May 1, 2017

When I visited a dozen news organizations on our Innovation Tour last November, there was a surprising consensus among the biggest players on the next big thing they had in focus:

Voice-activated news for devices like Amazon Echo and Google Home.

As was true for Virtual Reality/360-degree video a year earlier, the medium is so new that publishers are just beginning to figure out where to start. But they believe that the eventual opportunity could be huge.

Related training: Poynter 2017 Innovation Tour — inside America’s leading news organizations

Checking back six months later, I find that the art and science of voice-activated news is still in its infancy. No consensus business model has emerged. That probably explains a dearth of coverage to date of voice as a media story.

But lots is going on beneath the surface:

NPR is the leader in voice news, coming out of the gate.

If you are a novice Echo user like me and ask Alexa for the latest news, you will get a current NPR summary read in a familiar announcer’s voice.

That’s by design. NPR secured a coveted spot as default provider. (A more sophisticated user could pick from a variety of other options and move a favorite to the top spot in its place).

Joel Sucherman, NPR’s senior director for digital products, explained the match in a phone interview. “You could say that radio is the original screen-less news experience,” he said.  “So it’s…natural…and the Echo can be something special now that people often no longer have a radio in the kitchen.”

Preliminary metrics suggest fast adoption of the in-home devices, he added, and show that those interested in news will return regularly for updates, daily or more often, as use of the voice option becomes a habit.

In-home is also a match to NPR One, an app which allows users to get a current summary of news on demand with a customizable supplement of podcasts and other features.

Even before talking to Sucherman, I had guessed that a human broadcaster would be preferred to the synthetic text-to-speech voice of Alexa.

His colleague, senior developer Ha-Hoa Hamano, said that proved to be doubly true.  “At first we had an introduction to NPR One in Alexa’s voice,” she said, “and that was just a hand-off.  We swapped that out with a familiar NPR voice, and (that) did a lot better.”

Like others in the space, Sucherman sees lots of work ahead, “to try to do more than just a command-and-response.” But he said that he is confident this is “more than a passing fad. It will be important.”

More complex news conversations will spawn broader competition.

The current generation of news summaries is just the start, said Francesco Marconi, a senior strategy manager who heads development of voice for news for the Associated Press.  With deep experience in broadcast, AP is well-positioned to write “a suite of top stories” to be spoken by Alexa or a human voice, he said.

But as time goes on, Marconi continued, “the holy grail is to have structured conversation with the device…No way those will be served by a human voice.”  The dialogues will instead rely on answers generated by artificial intelligence and rendered by machine speech.

Futurist Amy Webb offered this example in a column in the current issue of Nieman Reports, making the case that news organizations need to up their voice and artificial intelligence game:

Voice will offer a fluid connection to journalism. It could elevate the role news organizations play in our everyday lives from informational to indispensable. Beyond just asking what’s the news, we’ll be able to interrupt to ask more questions and to better understand the context. Wait, who’s Assad, exactly? Why is he such a bad guy? Has he ever had meetings with Donald Trump? What did they talk about? Who else was in the room? Why should this matter to me?

Voice response to news questions pulling information from databases is close at hand, Marconi said What’s the crime rate in my city?  How does it compare to others the same size?  Is that up or down from 10 years ago?

The AP has already invested in artificial intelligence, machine-producing thousands of routine stories on quarterly financial reports and minor league ball games. So that next step is an easy match.

A little further down the road, Marconi said, the devices may offer a machine version of empathy.  “It may be able to tell from the tone of your voice whether you are a happy or sad and tailor a response appropriately.”

Amazon offered a non-news preview of that feature this week, announcing that the “Echo Look” will have an enhanced version equipped with a $200 camera that can give users a real-time critique of outfits they’re trying on as they’re getting dressed.

Local news companies are experimenting with voice, too.

As with VR/360, one wonders whether what makes sense for big, resource-rich organizations has any relevance to a small-town newspaper or local broadcast station.

The companies that own those properties are trying to answer that question, too.. In fact, the GateHouse chain of newspapers is already producing voice local reports on 400 websites.

The process is straightforward, said Bill Church, GateHouse’s senior vice president for news  With help from the company’s news center in Austin and Amazon’s technical team, the websites write summary feeds that can be read by Alexa.  No human review is needed.

Now, Church added, local sites are getting the added option of having the summaries read by a local broadcaster. GateHouse’s initial deal was with Amazon, but the company is opening discussions with Google, as are national news providers, to be on Home as well.

And there is potential, he said, as with vertical websites and newsletters, to have specialized reports with potential for a national audience like one on Alabama football at GateHouse’s Tuscaloosa News.

While initial traffic has been encouraging and GateHouse wanted a first-mover advantage, Church said, the effort really is still a learn-as-you-go experiment.  “It’s product first, then audience, and monetization later,” Church said.

The most fundamental of business model basics are up in the air.

Sucherman told me that in NPR’s big front-of-the-line placement deal with Amazon, no money is changing hands.

As the practice ripens, will the platform companies pay publishers and broadcasters for content on the cable TV model? Will content providers pay to be carried or for a prominent position?  Or will the two split advertising revenues if the voice medium ends up having the equivalent of video pre-rolls?

News has a place but will need to compete for time and mind share with other uses of home devices.

Playing music or turning on lights and appliances may turn out to be much more popular.  (Alexa had no trouble when we asked her recently for a selection of tracks from our across-the-street neighbor Charlie’s rock band, We the Kings).

I take heart from the Amazon promos for the Echo that lists news among top uses for the voice service.

Amazon in February announced that in the product’s short, two-year lifespan to date it recently passed the 10,000 mark what it calls “skills” (I would say functions or applications).  There are a host of news organizations not discussed in this piece which have taken a spot in the space and are doing heavy R & D to expand:  national broadcast networks, local radio and television stations, the New York Times and, natch, the Jeff Bezos-owned Washington Post.

For all that promising activity, I would concur with Webb that news leaders need to be sure to secure a seat at the voice table early. Otherwise, if they come later to the party, as with Facebook and Google, they will be forced to sort through a structure developed without much of their input.

Support high-integrity, independent journalism that serves democracy. Make a gift to Poynter today. The Poynter Institute is a nonpartisan, nonprofit organization, and your gift helps us make good journalism better.
Rick Edmonds is media business analyst for the Poynter Institute where he has done research and writing for the last fifteen years. His commentary on…
Rick Edmonds

More News

Back to News