What hold rate measures
Hold rate is a video ad metric for how well an ad keeps the viewers its opening caught. The common formula is ThruPlays, or 15-second views, divided by 3-second video plays, though some teams divide by impressions instead and others pick a different second mark. Whichever version you use, it isolates the body of the video from the hook.
It is the partner metric to hook rate, which measures how many impressions turn into 3-second views. Hook rate grades the first seconds; hold rate grades everything after. A video can score well on one and badly on the other, and the pairing tells you which half of the edit to fix. Outside advertising the phrase has unrelated meanings; this entry covers the video metric only.
How teams use it
Set hold rate up as a saved custom metric so every variant in a round shares one formula. Then read it in sequence: hook rate first, hold rate second, cost per result last. That order mirrors the viewer's path from stopping, to staying, to acting.
For creator content, low hold rate with a good hook usually means the video promises something in the opening and then wanders. Common causes are a long personal preamble, a product that appears too late, or a demo that repeats itself. A tighter edit from the raw footage costs far less than a reshoot.
Worked example
Two creator videos for a cleaning spray each get about 40,000 impressions. Video A has 12,000 3-second plays and 1,800 ThruPlays, a hook rate of 30% and a hold rate of 15%. Video B has 8,000 plays and 2,400 ThruPlays: a weaker hook at 20%, a much stronger hold at 30%.
The team does not simply pick one. It cuts video A's opening onto video B's body, since B's middle section, a single continuous stain removal shot, keeps people watching. The combined edit becomes the new control for the next creative testing round.
Common mistakes
Mixing definitions across reports is the usual problem. A hold rate built on impressions will always look smaller than one built on 3-second plays, and comparing the two leads to wrong calls. Write the formula into the column name.
The second mistake is judging long and short videos on one scale. A 15-second ad only reaches a ThruPlay by finishing, while a 60-second ad has more room to lose people. Compare within a length band, set targets from your own account median, and draft tighter middles with the script generator.
How it's used
"Hook rate is fine at 31% but hold rate is 9%; the middle drags, so recut before we kill it."
"Our hold rate column is ThruPlays over 3-second plays. Please don't switch it to impressions in the deck."