this five-step audit on your account this week:

1. Pull your top five ads by spend from the last 90 days.

2. Screenshot frame one of each and write out all four layers next to it.

3. Flag any layer that repeats another or could appear in a competitor’s ad.

4. Rewrite the weakest layer on your top two ads and launch them as new creative, separate from the originals.

5. Add a four-layer hook section to your creator brief template so the next round starts complete.

If you want a second set of eyes on the brief, our creative direction team starts there and writes down every layer with you.

What to do next

Start with step one today. It takes about 30 minutes, and most teams find at least one ad where the overlay and the voice say the same thing.

If you’d rather hand the work to operators, book a free audit at commnty.com/booking/free-audit. It takes 30 minutes and there’s no sales pitch. We open your account, go through it line by line and show you what’s working and what isn’t.