Hook Rate: Formula, Meaning, Fixes

Hook rate explained: the formula for Meta, TikTok and YouTube, why benchmarks mislead, and the production fixes that lift the first seconds of a video ad.

Published 2026-10-30 · Video Marketing · Neverframe Team

Most video ads are judged before the product shows up. On YouTube the skip button appears at five seconds, on TikTok a view is counted at two, and on Meta the standard reporting threshold is three. Hook rate is the number that tells you how many people made it past that first gate.

In plain terms, hook rate is the share of impressions that turned into a short view in the opening seconds of a video ad. The usual formula on Meta is 3-second video plays divided by impressions, multiplied by 100. On TikTok the closest equivalent uses 2-second video views. On YouTube the skippable in-stream format makes the five-second mark the natural cut-off. A low hook rate means the first frames are losing people. A high hook rate with weak results further down means the opening works and the middle of the ad does not.

What Is Hook Rate?

Hook rate is an advertiser-built ratio. None of the big ad platforms reports a column with that exact name by default, which is why you see it computed in spreadsheets, custom columns and reporting tools. The ratio takes the metric each platform uses for an early view and divides it by impressions.

On Meta the numerator is the 3-second video play. Jon Loomer's glossary of Meta advertising terms gives the definition used by most media buyers:

"A 3-second video play is recorded when someone views a video for 3 seconds or views 97% of a video that is less than 3 seconds long."

The same metric is what data connectors pull from the API. The Fivetran documentation for its Facebook Ads connector says it plainly:

"3-second video views are available through Facebook's Graph API using the video_view metric in the action_type field."

Hook rate therefore covers a narrow moment: the decision to keep watching once the ad starts moving in the feed. It says nothing about whether the viewer understood the offer, remembered the brand, or clicked. Those questions belong to other metrics, which come later in this guide.

The term comes from performance marketing teams who run many variations of the same ad and need a fast signal on the opening. Our guide to performance creative for video ads explains how that testing culture grew. Hook rate is one of its simplest tools, and one of the most misread.

The Hook Rate Formula

The formula fits on one line:

Hook rate = early video views / impressions x 100

What changes from platform to platform is the definition of an early view. Here is how the numerator maps across the three channels most brands buy.

A quick worked example. A Meta ad gets 40,000 impressions and 10,000 3-second video plays. Hook rate is 10,000 divided by 40,000, so 25%. The same creative on TikTok might show 12,000 2-second views on 40,000 impressions, a 30% hook rate. Those two numbers cannot be compared directly, because the second one uses a lower bar. Keep that gap in mind whenever someone quotes a hook rate at you.

Hook rate formula searches often ask for a "hold rate" too. Hold rate is the companion ratio that measures how many hooked viewers stayed longer. Teams compute it in different ways: ThruPlays divided by impressions, 15-second views divided by 3-second plays, or a quartile metric divided by early views. Whatever definition you choose, it has to stay the same in every report. Changing the denominator halfway through a quarter destroys the trend line.

How Each Platform Counts the First Seconds

TikTok: two seconds, one session

TikTok publishes its definitions in the Ads Help Center, in an article on video play metrics last updated in November 2025. The two-second view is the base layer:

"Number of times your video starts playing for at least 2 seconds in an impression session. Replays are excluded."

The session logic is worth reading closely, because it affects the denominator as well:

"If the person swipes away from the video and swipes back, it is counted as 2 impressions, and the new video view will be counted again with the new impressions session."

TikTok also reports a six-second view with a broader trigger. The help center describes it as the

"Number of times your video starts playing for at least 6 seconds in an impression session, played in full if it is less than 6 seconds, or received at least 1 engagement within the first 6 seconds."

That engagement clause turns the six-second view into a mixed count of watch time and interaction, since a like in second four counts. TikTok goes further for advertisers buying focused view products, where the definition reads: "The number of times your video was viewed for at least 6 seconds, or viewed in full if it is less than 6 seconds long, or the person initiates any interactions within the first 6 seconds." The listed interactions include "Like, Follow, Share, Click, Hashtag Click, Music Click, Anchor Click, and Interactive Add-on Activity Click." For hook rate, most TikTok buyers use the two-second metric and treat the six-second metric as an early hold signal. Brands working with creators on boosted posts face the same counting rules; our TikTok Spark Ads guide covers how those placements are bought.

YouTube: the five-second skip

YouTube does not have a three-second column. The structure of the skippable in-stream format does the job instead. The Google Ads Help page on video ad formats says:

"Viewers can skip the ad after 5 seconds."

