Hook Rate & Hold Rate: Diagnose Your First 15 Seconds

1. The two numbers: 3s Hook Rate vs. 15s Hold Rate

Hook Rate = 3-Second Video Plays ÷ Impressions (target: 30%+ on paid social). Hold Rate = 15-Second Plays ÷ 3-Second Plays (target: 25%+). In Fyxer AI's 300-Marble Filter (29s), pouring 300 marbles into a strainer drives the 0–3s thumbstop, while lifting the strainer to reveal 3 urgent marbles bridges into the 15-second UI hold.

Evaluating Hook Rate alone leads to clickbait. Always pair 3-second thumbstop with 15-second hold so your opening visual directly sets up the product demo.

2. High Hook Rate, Low Hold Rate ('Bait & Bounce'): fix the 5s bridge

When Hook Rate is above 30% but Hold Rate drops below 20%, viewers liked your opening visual but bounced during a slow transition. Look at how the All-in-One AI Hot-Dog cut (18s) avoids this: it opens on a ChatGPT-stamped brioche bun (00–04s), layers the model 'toppings' in six seconds (04–10s), and cuts straight into the software UI at second ten.

If your viewers drop off between seconds 4 and 10, delete any introductory throat-clearing and cut directly from the visual prop into the live product action.

3. Low Hook Rate, High Hold Rate ('Hidden Gem'): swap the 0–4s stem

When Hook Rate sits under 20% but Hold Rate exceeds 30%, your product demonstration works—too few people are stopping to see it. In Fyxer AI's 30-Email Showdown (36s), the split-screen race ('PAM [FYXER] VS. JOHN [OLD SCHOOL]') gives the mid-roll demo a high-contrast prizefight opening in seconds 00–05.

Never throw away a high-hold product demo. Keep seconds 05–30 untouched and splice on three new 00–04s visual stems.

4. Sustaining mid-roll hold with rhythmic clicks: YDV's folder UI

The YDV Lookbook cut (18s) sustains hold through a visible 3-step progression: three floating folders ('GREY / BLACK / WHITE') clicked on the beat at seconds 03, 08, and 11.

Give the viewer a visual progress bar—three folders to click, three prompts to test, or 30 emails to clear—so they stay until the sequence finishes.

5. Compressing setup and payoff into 10 seconds: Google Pixel UK

Google Pixel UK's 10-second cut splits runtime down the middle: 00–05s shows the raw handheld pub rig ('the bts') and 05–10s reveals the stabilized final shot ('the result').

When testing cold audiences with short attention spans, a 10-second two-act cut removes mid-roll drop-off entirely.

• High Hook + Low Hold (Bait & Bounce): cut the 04–08s monologue and show the UI earlier
• Low Hook + High Hold (Hidden Gem): keep the 05–25s body and test 3 new 00–04s visual stems
• Low Hook + Low Hold (Full Reset): switch macro format (e.g., split-screen race or cursor UI)
• High Hook + High Hold (Scale Winner): film fresh 00–04s stems before frequency fatigue hits

6. Modular stem testing protocol

Isolate your opening variables cleanly when running creative tests:

• 1. Lock one 15–20s product demonstration body and CTA across all variants
• 2. Film 3 to 4 distinct 00–04s visual formats (prop metaphor, split screen, cursor UI, skeptic)
• 3. Keep the 5-second handoff frame identical so every stem flows smoothly into the body
• 4. Wait for 2,500+ impressions per variant before comparing 3s Hook Rate
• 5. Promote stems that lift 3s Hook Rate without hurting 15s Hold Rate