Formi Book a short meeting
← All articles

The hook in a short video: what has to happen in the first seconds

8 min read

The hook is a short video's first line and first shot, and it decides whether anyone watches the rest. In our experience the hook that works best is an everyday, spoken and specific question that the viewer answers in their head before they get the chance to swipe on: "Would you have known these questions about Finland?" gathered 106,500 views for Dripnord and "Guess how much Theo's shopping cost?" 183,000. A logo, a greeting and ad language are the three things that lose the viewer before the third second.

Why do the first seconds decide a short video?

Because the platforms distribute a video according to how many people watch it to the end. TikTok itself says that in its recommendation system a strong signal, such as a user watching a longer video from start to finish, carries more weight than a weak signal, such as the viewer and the creator being in the same country (TikTok Newsroom). According to Instagram, the most important predictions for Reels are how likely the viewer is to share the video, watch it all the way through, like it and go to the audio page, and most of the Reels feed comes from accounts the viewer does not follow (Instagram, 2023).

For advertisers, TikTok gives the same advice in numbers: the core of the content within the first 3 seconds and the hook within the first 6 seconds, so that watch time gets longer (TikTok: Creative best practices). In an organic video there is less time. The viewer does not read the first sentence to the end, but recognises straight away whether this is an ad or a person saying something interesting.

In Finland there is plenty of this audience: TikTok reaches 1.62 million people aged 18 and over, or 35.1% of adults (DataReportal, Digital 2026: Finland). None of them follow you yet, so the video cannot lean on the brand. It has to earn the first second.

What kind of hook works with a Finnish audience?

Spoken, specific and addressed to the viewer. In all the videos below, the first line is written the way it is said, not the way it would be written in an ad.

  • Dripnord: "Guess how much Theo's shopping cost?" 183,000 views. "Would you have known these questions about Finland?" 106,500 views. "What's the dumbest clothing purchase you've ever made?" 72,100 views.
  • Sneak.fi: "Will these two ever see each other again after this video?" is the account's most-watched video, 1.2 million views.
  • realme Finland: "How many times a week do you clean your phone?" gathered 35,800 views with no ad money, while the account's other unboosted videos stayed at 750–1,500 views.

They have three things in common. The line is a question the viewer has their own answer to, so they stay to hear what the others say. It is in spoken Finnish, "oisitko", "sun", "makso" instead of the written forms, which tells you straight away that the speaker is a person, not a brand. And it is specific: not "how much do clothes cost" but "how much Theo's shopping cost".

A specific question sounds like the answer really exists. It also gathers comments, because many people write their own answer before the video is even over, and comments are one of the interactions TikTok says it weights in its recommendations. So the question is not just the opening. It is what the viewer does during the video.

The same applies when the product is a phone. The engagement rate of the other 50 videos on realme Finland's account was 0.10%. Our five test videos opened with an everyday question, such as "When did you last clean your phone?", which the viewer answers in the comments and the phone's feature answers on screen, and the engagement rate of the trial period was 0.88%. The phone was the same, the opening was not.

Which hook types work, and when?

We have produced over 1,000 videos, and the hooks fall into a few types. Below are the ones that recur in our clients' most-watched videos, and the situation each one suits.

Hook typeExampleViewsWhen
Guessing question"Guess how much Theo's shopping cost?" (Dripnord)183 000When there is a number or a price at the end of the video that the viewer wants to know. Celebrity shopping trips and shopping baskets.
Trivia question"Would you have known these questions about Finland?" (Dripnord)106 500Street quizzes. Questions that look easy but that people get stuck on.
Confession"What's the dumbest clothing purchase you've ever made?" (Dripnord)72 100Street interviews where the respondents are the entertainment and the product is in the background.
A question about the viewer's everyday life"How many times a week do you clean your phone?" (realme Finland)35,800 with no ad moneyWhen the product relates to an everyday habit and the answer is shown with the product. Brings comments.
Open ending"Will these two ever see each other again after this video?" (Sneak.fi)1.2 millionChallenges and situations where the viewer stays to see how it turns out.

One type is missing from the list on purpose: the prize stunt. A question that pays money for the right answer brings viewers who watch the stunt, not the brand. We do not use them. More on designing street interview questions in the article street interview videos.

How do we write the hook before the shoot day?

We script every video before the shoot, and the hook is written first. You approve the scripts before the camera rolls. The process is the same at Dripnord's warehouse and in realme's product tests:

  1. One thing per video. Before the hook, we decide what the one thing is that the viewer gets. If there are three, there are three hooks and three videos.
  2. Several hooks for the same concept. We write several options for the opening of the same video and shoot them on the same shoot day. One is chosen in the edit, and the rest remain as ad versions.
  3. Reading aloud. If a line sounds like an ad when read aloud, it is rewritten the way the presenter would say it to a friend. "Discover the new collection" becomes "why is everyone buying this right now".
  4. The first shot gets straight to the point. The product or the situation is visible straight away, not only after a lead-in. In a street interview the question is asked in the first second, and the interviewee's reaction fills the second one.
  5. On-screen text repeats the question. The same line is written on screen so that the hook works without sound. TikTok recommends 5–10 words per second when text is used (TikTok: Creative best practices).

