Converting your text-based, long-form content into audio versions sounds like an easy – and smart – thing to do.
After a few months of working with an AI-based audio-conversion tool on the Content Marketing Institute blog, I still think it’s an intelligent move, especially given the rise in voice-activated devices and audio-delivered search results. But I now know doing it well requires more than a simple click of a button.
We began our audio quest for several reasons. In a reader survey, many of you said you wanted alternative ways to consume our content. Audio versions make the articles accessible to a broader audience, from those with visual impairments to those who consume content while commuting, working out, or cooking.
Gini Roberts, senior director of media and sponsorships for Convince & Convert, points out another audience segment that benefits from audio content: Multilingual visitors who prefer to listen to (rather than read) content that’s not in their native language.
When Convince & Convert started offering audio blogs in 2015, the team would read and record the audio directly. They’ve since moved to a WordPress plug-in that requires minimal quality control.
The variety of tools available today means you can choose what makes sense for your goals. The options range from basic text-to-speech programs to platforms that let you customize the audio experiences.
If you choose the plug-in route, select one that’s reliable and keep it updated, advises Michelle Saunders, director of content at Convince & Convert. She also offers this tip: “Don’t have audio auto-play. It not only takes control away from users, but it also takes up unnecessary bandwidth, slows speed, and negatively impacts SEO.”
If you decide on a platform that lets you customize the audio experience, pay attention to these factors to enhance the audio version of your blog or other long-form content.
1. Dedicate team member for quality control
If your audio content contains mispronunciations, you risk confusing or alienating your audience. Assigning responsibility for quality control guarantees the audio will offer at least a satisfactory experience.
Fox News recently started audio versions of its articles. You can hear how a lack of quality control affects the audio experience in this piece about the worst Marvel movie.
The mistakes start with the headline: The Worst Marvel Movie Yet Is Also the Cringiest. The last word caused the problems. The audio player pronounced “cringiest” with a hard “g”. The audio player also uses a monotone robotic voice. Around the 1:35 mark, you can hear it mess up a link in the text.
It seems no human working on the site actually listened to the audio. In this case, the execution mistakes make the audio less than helpful — and will likely dissuade listeners from seeking other content from the site.
For a good user experience, someone should preview the audio file to fix errors and pronunciations. This person also should tell the audio tool what to read – and what not to read.
In the following example, the article’s audio call to action is highlighted in yellow, which tells the audio tool to read it. However, the related articles with links at the end are in purple, indicating that the audio tool should not read the short code CTA or the list.
2. Build a pronunciation library
No audio AI tool knows how to pronounce all the words in the world. To help it understand your content, you can build a pronunciation library. Perhaps you have products that are hard to pronounce just based on their letters or authors’ or subject matter experts’ names that aren’t easily read.
For example, our tool didn’t know how to pronounce the last name of frequent CMI author Ann Gynn. I added brackets before and after “g,” so the player would recognize the hard G. It’s still not perfect, but it’s close enough. (Ann tells people her last name sounds like the second syllable in “begin,” but that explanation doesn’t work with an AI tool.)
If you’re stuck on how to phonetically spell out a word, free tools like HowToPronounce.com can help. But please know it’s a trial-and-error process to get it right.
You can also ask authors and subject matter experts to create an audio version of their name. Easy and free tools like Namedrop.io (registration required) let people record their pronunciation and provide a link they can share. I added the “Hear my name” link to my email signature (shown in the image below) so no one has to wonder how to pronounce my name.
3. Make a full production
Create an audio experience rather than a transcription-like reading. Add intro and outro music as audio signals for the audience.
Customize the audio player. You can add a thumbnail of the article’s primary image or your brand’s logo. You can even personalize it with your brand’s color palette on some platforms.
Listeners are used to having options for the speed at which the audio plays, so make sure to look for a player that offers that feature.
4. Customize the voices
Most AI audio tools let you pick from a menu of voices. But some let you clone a real person’s voice, so the experience sounds more authentic and natural.
Ask your regular contributors to record themselves reading something (an article you have already published, for example) for three to five minutes. They should use a voice style that works with their content. For example, an upbeat, playful tone will result in an upbeat, playful clone. Ensure they use a high-quality microphone and record in a quiet place without background noise.
Before your audio provider clones their voice, get them to sign a voiceover licensing agreement. Include terms that clarify who owns the rights, title, and interest for the performance and all intellectual property rights arising from the performance, including any copyrights. Also, make sure that the AI company agrees to use the performance only to narrate literary works or materials owned or otherwise provided by your company. I’ve included a free AI voiceover legal template (registration required), but you should consult with your organization’s legal team, too.
5. Tell the audience
Make the audio version easy to find on the web page. In CMI’s case, the audio version appears just below the article image and byline.
Let listeners know that the voice isn’t the author’s (or any human’s), especially if you use a tool with realistic (non-robotic) voices. For most CMI articles that include audio, the intro states, “Synthetic voice and sound recording created by Everlit.”
If the voice is the cloned version of the author’s, the disclosure states, “Synthetic voice and sound recording created by Everlit based on the voice of [insert cloned voice name of person] with their permission.”
Go beyond standard transparency to ensure your audience knows the audio option exists. Promote the new feature in your newsletters and on your social channels. (You could even write a blog article about it.)
6. Tie to the business purpose
Just because your audience is listening doesn’t mean they can’t take an action that connects your content to the business goals.
Insert an audio call to action into your content. Structure the CTA verbiage so listeners know what they’re getting before you provide how they can get it. For example, don’t say, “Subscribe now for industry news.” Instead, say, “For industry news, subscribe now at www.URLLinkHere.com.”
CMI uses this CTA for the audio versions: “You can learn so much more from the voices of experience in content marketing. To subscribe to the daily or weekly updates, visit www.News.ContentInstitute.com/Subscriptions or click the orange subscribe button at the top of any page.”
“Having a small budget or small team is no longer a barrier to implementing an audible option for your site,” says Convince & Convert’s Gini Roberts.
And converting long-form text into audio versions expands the distribution options for your content. You can deliver a format that works better all the time for some audience members and some of the time for those who prefer audio when they’re traveling, multi-tasking, etc.
Give the same attention to detail to your audio content as you do to your text-based articles, and you’ll build an audience that wants to keep listening.
All tools mentioned in this article were suggested by the author. If you’d like to suggest a tool, share the article on social media with a comment.
HANDPICKED RELATED CONTENT:
Cover image by Joseph Kalinowski/Content Marketing Institute