This story is over 5 years old.

Longreads Or Whatever

In the Streaming Era, Should Musicians Be Thinking Like Startups?

In the war to win streaming, services are highlighting their analytics features to appeal to artists. Is it time for those artists to get on board and disrupt themselves with data?

Images by Paul Raffaele

Streaming is the future of music. It’s a controversial subject, but it’s a foregone conclusion. Many people, particularly artists, say that digital streaming promotes a view of music as a cheap commodity, while its advocates see it as a rebirth for an industry that’s often slow to react to change. Still, there’s no doubt that it’s happening. Apple Music’s June arrival was just the latest nail in the coffin for music sales. Spotify is now valued at $8.5 billion. Yet even as streaming becomes the dominant industry model, the streaming services are tangling with questions about its long-term profitability. It’s still not clear if paying subscribers will come on board fast enough or in large enough numbers for the existing services to succeed. And even if the model becomes profitable for the streaming companies, can it be made similarly sustainable for artists (and their record labels)? Arguably, none of the major streaming services have yet found an answer to all these questions.


From a PR perspective, that last question is the one streaming services most need to answer. Artist payment, if not the core financial problem in music streaming’s future, is certainly its core ethical one. Many musicians are angry about the services’ lofty public statements about community-building that often seem to contradict reality. Taylor Swift’s recent appeal to Apple Music to pay musicians for songs its trial users listen to was a positive sign, but the last few years have seen other, less well received, public appeals from artists about paltry payments. Damon Krukowski of Galaxie 500 wrote in 2012 that his band received 21 cents for 7,800 spins of one song on Pandora and would need 312,000 plays to earn the same in royalties as the sale of one LP. Last fall Aloe Blacc revealed that he’d made less than $4,000 on Pandora up to that point for his co-writing credit on Avicii’s summer 2013 mega-jam “Wake Me Up!” Streaming services need to address artists’ perceptions about whose side they’re really on, and they know it.

“We talk a lot about consumers not necessarily understanding the benefits of streaming, but there’s a lot of education to do in the artist community,” said Joe Armenia, the head of artist and label relations at Rdio. “You can’t underestimate the value of getting artists to understand us.”

To win artists over, the competing services have to show that they’re including artists meaningfully in their plans. And in an environment in which most services pay less than one cent per stream that likely means showing they have value to musicians that’s more than monetary. Unsurprisingly, technology is at the core of these plans: Across the board, the streaming companies have started to talk up their efforts to offer artists access to listener data and tools to analyze it.


Continue below…

“In the artist community, there are already people who get the value of using data,” said Chris Phillips, Pandora’s chief product officer. “There’s another big chunk of the industry that doesn’t get it.” At Pandora, Phillips has helped spearhead the Pandora Artist Music Platform, an artist analytics program that launched last October after years in the making. Musicians can use it to peruse information about their songs and listeners, looking not only at play counts but geographical, social, playlist, and other data, all laid out in reports on a user-friendly interface. They can use this data to plan tours around areas with high concentrations of fans, form setlists based on song popularity, interact with fans over social channels, and more. The major streaming services all have similar platforms either in operation or in the works. The possibilities are definitely enticing, but it’s the rhetoric the companies use that’s more intriguing: They’re pitching data analytics to musicians like an analytics firm might to an unimaginative business. They’re suggesting using analytics to build a strategy in the same way a startup might.

“Ultimately, we would very much like it if the first thing musicians did when they woke up in the morning was to log into Pandora and have a look at AMP,” Pandora co-founder Tim Westergren told VentureBeat when AMP launched last October. Westergren’s vision fulfilled would see bands behaving like analytical firms making data-driven decisions—thinking like startups to “disrupt” the problem of their music’s lack of popularity. Artists often cater to their poppy, populist side, but data gives that a whole new dimension. If you know exactly what an audience in one place, or one demographic, likes best, do you then start calibrating your musical direction in response?


Currently, streaming services report that they pay 50 to 70 percent of their income for song rights, either to labels or to royalty distributors like SoundExchange. Only a fraction of those payments trickle down to artists. So, while artists are right that they’re not seeing fair compensation for their music, it’s also true that the business model hasn’t been raking it in for the streaming services. The biggest problem—the lack of money the services make—has no simple solution. The music industry on a whole has seen profits dwindle annually for the past 15 years, and it’s possible much of that won’t be reclaimed.

