Optimizing for audio

NPR’s finding¹ that 20% of online listening now comes from voice-activated, Artificial Intelligence (AI) systems means that more people are ditching screens and opting for personalized and convenient ways to consume media.

News publishers are now faced with a different dissemination process and need to keep up the pace to be part of this technological wave.

Nic Newman, senior research associate at the Reuters Institute, supplied key findings on news integration and use of voice services like Amazon Alexa, Apple HomePod, and Google Assistant the Future of Voice and the Implications for News report.

1 in 10 US adults (14%) now use voice systems (UK 10%, Germany 5%) and numbers are on the rise. However, only 1% of users in the UK rate news as the most valued feature.

Voice services are in the early stages of development, and so far there is little data on how news is being adopted through these products. So, we have highlighted the issues raised in the report and the potential for news publishers and broadcasters.

Less is more

Nic Newman, Reuters Institute for the Study of Journalism

Many news publishers have not adapted well to audio listening. The Voice report demonstrates that people using voice systems are generally frustrated with:

1. unnecessarily long updates
2. poor quality of production
3. lack of regular updates
4. use of synthesised voices, which can be difficult to listen to for extended periods.

News distributors like The New York Times The Daily and the Economist have improved there audio briefings, due to customer criticisms. Reducing the length of updates proved that less is more, as regular users felt more informed with shorter updates (56% in the US and 45% in the UK). This could help increase audience reach and loyalty as people are more likely to opt for a channel that supplies a general, concise update as news consumption is becoming increasingly fragmented across many providers.

Embrace change and experimentation

Considering that voice systems are new, it makes sense for news publishers to be more creative and experiment further with how their audio can be experienced.

The report findings show that voice system users prefer a more human-like voice instead of a ‘roboticized’ translation of current content. Just as the platforms themselves are experimenting with this (Google Assistant’s John Legend voice), news publishers are also testing what works.

The Guardian Voice Lab funding will allow them to test the acceptable range of synthesised voices and interactive formats.

BBC News cut a 45 minute radio interview, mapped answers to different queries, and turned it into an Amazon skill to test with users. Experimentation is key to discovering what works for the audience.

Platforms as competition

Default settings on voice system products are in-built by the platform, which might mean that some news providers have an advantage when it comes to building voice audiences.

• BBC News dominates UK usage by two-thirds (64%)
• (Sky News, 19%)
• both CNN and NPR in the US at 28% each

All of this means that other brands may find it challenging to reach an audience even before innovating and experimenting with how their news is projected.

As many news providers haven’t started using an audio outlet –– and with defaults already in place ––the field could be distinguished ‘by heavy winner-takes-all dynamics’.

However, platforms don’t have to be an obstacle to listening and selecting which news brands are chosen for the user, voice could be an opportunity for news distributors to wrestle users off the two big platforms.

News publishers have a chance to shape the way news is consumed and enhance their monetization through voice technology –– if they see platforms as competition. The Google Digital News Initiative could be one way to use funding to establish tools for cross-platform content creation and other ways to optimize news for audio and voice.