In a street interview the hook is the question itself, and it cannot be patched in the edit. The question has to look easy, so that the viewer starts answering, and be just hard enough that the person on the street gets stuck. Then the viewer stays to see whether they could do better themselves. "Would you have known these questions about Finland?" is built exactly like this, and Dripnord's street quiz gathered 106,500 views with it.

Which mistakes lose the viewer before the third second?

These are the ones we see most often when we go through a new client's channel in the first meeting.

  • A logo or an intro. The brand is not a reason to stay. The logo belongs at the end, if anywhere.
  • A greeting and an introduction. "Hi, we're X and today we'll show you" is three seconds during which the video has already been skipped.
  • Ad language. "High-quality", "affordable" and "available now" say straight away that the speaker is a brand. People do not talk like that.
  • A promise the video does not keep. If the hook promises a reveal and it never comes, the viewer leaves partway through, and that is exactly the signal that weighs most in distribution.
  • Three questions in one. The viewer has no time to answer any of them. One question, one answer.
  • A question nobody wants to answer. "What do you think of our brand" is not a hook. "What's the dumbest clothing purchase you've ever made" is, because everyone has one.
  • Building on a trending sound. On a business account the commercial music library is limited, and if the concept falls apart without the track, the problem is the concept.

How do you measure whether a hook works?

From the retention curve, not from likes. TikTok's and Instagram's analytics show you at what point viewers leave. If the curve drops steeply in the first seconds, the problem is the hook. If it drops in the middle, the problem is the content, and the hook has done its job.

We compare different openings of the same concept with each other, not different concepts. On realme Finland's account, our video published with no ad money gathered 35,800 views, while the account's other unboosted videos stayed at 750–1,500 views. The audience and the account were the same, the content and its opening were not.

Volume is part of the measuring. In Sneak.fi's Black Friday month of 2025, 47 videos were published, meaning 47 different openings for the same product and the same audience, and the month gathered 9.87 million organic views. The record did not come from one perfect hook but from the fact that weak openings showed up in the data immediately and the next video was written on the basis of them.

On day 30 we go through which hook type worked with your audience, and the next month is written around that. The best videos are turned into ads as they are: Spark Ads keeps the original post, so a tested hook goes into the ad without a new cut. How that is done is in the article Spark Ads guide. What the algorithm measures after the hook is explained in the article TikTok algorithm 2026.

Does the same hook work on TikTok, Reels and Shorts?

It does, and we publish the same cut on all three. The distribution logic is the same: TikTok weights watching to the end, and for Reels Instagram predicts precisely whether the video is watched all the way through and whether it is shared. The difference is in how long a video lives, not in what the opening has to be like. YouTube Shorts keeps gathering views the longest, because the content is also found through search weeks after publishing, and there a question-shaped hook works as a title too.

One limit: the hook must not lean on the platform's own features. "Would you have known these questions about Finland?" works everywhere, because it points to what is in the viewer's own head, but "double tap" belongs to one platform. The hook refers to the viewer's life, not to the app they happen to be in right now.

Want us to look at the openings of your videos?

Book a 30-minute meeting on Google Meet. We go through your channels and tell you straight what we would do differently in the first three seconds, whether or not you buy anything. Dripnord's whole journey since October 2025 is on the page Dripnord case study, and Sneak.fi's Black Friday month, 47 videos and 9.87 million views, on the page Sneak.fi case study.

Frequently asked questions

How long does the hook need to be?

One line, said within the first second. TikTok's own advice to advertisers is to present the core of the content within three seconds and the hook within six, and in an organic video there is less time.

Does the hook have to be said out loud, or is on-screen text enough?

Both. The presenter says the line and the same text is on screen, so the hook works without sound too.

Does the same hook work on TikTok and Instagram Reels?

It does. Both platforms weight whether the video is watched to the end in their distribution, so we publish the same cut on both and compare the results.

What if our product is boring?

The hook is not about the product but about the viewer. Cleaning a phone is boring, but a question about how many times a week the viewer cleans their phone gathered 35,800 views for realme Finland with no ad money, while the account's other unboosted videos stayed at 750–1,500 views.

How many hooks should you shoot for the same video?

Several, on the same shoot day. The best one is chosen in the edit, and the other openings remain as ad versions that are tested in paid distribution.

How long before we know which hook works for us?

The first direction shows on day 30, when we go through the first month's videos. After a three-month pilot you know which hook types work with your audience.

Can the hook be changed afterwards?

Yes. The same material is cut with a new opening, and one round of revisions with the editor is part of the production. In ads, different openings are tested from the same video.

How much does testing hooks cost?

It is part of the production, not a separate invoice. The price is built around your overall package: content volume, channels and media budget.

See also

Sources

  1. TikTok Newsroom: How TikTok recommends videos #ForYou (18 June 2020)
  2. Instagram: Instagram Ranking Explained (Adam Mosseri, 31 May 2023)
  3. TikTok for Business: Creative best practices for performance ads
  4. DataReportal: Digital 2026: Finland

Want the same result for your brand?

Book a short meeting