Video Ad Hook Rate and Hold Rate Benchmarks (2026) — What the Numbers Mean and What to Do Next

# Quick summary Hook rate is the share of impressions that convert into a 3‑second view. Hold rate is the retention curve that follows — percent of those viewers who keep watching at 25%, 50%, 75% and 100%. Both metrics together decide whether your CPMs buy attention or waste it.

# What the benchmarks say (2026)

• Meta: hook rate roughly 25–40%.
• TikTok: hook rate roughly 30–50%.
• YouTube pre‑roll: hook rate roughly 15–30%.
• Instagram Reels: hook rate roughly 30–45%.
• LinkedIn: hook rate roughly 15–25%.

The sharpest viewer loss occurs between the 3‑second mark and the 6‑second mark on every platform.

# Why these two metrics matter

# How to diagnose your bottleneck

• Low hook, low hold: first 3 seconds fail to stop the scroll. Rework opener.

# Concrete editing and production tactics

• Reopen the first 3 seconds: start on action, question, or outcome instead of exposition.
• Deliver a second reason to stay by the 6‑second mark: introduce a new visual, a clear value promise, or an open question.
• Cut faster: reduce long static shots in the first 6–10 seconds.
• Use captions and on‑screen text to communicate the hook without sound.
• Vary camera angle, pacing, or color at the 6‑second point to reset attention.
• Place unavoidable branding after the hook if it competes with the opener.
• Run A/B tests that swap only the first 3–6 seconds to isolate hook performance.

# Platform considerations

• Meta and Instagram: prioritize the 3‑second hook and 25%/50% checkpoints in Ads Manager reporting.

• LinkedIn: plan for lower raw hook rates and optimize messaging for professional relevance early in the ad.

# Verdict Treat hook rate and hold rate as separate, actionable diagnostics. If your campaign underperforms, start by comparing impressions → 3‑second views → retention milestones to find whether the opener or the middle is the problem. Apply tight, time‑bound edits focused on the first 6 seconds, then iterate with small tests.