“The economics are different,” Chris Becherer, head of product for Rdio, acknowledged. “We’re still getting a lot of people into the model. It’s going to take a while for the royalty checks to get back to the point that people were used to from the old days.” The streaming model itself is partially responsible for that. Illegal (free) music downloading, which flourished in the pre-streaming era, is in the DNA of these companies. It’s evident both in the Napsterish on-demand model and in the prices, kept low by the reality that music’s still free if you look hard enough. Further cutting into profits are “freemium” ad-supported tiers that Spotify and others offer. They’ve been roundly criticized by subscription-only services, but they reason that since listeners are used to low or nonexistent prices, it’s hard to raise them without risking defections. However, if the model keeps allowing consumers to pay little to nothing for access, there’s just not enough money to go around.


This creates a problem—or, in rosy industry parlance, an “opportunity”—for streaming services, where they want to deliver more value to artists, only it can’t actually be in the form of straight-up royalty payments. This is where that rhetoric that treats artists as business users of streaming services’ data starts to make sense. In theory, artists supplement royalty payments with less tangible value, sort of like how an Uber driver supplements her fare payments with the value of having the flexibility to choose her own hours and fares. Instead of just making music and getting paid, artists become active participants in using the service to create value for themselves.

“I'd love for the data from streaming services to be provided to the artists themselves,” Simon White, who manages Phoenix and Hudson Mohawke, told Noisey. “I think data mining is going to become a key part of the business for us going forward. Having the ability to be able to pinpoint where and who is listening to your music, along with being able to speak directly to them is an incredible tool for us—particularly if you work with the kind of music which doesn't necessarily penetrate mass media. I'd welcome all of the streaming services opening their back doors and enabling the artists to have access.”

Music and tech have become increasingly intertwined, especially in the rise of music-focused data analytics startups. Companies like Semetric (acquired by Apple in January) and Next Big Sound (bought by Pandora in May after previously working with Spotify) offer access to public social interactions, record sales, concert data, and streaming numbers that can provide unprecedented information about song performance and listener bases. This data is then presented to artists on proprietary software platforms like Pandora AMP.


“If you know when, where, and in what context fans are listening to your music then you can give people an amazing experience at the right time of day in the right place,” said Spotify spokesperson Graham James.

Spotify / Screenshot

Many artists have gotten on board: Pandora reports that thousands of acts, from chart-toppers to indie bands, have claimed their band’s identity on AMP; Rdio offers similar numbers for its artist account offering; SoundCloud states that it has over 20,100 partners on its partner program, On SoundCloud. And it’s not just the streaming companies acquainting bands with data and using it to build more targeted followings—on the record label side, there’s 300 Entertainment, which uses access to all of Twitter’s music data to develop new artists, and is credited for helping popularize acts like Migos using innovative social media tactics.

While the industry’s data push is undeniable, its value for artists remains hard to quantify or predict. For artists new to merging tech and music, it’s likely hard to grasp if and how data can help you, now or in the future. And how much do you let a company’s analytics program, which it uses in hopes of raising its own profits, influence how you operate?

“That sounds like death,” Thomas Arsenault, who makes music under the name Mas Ysa, told Noisey when asked about leveraging data to improve his career. “I'm probably wrong-headed in thinking this. But I don't read comments or look at how many likes or plays I get on anything. I don't read interviews. I don't want to know where what is being taken by who. Again I'm not very successful, so maybe I should. I think it might be a part of being a professional in 2015 to do that kind of thing. I have no idea.”


In this model, data is king and makes better decisions. Ideally for the streaming services, artists who have been educated on their platform leverage user information to “marry art and science,” as Pandora’s Phillips suggested, drawing long-term value from the available technology and ultimately making more money off of their music. It also gets artists active on the service, in turn helping to strengthen the service’s cultural cachet and user base, which increases revenue.

As artists become streaming service users, they’re more directly responsible for how much value they get from them. That works both ways: A band that’s good at using data to make decisions could find innovative shortcuts to more popularity and money, while a band that doesn’t want to use analytics, or uses them poorly, finds itself out of luck and with less money than it would’ve had under fairer royalty payments. Of course, if it seems likely that fairer royalty payments may never come, adopting a streaming service’s platform starts seeming like the better option.

