A Newsletter about Big Ideas in Music and Technology, by Cherie Hu
This is issue XXX, published on XXXX, 2019
Happy Thursday everyone! :)

I'm writing this from London, where I just moderated a conversation with Troy Carter on music and tech at Founders Forum. Yes, if you've been following my travels, you may have noticed that I also moderated Troy's keynote at MIDEM in Cannes last week.

Both conversations touched upon the future of music distribution and how technology can better serve entrepreneurial artists in their careers. You can watch the MIDEM conversation on YouTube and/or read a brief recap on Music Ally; I'll also be publishing a more personal reflection on both experiences for $15+/month supporters on my Patreon page in the coming days.

Now, onto today's topic...
What Stem’s upheaval reveals about music distribution’s new “middle tier”

Stem is growing up, and not everyone is happy.

On May 31, 2019, the L.A.-based music distribution platform — known for working with the likes of Frank Ocean, Anna Wise and Childish Gambino — announced without forewarning that it would be dropping the majority of the tens of thousands of artists distributed through its system by the end of July. The move comes as Stem transitions towards a new model, titled Stem Direct, in which the company will instead serve only a select number of artists with already-established teams and a demonstrated level of streaming momentum.

Each qualifying artist will get a dedicated account manager at the company who can assist with playlist pitching, marketing strategy and general administrative tasks. In exchange for these more hands-on, bespoke services, commission charges for all future clients will increase twofold, from Stem’s original rate of 5% to the new rate of 10%. (Existing Stem users get a discounted 8% rate on future catalog, with the 5% rate remaining for all previous catalog.)

For the artists who don’t qualify for Direct, Stem is no longer accepting any new uploads (“as of May 31,” which they announced... on May 31) — and is asking them either to opt in to migrate their catalog to their official partner TuneCore, or to download their data and handle migration to another distributor themselves.

A rep for Stem tells me that only around 5% of the company’s total streaming volume, and less than a quarter of total paying accounts, will be affected by the change. Note, however, that the number of paying accounts is not the same as the total number of artists uploading content to Stem; for instance, Red Light Management has one paying account on Stem, on behalf of over 300 uploading clients. The Stem rep confirms that the majority of the company’s uploading artist base is included in the 5% of streaming volume that will be forced to migrate.

Several artists expressed shock, fear, frustration and/or disappointment in the wake of Stem’s changes, because many of them had entrusted their catalog to what they saw as one of the friendliest distribution platforms for the music industry. From one of the lowest commission fees in the market (5%), to backend royalty-split capabilities (which most distributors still don’t have, for complicated legal reasons), to monthly payments (most distributors still pay out quarterly) and a sleek visual brand, Stem seemed to offer a product built for the modern musician with modern needs.

“Stem was the best option for labels releasing original music that weren’t making much money,” says Ben Briggs, an independent artist and label owner known for his remixes of video-game music, who did not qualify for Stem Direct. While Briggs’ covers and remixes have garnered millions of streams combined on Spotify, they are all distributed through Soundrop, and most of his own original catalog lives on DistroKid; he had used Stem only for releasing artists on his label Tiny Waves.

“We aren’t getting huge returns on investment yet for our original music, and wanted to maximize the amount of money going back to the artists,” Briggs tells me, when asked about why he had previously turned to the service. Hence Stem’s 5% revenue-share was ideal for Tiny Waves, whereas TuneCore’s annual fee of $9.99 per single and $29.99 per album (rises to $49.99 per album after the first year) would mean a loss for many tracks on the label’s roster.

Why did Stem decide to let go of the majority of its customers?

Three main reasons come to mind: finances, quality control and market opportunity.

As I’ve written before, the business of music distribution is both competitive and low-margin, and many distributors have shut down or consolidated over the past five years due to unforeseen backend and customer-service costs. Moreover, companies like Stem that work on a percentage revenue-share rather than flat-fee model arguably need high-performing hits more than they need a high number of users in order to succeed.

Current Stem employees tell me that the company was never meant to be a “long-tail” music distributor, anyway. While they were catering to independent artists more broadly, the only ways you could access Stem were by applying directly through the site or by receiving an invite code from an existing client. One employee, speaking on the condition of anonymity, tells me that the Stem “gained a lot of users who came in through unofficial ways” as a result of these invite codes, “which meant that the platform ended up being used in ways we didn’t intend for it to be.”

That said, Stem’s ambitions in its early days were to be more or less open and neutral, rather than closed and selective — at least according to the company’s co-founder and former VP of Product Development Jovin Cronin-Wilesmith, who left Stem in October 2018.

