HOOK & RETENTION BLOG / 8 MIN READ

What should the first 3 seconds of a video be?

They should give one clear reason to keep watching, and almost nobody does. Here is what belongs in those three seconds, what to cut, and how to tell if yours actually earns the next ten.

3sto earn the view
1reason to stay
0sspent on a logo
0–100craft score

By Thomas, founder of CutScore · Updated June 2026

HOOK CHECK · first_3s.mp4
A presenter leaning into a microphone at the very start of a take, the three seconds where a video either earns the next view or loses it.
CRAFT SCORE
FIXES ADVISED
the three seconds that decide it
Logo sting at the front · cut the first 2.4s00:00
Payoff buried · move from 00:3800:38
Clear subject in frame · face, well lit
The 30-second answer The first three seconds of a video should give the viewer one clear, specific reason to keep watching, and nothing else. That means leading with the most interesting thing you have: the payoff you promised in the thumbnail, a sharp question, a visible stake, or the surprising result, on screen in the first frame. No logo sting. No slow fade-in. No "hey guys, welcome back." If a stranger could not tell what they get from staying, the opening is not finished. That is the single test, and it is exactly the hook CutScore checks first.
WHY THREE SECONDS

A feed is a brutal place. Your video does not open on a patient cinema audience. It autoplays into a thumb that is already moving, next to a hundred other clips fighting for the same half second of attention. The viewer is not asking "is this good." They are asking "is this for me, right now," and they answer before they know they have answered.

I have shipped openings I was proud of that did nothing. A clean logo animation, a soft music swell, a friendly "so today I want to talk about." Watched back on my laptop it felt warm and professional. Watched in an actual feed, on a phone, it was three seconds of me asking the viewer to wait. They do not wait. The thumbnail made a promise and my opening spent its first breath on housekeeping.

So the question is not "what looks nice at the start." It is sharper than that: what gives someone a reason to still be here at second four. The first three seconds are not an intro. They are an audition, and the audience is holding the skip button.

WHAT BELONGS IN THE FIRST 3 SECONDS

What should actually go in the first 3 seconds.

Pick one of these and lead with it. You do not need all of them, and stacking them usually makes the opening noisier, not stronger.

Open withWhat it doesBest for
The payoffShows the result or the best moment up front, then earns the right to explain how you got there.tutorials, builds
A sharp questionNames the exact thing the viewer is wondering, so staying feels like getting an answer.explainers, essays
A visible stakeMakes it clear something can go right or wrong, so the viewer wants to see which.vlogs, challenges
A pattern breakAn unexpected image, motion or sound that stops the scroll because it does not fit the feed.shorts, reels
A concrete promiseStates plainly what the viewer walks away with, in their words, not yours.how-to, reviews
A confident faceA clear, well-lit subject already talking, no warm-up, no settling in.talking-head
The one rule under all of themWhatever you open with has to be true. A hook that promises something the video never delivers buys you three seconds and loses you the next time you post. The strongest opening is a real moment from later in your own video, moved to the front.
DOES YOUR HOOK LAND?

CutScore watches your opening and flags a weak hook with a timestamp, then tells you what is missing and where your real payoff is hiding.

Join the waitlist
WHAT TO CUT

What does not belong in the first 3 seconds.

The logo sting and the intro animation

Nobody has ever stayed for a video because the intro animation was good. A logo at second zero spends your single most valuable moment on something the viewer did not come for and will not remember. If you love your branding, show it later, under the first cut or around a natural break, where it costs you nothing. The front of the video belongs to the viewer, not to you.

The slow fade and the throat-clear

A fade from black is a confident move in a film with a captive audience. In a feed it reads as dead air, three seconds where nothing has happened yet. Same with the verbal version: "hey guys, welcome back, so today." That is you clearing your throat on the viewer's time. Start mid-thought, mid-action, already moving. You can always say hello later, once they have decided to stay.

The same video playing on a phone and a larger screen at once, a reminder that your first three seconds get judged on a small autoplaying feed before anything else.
Your opening is judged on a small, autoplaying, muted feed, not on the timeline where you cut it. Photo: Andrea Piacquadio / Pexels.

Context the viewer did not ask for yet

You feel like the story needs setup. It usually does not, not at the front. Backstory, disclaimers, "before I start, a quick word," all of it can wait until after the hook has done its job. The order that works is payoff first, context second. If your opening explains who you are before it shows why anyone should care, the order is backwards.

