3 days free on every plan · Try any plan before you pay.

Start free trialStart

BlogCreative testing

UGC ad hooks: how versions of the same video compared on views

We compared 26,423 posts of 212 videos, version by version. Reworded on-screen hooks got similar views; versions in different frames differed more.

By Abondy ResearchPublished 10 min read

In short: In our data, versions of a video with reworded on-screen hooks got similar views. Across 26,423 posts of 212 uploaded videos, we compared each version with the other versions of the same video on the same app. Hooks phrased as a question, kept short, or saying "you" each landed within about 5% of the rest. In the set-by-set test, none led in clearly more than half the sets (the best was 56.8% of 74). Views differed more by the frame around the video. Versions shown as a social post card, with the hook as its headline, beat the same video's other versions in 65.7% of 137 comparisons. Picture-in-picture versions trailed, ahead in only 43.4% of 76. So test different hook ideas, not different wordings of one idea.

Most lists of UGC ad hooks give examples with no results attached. We have something different: thousands of versions of the same videos, each with its own words on screen, posted by real people on their own accounts. Because the footage stays the same inside one video, we can compare words and frames without the video itself getting in the way. The catch is that our versions reword one hook rather than trying new ones, and the section on what we compared explains why.

What is a hook in a UGC ad?

The hook is what makes someone stop scrolling: the first scene, line or words on screen. TikTok's help page on creative best practices for performance ads (last updated June 2025) says: "Prioritize your hook in the first 6 seconds to boost engagement and increase watch time. You can do so by creating suspense, surprise, or other emotions." It adds: "Introduce your content proposition in the first 3 seconds for better recall and awareness," and, for text on screen, "We recommend displaying 5-10 words per second when using text" (TikTok Ads Help, read November 21, 2026).

TikTok's May 2025 article on creative effectiveness says the best ads experiment "with story arcs and fresh creative hooks". It also says "iterating upon a central concept can give brands the volume they need for a prolonged campaign, without having to come up with thousands of creative ideas" (TikTok for Business, read November 21, 2026). That is the trade-off this post measures: new ideas versus new versions of one idea.

What did we compare?

Each uploaded video was cut into many versions. Every version got its own words on screen and its own frame or crop. The words are rewrites of one starting line (sometimes a few near-identical lines from the video's editor), written to keep its meaning and voice, usually at a similar length, with all-caps lines kept in capitals. So inside one video, the hooks are different wordings of the same hook. They aren't different opening ideas.

For each post we asked how it did against the other versions of the same video on the same app. That removes the biggest source of noise: some videos and campaigns simply get more views than others. Each post was compared on a log scale, so one big post can't swing the result. A ratio of 1.05 means versions with that feature got about 5% more views than the same video's other versions.

We ran it three more ways as checks. We split the videos into two random halves and redid the sums in each. For every video and app with at least 5 versions with the feature and 5 without, we asked whether the versions with it had the higher median (ties left out). And we redid the sums without the 11.3% of posts that never gave us a view count; those are stored at 0 views, but their real count is unknown.

Did rewording the hook change views?

Not in any steady way. These are the words on the video, across 666 video-and-app sets with 20 or more posts each.

Words on screenPosts with itViews vs the same video's other versionsRead posts onlyHalf AHalf BSets where it was ahead
Starts in lower case3,6421.04×1.04×1.04×1.04×53.0% of 132
1 to 5 words6,2671.03×1.03×1.07×1.00×51.9% of 316
Says I, me, my or we14,9511.00×1.00×0.97×1.04×48.6% of 395
10 or more words5,8391.00×1.01×0.99×1.00×51.2% of 217
6 to 9 words14,3170.98×0.97×0.96×1.01×47.7% of 465
Has a question mark4,4660.97×0.97×0.98×0.96×53.1% of 209
Says you or your1,4120.97×0.99×0.95×0.99×56.8% of 74
Has an exclamation mark1,1780.92×0.92×0.88×0.95×Too few to compare (28)

Read across a row. Starting in lower case came out about 4% ahead in both halves, but was ahead in only 53.0% of sets. Questions came out about 3% behind on average, yet were ahead in 53.1% of sets. When the size and the direction change with the test, the honest reading is no clear effect. Exclamation marks were behind in both halves, about 8% overall. But only 28 sets had enough versions both with and without one to compare directly, so treat that as a hint at most.

Some features couldn't be tested at all. Every video kept one style for capitals, so no video had versions both ways, and only 20 sets had enough versions with and without an emoji. That matters, because the raw numbers look striking: of the posts we read, those whose words were all in capitals reached 100 views 48.1% of the time, against 40.2% for the rest. But that gap compares different videos and campaigns, not capitals. It's the trap in most hook advice: comparing hooks on different videos mixes the hook up with everything else about the video.

Did the frame around the video change views?

Views differed more by frame than by wording. Our video model Flagship puts a video's versions in different frames. We compared frames the same way, inside 254 video-and-app sets of Flagship versions with 20 or more posts (9,439 posts from 76 uploaded videos).

