Ad Hooks: Writing the First Three Seconds

Most of the money you spend on paid social gets decided before anyone hears your offer. It happens in the first three seconds, when your ad hooks either stop the scroll or get skipped. We've run enough accounts to say this plainly: a mediocre offer with a great hook beats a great offer with a weak one, almost every time. Thumb-stop rate and three-second view rate quietly gate everything downstream, so ad hooks are where we put a disproportionate share of our creative effort.

This isn't about being clever. It's about earning the next two seconds, then the next ten.

Why the first three seconds decide the rest

The math is unforgiving. On Meta and TikTok, a big chunk of impressions never make it past three seconds. If your hook loses 70 percent of viewers up front, then your body copy, your social proof, and your call to action are speaking to a tiny remainder. You paid for the full audience anyway. That waste shows up as a high CPM relative to results, which is really the algorithm charging you for attention you failed to hold.

Strong hooks also change how the platform prices you. Higher hold rates feed better engagement signals, effective CPMs come down, and the algorithm widens the audience it's willing to show you. So a good hook isn't only a creative win. It's a media-buying one, and it's one of the first things we reach for when the goal is lowering CPA without touching the offer.

Hook types that actually earn attention

There's no master list that works forever. But a handful of hook shapes keep performing across accounts. Treat them as starting points, not scripts to copy word for word.

• The callout. Name the person you want. "If you run Meta ads and your CPA doubled this quarter, watch this." It filters, which feels backwards, and yet a narrow callout beats a wide one because the right person leans in.
• The contrarian claim. "Stop split-testing audiences. It's wasting your budget." You're picking a small fight with conventional wisdom, and disagreement holds attention.
• The visual pattern break. Something on screen that doesn't belong in a feed: a product used the wrong way, an odd close-up, motion in the first frame before any words land.
• The problem snapshot. Show the exact frustration in frame one. Tangled cables. A cluttered inbox. People recognize their own problem faster than they parse a benefit.
• The result tease. Open on the outcome, then rewind. "This ad set spent 4,000 dollars and returned 18,000. Here's the structure." Curiosity does the work.

The mistake we see most is a hook that describes the brand instead of the viewer. Nobody stops for "We're a premium skincare company." They stop for "You're washing your face wrong, and it's why nothing sticks."

Mine reviews and support tickets for angles

The best ad hooks are rarely invented at a desk. They get lifted, almost verbatim, from the language customers already use. We start every creative sprint by reading, not writing.

Product reviews are the obvious source, and the three-star ones are gold because they mix praise with a real objection. Support tickets and chat logs are better still. That's where people describe their problem in their own unpolished words, before any marketing has sanded it down. Sales call recordings, Reddit threads, and the questions section on competitor listings all work the same way.

What we're hunting for is specific phrasing. A customer who writes "I gave up on retinol because my skin peeled for a week" just handed you a hook. So did the person who wrote "I stopped running ads myself because I couldn't tell which campaign actually made money." That second one is an attribution problem, and it connects to how we think about measuring what actually drives revenue. Copy the emotion and the specificity, not just the topic.

A quick way to run this:

• Pull 50 to 100 reviews and tickets into one doc.
• Highlight any phrase that names a fear, a failed past attempt, or a moment of relief.
• Cluster the highlights into three or four recurring angles.
• Write five hooks per angle, using the customer's words where you can.

Test hooks in isolation, not buried in the ad

Here's where most testing goes wrong. Teams change the hook, the middle, the offer, and the CTA all at once, then can't tell what moved the number. If you want to learn anything about ad hooks specifically, you have to isolate them.

The clean version: hold the body of the creative constant and swap only the first three seconds. Same product footage, same voiceover after the opening line, same end card. Now the three-second view rate and hold rate tell you something real about the hook alone. We treat this as a core rule inside our creative testing framework, and it pairs with how we approach account structure on Meta so tests don't cannibalize each other.

Watch the right metric at the right stage. Hook quality shows up first in thumb-stop rate and three-second views, not in ROAS, so judge it there before you judge it on revenue. A hook can win on hold rate and still lose on conversions, which usually means the hook wrote a check the rest of the ad couldn't cash. That's useful too. It tells you the promise and the payoff are out of sync, and the fix lives in the body, not the first frame.

Give each hook enough impressions to mean something. A few hundred views is noise. We want a few thousand before we call a winner, and we kill the clear losers fast so budget flows to the variants doing the work.

Where to start this week

Pick one product. Read 50 reviews and 50 support tickets, pull the ten phrases that make you wince because they're so honest, and turn them into ten hooks against the same base creative. Run them isolated, watch hold rate first, and cut ruthlessly. You'll usually find one or two that beat your control by a margin that pays for the whole exercise.

If you'd rather have a second set of eyes on it, we run a free ad account audit where we pull your current creative, look at hold rates by ad, and show you where attention is leaking. It's the fastest way to see whether your hooks or your bodies are holding you back.

FAQ

How long should a hook actually be? Three seconds of screen time, usually one spoken line or one strong visual beat. If the viewer can't tell what's in it for them by second three, it's too long or too vague.

Should the hook match the rest of the ad, or just grab attention? It has to match. A hook that overpromises will lift your three-second view rate and tank your conversion rate, because the body can't deliver what the opening sold. Grab attention with a promise you actually keep.

How many hooks should we test at once? Enough to learn something without splitting budget into dust. We usually run four to six hooks against one base creative, give each a few thousand impressions, and cut the losers quickly.