“I apologize to all the Stem artists being removed today,” Cronin-Wilesmith wrote on May 31, in reaction to the company’s announcement. “This was never the original intention of the mission.” Over the phone, Cronin-Wilesmith reaffirms to me that Stem’s latest changes were “not something in the works when [co-founder] Tim [Luckow] and I were still there.” Instead, the original product roadmap at the company’s outset consisted of an unbiased, self-serve suite of tools for efficient payments and data-driven insights across multiple revenue streams for artists — using distribution of recordings as a starting point to expand to other sources including publishing, live and merch.

“I can count at least 100 conversations early on where we said that we needed to ‘be Switzerland’ and be as unbiased as possible, to ‘be the ocean and not the island,’” says Cronin-Wilesmith. “I don’t think we wanted to get into the game of being tastemakers.”

In contrast, there was arguably a highly subjective, tastemaking aspect to Stem’s decisions about which artists to keep for its Direct program. Absolute streaming numbers mattered only up to a certain point; what was arguably a more important dealbreaker was a qualitative evaluation of an artist’s sound.

For instance, a source close to the situation tells me that Stem is booting off some accounts “with large amounts of content that did generate revenue, but that Stem didn’t think they could promote and pitch effectively” — including electronic instrumental tracks. This explains why some electronic artists with Stem-distributed tracks that surpassed one million streams within just a few months, such xJK’s “Fuse” (ft. Zane Schaffer), weren’t ultimately selected for Direct. In contrast, artists like Bravo — who has just one Stem-distributed track that gained around 100,000 streams in a few months, sources say — were selected, likely due to a combination of their sound and an evaluation of their surrounding team.

Looking at Stem’s future roadmap, it’s become increasingly clear that they want to build tools for teams, rather than merely for individuals. And the company isn’t afraid to admit that this part of their business is motivated explicitly by money; their website claims that only “labels currently earning over $1M per year or [that] have direct deals with platforms” can qualify for their Teams product. In this vein, the success of Teams would position Stem as a competitor to larger, major label-owned “independent” distributors such as Caroline, The Orchard and Warner’s ADA, in addition to smaller competitors like AWAL and EMPIRE.

[Screenshot from Stem’s website advertising their Teams product, and being explicit about money.]

Kara Nortman, Partner at Upfront Ventures and one of the leading investors in Stem, tells me that Stem’s long-term roadmap has the potential to reach even beyond music distribution, perhaps into becoming an all-purpose Intuit-like service for independent creators across media and entertainment. “Right now, Stem is focused on building a platform for tracking finances and helping artists recoup expenses, but over time you could plug in anything you wanted as a company, from taxes to insurance,” says Nortman. This could be a game-changer because it’s still surprisingly hard for artists to get a clear, one-stop-shop picture of their finances — especially because a large number of them are dividing their catalog among several distribution platforms that each pay at different frequencies and have different reporting standards.

Sources say that Stem is also now providing digital distribution, payment and accounting services for video-first media companies like BroadbandTV and Worldstarhiphop, with more diversified clients in the pipeline. This brings Stem even closer to competing directly with The Orchard, which has its own in-house film/TV division, and is a generally shrewd move as the boundaries between music and other creative industries continue to blur, both culturally and commercially.

Visualizing distribution's new "middle tier" — and where Stem fits in

Regardless of what you think about Stem, arguably one of the most important takeaways from the change in their business model is the danger of treating “artists” as one homogenous, unified group.

Stem’s official blog post announcing the changes espouses its commitment to an “Artist-First, Team-Oriented Future.” Yet as I’ve argued in the past, the phrase “artist-first” carries a different definition not just for every company, but also for every artist. That may have been a useful catch-all term in the early stages of digital distribution, when independent entities had fewer options in the market, but now it’s a rather vague and unhelpful placeholder with respect to informing actual product development for a specific consumer segment. As Nortman tells me: “You can’t focus on music hobbyists, bigger labels and the growing middle class of artists at the same time. If you try to support too many segments and push out different features for all of them, you end up helping no one.”

All music distributors, including but not limited to Stem, are learning this the hard way. A few years ago, it felt to me like all distributors were offering more or less the same business model and product — minus a few dollars or percentage points — with the blanket goal of “empowering artists to be independent.” But nowadays, and particularly with Stem’s transition, the segmentation of distribution is becoming much more explicit and intentional.

In particular, my sense is that Stem 2.0 marks an expansion of the "middle-tier" distribution market — i.e. companies that serve neither completely new/DIY artists nor major-label/mainstream artists, but rather artists in the middle with an already-established income and distinct sets of goals and needs that current technology is under-serving.

