The 3-second hook rule, traced. What TikTok and Instagram measure and which statistics have no source

KEY TAKEAWAYS

Meta hook rate divides by impressions, TikTok view rates and Instagram skip rate divide by views. They are not comparable.
The 71% figure describes people who kept watching in a simulated feed, not people who left.
TikTok's 90% in six seconds figure is about ad recall, from an internal analysis with no published method.
Instagram has published no skip rate benchmark. Use your own median.

Every guide to short video says the first three seconds decide the outcome. Most attach a number to it. We traced the numbers. Some lead to a real study that says something different. Some lead nowhere. The platforms, meanwhile, measure the opening in three different ways. This article sets out what each one counts, which statistics hold up, and how to read your own figures.

The 3-second rule is three different measurements

The phrase suggests one metric. There are at least three, and they do not share a denominator.

Where	Metric	What is counted	Divided by	Who defines it
Meta ads	Hook rate	Plays that reach 3 seconds	Impressions	Advertisers. Meta supplies the two inputs, not the ratio
TikTok ads	2-second and 6-second views	Views lasting more than 2 or 6 seconds	Views in TikTok One. Shown as counts in Ads Manager	TikTok
Instagram organic	Skip rate	Views where the person left within the first 3 seconds	Initial views of the Reel	Instagram

On Meta, hook rate is a ratio advertisers build themselves. Motion, Billo and AdRiseLab all give the same formula: 3-second video plays divided by impressions. SparkUGC and AdMakeAI, citing Meta's help page, report that a play counts at 3 seconds or at 97% of the video's length if it is shorter, with replays excluded. We could not open Meta's page directly, so treat that detail as reported.

TikTok counts earlier. Its reporting documentation defines 2-second views as the views that lasted more than 2 seconds divided by views, and does the same at 6 seconds. It adds that these appear as percentages in TikTok One and as counts in Ads Manager. Advertisers who want a TikTok hook rate usually divide 2-second views by impressions, according to Billo and SparkUGC. TikTok's focused view, on a help page updated in August 2026, is a watch of at least 6 seconds, or 15 on the longer setting, or an interaction within the first 6.

Instagram's organic metric runs the other way. Skip rate arrived in Reels Insights in August 2025 and replaced View Rate, which TechCrunch had reported in January 2025 as the share of people who kept watching past three seconds. Social Media Today's report of the announcement gives the new definition: the share of views from people who skipped during the first three seconds. Fanpage Karma adds that the base is initial views, meaning the first play in a Reel session.

So a 30% on one platform is not a 30% on another. One is a share of everyone the ad was served to. One is a share of people who already started watching. One counts leavers, not stayers. SparkUGC makes the same point for paid media: benchmark inside a platform, never across.

Where the 71% statistic comes from

The most repeated line is that 71% of viewers decide within three seconds whether to keep watching. Dynamoi states it with no source. LeadsOff credits Manychat. Animoto links to a syndicated copy of an article by EyeSee, a research firm.

EyeSee's own page describes a two-wave global study: 60 ads from 20 brands, 7,280 respondents, tested in simulated TikTok feeds, as standalone ads and through surveys. The page carries no date and does not say who paid for the work. Its finding is that 71% of TikTok users "stop and watch the first three seconds of the video". It also reports that viewers saw 56% of each sponsored video on average.

That is the reverse of the popular version. The study measured the share who stayed for three seconds, in a simulation, with recruited respondents. By our calculation the remaining 29% did not. It says nothing about a decision point, and nothing about live feeds.

Compare it with field data. Billo published an analysis on August 6, 2026 covering 88,329 sales-objective Meta video ads that ran from January to June 2026, with $122 million in spend. The average hook rate was 25.44%. The same three seconds produce 71% in a simulated feed and about 25% of impressions in live ads on another platform. Neither figure is wrong. They answer different questions.

Which hook statistics hold up?

