Why captions matter so much
In a 2019 survey of 5,616 US adults by Verizon Media and Publicis Media, 69% said they watch video with the sound off in public places, and 80% said they are more likely to watch a whole video when captions are available. Feeds on Instagram, Facebook, LinkedIn and X start videos muted, and browsers generally let a video start on its own only when it is muted, so the same is true of most autoplaying videos on websites.
There is also the plain matter of access. People who are deaf or hard of hearing cannot use a product video without captions, and the Web Content Accessibility Guidelines (WCAG 2.1) list captions for recorded video among their most basic requirements, at Level A. If your customers include schools, public bodies or large companies, they may be required to buy from vendors whose materials meet those guidelines.
Two kinds of captions
Burned-in captions, also called open captions, are part of the picture. They show on every platform, in every player, with no setting to turn on, and you control exactly how they look and where they sit. Their drawback is that nobody can turn them off or translate them, and a search engine cannot read them.
Subtitle files, also called closed captions, are text files that travel beside the video. The player shows them when the viewer asks for them. They can be switched off, offered in several languages, and read by YouTube and other platforms to understand what your video is about.
For a product video the practical answer is both: burned-in captions for the muted feeds and autoplay, and a subtitle file for YouTube, LinkedIn and your own player.
What a subtitle file looks like
The two common formats are SRT and WebVTT. Both are plain text: a time range and the words to show during it.
An SRT file numbers each caption and uses a comma before the milliseconds:
1
00:00:02,900 --> 00:00:06,800
Reports that write themselves
WebVTT, the format web browsers use natively, starts with a WEBVTT line and uses a full stop instead of the comma. Most platforms accept SRT; YouTube accepts both, along with several other formats. Converting one to the other is a matter of changing those details.
Headline captions and word-for-word captions
There are two styles of caption, and product videos use both.
Headline captions put a short phrase on each screen, three to eight words that state its point: "Every level on one chart". They are designed to be read at a glance by someone scrolling past, and they work even for viewers who never turn the sound on. Most product videos show one per screen, in the same place throughout.
Word-for-word captions show what the voice says, as it says it. They are what a deaf viewer needs and what a subtitle track in a player is expected to contain. On tall 9:16 videos, a popular style shows the spoken words a few at a time in large type, highlighting the word being said.
ShotDemo draws a headline caption on every screen, in one of three styles, and on 9:16 it can add the spoken words a few at a time over the lower part of the screen. The subtitle file that comes with each video carries the on-screen captions with their timings. For a word-for-word track, YouTube's automatic captions of the voice are a good start; check the product names, which automatic captions often get wrong.
Writing captions that read well
- Keep lines short. Broadcasters and streaming services generally allow about 32 to 42 characters per line, and at most two lines on screen at once.
- Give people time to read. A caption should stay up at least a second and a half, longer for longer lines. If it flashes past, it is too long for its moment.
- Break at natural pauses. Split lines between phrases, not in the middle of a name or between a number and its unit.
- Spell names the way they are written. If you spelled a product name out for the voice, make sure the captions still show the real name. A caption reading "O-I Data" instead of "OIData" looks careless.
- Write numbers as numbers. "3 minutes" reads faster than "three minutes" in a caption, even if the voice says the words.
- Match the voice's meaning. A headline caption can be shorter than the spoken line, but it should never say something different.
Where captions go in each shape
In a wide 16:9 video, a caption row along the bottom is conventional and rarely covered by anything.
In a tall 9:16 video, the bottom of the frame is where Reels, Shorts and TikTok draw the account name, the post text and the sound, and the right-hand side holds a column of buttons. Captions placed there are covered. Put them higher, roughly in the middle third of the screen, or at the top, and keep them clear of the right edge. The video sizes guide describes these zones.
Wherever they go, captions must not sit on top of the part of your interface the line is talking about. Keep the caption row separate from the screen itself, so the two never fight for the same space. In ShotDemo every frame ends above the caption row, in every shape, so a caption never covers the product.
Contrast and style
Captions must be readable over whatever is behind them. White text over a bright spreadsheet disappears. The reliable options are a solid or semi-transparent box behind the text, or a strong outline around each letter. Use a plain, heavy typeface, large enough to read on a phone held at arm's length, and keep the same style for the whole video.
What each platform does
- YouTube accepts uploaded subtitle files in SRT, WebVTT and other formats, and also generates automatic captions from the voice, which you can edit. Uploaded captions are shown when viewers turn captions on.
- LinkedIn can create captions automatically, currently in ten languages, and lets you review them before posting.
- Facebook accepts uploaded SRT files for videos posted from a Page.
- Instagram and TikTok can generate captions automatically, which viewers can turn on. Burned-in captions remain the safest way to control what muted viewers see.
- Your own website: HTML video players accept a WebVTT track, which viewers can switch on from the player's controls. The example videos on this site work that way.
A caption checklist
- A headline caption on every screen, three to eight words, saying that screen's point.
- Captions burned into every version that will play in a feed or autoplay muted.
- A subtitle file for YouTube, LinkedIn and your own player.
- On 9:16, captions above the bottom band and clear of the right edge.
- High contrast, one style throughout, product names spelled correctly.
- A final watch with the sound off: does the video still make sense?
That last test is the one that matters. If a muted viewer can follow your video from start to finish, the captions are doing their job.