Calculating the exact market size of this “middle tier” remains an ongoing challenge, of course. We do have a ballpark idea of the DIY sector’s growth with respect to revenue: According to MIDiA Research, annual revenue from unsigned acts reached $643 million in 2018, growing three times faster year-over-year than both independent and major labels.

Pinpointing the number of artists in the “middle tier” is a whole other inexact science altogether. In January 2018, Kobalt CEO Willard Ahdritz claimed that his company could play a leading role in growing the number of artists in the “middle class” — admittedly a vague category with no clear bounds from Ahdritz himself — from 5,000 to 100,000 within five years.

Stem has even more ambitious projections on the organizational level. One employee tells me that there are an estimated 10,000 to 30,000 companies globally that, like mtheory and 88rising, “look like labels but are a hybrid of a label, management company, media company and sync agency in terms of the services they provide, and are acting independently as opposed to doing an imprint deal with a major label.” Stem is betting that this market will grow to 100,000 label-services entities within the next five years, and that Stem itself can then become “an out-of-box backend solution for distribution and payments for these companies, so they can keep focusing on strategy and creative,” says the employee.

For this newsletter, I tasked myself with visualizing this burgeoning middle tier, in relation to traditional major-label versus long-tail distribution deals. This chart is not at all exhaustive, but rather aims to give a sense of structure to a landscape that historically had none.


Some points to highlight:

  • In general, “closed” means available only to qualifying, hand-curated clients, while “open” means available to anyone who signs up and/or pays for an account. For instance, in-depth analytics are available for all paying SoundCloud Pro artists as “open software,” while the hands-on label services for AWAL, Stem, EMPIRE and Ditto are all invite-only, i.e. “closed services.”
  • Distributors that emphasize closed services (i.e. look more like labels) tend to rely on a percent-revenue-share business model. In contrast, those that emphasize open software and services tend to rely on a flat-fee model (with the exception of companies like UnitedMasters, which is perhaps why they’re still struggling to gain traction after so many months of operation).
  • The leftmost segment refers to major labels. From top to bottom, the pictured logos are for Universal Music, Sony Music and Warner Music. In these deals, the emphasis tends to be much more on services than on technology, although recent news like Sony Music’s real-time payment capability and Warner Music’s acquisition of Sodatone may be trying to change this.
  • The rightmost segment, consisting of distributors that rely on maximizing their market share of the long tail to sustain their businesses, is the most crowded and most competitive segment of all. Differentiation in this segment will become more and more difficult as the underlying mechanics of distribution overall become more commoditized.
  • The second column from the right features companies with a mix of open software and closed services. For instance, anyone can sign up for a Ditto account, but you need to apply for access to their bespoke promotional services. Anyone can upload music to Amuse, but their in-house record label is much more selective.
  • People thought Stem wanted to move towards the right of the curve as it grew, when in reality it’s now actually traveling further toward the left — competing with major label-owned independent distributors in how they cater to establish teams.
What I hope this chart drives home is that it’s no longer appropriate to talk about the music industry simply as a binary between the “mainstream” and the “long tail,” whether in the context of business models or product development. Artists in the “middle tier” are de facto small businesses with teams of employees, and need tailored small-business management tools just like in any other industry, rather than being bucketed into the same category as hobbyists using long-tail-oriented services.

One Stem executive, speaking on condition of anonymity, hopes the distribution platform will have a similar impact on artists and music companies as Shopify had on e-commerce. “Before Shopify, it was really expensive for any brand or retailer to start an e-commerce store,” she tells me. “They had to use more labor-intensive tools like Magento and build out an entire in-house engineering and product team to help. Then Shopify came along, and essentially enabled any brand to start their own store overnight. What we believe, and what we’ve seen a little bit already, is that if we can alleviate the burden and expense of handling distribution, reporting and payouts for music, we can enable more businesses in the market to grow on their own and become labels overnight.

Hopefully the wider change signaled by Stem’s decision is net-positive, not only in how it will give more people in the world the foundation to start a record label at low cost, but also in how it will encourage clearer, two-way communication between artists and services companies about whether a given product offering is really right for both sides. After all, as an artist, manager or record label, you can’t run your business without the appropriate set of tools — and as a distributor, you can’t run a good business if you don’t actually know your own customers. 🌊
✨ If you’d like to support even more thoughts and conversations on music and tech, I encourage you to become a paying member of the Water & Music ecosystem on Patreon. For as little as $1/month or as much as $40+/month, you can access a wide range of perks including:
  • A closed, members-only Discord server, consisting of updates and analysis on the most important music and tech news every weekday
  • Voting rights on future newsletter issues
  • Exclusive longform essays
  • Previews and bonus material for my freelance articles
  • Monthly video hangouts with me
  • Live meetups and events (forthcoming!!)
…and much more! Thanks so much for reading! ✨
Follow-up from Issue #53: More notes on B2B music streaming

I wanted to dedicate a brief section to following up on my previous newsletter issue, which provided an in-depth look into the B2B/in-store music streaming business and unveiled the new Apple Music for Business product. A handful of readers provided some helpful clarification and context after the fact, which I thought I would share here.