Claim as usually repeated	What we found	Verdict
71% decide in the first 3 seconds	EyeSee study, 7,280 respondents, simulated feed. 71% watched the first three seconds	Real study, reversed meaning
90% of the impact is in the first 6 seconds	TikTok Marketing Science, an internal meta-analysis. 50% of impact in 2 seconds, 90% of ad recall and about 80% of awareness in 6	Real, but about recall, with no published sample or method
65% who watch 3 seconds watch 10, and 45% watch 30	A Facebook business post, as quoted by Instapage	Describes people who stayed, on Facebook's scrolling feed
47% of a video campaign's value arrives in 3 seconds	Facebook and Nielsen study, reported by TechCrunch in February 2016	A decade old, in-feed video
65% of viewers leave in the first 3 seconds	Stated by Go-Viral with no citation	No source found
Videos need 65% to 70% retention at 3 seconds to be promoted	Stated by Dynamoi with no citation	No source found

Two notes on TikTok's figure. One TikTok post credits it to its Value of a View meta-analysis, dated 2021. Another credits a creative guide dated 2020. And the measure is brand recall and awareness, not purchases. It supports putting the brand and the proposition early. It does not show that a sale is won or lost by second three.

The Facebook figures are conditional. They describe what people who had already watched three seconds went on to do. They have been repeated for years as if they described everyone.

What have the platforms said themselves?

Less than the guides imply. TikTok's creative best practices page, last updated in June 2025, asks advertisers to put the hook in the first 6 seconds and the proposition in the first 3. It gives no statistic. Note the wording: the hook window is six seconds, not three.

Adam Mosseri, head of Instagram, has called the first three seconds of a Reel very important and named them the hook, in remarks reported by Inman. In the same remarks he said this does not require bait or a cliffhanger, and he described sends as one of the biggest factors in what gets promoted. Hootsuite's summary of his guidance, updated on October 2, 2026, lists watch time, likes and sends as the top signals.

Instagram has published no skip rate benchmark. Metricool and Volanea both say so. The bands in circulation come from vendors. Retensis rates under 15% as exceptional and above 40% as a problem, with no dataset. Babbleboxx calls under 30% to 40% healthy and credits unnamed industry trackers. A debunking piece from Silma in July 2026 found no source at all for several other common hook claims.

Does a strong opening predict results?

Here the sources disagree, and none settles it with data.

Billo's guide says creatives with strong hook rates beat their peers on every measure it tracked. The same guide admits some of its highest-hooking ads barely broke even. Ryze states that ads which hook viewers have 40% lower acquisition cost and shows no breakdown. On the other side, AdRiseLab calls hook rate a triage metric that says nothing about revenue, Motion warns that optimising for a single attention metric pushes teams toward clickbait openings, and AdMakeAI relays a practitioner whose ads under 20% beat ads at 60% on cost per acquisition.

Billo's 88,329-ad dataset would be the place to test this. The report does not relate hook rate to cost or return. We found no published correlation with a stated method.

Length points the same way. Buffer's analysis of 1.1 million TikTok videos, reported by eMarketer in March 2025, found that videos over 60 seconds had 43.2% more reach than those of 30 to 60 seconds. If the opening were the whole story, length would not matter.

What the evidence supports is modest. A weak opening caps everything after it, because few people see the rest. A strong opening guarantees nothing. That makes the metric a diagnostic.

How to read your own numbers: a decision table

Use two readings. Opening means hook rate on Meta, 2-second views on TikTok, or the inverse of skip rate on Reels. Hold means a later point: 15-second plays over 3-second plays on Meta in Motion's definition, 6-second views on TikTok, the retention chart on Reels. Compare each with your own median, not a published band.

Opening	Hold	Outcome	Likely problem	What to change
Below your median	Any	Weak	The first frame or first line	Reshoot or recut the opening only
Above	Below	Weak	The opening promises something the body does not deliver	Rewrite the body, or make the opening honest
Above	Above	Weak	Offer, landing page or audience	Stop editing the video
Below	Above	Strong	A narrow hook that filters for buyers	Leave it alone
Above	Above	Strong	Nothing	Scale, and reuse the structure

The fourth row is the one hook-first advice gets wrong. An opening that turns most people away can still be the right one.

A five-step method

Pick one metric per platform and write down its definition. Include the denominator. Do not mix TikTok One percentages with Ads Manager counts.
Build your own baseline. Take the median of your last 20 to 30 videos per platform and placement. Outside figures are context. Billo's monthly average moved from 26.52% in March to 23.99% in June, a swing of about 2.5 points across the whole dataset.
Wait for enough impressions. By our calculation, a 25% hook rate measured on 1,000 impressions carries a 95% range of about plus or minus 2.7 points. At 5,000 it is about 1.2. Small gaps between two cuts are often noise. The same logic applies to conversions per concept.
Read opening and hold together. Use the table. Change one part of the video at a time.
Decide on outcomes. For ads, that is cost per acquisition under one consistent attribution setting. For organic Reels, look at sends and watch time, the signals Instagram names.

