Retention curve: the number, and what it hides
The graph of how many viewers are still watching at each second of the ad.
Why retention curve matters
The shape tells you where the ad breaks. A cliff in the first two seconds is an opening problem. A steady slope is a pacing problem. A cliff in the middle is usually the moment the ad starts selling.
What this looks like with real numbers
A curve that sits flat until second seven and then falls off a cliff almost always means second seven is where the ad starts selling. Either the transition is too abrupt, or the promise made in the opening was not the promise the ad went on to keep.
What it does not tell you
A drop is not automatically bad. Ads that filter hard early — naming the price, naming the audience — lose viewers on purpose and can still be the most efficient thing in the account.
The mistake people actually make
Treating every drop as damage. An ad that names its price at second three loses viewers on purpose and can still be the most efficient thing in the account.
Questions, answered.
How do you calculate retention curve?
Viewers remaining at second N ÷ viewers at second 0. The graph of how many viewers are still watching at each second of the ad.
What goes wrong with retention curve?
Treating every drop as damage. An ad that names its price at second three loses viewers on purpose and can still be the most efficient thing in the account.
What does retention curve not tell me?
A drop is not automatically bad. Ads that filter hard early — naming the price, naming the audience — lose viewers on purpose and can still be the most efficient thing in the account.
Change the creative, not the dashboard
The graph of how many viewers are still watching at each second of the ad.