Hook rate is not click-through rate
Hook rate is who stayed through the open. CTR is who clicked a thumbnail. Feeds score the first; browse and ads score the second.
Hook rate is not click-through rate. Hook rate is the share of people shown a video who are still watching a few seconds in. Click-through rate is the share of people shown a thumbnail, a title, or an ad unit who clicked. One grades the opening of a file that is already playing. The other grades a still that has not started. Mixing them is how a team rewrites a hook that was fine and leaves a thumbnail that was not.
The open is not a click
In a vertical feed the file is the impression. TikTok, Reels, and Shorts autoplay. The viewer does not opt in with a tap on a still. They are already in the first frame. The audition is whether they stay. Hook rate isolates that stay. Three channels land at once: the first picture, the first burned line, the first sound if anyone has it on. A strong visual under a vague caption is a wasted open. A sharp line too small to read at phone size is the same miss.
That is why hook packs exist. Several openings, one body, one variable. You are measuring the first seconds, not the whole cut, not the offer, not the account size. Swipe-away is the inverse event: the thumb replaced you. After 24 August 2026 a YouTube Short that is swiped at second one can still write a public view, because the public counter now starts at the first playback frame. Engaged views kept the old meaning. Studio still computes average view duration and retention on the engaged side. Hook rate sits with that stay family. Public starts do not.
CTR does not live in that window. A Short in the feed did not earn a click to begin. A Reel that autoplays did not earn a click to begin. Treating the like button or the profile tap as "the CTR of the hook" invents a metric the surface does not print. The open either held or it did not. That is the number.
Mute is the default of the test. Burn the claim into frame one. The AI caption generator is for that job: speech type in pixels, not a sidecar a muted scroller will never enable. If the first second only works with sound, the hook rate you are reading is a sound-on minority.
CTR is a different room
Click-through rate is a browse and ads object.
On long-form YouTube, impressions in Studio are thumbnails shown in browse, suggested, search, and related. CTR is clicks divided by those impressions. The still, the title, and the duration badge did the work. The file has not started. A 12% CTR with a cliff at second two is a thumbnail that over-promised. A 2% CTR with a flat retention curve is a still that never got the people who would have stayed. Those are opposite diagnoses. Hook rate cannot see the first one. CTR cannot see the second.
Paid social has its own click. Demand Gen, Spark ads, and in-feed units print CTR on the destination tap: shop, swipe-up, watch more. That click is an ask. It is closer to a CTA card than to an opening. Front-loading the shop line into second one will often raise CTR on the unit and crush hook rate on the organic twin of the same file, because the card covered the reason to stay. Do not read the ad CTR as proof the open is good. Read the three-second hold on the same creative.
Search CTR is a third room again. A how-to that ranks can have a high title CTR and a weak open if the first thirty seconds are a greeting. YouTube long-form still has a player. People who clicked expected a watch. Hook rate on that surface is the early slope of the retention curve, not the search click that got them there.
Do not average these CTRs. Browse CTR, ad CTR, and search CTR do not share a denominator with each other, and none of them share a denominator with hook rate.
Why teams glue the two together
Dashboards sit next to each other. A weekly sheet with "CTR" in one column and "3-second views" in the next invites a single story. The story is usually wrong.
A creator sees Shorts public views up and CTR down, and recuts the open. The CTR that fell was long-form browse CTR on a different file. The Shorts number that rose was first-frame starts after the August 2026 cutover. Neither number was hook rate.
A media buyer sees a high outbound CTR and scales the cut. Organic hook rate on the same opening is poor, because the paid unit had a product tile the organic file does not have. The click was the tile. The open was never tested.
A brand sees a low hook rate and changes the thumbnail. There is no thumbnail in the Reel. The low number was the first caption line sitting in chrome. Safe area is that collision. CTR will not move when you fix it. Hook rate will.
The honest split:
- If the surface autoplays, score hook rate (and swipe-away) on the opening window you named in advance. Two or three seconds is a common window. The exact second matters less than using the same one on every variant.
- If the surface is a still that must be clicked, score CTR on that still. Then, separately, score hook rate on the people who arrived.
- If public views rose and engaged views did not, you were exposed and skipped. That is a hook story, not a CTR story.
Captions for silent autoplay is the craft that feeds hook rate. CTR does not care whether the first line is burned. Hook rate does.
Measure the opening on its own
Pre-commit the metric before either variant goes up.
For a feed test, the metric is hook rate or the platform's viewed-versus-swiped number at a fixed 24-hour checkpoint. Views are the lottery. Likes write after a stay. Comments write after a stay. CTR, if the platform even prints it on that post, is often a profile tap or a link tap — a different job. Scoring hooks at 24 hours is the timing rule, because a promotion window is short and creative fatigue is real.
For a YouTube long-form test, print two rows: browse CTR, and the percentage remaining at 15 or 30 seconds. Do not let a winning thumbnail hide a losing open. Do not let a winning open hide a still nobody clicked.
For an ad test, print outbound CTR next to three-second view rate, not instead of it. The buy may be CPM. The creative still has a hook. CPM vs CPV is the buying model. Hook rate is still the open.
Generation makes it cheap to test the wrong thing. Five opens that are the same smile with a different adjective will not move hook rate. Change the picture. Burn a different claim. Watch the first second on mute at phone size. If you would swipe, the feed will. The UGC video generator is for that pack, not for a fifth synonym.
Hook rate is not click-through rate. Stay versus click. Feed versus still. Keep the columns apart.
Keep the open measurable in Versely.
FAQ
Can I use CTR as a proxy when hook rate is missing?
Only on surfaces where a click is required to start the file, and only as a proxy for the still. Once playback starts, you still need an early-hold number. A homemade "CTR" on a Reel is usually a profile tap. That is not the open.
Why did my Shorts views rise while CTR fell?
They may not be about the same object. Shorts public views after 24 August 2026 include first-frame starts. Studio CTR on long-form is thumbnail clicks. Put engaged views and hook rate in the report before you recut either file.
Does a high ad CTR mean the hook is working?
It means the unit earned a tap. The tap may be the destination button, the product tile, or the headline, not the first three seconds of picture. Read three-second hold on the same creative before you scale the open.