Where this leaves the rule

Open strongly. Both platforms ask for it. But the precise percentages attached to the rule are mostly borrowed from a simulation, a brand recall analysis and a ten-year-old feed. Your own median is a better benchmark than any of them.

The same habit applies wherever a number gets repeated until it sounds official, including in how brands are measured in AI answers. If you want that kind of tracing applied to your own visibility, join the waitlist.

Questions

Where does the claim that 71% of viewers decide in the first 3 seconds come from?

It traces to an undated article by the research firm EyeSee. That study tested 60 ads from 20 brands with 7,280 respondents in simulated TikTok feeds and found that 71% of users stopped and watched the first three seconds. The popular version reverses the meaning and applies a simulation result to live feeds.

What is a good hook rate on Meta ads?

Billo measured an average of 25.44% across 88,329 sales-objective Meta video ads that ran from January to June 2026. Vendor guides commonly call 30% to 40% good. Meta publishes no target. Compare each ad with your own account average on the same placement before using any outside figure.

What is a good skip rate on Instagram Reels?

Instagram has not published one. Skip rate is the share of views where the viewer left within the first three seconds, and it has been in Reels Insights since August 2025. The bands you see quoted, such as under 30% is healthy, come from vendor articles with no dataset behind them.

Does a high hook rate mean an ad will convert?

No source we could open shows that with a stated method. Hook rate tells you whether the opening held attention. Several practitioner guides report high-hook ads that lost on cost per acquisition. Treat it as a way to find what to fix, and judge winners on conversions.

Sources

TikTok Business Help Center, Creative best practices for performance ads ads.tiktok.com
TikTok for Business, Resonance, A Key Factor For Ad Effectiveness ads.tiktok.com
TikTok for Business, Creative Best Practices for TikTok Ads ads.tiktok.com
TikTok Business Help Center, About TikTok One Project Reporting ads.tiktok.com
TikTok Business Help Center, About the video views objective and focused view optimization ads.tiktok.com
EyeSee, Tap into hidden potential with new TikTok creative insights eyesee-research.com
Billo, Video Ad Hook Rate Benchmarks H1 2026 billo.app
Billo, From Hook Rate to Hold Rate, Video Metrics Growth Teams Track billo.app
Social Media Today, Instagram Adds Retention Metrics for Reels socialmediatoday.com
TechCrunch, Instagram gives creators more insight into their Reels performance techcrunch.com
TechCrunch, Facebook Lures Video Creators With Minutes Watched And 10-Second View Metrics techcrunch.com
Inman, Head Of Instagram Reveals Algorithm Tips For 2025 inman.com
Hootsuite, Instagram algorithm tips for 2026 blog.hootsuite.com
Fanpage Karma, Instagram Insights Update fanpagekarma.com
Metricool, Instagram Reel Analytics, Retention and Skip Rate Explained metricool.com
Volanea, Instagram Reels Skip Rate, Improve Your Hook volanea.com
Retensis, Instagram Reels Skip Rate Benchmarks 2026 retensis.com
Babbleboxx, What's a Good Instagram Skip Rate? babbleboxx.com
Silma, Most Instagram hook advice is made up silma.app
Animoto, Why The First 3 Seconds of Video Matter More Than the Next 30 animoto.com
LeadsOff, The TikTok Hook Formula leadsoff.com
Dynamoi, What is the 3 second rule on TikTok dynamoi.com
Go-Viral, The First 3 Seconds, How to Hook Viewers on TikTok and Reels go-viral.app
Instapage, How Closed Captioning Facebook Videos Can Improve Viewership instapage.com
Motion, Key creative performance metrics motionapp.com
AdRiseLab, What Is a Good Hook Rate on Meta Ads? adriselab.com
SparkUGC, Hook Rate Benchmarks for Meta and TikTok Ads (2026) sparkugc.com
AdMakeAI, Hook Rate, The Formula, Real Benchmarks, and a Calculator admakeai.com
Ryze, Meta Video Ad Benchmarks, ThruPlay, Hook and Hold Rate 2026 get-ryze.ai
eMarketer, Longer TikToks perform better, presenting new opportunities for advertisers emarketer.com

Esh

Founder, Ethos Engine. Builds the Digital Authority Score and runs the placement marketplace. esh@ethosengine.io

NEXT STEP

See where your brand stands.

Your Digital Authority Score, who AI recommends for your target questions, and the placements that close each gap. Your first report is free.

Get your free report →