FramePostsViews vs the same video's other versionsRead posts onlyHalf AHalf BSets where it was ahead
Video inside a YouTube page3021.45×1.35×1.14×1.60×61.3% of 31
Social post card, hook as the headline1,7601.11×1.13×1.05×1.16×65.7% of 137
Card on a plain background6721.10×1.17×1.11×1.10×53.0% of 83
Chat or DM screen9970.98×0.91×0.92×1.03×53.1% of 130
Close-up crop, full screen5930.92×0.99×1.22×0.80×51.8% of 56
Wide video placed on a page4,2590.94×0.96×0.99×0.90×44.3% of 149
Picture-in-picture8560.86×0.81×0.79×0.89×43.4% of 76

Here the per-set test had to use 3 or more versions with the frame and 3 without, because each frame is a smaller share of a set. The social post card is the steadiest result: ahead in both halves, with and without the posts we never read, and in 65.7% of sets (65.5% of 145 counting only posts we read). In that frame the words on screen become a post's headline, so the hook is the first thing a viewer reads. Picture-in-picture trailed in both halves and in most sets.

The frames didn't do the same on every app, though. This table counts only posts we read, because about half of our Flagship posts on TikTok (47%) never gave us a view count. Counting those as 0 moved some TikTok cells a lot: chat screens to 1.36×, the close-up crop to 0.92×.

FrameTikTokYouTube ShortsInstagram ReelsFacebook
Social post card0.82×1.07×1.32×1.32×
YouTube page1.54× (45 posts)1.03×1.41×2.65× (38 posts)
Card on a plain background0.79×1.52×1.19×0.90×
Chat or DM screen1.12×0.91×0.80×0.94×
Close-up crop1.27×0.81×1.23×0.94×
Picture-in-picture1.00×1.08×0.65×0.54×
Wide video on a page1.04×0.91×0.94×1.05×

Every cell has at least 38 posts, but many have under 200, so single cells are noisy. The pattern is still clear: no frame won everywhere. The social post card trailed on TikTok, where the YouTube page and the close-up crop did well, and picture-in-picture did fine on TikTok and Shorts but poorly on Reels and Facebook. A mix of frames, checked app by app, beats betting on one.

What should you test instead?

  1. Test hook ideas, not rewordings. A different first scene, a different claim, a result shown first. Versions with reworded lines landed within a few percent of each other in most of our tests.
  2. Compare hooks on the same footage, on the same app. Otherwise the hook gets mixed up with the video, the campaign and the account.
  3. Give each hook enough posts. Single posts are very noisy. Our numbers on how many posts a fair test needs are in ad creative testing without an ad budget: about 20 posts per creative, and a winner only when the gap is about 2 times.
  4. Try more than one frame per app. Versions in different frames differed more than versions with different wording, and the best frame changed from app to app.
  5. Follow TikTok's timing. Hook in the first 6 seconds, proposition in the first 3, and text at a readable 5 to 10 words a second, per TikTok's help page above. We didn't test timing ourselves.

Where Abondy fits

Flagship 1.2 makes many versions of one upload, each with its own hook text on screen, its own frame, crop and caption, and real people post them on their own accounts. To test hook ideas rather than wordings, upload each idea as its own video and compare their medians on one app. You can see versions it has made on Explore. Each plan includes free videos to make versions from: 100 a month on Plus, 1,000 on Pro and 2,500 on Ultra. A post costs 12 credits, about 16¢ to 20¢ a post, and since November 20, 2026 you're charged only for posts that reach 100 views within 24 hours of the link being handed in. Plans start at $20 a month. We make Abondy, so weigh our part with that in mind.

How we counted

  • Sample: counted posts made from August 29 to November 12, 2026 (New York time), pulled on November 21, 2026. Every post was at least 7 days old when pulled.
  • Words: 666 sets (one uploaded video's versions on one app, 20 or more posts each), 26,423 posts, 212 uploaded videos, 9,576 different on-screen captions, from 6 campaigns. One brand's campaign accounts for 45% of these posts, and Abondy's own two campaigns for about a quarter (24%).
  • Frames: Flagship versions only: 254 sets, 9,439 posts, 76 uploaded videos, 4 campaigns. Two brands' campaigns account for nearly nine in ten of these posts, and Abondy's own for about a tenth.
  • Who's in it: mostly gaming clips, largely Roblox, posted on small accounts. The words were rewordings of one line per video, most of them written by our tools.
  • Never read: 11.3% of the posts (14.4% of the Flagship ones, 47% of the Flagship ones on TikTok) never gave us a view count and are stored at 0. The "read posts only" columns and the table by app leave them out.
  • Views: each post's count when we last read it, not a lifetime total. A few posts were last read before they were 7 days old.
  • Exclusions: Abondy's own test accounts, posts whose views were confirmed as bought, and posts that were flagged, rejected or deleted.
  • Instagram: read the Reels numbers as a floor. Some counts come from each post's public page, and in a 16 Nov 2026 check of 37 posts that page showed about a third of the real count.
  • Associations, not proof: versions differed in more than one way at once, and sets share campaigns, so they aren't independent. These numbers describe our posts, not what a hook of yours will do.

Questions people ask

What are UGC ad hooks?

The opening of a UGC video that makes someone stop scrolling: the first scene, the first line, or the words on screen. TikTok's help center advises putting the hook in the first 6 seconds and the main point in the first 3.

Does changing the text hook change views?

Versions with a reworded hook got similar views in our data. Questions, short lines, "you" and lower-case starts each landed within about 5% of the same video's other versions, and none led in clearly more than half the sets. Views differed more by the frame around the video.

How long should a hook be on screen?

TikTok recommends 5 to 10 words per second for text. In our data, hooks of 1 to 5 words got 1.03 times the views of the same video's longer versions, close to no difference.

What is the best frame for a UGC video?

None won on every app. Social post cards with the hook as the headline came out ahead of the same video's other versions overall and on Reels, Facebook and Shorts, but trailed on TikTok. Picture-in-picture trailed overall.

Read next