What Makes a CPI Video Work in the First Three Seconds

The first three seconds of a CPI video determine whether anyone watches the rest. In a feed environment, the viewer is scrolling. Your video has three seconds to create enough curiosity or recognition to stop the thumb. Everything after second three, the gameplay, the payoff, the call to action, depends on winning those first three.

Why three seconds is the number

Platform data across Facebook, TikTok, and ad networks consistently shows that most viewers decide within the first three seconds whether to continue watching or scroll past.

A video with a strong opening and a weak middle outperforms a video with a weak opening and a strong middle, because the strong middle is never seen. The opening is a gate, and the gate determines the audience for everything after it.

This is measurable. The video watch rate, specifically the percentage of viewers who watch past three seconds, is the first metric to check on any CPI creative. A watch rate below twenty-five percent means the opening is failing, regardless of how good the rest of the video is.

What the opening must show

Two things in three seconds. Not one, not three.

The core action. What the player does. Not the result, not the consequence, but the physical action. A finger tapping, swiping, dragging, or drawing. The viewer must understand the interaction in a glance.

A visual that is unlike the surrounding feed. The feed is photos, text, and other ads. Your three seconds need to look different from all of them. High contrast, unusual colours, or a visual that does not match the context of the platform. The goal is pattern interruption, not beauty.

Both must be simultaneous. An opening that shows the action but looks generic gets scrolled. An opening that looks striking but does not show gameplay gets clicks from people who do not want your game, which raises CPI.

The first frame is the thumbnail. If you pause the video at frame one and it does not communicate "this is a game and here is what you do," the opening needs work.

What the opening must not show

Five things that waste the three-second window.

• A logo. Nobody stops scrolling for a logo they do not recognise.
• A loading screen. Real or fake, it looks like dead time.
• Text explaining the game. Viewers do not read in the first three seconds. They decide visually.
• A menu or UI. Interface elements look like every other game ad.
• A slow build. Cinematic openings, dramatic pauses, and fade-ins lose the viewer before the game appears.

Every one of these can appear later in the video. None of them can appear first.

The thumbnail test

A fast test for any CPI video opening.

Take a screenshot of the video at the one-second mark. Show it to someone who has not seen the game. Ask two questions.

1. What kind of game is this?
2. What does the player do?

If they can answer both from a single still frame, the opening works. If they cannot, the opening does not communicate fast enough. The still frame is what the viewer processes in the fraction of a second before deciding to scroll or stop.

This test is cheap and immediate. Run it on every video variation before spending on the campaign. Two rounds of this test, each taking five minutes, are worth more than optimising the video after the spend.

Three patterns that work

These are not the only openings that perform, but they are the most reliable starting points.

Pattern one: the impossible attempt. Show the player trying something that looks easy and failing in an unexpected way. The viewer watches to see the resolution. This works for puzzle games, physics games, and anything where failure is visual.

Pattern two: the satisfying action. Show the core action at its most visually satisfying. Slicing, sorting, popping, building. No context, no setup, just the action. This works when the action itself is the appeal.

Pattern three: the before and after. Show a messy or broken state, then cut to the transformed result. The gap between the two creates curiosity about how it happened. This works for decoration, cleaning, makeover, and building games.

All three communicate within a single second what kind of experience this is. The remaining two seconds of the window confirm it and add a question that the viewer wants answered.

The broader hook structure is in Ad Hook Four Beats, and these three patterns all fit within the first beat.

Testing two versions fast

You do not need a large budget to test CPI video openings.

Step one. Create two versions of the same video with different openings. Same game, same gameplay footage, different first three seconds.

Step two. Run both on the same platform targeting the same audience. Budget: $50 to $100 per variation for two days.

Step three. Compare three-second watch rate. The version with the higher watch rate has the stronger opening, and that is the version to develop further.

Do not test more than two at once on a small budget. Two gives you a clean comparison. Three or more splits the data too thin to be reliable at low spend. The testing method scales with Creative Testing on $500.

If both versions perform below a twenty-five percent three-second watch rate, the concept visual may be the problem rather than the opening execution.

Common failures

Four patterns that consistently underperform.

The first failure is the most common. A CPI video showing typical genre gameplay without a visual twist performs at or near the genre average, which is usually too expensive for an unproven title. The opening needs to show something the viewer has not seen in the last ten ads they scrolled past.

What we would do

Test the opening separately from the rest of the video. Create three opening variations and attach the same body to all three. Run them for two days and measure three-second watch rate only. The winning opening goes into the full creative test.

Use the thumbnail test before spending anything. Two rounds of five-minute tests with people who have not seen the game catch the obvious failures that budget would have measured expensively.

Record the opening at the highest possible contrast and the simplest possible framing. A close crop of the core action against a clean background outperforms a wide shot of the full game screen in almost every test we have run.

The short version

• The first three seconds determine whether anyone sees the rest of the video.
• Show the core action and a visually distinct frame simultaneously in the opening.
• Never open with a logo, loading screen, text, menu, or slow build.
• Run the thumbnail test: pause at one second and ask what the game is and what the player does.
• Test two openings against each other for $50 to $100 per variation over two days.
• A three-second watch rate below twenty-five percent means the opening is failing.

Screenshot your current CPI video at the one-second mark and run the thumbnail test today. If you want CPI creatives that clear the three-second gate, talk to us.

Related reading: Ad Hook Four Beats, Creative Testing on $500, and CPI Benchmarks by Genre.