The BBC has launched an AI-powered synthetic voice called Project Songbird which can read out its online articles, The Drum reports. “Project Songbird offers a ‘third way’ to consume BBC digital content, slotting in neatly alongside text and video news offering deepening the relationship we have with our audiences,” said Errol Baran, Global SVP for Business Development and Innovation at the BBC.
We have seen through the success of our podcast business across World Service that the demand for audio and audio based propositions shows no signs of slowing down and this allows our readers to be able to multitask should they wish to do so.Errol Baran, Global SVP for Business Development and Innovation, BBC
“Audio has kind of quietly emerged”
The success of the podcast format has sparked a growing interest among publishers in the opportunities offered by audio. The BBC has now joined the group that is looking beyond podcasts to integrate audio more widely across their content.
Converting text-based articles into audio gives readers an additional way of consuming the content. The strategy is paying back for several publishers who have implemented it.
“While offering articles in an audible format is nothing new, this type of content is currently enjoying a massive boom, as time-strapped audiences look for convenient ways to consume online media,” says journalist Charlotte Ricca. “In the US alone, spoken word’s share of audio listening has increased 30% over the last six years and 8% in the last year alone.”
Audio articles are making quite a noise in publishing. Podcasting may be hogging the headlines. But creating features and news that your audience can listen to is an audio opportunity that should not be ignored.Charlotte Ricca, Independent Media Reporter
“Audio has kind of quietly emerged…as kind of the preferred way for people to consume content,” Jim Bodor, MD, Digital Product Strategy, HBR told NiemanReports. “If audio does become, maybe not the dominant currency of the internet, but if it becomes an even more significant piece, then we have to be prepared for that.”
“Surprised and pleased to see steady adoption and use”
What makes it even more interesting is that publishers are taking different approaches in presenting the audio version of their articles for readers. Some produce it using voice performers (The New York Times/Audm) or journalists (Zetland); others rely on tools that convert text into computerized speech (The Washington Post).
The New York Times has acquired Audm, a company that uses audiobook narrators to convert text-based articles into audio. It had been working with the company since last fall, creating audio versions of select longform articles before acquiring it in March. “This is a way in which long-form journalism can fit better into your life,” Stephanie Preiss, the publisher’s VP of TV and Audio told WSJ.
Audm is also used by The New Yorker, Wired, Rolling Stone, Vanity Fair, The Atlantic, and BuzzFeed News among others. Subscribers pay $7.99 per month, or $56.99 per year, to access the entire library. The app is available to download for iOS and Android devices.
The Washington Post differs from the Times in that it uses computer generated voices. The publisher has been experimenting with audio articles since 2017 and made all its articles available in the audio format on its Android and iOS apps in July. It uses the text-to-speech features on Android and iOS mobile operating systems to do so.
When we first began testing audio articles on our Android app earlier this spring, our thinking was that this feature would appeal to readers during their commutes. We’ve been surprised and pleased to see steady adoption and use over the last month as many people continue to work from home.Leila Siddique, Senior Product Manager, The Washington Post
Subscribers who listen to audio articles spend 3x times on the app
While the Post’s strategy is to give its audience the option to listen to any of it’s articles, the Times is focusing on quality (which limits the number of articles available as audio).
NYT’s Preiss says, “We’re interested in building products worth paying for. We feel like ‘worth paying for’ doesn’t include computer-sounding voices.”
The Post’s Managing Editor Kat Downs Mulder comments, “We are excited about the habit-forming nature of audio. But it’s hard to build those habits if you only offer a limited number of articles.
“Our readers know they always have an audio option, across all articles, which really reinforces its availability.”
Mulder adds that subscribers who listen to audio articles spend 3x more time in the Post’s apps compared to those who don’t.
We know people love listening to news – our podcasts show that. Opening up more audio content and making it easier for them to engage with The Post helps draw them into our ecosystem even further. That improves retention rates.Kat Downs Mulder, Managing Editor, The Washington Post
“Important to the success of our audio products”
Meanwhile, the Danish startup Zetland uses its journalists to read the articles. “The personal voice of the journalists is quite important to the success of our audio products,” says Zetland CEO, Tav Klitgaard. “We originally tried using professional voice actors, but it sounded too perfect.
“We like to build what we call a ‘human product’, including all the glitches and quirks of human beings. It is important for us that our content conveys honest passion and curiosity, and we’ve found that to be done best when the journalist narrates.”
The approach appears to be working for them as they have grown 650% since launching the audio model in early 2017.
It turned out that transforming into audio had an enormous impact for us. It has improved retention and member satisfaction. Listeners consume more, they stay longer with each story and it has also brought us an increased feeling of loyalty and commitment.Sara Alfort, Journalist and Community Organizer, Zetland
However, Zetland publishes only three longform articles a day, which makes it possible to have all of them read by its journalists. It may not be possible for publishers with larger outputs.
“Sizable audience, editorial prestige and a potential revenue stream”
The BBC’s Project Songbird appears to have hit upon a solution to this conundrum. The technology can read articles in different tones. “The tone of the voice is also part of our unique selling point and we have gone to a lot of effort to make it warm and welcoming,” explains Baran. “Our ambition is to develop it further so it’s appropriate for the wide variety of different content we offer.”
“In the future, there could be different voice ‘fonts’ which allow for a sliding scale of tones; serious for hard news, elevated for sport, relaxed for features or a mix of them all depending on the individual story,” he adds.
Project Songbird has integrated cognitive and behavioral software which enables it to learn the consumption patterns of its readers. That allows it to prioritize content “based on user preference, while maintaining the BBC’s curation of balanced journalism,” says Baran.
“Qualities that make audio engaging and valuable”
How different publishers approach their audio strategies will obviously depend on their unique requirements and capabilities. What seems likely is that in the long run, audio can help build engagement and loyalty, as well as attract new subscribers and generate ad revenue.
“The success of podcasts has shown that audio journalism, much of it delivered digitally, has a sizable audience, editorial prestige (the first audio Pulitzer will be awarded this year), and — through podcast advertising or subscriptions to audio apps—a potential revenue stream,” writes Gabe Bullard for NiemanReports.
Narrated articles, which exist in between podcasts and text, can achieve these ends, too. The qualities that make audio engaging and valuable are present in both podcasts and narration.Gabe Bullard, Senior Editor, WAMU 88.5