A muddy, quiet, or off-frame first shot

The first three seconds are a craft check too, not just a writing one. If the opening shot is underexposed, the audio is quiet, or the subject is half out of frame, the hook fails before the words land. Roughly half the feed is watched on mute, so the image and the on-screen text carry the opening on their own. A great line buried in a dark, silent, badly framed shot is a great line nobody reads. Picture and sound are the other half of what we analyze, and they matter most exactly here.

RATHER SEE IT THAN READ IT?

Here is a real CutScore coaching report, hook included: the opening scored, with the timestamp where retention drops and the exact fix.

See a sample report
SHORT ON TIME

If you only fix three things.

Most weak openings fail for the same three reasons. Fix these and your first three seconds jump from filler to function.

1
2-MIN FIXNARRATIVE
Delete everything before the first real moment
Open your timeline and find the first second where something actually happens. Cut everything in front of it: the logo, the fade, the "hey guys." Nine times out of ten the video is better starting two or three seconds later than you thought.
How Scrub to your first genuine beat, set an in-point there, and trim the front. See the hook.
2
EDITNARRATIVE
Move your best moment to the front
If your strongest line or result lands at 0:38, a slice of it belongs at 0:01. Lead with the payoff, then promise the rest. This single re-cut does more for whether people stay than any change to your thumbnail.
How Copy two seconds of your best moment to the head of the timeline as a cold open, then return to the start of the story.
3
QUICKIMAGE
Make the opening shot clean and loud enough
The hook is also a craft check. Make sure your first shot is well exposed, the subject is in frame, the audio sits near −14 LUFS, and any opening text is large and high-contrast. A strong line in a dark, quiet, off-frame shot fails before it is heard.
How Watch the first three seconds on a phone, on mute, then with sound. If you cannot tell what you are looking at, fix the shot.
HOW TO TEST YOUR OPENING

How do I know my first 3 seconds work?

OPTION 01

The thumb test, on yourself

Watch your opening on a phone, in a feed, sound off first then on. Ask one question: is there a concrete reason to still be here at second three? The trap is that you already know the payoff, so be honest. Works best on a video you have not touched in a day.

OPTION 02

Hand it to a stranger

Show someone the first three seconds and stop. Ask what they think the video is about and whether they would keep watching. If they shrug or describe your logo, the hook is not landing. This is honest and free, and a little uncomfortable, which is how you know it is useful. See how to get honest feedback.

OPTION 03

Let a coach time-stamp it

Hand the file or a link to CutScore. It checks the hook against the craft around it, flags a weak opening with a timestamp, and points to where your real payoff is sitting later in the cut. One 0 to 100 score, with the evidence. See a sample report.

How CutScore checks your hook CutScore is an AI video quality coach for pre-publish QC. It reads your opening the way a viewer would, then checks the craft underneath it: is the first shot exposed and in frame, is the audio near −14 LUFS, is there a clear reason to stay before second three, or does the video open on a logo and a throat-clear. You get one 0 to 100 score, the evidence behind it, and a prioritised list of fixes, the hook included, before anyone else sees it. It judges the craft of the video, not your tags or thumbnail, so it sits next to a growth tool rather than competing with one. More on the method and the standards.
QUESTIONS

Frequently asked.

One clear reason to keep watching, delivered fast. That means the most interesting moment you have, a sharp question, a visible stake, or the payoff you promised, on screen in the first frame. No logo sting, no slow fade, no throat-clearing intro. If a stranger could not tell what they get from staying, the opening is not doing its job yet.
Because most feeds autoplay and most of your drop-off happens at the very start. A viewer decides whether to stay before they have consciously decided anything. The first three seconds set the thumbnail's promise against reality, and if the reality is slower or vaguer than the promise, the thumb keeps scrolling. Win those seconds and the rest of the video gets a fair hearing.
No. A logo sting at second zero spends your most valuable seconds on something the viewer did not come for. Open with the content, then show your brand later if you must, around a natural break or under the first cut. Nobody has ever stayed for a video because the intro animation was good.
Watch your own opening as if you were thumbing past it in a feed, with the sound off first, then on. Ask one question: is there a concrete reason to stay by second three? Better, hand it to someone who has not seen it, or to a tool like CutScore that flags a weak hook with a timestamp and tells you what is missing.
EARLY ACCESS

Stop losing viewers in the first three seconds.

CutScore checks your hook against the craft around it and tells you exactly what to fix, with the evidence to back it up. Join the waitlist for early access.

Join the waitlist