Hold rate: the number, and what it hides
The share of people who stayed to the end once the ad had hooked them.
Why hold rate matters
Hook rate and hold rate fail for different reasons and need different fixes. A strong hook with a weak hold is a pacing or payoff problem in the middle. A weak hook with a strong hold means the ad is good and nobody is seeing it.
What this looks like with real numbers
An ad gets 20,000 three-second plays and 3,000 completions: a 15% hold rate. A second ad in the same set holds 28% but only hooked 8,000 people. The second is a better ad trapped behind a worse opening, so re-cut its first two seconds rather than starting again.
What it does not tell you
Longer ads will always hold worse than short ones. Compare like with like, or you will conclude that 15-second ads are better than 45-second ads when you have only measured their length.
The mistake people actually make
Comparing hold rate between a 15-second ad and a 45-second one, then concluding short ads are better. You have measured length, not quality.
Questions, answered.
How do you calculate hold rate?
Video completions ÷ 3-second video plays. The share of people who stayed to the end once the ad had hooked them.
What goes wrong with hold rate?
Comparing hold rate between a 15-second ad and a 45-second one, then concluding short ads are better. You have measured length, not quality.
What does hold rate not tell me?
Longer ads will always hold worse than short ones. Compare like with like, or you will conclude that 15-second ads are better than 45-second ads when you have only measured their length.
Change the creative, not the dashboard
The share of people who stayed to the end once the ad had hooked them.