Billing follows a different clock. On the same page Google explains:

"With CPV bidding, you pay when a viewer watches 30 seconds of your video, or the entire duration of the video if it's shorter than 30 seconds, or interacts with your video, whichever comes first."

So the hook on YouTube has two jobs. It has to survive the skip button at five seconds, and it has to carry enough interest for the view to count at thirty. Google's own reporting page on view metrics names the drop pattern directly:

"For example, it's common to find a steep drop off of users after 5 seconds for in-stream ads when the skip button appears for users."

Google also connects the view rate to the opening of the creative:

"A low TrueView view rate might suggest that the initial portion of your creative wasn't compelling enough."

The same page sums up what the view rate tells a creative team in one sentence: "It attests to how compelling your ad is." In other words, the ratio of views to impressions is read by Google itself as a verdict on the creative.

For YouTube, the practical hook rate is the share of viewers still watching after the skip point, which you read from audience retention in YouTube Analytics and from quartile reports. Google's own format table describes the skippable in-stream ad in one line that doubles as a definition of hook rate: "Users can skip so you can understand attention". Shorts ads follow a separate rule, and the swipe logic runs the opposite way to TikTok's: "If a viewer sees your ad, swipes away, and then swipes back to the exact same ad in the feed, it does not count as an additional impression or view." In-feed placements have their own trigger too, with a view counted when "A viewer watches the video autoplay inline for at least 10 seconds." Google counts a Shorts ad view when "A viewer watches 10 seconds or until the end of the video, whichever comes first." Our YouTube ads production guide covers the format choices in more depth, and the pre-roll ads guide explains where skippable placements sit next to the other in-stream options.

Meta: three seconds, with or without sound

On Facebook and Instagram, the 3-second video play is the default early metric and the one most hook rate dashboards are built on. Feed placements autoplay, often without sound, so the first three seconds are almost always watched on mute. That is a production constraint as much as a reporting one. If the opening depends on a spoken line, a large part of the audience never hears it.

Hook Rate Is Not Viewability

A common mistake is to treat hook rate and viewability as the same thing. They measure different events. Viewability asks whether the ad had a chance to be seen. Hook rate asks whether the person chose to keep watching.

The industry standard for video viewability comes from the Media Rating Council. When the MRC extended its guidelines to mobile, its announcement stated:

"A mobile Viewable Impression requires a minimum of 50 percent of pixels in view for one consecutive second for display and two consecutive seconds for video."

The same guidelines acknowledge how fast feeds move. For display ads in mobile newsfeeds, the MRC allows separate reporting of "Sub-Second Impressions," described as "mobile (display only) ad impressions that appear in a newsfeed where 50 percent or more of the ad's pixels are in continuous view for more than 0.5 seconds but less than the one second (post-render) minimum threshold for a Viewable Impression." Feeds are judged in fractions of a second, and video openings live under the same pressure.

Google's Campaign Manager help page applies the same rule to video:

"Video ads are considered viewable when at least 50% of the ad is visible while the video is playing for at least 2 consecutive seconds."

Notice how close the two-second viewability rule sits to TikTok's two-second view. They are still different counts. A viewable impression can happen while someone scrolls past with half the ad on screen. A two-second view in a full-screen feed is closer to an active choice. When a client report mixes the two, the hook rate looks better than it is.

On that last row, Google's view metrics page adds a useful line: "Unlike views, engagements are counted after 10 seconds of viewership." Between the five-second skip and the ten-second engagement mark sits the stretch where a hook either turns into attention or fades.

Why Hook Rate Benchmarks Are Mostly Noise

Search for "good hook rate" and you will find tables full of percentages. We checked the published ones while writing this guide and found the same problem in all of them: they rarely state the numerator, the placement, the audience temperature or the ad length. A 2-second TikTok view and a 3-second Meta play on a Reels placement are different events. A cold prospecting audience and a retargeting list behave differently in the first seconds. A six-second bumper and a sixty-second brand film have different openings by design.

Even Google, which owns the data, attaches a warning to its own creative guidance. The YouTube ABCD reference guide says:

"This guidance is backed by correlational data, but does not guarantee future performance."

Google is just as careful with its own view rate data. The help page notes: "While the strict definition of view rate is the same across video formats, the individual formats have different criteria for what is considered an Impression or a view." On Shorts it goes further: "The Shorts watch experience is inherently different from in-stream because all content (organic and ads) is immediately swipeable. As such, we expect performance metrics on Shorts to vary from current in-stream expectations."

Apply the same caution to hook rate tables. A more useful reference is your own baseline: the hook rate of your best-performing current ad, in the same placement, against the same audience type. A new variant either beats that baseline or it does not. That comparison is honest and repeatable, and it uses data you already own.