In the original sendout, I had written that “it’s unclear whether PlayNetwork has the proper mechanical licenses to legally white-label its services to Apple Music for Business, which could explain why they’re so quiet.” One reader involved with the project let me know that PlayNetwork is indeed securing both performance and mechanical licenses to operate not just the Apple service, but also their own core business at large. Moreover, PlayNetwork formerly relied on Music Reports (MRI) for licensing and royalty processing, but are now building technology to do the job all in-house, which allegedly has led to improvements in matching metadata.

While Soundtrack Your Brand now has direct deals with labels and publishers (and is working with MRI), PlayNetwork actually handled all of the company’s content acquisition and licensing until early 2019. In other words, one of SYB’s biggest competitors was formerly a partner.

Finally, one important point about Sonos that I didn’t mention in the original issue is that Soundtrack Your Brand is not the speaker company’s first B2B music partner. Sonos has had integrations with SiriusXM Music for Business, Rockbot, Mood Media and Custom Channels for several years, making SYB a rather late mover.

My writing elsewhere

I wrote my debut feature for Resident Advisor about artists who are using machine learning and speech synthesis to push the boundaries of music production by creating artificial clones of themselves. Such projects include Holly Herndon and Mat Dryhurst's DIY voice model SPAWN, Reeps One's collaboration with Dadabots and Bell Labs and Actress' formulation of his AI "sprite" Young Paint. This is one of the first pieces I’ve written in a while that focuses more on creative processes than on business models, which felt refreshing!

On my Patreon page, I wrote an exclusive essay about the rise of music podcast spam and its potential impact on Spotify's content ecosystem. It elaborates on my unexpected discovery of a fake Cardi B podcast in Spotify's recommendations, which I first exposed on Twitter.

Be on the lookout: For Merlin's upcoming annual report, I wrote a case study on Secretly Distribution that dives into a wide variety of topics, including but not limited to how Secretly is differentiating itself in a crowded distribution landscape and how distributors overall can play a more proactive role in paying artists and labels fairly. The report will be launched during the Indie Week conference next week in NYC, and I'll be sure to share it in my next newsletter. :)

What I’m listening to is a collection of 25 pieces of algorithmically- and infinitely-generated ambient music developed by Alex Bainter, and has become an unexpectedly wonderful tool to help me focus during long spouts of writing. My favorite piece is “aisatsana (generative remix),” a piano instrumental that tries to extend Aphex Twin’s song of the same name ad infinitum, building on the same melodic and rhythmic patterns. The result feels unusually humanistic and organic considering its hard-coded roots. (Bainter wrote a great blog post in November that dives into more of the engineering behind the scenes!)

Abstract Orchestra, a self-described “hip-hop big band” from the U.K., has released two impeccable tribute albums to Madvillain, the renowned collaborative project of MF Doom and Madlib. I’ve been listening through Madvillain Vol. 2 nonstop; the band’s interpretations of “Meat Grinder” and “Figaro” are particularly powerful, while retaining all the musical nuances of the original samples. Many of the tracks incorporate improv from the group itself, and almost all of them were recorded in one take with few overdubs.

99% Invisible’s two-part series on “Sound and Health” — which dives into the profound impact of sound on our daily lives in cities and hospitals — has got me wondering why music companies aren’t getting more involved in urban planning issues. Particularly with respect to sound and noise policies, one would think governments could at least include the voices of professionals whose careers revolve literally around making sounds that resonate emotionally with people.


Thanks to the Study Hall listserv for tipping me on to the fact that Spotify is now hiring a News & Documentary Podcast Producer, based in New York. The job description seems to be unusually wide-ranging, baking all aspects of a full-time audio engineering position (“scheduling, recording, assembling, mixing, and publication of episodes, including recording interviews in the field and at events when necessary”) into a role that seems more focused on executive production. My hunch is that many upcoming podcast roles at Spotify will be this all-encompassing and unstructured, given the pseudo-startup nature of a new vertical in an otherwise very corporate environment. Also, this is a friendly reminder that Spotify is not, and maybe never was, a music company.

Copyright © 2019 Water & Music, All rights reserved.

Want to change how you receive these emails?
You can update your preferences or unsubscribe from this list.