Are smart speakers the next frontier for journalism - and is there any money to be made there? A panel of speakers at the March 2018 news:rewired explored the potential — and the challenges.
- Nic Newman, senior research associate, Reuters Institute for the Study of Journalism
- Peter Stewart, journalist and author
- Susie Coleman, software developer, Guardian Voice Labs
- Chris Stone, video executive producer, London Evening Standard
- Suchandrika Chakrabarti, founder, Freelance Pod, moderating
Reuters Institute for the Study of Journalism
Smart speakers are growing faster than mobile phones did in the same stage of their evolution. In the UK Amazon is dominant - but in Australia it's Google. It varies around the world, often based on who entered that market first.
But people are using them in very simple ways - very command and control: turn on lights, turn on music. They rarely use it for news - only 1% say it's the most important feature. Less than 1/5 of the audience are using flash briefings daily - and podcasts aren't doing well, either. People seem to prefer them as a more intimate experience, with headphones.
But it's not just about smart speakers - it's about voice interfaces. It's that that is likely to be really disruptive. Voice is spreading to screen-based devices, headphones, cars and is already on your phone. This is only going to grow.
Journalist & author
There are some routes to making money on speech - in-skill sales, for example.
Amazon offers developer rewards for the most engaging skills - but you might bet £1000 or you might get a sweatshirt. Amazon also has a prize for people who have good ideas for research new uses.
You can run adverts in your podcasts, flash briefings or streams.
You can use Echo for promotions and marketing. You can raise awareness, or improve an experience. The FT use it to enhance a weekend city profile. The NYT allows you hear and order media they're reviewing.
There are two boardgames that only work with Alexa. And subscription audio is coming soon.
The Evening Standard had an events-based skill which has been discontinued. It does have a flash briefing on both Amazon and Google's platform.
Google are developing a new version of the news experience, creating a personalised feed of sub-60 second single-topic stories, from different outlets. It's potentially a much better user experience - but with serious disadvantages for publishers. You're ceding control to the platforms, with consequent costs for discoverability and loyalty.
The Evening Standard is working with Google, producing a mix of content that will be part of the experiment.
They're running small and lean, but with a multi-disciplinary group. Audio is expensive - but it's cheaper than video! But there's lots of work to do. What should the Evening Standard sound like, for example?
This is really the beginning of a new dimension of the internet - the nascent audio web.
Guardian Voice Lab
The team is prototyping meaningful voice-first content experiences, to allow The Guardian to develop a strategy. The Year in Review was a gamified audio experience - that allowed you to discover The Guardian podcasts.
They get a lot of data from this experiment - but it's difficult to process meaningfully. Discovery is challenging. And building on a new platform is hard work - it keeps changing around you.
Next time, they wanted shorter, more habitual interactions. They want to launch sooner and iterate - and make use of existing Guardian content.
That led to Guardian Briefing. They're using synthetic voices to read the content, and machine learning to help determine how the slots are filled.
Q. What's holding publishers back from innovating?
Chris: If you're not a broadcaster - content creation. There's a lack of existing audio capacity. Many publishers can't afford to do it without an immediate return. However, audio production is cheap compared to video.
Susie: A human voice is nicer to listen to than a synthetic voice. The technology is getting better, as are the voices. It's terrifying how good the voices are getting. There will always be problems - differentiating 90210 and 90,201 when an actor dies, for example…
Peter: Publishers have been bitten before. Look at video three years ago. There's also a huge discovery problem. It's very difficult to find skills. And even after you find one, there's no reminder that you used it - or what it's called, or how to invoke it.
Nic: My bugbear is data - the platforms have been very secretive about what's going on. And getting data on your own product is hard.
Q. How can you promote your audio and increase discovery?
Peter: Podcasts are quite hard - you need to invoke the podcast platform and then the actual podcast. Discoverability is hard: getting people to use the app is not easy. It's five or six taps to find a flash briefing.
Chris: Obviously we have the advantage of having a newspaper and website. We have ads running on our websites and videos. We have articles in the paper. And we're getting coverage on social. Time magazine used a sticker on their Twitter page.
Q. To what extent is this a pull or a push process? Is it driven by the publishers or the platforms? And can audio become searchable?
Nic: It's a huge amount of push. The platforms want us to do stuff on screens, because they want to sell them, but there are few people using them.
Susie: Discovery is really, really hard. On your phone you can find it by searching the app store. On your speaker, if you get the invocation wrong, you get nothing. We need to be clear about telling people what they need to say to get the product.
Chris: Brand are exploring this space, too. That creates some opportunities for interesting relationships. Can you have a call-to-action in your flash briefing that hands off to a brand?
Peter: Google is transcribing all sorts of audio across the web. It will become more discoverable.
Elisabetta Gardini has also written up a summary of the session.