But there’s something else going on here. Educating artists about the potential value of analytics isn’t just about encouraging them to try something new. It’s also to get artists to accept the benefits of data and analytics as self-evident, to accept the reality that this is how things are now. It brings artists closer to the streaming services, making artists as dependent on the services as the services are on them. This co-dependence seems to be what the services are after—and it could be where the streaming market is lost or won.


Consider that each service’s data is tied to its user base and mode of operation; Spotify’s analytics aren’t Rhapsody’s analytics, which aren’t Pandora’s analytics. Should bands tailor their songs to the service? A recent essay by musician and professor Mike Errico suggests that musician-disruptors would be wise to write songs that barely cross Spotify’s 30-second threshold for payment. Spotify, he reasons, is effectively “incentivizing a change in song form,” so musicians would be shortsighted not to exploit this “loophole” by writing lots of 30-second songs and reaping the extra per-play dollars.a

Music technology has long played a role in dictating song length, going back to the limitations of early recording technology: 78s, and later, 45s, could only hold three to five minutes of music. If a band didn’t have a 45, it wouldn’t get radio play, giving musicians a firm cutoff for the length of a single. That’s a big reason so many hit songs are less than four minutes long, something that’s continued long after technology evolved. Just as artists have factored technological limits into songwriting in the past, why not continue writing to the medium by putting out songs that meet the bare minimum for payment? If new technology (despite its claims of liberation) continues to endorse limits on how songs are played and paid for, why not figure out how to use it to your advantage?

Would Spotify be down? We found out what happens in one instance of a band trying to put its own spin on Spotify’s payment plan when Michigan outfit Vulfpeck uploaded a completely silent album, “Sleepify,” to the service and asked fans to stream it while they slept. The band made $20,000 before Spotify caught on and banned the album; they said that if Spotify ever gave them the money, they planned to use it to tour. This is a gray area, since Vulfpeck seems to have technically followed Spotify’s rules; on the other hand, Spotify gets to decide what its rules are. As a precedent, though, this is a streaming service attempting to control the market and dictate its terms, while sending a message that if bands don’t accept them, their opportunities for exposure could evaporate.

Apple Music / Photo courtesy of Apple

No less a fierce supporter of artists’ rights than David Byrne recently declared that musicians are entrepreneurs. He said that cooperation and transparency are the keys to a better system and noted that people from all corners of the music business are working with these principles in mind. And to some extent, it may be working: A recent New York Times Magazine article argued that, based on income and employment statistics, musicians are actually doing better post-Napster, largely due to increased live revenue (something that streaming data could further improve). But there are pitfalls if bands are expected to behave like optimization-focused small businesses. It doesn’t change the fact that they’ll still depend on profit-seeking companies, which have little reason to help artists more than necessary, to be the ones to drive change. It also upholds the hierarchy where a few powerful corporate entities can control distribution and consumption, in basically the way that the music industry has always operated. And if, as an artist, you’re able to transcend that hierarchy, you just become Tidal. You stay in the system.

To this end, data can be used as a distraction. It can make opportunities seem more varied and bigger when they’re only slightly less narrow and small than before. Adopting analytics is a choice, but a company compelling artists to adopt analytics as a meaningful source of bottom-line income removes that choice. And analytics won’t fix the parts of the system—the lack of transparency, the entrenched hierarchies, the deep-seated discord—that have been broken for years. They could be part of a new music industry model, one that places less emphasis on artists buying into business practices and more on turning information flows into cash flows. That’s the great challenge facing the streaming companies, and they say they’re up to it. It will be really cool if they are. And who knows? Maybe in 25 years we’ll be talking about the moment this generation’s Dylan went data-centric.

The long-held ideal is that there should be a separation between the pursuit of art and the pursuit of commerce. But artists want to get paid for their art, and the streaming services want artists to adopt commercial practices. Negotiating such seemingly incongruous objectives will depend on a continued rise in subscribers, meaningful compromise between the many players, and being realistic about what data does. And that’s if things go as planned; listeners could always decide they only want to hear human music from now on and fuck up the whole thing for everyone.

Devin Schiff is a writer living in Chicago. Follow him on Twitter.