# Hold Rate

**Category:** metrics  
**Short Description:** Percentage of hooked viewers (3-second video views) who keep watching to the 15-second mark.  
**Last Updated:** 2026-05-29T12:00:00Z

## Definition

Hold Rate measures how well a video ad retains the viewers it has already hooked — the share of 3-second video views that go on to reach 15 seconds (or completion for shorter videos). Where Hook Rate (Thumbstop Rate) judges the open, Hold Rate judges the middle: it isolates whether the body of the ad earns continued attention after the scroll-stopping first frames, normalized to the audience that actually started watching rather than to total impressions.

## Formula

**Formula:** `Hold Rate = (15s Video Views / 3s Video Views) × 100`
**Result Unit:** %

Of the people the hook won (3s views), what share stayed to 15 seconds. Isolates body retention from hook strength.

## Calculation

**Formula:** `(15-Second Video Views / 3-Second Video Views) × 100`

**Explanation:** Divide the number of viewers who reached 15 seconds by the number who reached 3 seconds, then multiply by 100. Using 3-second views (not impressions) as the denominator isolates retention from hook strength: a low Hold Rate means the body of the ad is losing people the hook already won.

### Components

- **15-Second Video Views**: Number of views that reached 15 seconds (or completion for sub-15s videos)
- **3-Second Video Views**: Number of views that reached 3 seconds — the hooked audience

## Industry Benchmarks

| Segment | Typical Range | Median | Notes |
| --- | --- | --- | --- |
| Meta feed, DTC prospecting (15–30s) | 12% – 25% | ~18% | The honest mid-funnel range; the body of the ad must hold viewers the hook already won. |
| Meta Reels (any length) | 18% – 30% | ~23% | Full-screen immersion holds hooked viewers better than feed. |
| Longer video (30–60s), prospecting | 8% – 18% | ~12% | Longer runtimes shed roughly half of hooked viewers before 15s on cold traffic. |
| Retargeting (warm) | 25% – 40% | ~30% | Warm audiences are more patient through the body of the ad. |
| DTC Beauty / UGC, 15–30s | 18% – 28% | ~22% | Demonstration and UGC formats hold attention through the body better than studio shots. |
| Video under 15s (mechanically inflated) | 50% – 80% | ~65% | For sub-15s videos the 15-second threshold resolves on completion, so Hold Rate approaches completion rate and is a weak quality signal. |

**Sources:** Motion Creative Benchmarks 2026, Motion / Foreplay 2025, Triple Whale / Atria 2025, Foreplay 2025, Triple Whale + Common Thread Collective composite 2025, Meta Business Help Center

## Examples

- A Meta video with 10,000 3-second views and 2,000 15-second views has a 20% Hold Rate
- Hook Rate 35% but Hold Rate 9% — the open is winning attention the body immediately loses
- Reels variant holding 28% of hooked viewers vs a feed cut holding 16% of the same audience

## How AdSights Helps

**Tracking Hold Rate:** AdSights plots the full retention curve of every variant and reads Hook Rate and Hold Rate together, so teams can tell a hook problem (low 3-second capture) apart from a hold problem (strong hook, collapse in the body). It pinpoints which scene, voiceover beat, or pacing change between seconds 3 and 15 correlates with the drop, and — because it ties retention to downstream click and revenue outcomes — it flags when a higher Hold Rate is real attention versus merely cheaper, lower-intent views. Teams use this to rebuild the middle of an ad without touching a hook that's already working, and to length-match new briefs to the retention shape that converts.

## FAQs

### What's a good Hold Rate?

On Meta feed for cold DTC prospecting (15–30s videos), 12–25% is healthy, above ~30% is strong, and below 10% means the body of the ad is losing viewers the hook already won. Reels and warm retargeting run higher. Always compare Hold Rate within the same length bucket and audience temperature — the number is meaningless across mismatched contexts.

### How is Hold Rate different from Hook Rate?

They're a paired diagnostic. Hook Rate (Thumbstop Rate) is 3-second views ÷ impressions — it measures whether the open stops the scroll. Hold Rate is 15-second views ÷ 3-second views — it measures whether the body keeps the people the hook captured. A high Hook Rate with a low Hold Rate is the classic 'great open, weak middle' pattern: fix the body, not the first frame.

### Is Hold Rate the same as ThruPlay Rate?

They use the same 15-second threshold but different denominators. ThruPlay Rate divides ThruPlays by total impressions, so it blends hook strength and retention into one number. Hold Rate divides 15-second views by 3-second views, so it factors out the hook and reports retention alone. Use ThruPlay for cost/efficiency framing and Hold Rate when you specifically want to know whether the middle of the ad is the bottleneck.

### There's another definition of Hold Rate — which is right?

Some teams define Hold Rate as average watch time ÷ video length (the average percentage of the video viewed). That version captures overall attention but conflates length and pacing, and it can't separate hook from body. The 15-second-view ÷ 3-second-view definition used here is the more diagnostic one for short-form social video because it normalizes to the hooked audience. Whichever you adopt, define it once and stay consistent across the account — don't mix the two in the same report.

### How do I improve a low Hold Rate?

Because the hook is already doing its job, focus on the body: tighten pacing between seconds 3 and 15, move the value proposition or product reveal earlier, cut dead air or slow scene transitions, and length-match the edit to the attention curve rather than forcing a long cut on cold traffic. Re-shooting the first frame won't help a hold problem — it's a middle-of-the-ad problem.

### Does a falling Hold Rate mean creative fatigue?

It can. If Hook Rate holds steady but Hold Rate declines on the same audience over time, viewers who recognize the ad are bailing once they place it — an early fatigue signal that often precedes CTR decay and CPM creep. Pair Hold Rate trend with frequency and CTR-decay to confirm fatigue before refreshing.

## Related Terms

### Similar Terms

- **[Thumbstop Rate (TSR)](/resources/glossary/metrics/thumbstop-rate-tsr)**: Hook Rate's companion — TSR measures the open (3s views ÷ impressions); Hold Rate measures retention after the hook (15s views ÷ 3s views)
- **[ThruPlay Rate](/resources/glossary/metrics/thruplay-rate)**: ThruPlay is measured against impressions; Hold Rate normalizes the same 15-second threshold to the hooked (3-second) audience
- **[Video Completion Rate (VCR)](/resources/glossary/metrics/video-completion-rate-vcr)**: VCR measures full completions; Hold Rate measures retention to the 15-second checkpoint

### Component Terms

- **[Creative Fatigue](/resources/glossary/creative/creative-fatigue)**: A steadily falling Hold Rate on an unchanged audience is an early creative-fatigue signal

## Related Resources

- [Video Drop-off Rate Calculator](/resources/tools/calculators/video-dropoff-calculator) - Map the retention curve between the 3-second hook and the 15-second hold checkpoint to see exactly which scene is shedding hooked viewers.