Hold rate answers the second of the two questions every video ad has to pass. Hook rate tells you whether the opening earned attention; hold rate tells you whether what followed deserved it. Read alone, either can send you to the wrong fix.
The pairing is what makes it diagnostic. High hook with low hold means you bought attention the rest of the ad did not pay off — usually a promise in the first three seconds the body never delivers on. Low hook with high hold is the more encouraging failure: the content works, almost nobody is seeing it, and a re-cut opening on existing footage can recover the asset for the cost of an afternoon.
It is also the metric that exposes length problems honestly. A sixty-second ad with a fifteen percent hold rate and a twenty-second ad with a fifty percent hold rate may deliver similar total watch time, but the shorter one is reaching completion — and completion is what most platforms reward in downstream delivery.
Formula
Hold Rate = Completed Views ÷ Views That Started × 100
Denominator matters enormously. Dividing by impressions instead of by views that started conflates hold rate with hook rate and produces a number that moves for the wrong reasons. Keep the two separate or neither is diagnostic.
Why Hold Rate matters
It tells you whether to rewrite the ad or re-cut its opening — two very different costs. Without it, a poor result sends teams into full re-production when the footage may be fine, and it is the clearest available signal that an ad is longer than its argument justifies.
When length is the problem
A forty-five second product film holds eighteen percent of the people who start it, and watch-time data shows a cliff at around twelve seconds — exactly where the ad stops demonstrating and starts listing features. Cutting everything after the demonstration produces an eighteen-second version holding fifty-one percent. Nothing new was filmed and nothing was rewritten; the ad simply stopped after it had made its point.
Benchmarks
- Weak for a short-form ad — the body is not landing
- Under 20%
- Workable for 15–30 second creative
- 30–45%
- Strong, worth building variants from
- Above 50%
Ranges drawn from Digital Squad client accounts and published industry data. Treat them as orientation, not targets — your category may differ substantially.
Common mistakes
Reading it without hook rate
Hold rate on its own cannot distinguish a great ad nobody watches from a mediocre one with a strong opening. The two numbers are only useful together, and teams that track one usually track the wrong one.
Comparing across ad lengths
A fifteen-second ad will almost always out-hold a sixty-second one, which says nothing about quality. Compare hold rate within a length bracket, or you are just rediscovering that shorter videos get finished more often.
Chasing completion at the cost of the argument
You can lift hold rate by cutting an ad to nothing. If completions rise while conversions fall, you removed the part that persuaded people — the metric is a diagnostic, not a target to optimise directly.
Ignoring where the drop-off sits
The aggregate percentage is far less useful than the retention curve. A cliff at a specific second points at a specific edit decision; a gentle slope means the ad is simply long. Only one of those has an obvious fix.
Where we work on this