Text overlays and captions in video ads: design for sound off

Most of your video ads are watched without sound. Adding captions after the edit misses the point: sound off is the starting condition, not the exception.

A large share of your video ads gets watched without sound: on the train, at the office, on the couch next to a partner watching a series. The consequence is simple but rarely acted on: your video has to work completely without audio. Captions and text overlays are therefore not a finishing touch, but a design decision you make before the first second is edited.

Why is sound off the starting condition, not the exception?

Feeds default to mute and most viewers leave it that way. Build your video around a voice-over and slap subtitles on as the last step, and you have essentially made a radio spot with moving pictures. The muted viewer sees a talking head, a wall of tiny subtitles and no reason to stick around. The scroll continues, and your hook rate tells you exactly what went wrong the next morning.

So flip the design around. Start with the question: if someone sees this video on mute, do they understand within three seconds what this is and why it concerns them? Audio then becomes an enrichment for the minority who turn the sound on, instead of a carrier the whole ad leans on.

What is the difference between captions and text overlays?

The two get lumped together, but they do different work. Captions make spoken word readable: they follow what is being said, literally, in a steady rhythm, in a fixed spot. Text overlays are editorial choices: the headline in the first second, the core benefit displayed large, the price, the call to action. Overlays summarize and steer, captions transcribe.

• Captions: every spoken sentence readable, in short lines of a few words, always inside the safe zone of the frame.
• Overlays: the message in blocks, large enough to read on a phone in a single glance.
• Together: overlays tell the story to the silent viewer, captions fill in the details for whoever reads along.

A strong video ad uses both layers deliberately. The silent viewer should be able to follow the story on overlays alone. Whoever turns the sound on or reads the captions gets the depth. That way the same ad works for two viewing behaviors at once.

How do you style captions that work in the feed?

Readability beats aesthetics, every single time. That means high contrast between text and background, a sturdy typeface and a size that reads comfortably on a phone screen without zooming. Work with short lines of a few words instead of full sentences, and highlight the key word of each line so even the fleeting viewer catches the essence. Keep the position consistent and stay out of the zones where the platform interface overlaps the video.

Within those rules, style is welcome to be on brand: a signature color, a recognizable typeface. But the moment design touches readability, readability wins. Nobody ever bought because the subtitles were beautiful, and nobody ever scrolled away because the captions were too clear.

If you do not dare to watch your video without sound, it is not finished yet.

What does text pacing do to your watch time?

Pacing is the invisible force behind text that works. Text that stays on screen too long feels slow and invites the scroll. Text that switches too fast loses the viewer. The rule of thumb: every text change coincides with a visual change or a new beat in the story, so the eye and the message move in the same rhythm. That creates the sense of momentum that holds the viewer without them noticing why.

The first three seconds deserve the most attention here. For the silent viewer, the opening text is the hook: it has to land the problem or the promise in one line, large and impossible to miss. Everything that follows only matters if that first line did its job.

How do you test whether your video survives sound off?

The simplest test costs nothing: watch every new video yourself on mute, on your phone, in one sitting. Do you understand within three seconds what it is about? Can you retell the offer without having heard a single word? If not, the video is not finished, no matter how good the voice-over is. Then look at hook rate and average watch time per variant in your data: differences between versions with different opening texts or different pacing tell you exactly which text choices work.

Treat text as a variable you test, just like hooks and angles. The same video with a different opening overlay is a new test, and often a surprisingly cheap way to give a fatigued winner a second life.

Conclusion

Sound off first is not a trick but a design principle: the story runs on visuals and text, audio reinforces. Captions make the spoken word readable, overlays carry the message, and pacing holds the viewer. Choices like these decide whether a video ad gets watched or scrolled past, and they sit at the heart of how we build video ads that need to convert instead of merely look good. Want to know whether your videos hold up with the sound off? Book a call and we will gladly take a look with you.