mart speakers are already making inroads into the most currently commonplace listening mode: broadcast radio. From Pandora to Beats 1, the short history of streaming platforms has been marked by mimicry of radio’s free, passive mode of music circulation. Amazon’s Luke started his career as a radio program director, for massive rock stations like Chicago’s Q101 and Dallas’ The Edge, and his early programming initiatives for the Echo were clearly drawn from this experience. Last November, Echo owners who said, “Alexa, play the U2 Experience,” were dropped into a live broadcast that mixed tracks from 2017’s Songs of Experience with band interviews. U2 described it on their website as a “new type of radio.” Asking Alexa to “play The Soundboard” earlier this year cued up a live, career-spanning Elton John program. Daily programs like “Today in Music” and “Song of the Day,” which launched complete with their own specific voice commands, also suggest the strong influence of broadcast radio programming’s liveness.
All this makes the radio industry nervous, and with good reason. The NPR/Edison study backs up Jacobs’ trepidation, reporting that 39 percent of smart speaker users are now spending time listening to these devices rather than broadcast radio. When Amazon isn’t replicating radio programming and simulating its experience, the company is relying on its role in the promotional ecosystem to accommodate voice-specific requests. When Redington noticed that Echo users were asking for “the latest song” by an artist, he realized that simple release date metadata wasn’t enough to serve up the proper result. “We actually had to understand which song is being played at radio, so radio impact date became really important to us.”
When Australian indie rocker Courtney Barnett was featured on Amazon’s “Today in Music” program, Jessica Page, the director of digital at Barnett’s label Mom + Pop, says that Barnett’s sales and streams increased immediately. Amazon’s built-in listening audience is significant: One research firm put Amazon Music’s subscription numbers at 16 million last October, good enough for third place behind Spotify and Apple, and a source told Variety in March that those numbers are steadily climbing, and increasingly dependent on Echo integration. Page notes that it’s too early to tell if this kind of placement can drive the same sort of visibility or engagement as, say, a spot near the top of a prestige Spotify playlist. But until Apple, Google, and others launch their own specific programming initiatives for voice, Amazon is, in Page’s words, “another, for lack of a better term, box you can check for more visibility.”
For Page, as with other digital strategists at record labels, the smart speaker world has been built independent of their wishes, and it’s up to them to make it work for their roster. This strategy increasingly involves the incorporation of lyrics-as-metadata. During the digital era, song lyrics have re-emerged as a commodity in their own right. The crowdsourced platform Genius has integrated with Spotify and Google Home, while Toronto-based LyricFind, which dubs itself as “the world’s leader in legal lyric solutions,” operates on the backend, licensing lyrics from music publishers to work with Pandora, Deezer, and Microsoft platforms, among others. It seems that lots of smart speaker users are requesting songs via snatches of overheard lyrics, which requires a new level of metadata specificity. Lyric copyrights are typically owned by music publishers, and labels don’t usually make any money from lyric licensing. But the promise of smart speaker integration, LyricFind founder Darryl Ballantyne says, has triggered a shift: Any metadata element that could help their music to rise above the digital din helps. “Even though the labels aren’t getting paid by us, having the lyrics available gets them paid more from other people,” Ballantyne says. Page agrees: “The ability to find lyrics and match them with a song will lead to more streams and more sales.”
While labels are playing catch-up with lyrics-as-streaming-metadata, the technology companies are pitching their products toward the type of music fan who would ask for “the hipster song with the whistling.” Amazon, Apple, and Google aren’t going to sell millions of smart speakers by aiming their products toward music obsessives, especially when casual fans are much more amenable to algorithmic programming. This raises old issues for smaller players in the music industry, though. “For an indie label, the question always is: How are you going to convert listeners who are just going to say ‘play some music’ and get them to listen to music they haven’t heard before?” says Ninja Tune’s Slattery. The answer he ventures—even more algorithmic “discovery” engineered by the platforms themselves—isn’t the most anxiety-soothing. “As a label, you’re at the mercy of infrastructure created by tech companies,” he confesses.
Indeed, many of the most pressing issues of the streaming music economy—artist compensation, statistical transparency, sexism—remain untouched, if not deepened, by the rise of the smart speaker. Moreover, as Amazon, Apple, and Google continue to carve out their spaces in the voice marketplace, music consumers and musicians alike will continue to fight against the companies’ preferred walled-garden approach to exclusivity. And though there’s no real reason to sympathize with Tidal or Spotify, the idea that the smart speaker industry might become the exclusive province of massive firms with enough capital to experiment (and huge captive audiences to use as guinea pigs) is significant reason for pause, no matter how little one is interested in owning the devices. A world in which three of tech’s “frightful five” become the equivalent of the major labels, with exclusive holdings in hardware and software, and plenty of incentive to lock competitors’ products and content out of their systems, is a chilling idea, and not as far-fetched as it might seem.
Most music fans don’t automatically want to use a smart speaker to listen to music. They have to be trained to interact with virtual assistants, in the same way that they had to learn to swipe instead of type. The list of sample questions and commands that comes with every smart speaker does not simply tell you how to use the thing, but how to interact with it, employing something close to natural language. For years, speech recognition researchers have understood that talking to voice interface requires the same psychological and social resources as other forms of speech. They’re also uniformly female voices, which activates the trope of subservient women that far predates recorded music.
Beyond the clunky voice interfaces that could only understand robotic utterances a few years ago, smart speakers facilitate a more informal, even humorous level of human-computer interaction. Like the “What’s happening?” prompt on Twitter, or Facebook’s “What’s on your mind?”, this kind of performed intimacy smooths over the bigger project of continuous, ambient data collection. Smart speakers may or may not revolutionize the recorded music industry, but these personable gadgets seems designed for a much bigger project, for which music might merely provide an enjoyable entry point: generating goodwill not toward faceless corporations, but to the dulcet voice in the living room promising a world of constant, friction-less, surveilled consumption.