There is a second reason to be careful. Hook rate rewards anything that stops a thumb, including openings that have nothing to do with the product. A loud noise, a shocking image or a misleading text card can lift the ratio and wreck everything after it. Watch the next metric down before celebrating a high hook rate.

What Moves Hook Rate: The Production Levers

Most of what lifts hook rate is decided before anyone opens the ads manager. It is decided in the script, the shot list and the edit. Google's ABCD framework, built from Brand Lift research on YouTube in-stream ads, names the levers in its reference guide:

"Framing, pacing, people, and audio are particularly important in an ad's opening moments for hooking your audience."

Framing. The ABCD guide recommends: "Have tightly-framed shots of people or product at the beginning of your ad." Wide establishing shots are a film habit. In a vertical feed on a phone they read as empty space. A close shot of a face, a hand or the product gives the eye something to land on in the first frame.

Pacing. A static opening with a slow fade from black is the most common reason we see for a weak hook in the edits clients bring us. The first cut should arrive early. The first frame should already contain motion or a clear subject, because the platform starts counting from that frame.

People. The guide is direct on this point: "If your video features people, open with them on-screen. Having them speak to the audience directly can also have a positive effect." It adds that "People help hook viewers in an ad's opening moments and facilitate connection." On Meta, where much of the audience watches muted, a person speaking needs on-screen text or captions so the line still lands.

Audio and surprise. Google also recommends that brands "Use surprising, unexpected, or delightful images and visuals to break through the clutter early on." Surprise works when it is tied to the product or the problem it solves. Surprise that has nothing to do with the brand lifts the hook rate and leaves the viewer confused by second six. The ABCD guide makes the brand side of this explicit: "Integrating the logo into the storyline or featuring it on a product can be effective for increasing ad recall." It also names what keeps people past the opening: "Storytelling styles, messages, people, and audio levers are key to keeping people watching."

Two more levers come from the format itself. Text in the first frame matters on muted feeds, so the opening line should work as a caption. Aspect ratio matters too: a 16:9 master cropped badly to 9:16 often cuts off the very subject meant to hold attention. Plan the vertical framing at the shoot, before anyone opens the edit.

A Diagnostic Framework: Reading Hook Rate With Hold Rate

Read hook rate next to a hold metric and a business metric. This matrix is the one we use when a client asks why an ad is not working.

When the drop happens later in the ad, Google's guidance applies: "If you find a steep drop off at certain points, figure out what it is about your ad creative that might be leading to that (outside of expected behavior)."

Strong hook, strong hold and weak conversions usually points away from the creative. Recutting the opening in that case just burns budget. The last row has its own risk: winning ads wear out. Our ad fatigue guide covers how to spot the decline before it shows up in cost per result.

A checklist before you blame the hook

Run through these points before rewriting an opening:

• Same numerator everywhere? Confirm that every report uses the same early-view metric for the same platform.
• Same placement? Compare Reels with Reels and Stories with Stories. Feed and in-stream placements behave differently.
• Same audience temperature? Prospecting and retargeting hook rates should live in separate rows.
• Enough impressions? A few hundred impressions will swing wildly. Wait for a volume where the ratio settles.
• First frame checked on a phone? Watch the ad muted, at feed size, before reading any numbers.
• Quartiles checked? Google's advice reads: "Quartile reporting can be used alongside your other metrics, such as view rate and audience retention (found in your YouTube Analytics data) to check roughly where in your video users are dropping off." The drop point tells you more than the average does.
• Thumbnail or cover frame? On in-feed placements the cover frame does part of the hook's work.

How to Test Hooks Without Rebuilding the Ad

The cheapest way to improve hook rate is to keep the body of a working ad and swap the first few seconds. One shoot can feed several openings if the shot list plans for it. A product close-up, a person speaking to camera, a problem shown on screen and a text-first opening can all be captured on the same day.

A simple testing structure looks like this:

Google gives the right posture for reading the result. Its view metrics page describes quartile reporting as data "intended to be directionally helpful to show you where viewers are dropping off". Treat the test the same way. A variant that wins on hook rate by a hair on a small sample is a hint worth retesting. Our guide to video creative testing for DTC brands covers sample sizes and test design in more detail.

Change one thing per variant. If the new opening also changes the music, the voice and the first text card, a better hook rate tells you nothing about which change did the work.

Hook Rate by Ad Length

The length of the ad changes what a good opening looks like, and it changes how much the hook rate matters.

