The hook library: first three seconds, catalogued

The problem this solves

The first three seconds decide whether a video ad exists. Feed platforms bill on attention, viewers grant it in fractions of a second, and everything after the hook only matters to the audience the hook kept. Ad teams know this arithmetic, which is why the hook is the most tested element in their creative - and why the hooks running in your category right now are a map of what the market has already paid to learn.

That map is public and almost nobody reads it, because reading it means watching the opening seconds of dozens of ads attentively enough to notice the pattern: what appears first, what is said first, what text lands on screen, whether the product or the problem leads. Done by hand it is an afternoon of squinting at a phone; the notes decay immediately as creatives rotate.

This mission reads the whole map in one pass. It collects the current video ads in your category from the ad libraries, analyzes the first three seconds of each - shown, said, and written, quoted exactly - and builds the catalogue: every hook pattern, who runs it, how common it is, and where the untested openings are.

How the mission runs

1. Collect the running video ads. The ad libraries supply the current video creatives across your category - competitors and adjacent advertisers - with direct media files. The set is deduplicated so each distinct creative is analyzed once.
2. Analyze every opening. Each ad's first three seconds get the dense read: a frame burst across the opening, the first spoken words, and every piece of on-screen text exactly as rendered. The hook is treated as its own artifact, separate from the ad behind it.
3. Classify the patterns. The openings are sorted into named patterns: problem-first, product-in-hand, bold-claim text card, founder-to-camera, before-and-after, price-led. Each pattern carries its examples, so the label always points back at real ads.
4. Read the distribution. The catalogue shows what dominates and what is absent: which patterns half the category leans on, which appear once, and which plausible hooks nobody is running. Saturation and white space are both findings.
5. Recommend the three tests. The mission closes with the three hooks worth testing for your brand: grounded in the catalogue, differentiated from the saturated patterns, and written as concrete opening descriptions a creative team can shoot or generate this week.

The prompt

This is the exact objective the agent receives. Swap the obvious placeholders for your own domain, segment or channel and run it as-is from the console, Slack, or the API.

What comes back

A hook library for your category: every current video ad's first three seconds decoded and quoted, the openings classified into named patterns with prevalence and examples, delivered as a CSV catalogue plus a short read on the white space - ending with three concrete hook recommendations ready for testing.

Make it yours

• Rerun monthly and track pattern drift: hooks rotating out of the category are as informative as the ones arriving.
• Split the library by platform, since the hooks that survive TikTok pacing differ from the ones that open Facebook feed ads.
• Chain it with the reference-to-brief playbook: pick the recommended hook, feed its best example through the decode, and walk away with the brief.

Frequently asked questions

Go deeper

• Ad library surveillance: reading the market’s paid conversation
• The creative intelligence loop: from ad libraries to launched tests
• The ad creative testing cadence that compounds