The First 3 Seconds Are The Product

The platform publishes the number and it is not generous. 63 per cent of the top click through videos hook inside 3 seconds, while the average decision on a phone is made in about 1.7 seconds. At 30 frames per second that leaves roughly 51 frames to prove the video is about what the title said it was about. The retention reading is sharper still. Strong creators hold 70 per cent or more on the intro, and anything under 40 per cent kills distribution outright. Distribution is not a reward for the content. It is a reward for the first 90 frames, before the content has had a chance to arrive. The number worth putting on a wall is 300. A video that stalls at 300 views did not stall in the content, it stalled in the hook, and 300 is about the size of the first test pool. That pool does not watch for 40 seconds to find out whether the middle was good. It decides inside 3 seconds whether the middle will get a viewer at all. So the opening is not a warm up and it is not a greeting. It is the product being sold, and it is sold in the first 3 seconds or it is never sold.

A hook is 3 layers stacked in 1 order, and only 1 of the 3 is spoken. The visual trigger runs from 0.0 to 0.5 seconds. Movement, contrast, a bold text overlay, a face with energy. It has to register before the audio does, because a large share of first views happen muted, and a silent frame is the only element that survives a muted feed. The verbal trigger runs from 0.5 to 1.5 seconds and it is 1 sentence. A promise, a curiosity gap, or a problem. Not a greeting. Not a 12 second introduction. Not a throat clearing. 1 sentence, delivered inside the first 1.5 seconds, because past that mark the viewer has already decided. The context lock runs from 1.5 to 3.0 seconds and answers 2 questions at once: who this is for, and what they get. Skip it and the hook has stopped a stranger whose expectations do not match the video, which is worse than never being seen. At 30 frames per second the whole structure is 90 frames. The first 15 are silent, the next 30 carry the sentence, and the last 45 name the viewer. Reorder them and the same words stop working. The hook is 3 layers in a fixed order, and only 1 of the 3 is spoken.

7 frameworks exist and they are 1 trade in different shapes. Shock to logic: stop doing this, it is the reason you are getting that. Specific result: 1 method, 1 number, 1 window. Open loop: wait for the last part. Hot take: the advice everyone repeats is the problem. Point of view: you post every day and the number does not move. Speed run: the whole thing inside 30 seconds. Proof first: open on the receipt instead of the explanation. The strongest structure in the library is a 3 step curiosity loop, taken from cases that ran between 8 and 15 million views. Line 1 compels line 2, and line 2 compels line 3. Step 1 is the context lean across the first 2 lines: name the topic so the right viewer self selects, then offer common ground, a pain point, or a mind blowing fact. Step 2 is 1 scroll stop interjection built on a single contrasting word: but, however, yet, although. In its best form it contradicts the advice the viewer already suspects is hollow. Step 3 is the contrarian snapback, which goes the opposite way from the lean while staying on the same topic. The bigger the shock, the bigger the snap. 5 upgrades sit on top. Visual hooks beat spoken hooks by roughly 100 times, so 3 to 5 words in large bold type carry the lean and the audio confirms it. Lead with the solved pain rather than the topic. Anchor the unknown to something the viewer already knows, so nobody feels stupid for staying. Compress speed to value, with a hit of value inside 4 seconds. Then hold the hook in staccato sentences, and let sentences lengthen only after the viewer has committed. The loop only works while it stays open. Close it and the viewer has their answer, and an answered viewer swipes. Open the next 1 instead.

The method behind a hook is an audit, not an inspiration. 8 checks, run against the first frame of a video that already has numbers on it. Context first, and it is the most common failure of all. Put the context in the first frame's text and in the first spoken words. No context means nobody cares, and nobody caring means a swipe. Challenge a common belief. This is the strongest single framework in the set, because 1 line is a claim and the other is a list. Lean at 45 and eat carbs every day beats 21 things I do not do to stay lean past 40. The question test: what question is planted in the viewer's head. If there is none, or if it is hollow, the hook is rewritten rather than re-cut. Safe zone and declutter. Text kept clear of the interface overlap, which sits on roughly the bottom 20 per cent of a vertical frame, type smaller than instinct suggests, 1 clear subject, no mashed layers on top of each other. First impression mismatch. A still photo and a carousel need a hook as much as a video does, because the feed they land in is not the feed they were made for. Story hooks. In a saturated niche the story and the emotion are the hook, not the visual. Do not let them solve it early. Obscure the reveal, blur the key element, do not name the payoff in the caption, because a viewer who has worked it out has no reason left to stay. Visual and text contrast. A scroll stopping image paired with a line that does not describe the image. When the real frame is cluttered, rebuild the frame rather than the sentence.

63 per cent of the top click through videos hook inside 3 seconds. The average decision on a phone takes 1.7.

The editing grammar is a table of numbers and none of it is taste. Cut every 0.5 to 2 seconds, with most cuts landing near 1 second. Captions are effectively mandatory: burned in, keyword highlighted, set between 36 and 60 point, centred or in the upper third, and never across the face. Beat sync runs at 110 to 120 beats per minute, which works out at a cut every 0.5 seconds, and the most used technique on the platform is the velocity speed ramp easing along an S curve rather than a linear push. Length is a reach decision. Reels between 7 and 15 seconds outperform longer cuts on reach, and a 12 second video watched twice beats a 40 second video watched once and abandoned. Completion rate is the metric, not duration. Orientation is settled. Vertical 9:16 is more than 60 per cent of all social video, and the bottom of that frame is not yours to use because the interface sits on top of it. Timing is where most creators lose. Audio trends peak inside 7 to 10 days. Format trends, split screen, text overlay, carousel, run 3 to 4 weeks. The entry window is the first 3 to 5 days, and a trend joined on day 12 is a trend joined after the audience has stopped looking at it. Red reads cheap and alarming. Pull the colour from the composition instead, and say the hook aloud within the first 1.5 seconds, so the muted viewer gets the text and the audio viewer gets both.

The library ends in a score rather than an opinion, because a hook that cannot be scored cannot be fixed. 3 elements plus a clean safe zone. Score 0 to 3 and the hook is broken: it gets rewritten, not shipped. Score 4 to 6 and 1 element is working while the other 2 are absent. Score 7 or above and all 3 elements are present, the safe zone is clean and the loop is still open. Below 7 the note is never make it better. It is name the missing element and rebuild that 1 thing. Here is the arithmetic the standard is built on. Built companies across 12 countries. Restructured a industrial group. Deployed 210 energy systems across Africa and Asia. Took a company through a Frankfurt IPO. 16 years of field work from rural Nigeria to Fukushima. Entrade built 200 wood gasification machines. Signed a personal guarantee on a company that folded on Christmas Day 2008, and lost savings in the 2013 Cyprus bail-in. 393 pages in paperback and Kindle, published 8 April 2026. None of it was found by a hook. All of it had to be explained to somebody who had 3 seconds to decide. The probability that a strong body behind a weak opening gets watched to its end is close to 0, because nobody reaches the body. The probability that the same material, opened with 3 layers inside 90 frames and a loop left open, gets watched to its end is close to 1. The difference is not budget and it is not charisma. It is 3 seconds and whether anybody designed them. Nothing downstream rescues a hook. The viewer is gone before the good part starts.