For six-second ads the metric is less useful, since an early view covers half the ad; our bumper ads guide explains how to build a full message into that window. For longer brand films, hook rate is a weak target on its own. A film that opens slowly on purpose can still win on hold and recall if the first seconds create curiosity. Judge long-form work on retention curves and brand lift, and use hook rate only to compare different openings of the same film.

How a Production Studio Builds for Hook Rate

Hook rate is a media metric, yet the fixes almost always land in production. That is why we plan openings at the script stage at Neverframe, an AI video production studio that works for brands and agencies. We make four formats, and each one handles the first seconds differently.

• Social ads are built around this metric. A social ads package is three 15-second ads with different hooks plus two 6-second bumpers, delivered in 9:16, 1:1 and 16:9. Three hooks on the same message is a ready-made test, with the variants already cut for the placements where hook rate is measured.
• Spots are 30-second ads with script, storyboard, licensed music, sound design and voiceover, delivered as a 4K master plus a vertical version. The vertical cut gets its own opening frame, planned in the storyboard, so the subject is not lost in the crop.
• Films are 60 to 90 second brand films about who you are and why, with an original script, a cinematic score, a 4K master and social cut-downs. The cut-downs are where hook rate applies: each one gets an opening built for the feed, separate from the slower opening of the master.
• Avatars are a digital double of a CEO or spokesperson who can record new messages without new shoots. For hook testing that means new openings can be written and produced without booking a person again, which suits the keep the body, change the first seconds approach described above.

On timing: a first version arrives within a few days (72 hours for agencies), and final delivery takes about two weeks, three to four weeks for films. Agencies can work with us under white label through our partner program. If your current ads have a hook rate problem, or you want a set of social ads built with three tested openings from day one, see the four formats or talk to us about your next campaign.

Common Hook Rate Mistakes

Comparing across platforms. A 2-second TikTok view and a 3-second Meta play are different events. Keep them in separate rows.

Optimizing the hook in isolation. A high hook rate with weak results downstream usually means the opening was disconnected from the ad. Read hook and hold together.

Ignoring muted viewing. On feeds, the first seconds are mostly watched without sound. An opening that relies on a spoken line needs captions or a visual that carries the same idea.

Testing too many changes at once. If every variant changes the opening, the music and the call to action, the result cannot be attributed.

Trusting other people's benchmarks. Published tables rarely state the metric, placement and audience behind the numbers. Your own baseline is the only fair comparison.

Reading results too early. Small samples swing. Wait until the ratio settles before declaring a winner, and confirm the result on cost per result before scaling.

Leaving a winner running forever. Every opening wears out with frequency. Plan the next round of hooks while the current winner is still performing; our article on AI video ads covers how teams produce variants at that pace.

FAQ

What is hook rate?

Hook rate is the percentage of video ad impressions that turned into an early view, usually measured in the first two to five seconds. On Meta it is typically calculated as 3-second video plays divided by impressions, on TikTok as 2-second video views divided by impressions. It tells you whether the opening of the ad makes people keep watching. It does not measure understanding, recall or conversions, so it should always be read together with a hold metric and a business result such as cost per purchase or cost per lead.

What is the hook rate formula?

The hook rate formula is early video views divided by impressions, multiplied by 100. The early view changes by platform: 3-second video plays on Meta, 2-second video views on TikTok, and the share of viewers still watching after the five-second skip point on YouTube in-stream. Because those thresholds differ, a hook rate from one platform should never be compared directly with a hook rate from another. Write down the exact metric used and keep it fixed across every report.

What is a good hook rate?

There is no single good hook rate that holds across platforms, placements and audiences. Published benchmarks rarely state which early-view metric, placement or audience they used, so the numbers cannot be lined up against each other. The reliable approach is to set your own baseline from your best current ad in the same placement and audience, then judge new variants against it. A variant that beats the baseline on hook rate and does not lose on hold rate or cost per result is a real improvement.

What is the difference between hook rate and hold rate?

Hook rate measures how many people keep watching past the opening seconds. Hold rate measures how many of them stay for a longer stretch, such as a ThruPlay, 15 seconds or a quartile of the video. Teams define hold rate in different ways, so the definition should be fixed in writing. Read the two together: a high hook rate with a low hold rate usually means the opening grabs attention but does not connect to the rest of the ad.

How do you improve hook rate on video ads?

Most improvements happen in production. Open with a tight shot of a person or the product, start the first cut early, show motion in the first frame, and make the opening line work as on-screen text for muted feeds. Shoot several openings for the same ad body so you can test them without rebuilding the ad, change one element per variant, and plan vertical framing at the shoot. Google's ABCD guidance on framing, pacing, people and audio is a useful checklist for the first seconds.