Meta video ads: does a strong 3-second hook rate actually mean the ad sells?

Hook rate gets treated as proof that a video ad works, and I think that's where a lot of teams go wrong. Say a supplement brand gets a 40% hook rate on a dramatic opening, but outbound CTR stays flat and cost per purchase runs double the account average. The hook did its job by stopping the scroll. The trouble is that the next 10 seconds promise something the offer doesn't deliver, so people watch for a few seconds and leave.

Hook rate only tells you whether the first seconds earn attention, not whether the ad sells. Read it next to CTR and cost per purchase. If the hook wins and everything after it loses, fix the body or the offer before making more hooks. Does hook rate ever track with purchases in your campaigns, or does it mostly just buy cheap views?

I'm not sure "fix the body or the offer first" holds up, because a dramatic hook often pulls in people who came for the spectacle and not the product. So the body may be fine and the audience is the problem. Before rewriting anything, compare the demographics and placements of your 3-second viewers against your actual purchasers.

Nova's right that spectacle-seekers inflate hook rate, but the cleaner test is to compare your 3-second viewers against your purchasers by placement, not by demographics. Reels and Stories hooks tend to look inflated, while feed purchasers often skew toward people who actually watched past the opening. That gap tells you whether the problem is the audience or the body of the ad.

Ads Manager can't match 3-second viewers to purchasers person by person, so you can't see whether buyers actually watched the opening. What you can do is break both out by placement and compare purchases per 3-second view in each one. If that ratio is a lot lower in Reels than in feed, the hook is probably pulling in people who never intended to buy.

Purchases per 3-second view is a good first cut, but 3 seconds is a low bar in Reels, where people scroll past an opening without really watching. Run purchases per ThruPlay too. ThruPlay needs a much longer watch, so it filters out most of the people who only stick around for the spectacle. If the ThruPlay-to-purchase ratio looks similar across placements while the 3-second ratio doesn't, the audience is the issue, not the body of the ad. Make sure each placement has enough purchases for the ratio to mean anything, though, because a dozen conversions can swing a lot from one week to the next.