Providing an accurate transcript can make your video and audio content more accessible for all users. It’s also a good practice for search engine optimization (SEO): Transcripts make your content more readable for search engine crawlers.
National Public Radio (NPR) began providing transcripts for its radio show This American Life in 2011. According to the nonprofit, about 7% of all unique web visitors viewed at least one transcript. Over 6% of unique visitors who came from search traffic landed on a transcript page.
But in order to realize those types of results, transcripts must be written for real human readers — not for search engines. Automatically generated transcripts might not be good enough, especially if your videos have multiple speakers.
In this article, we’ll provide a few best practices for writing transcripts for accessibility.
For video transcripts, include important visual information
Transcripts are useful for a wide range of users, including those with hearing and vision disabilities. Ideally, they provide an equivalent experience to video or audio: A person who reads the transcript will be able to understand the content just as well as any other user.
Make sure to include a description of all non-speech information needed to understand the content. Some quick examples:
- An interview begins with a slide that identifies the speakers and the subject of the discussion. Your transcript should include all of the text from the slide.
- A video shows a babbling brook while a narrator talks about clean energy. The transcript should briefly describe the babbling brook (in this case, “babbling brook" is descriptive enough to get the point across).
- A video podcast shows two people sitting in a room discussing baseball. You don’t need to describe the people or the room — but if one of the speakers mimics a batter’s swing, that would be relevant visual information to add to your transcript.
Note that this also applies to video content that does not include audio. In order to satisfy the requirements of the Web Content Accessibility Guidelines (WCAG), you must provide a text alternative for all non-text content.
Related: Examples of Text Alternatives to Non-Text Content
Don't include "filler words" unless they're important
Human speech isn’t perfect. We need time to think, so we naturally use filler words like “uh,” “ah,” and “um" when collecting our thoughts. We might also restart a sentence several times when deciding how to get our messages across.
In verbal speech, that’s rarely an issue — but in a transcript, filler language can be distracting and confusing. While some resources insist on verbatim transcripts that are 100% identical to the source material, we believe that edited transcripts are the better option for the vast majority of cases.
Remember, you’re writing transcripts for real humans, and as we’ve discussed in other articles, clear and concise language is better for accessibility. Filler language should only be included in a transcript when it affects the meaning of the content.
For example, imagine a podcast that features this line of dialogue:
SPEAKER: Um, what I think — what I’m saying is, uh, that you — you’ve got a great point there.
That contains quite a bit of filler language, which probably isn’t important. “You’ve got a great point there," is probably sufficient for relaying what the speaker wants to say, and it’s certainly easier to read.
However, if the speaker is stammering because he’s nervous or excited, the filler words might convey something important. Additionally, if you’re transcribing a scripted video, the filler words are part of the script and should be included.
As with many other accessibility issues, human judgment is important. Do your best to determine whether filler words are relevant.
Related: Web Accessibility Is About Progress, Not Perfection
Make transcripts more readable by including lists, images, and headings
Your transcript doesn’t need to be a massive wall of text. Sometimes, you can provide relevant context by including images or graphs that relate to the speaker’s point. For example, if the speaker is describing the features of a Golden Retriever, you could include an image of a Golden Retriever (with appropriate alternative text) to provide a better experience for visual users.
Likewise, you can (and should) break up large blocks of text with other elements. You’re writing for real people, so readability is important.
Some quick tips:
- For longer conversations, include a new subheading when the subject changes. Make sure that the subheading is descriptive. Learn the best practices for writing descriptive subheadings.
- Use HTML lists where appropriate. Make sure to use unordered lists when order doesn’t matter; use ordered (or numbered) lists when the order affects meaning. Learn more about ordered and unordered lists.
- Break up longer paragraphs. While there’s no hard-and-fast rule, try to keep paragraphs to about 2-3 sentences. Start a new paragraph any time the speaker changes the subject.
- Add hyperlinks where relevant. For example, if a speaker refers to a recent event, you could link to news coverage of that event. Be sure to follow the best practices for choosing link text.
- Many transcription services automatically include timestamps, but don’t include timestamps unless they add context.
Make sure your readers can find your transcripts
Finally, remember that your transcripts aren’t especially useful if people can’t find them. Put your transcripts on the same page as your video. If this isn’t possible (for example, you’re hosting the video on a website that has a character limit), provide a hyperlink to the transcript on a separate webpage.
By taking the time to transcribe your content, you’re making your website into a much better resource. If you’re already writing captions, you won’t have to do much extra work — and the effort will pay off.
For guidance with transcripts, captions, or another specific accessibility issue, send us a message to connect with an expert.
