Part 1 in this series can be found here.

I’m not a software engineer, but I am a tech nerd. I remember in the 2000s getting really excited about apps using crowdsourced data to surface collective intelligence. The Waze traffic app seemed like pure magic. Foursquare check-ins were on fire. Yelp was becoming a verb. Google and Ask Jeeves were iterating madly on site popularity rankings. Everyone was trying to harness the social collective brain and feed it dopamine.

I recall at some point telling my husband I’d gladly sacrifice some personal data privacy to never be served another ad for a straight romantic comedy. The competition for the Netflix Prize in 2006 sounded like it could actually make my dream come true.

I viewed this era as a fourth phase in the evolution of communications technology, where the “many-to-many” of the internet becomes the “many-to-one” of crowd-sourced intelligence. (Yes, I was a liberal arts major, why do you ask?) Surely, I imagined in my naïveté, these rapidly evolving technologies will lead to a state of perfect market bliss. Buyers and sellers will be quietly and benevolently matched up for blind dates with wildly successful outcomes. We would finally leave behind the brutality of obtuse, interruptive mass media advertising. Pretty please?

<Sigh> and <shrug>. Twenty-plus years later, my Instagram feed is really dumb. LinkedIn’s algorithm just dishes up shameless self-promotion. Foursquare shut down their City Guides. Yelp quit the restaurant business. Google Maps reviews are almost useless (especially here in Germany). And Netflix still pitches me straight romance shows. The only difference is now the algorithm chooses very specific thumbnail images: always men, usually shirtless. That’s no joke. I mean it is a joke, it’s also sad and true.

Is that all there is? I’ll leave it to others to post-mortem why the internet’s full shared-brain potential was never realized, with one caveat: I’m not sure dropping vague, gripey terms like “enshittification” is useful, outside of a collective cathartic wail for the broken promises of Web 2.0 technologies. Which brings me to music streaming.

Streaming is eating itself

If you read the first post in this series, you know that I’m a big fan of visual data. Recently this lovely but tragic graph from grabbed my attention:

It shows that the proportion of streamed music that is less than 18 months old has been steadily declining for the past 7 years. Fewer and fewer people are streaming new releases. Could it be true? I immediately started rifling through my collection of Luminate annual reports to verify it for myself, and I was quickly disappointed to find out that it is.

I don’t have much taste for conspiracy theories. There’s a whole constellation of economic forces at work creating this trend. Ben cites several in his post; many of them live outside the streaming side of the business. Still, it’s clear that the ghost in this machine has a systemic bias, and to me it looks a lot like an ouroboros that’s eating its own tail faster than it can grow a new one.

If you didn’t hear a song released in 2024, why would you start listening to it now? The algorithm isn’t going to suddenly add it onto your personalized playlists in 2026. How does this play out over time? Aside from the draconian cultural implications, it’s obviously an unsustainable business model. Eventually a self-consuming snake just turns into a plump little donut hole and all that we’ll be left with is “Blinding Lights” and “Shape of You” on endless repeat.

Unsurprisingly, I don’t think anyone has told the algorithm to consider the impact of its DJ choices on the macroeconomic outlook of the recorded music industry over a 30-year timeframe. The algorithm’s only job is to keep you listening, whatever it takes, which apparently means delivering more and more music that feels familiar and “safe.” I think I would have a hard time finding a single indie musician who wants that job.

Earlier this year, Secretly Group co-founder Darius Van Arman posted a moving story titled “Song of the Year You Didn’t Hear.” It’s about Lucie Lozinski, an unsigned artist who randomly reached out to him via LinkedIn, and one of her undiscovered songs became his most-played track of 2025. Lozinski’s music (Ski Team) is waayyy better than mine, and she does everything “right” in the streaming-era playbook. As an indie musician, that makes reading about the struggle to find her audience pretty tough.

Van Arman is careful not to turn Lozinski’s story into a morality tale; he just wants to share some great music you probably haven’t heard. But it’s hard not to read it as a case study in how new music discovery in the streaming age is broken.

I say this knowing full well that the music business is—and always has been—awash with songs that I could have loved but will never hear, and talented songwriters who will never quite “make it.” But I feel like the social compact we struck with streaming when it fired all the gatekeepers and “democratized” distribution was that the discovery algorithm would let us quickly sort through the inevitable flood of new bands and more easily find our truest loves. Instead, it feels like what the algorithm mostly spawned was a mini-industry in how to game it.

It’s not you, algorithm—it’s us. (It’s you.)

Believe it or not, some indie musicians don’t dream of growing up to be massive rock stars. In fact, I’d say that for the vast majority of us, “making it” in this industry simply means being able to spend most of our time doing music while still affording our rent and the cover charge at a good club once in a while.

Having easy access to 800 million potential fans on streaming is enough to persuade you that this kind of career is possible. Surely my music will resonate with some people somewhere; it should just be a matter of finding them, right? The algorithm makes that possible somehow, right? Plus, there are a bajillion websites and blog posts with step-by-step instructions on how to “crack” the algorithm, so it must be doable… right?

Good news: it is possible, and Geese showed us how. Apologies for the snark, but here’s my point: the reason that streaming’s social algorithm is so easily getting played by the industry is because the algorithm is all played out. This recent outcry and counter-outcry over a talented band gaming the system and winning wasn’t randomly timed, it’s a sign of the times.

If (a) the streaming revolution cut the music market to half the size it once was, and (b) new music’s slice of that pie is steadily shrinking year after year, and (c) the algorithm’s job performance review is “Needs improvement” but it can’t, then ipso facto: of course new bands and labels are going to take increasingly extreme measures to get heard. Not waving. Drowning.

I see this and other signs as symptoms of a musical zeitgeist that’s ready to move on: a communal realization that relying on Yelp-era Web 2.0 tech to fulfill our emotional musical needs simply isn’t working anymore—if it ever did. Perhaps after years of trying to iron out our differences, we’ve just finally concluded that we need more from this relationship than the algorithm is capable of giving.

Don’t feel bad, algorithm: it’s not you, it’s us. Let’s just be friends, okay?

Stay tuned for the last part of this story, coming soon. Subscribe now to get it sent to your inbox.

Whenever a new form of something begins to emerge, and we don’t know what to call it yet, we often just add “post-” in front of the old thing. Like “post-punk,” or “postmodernism.” It’s a pretty cheap semantic trick, so it can cause confusion. Do you mean the old thing is gone, as in “post-war”? Or is the old thing still there and just less salient, as in “post-industrial”?

So to be clear, when I suggest that the “post-algorithmic” era of music is beginning, I mean the latter. I don’t think streaming or algorithms are going away (even though Jimmy Iovine says Spotify is minutes from death). Rather, when I read about current happenings in the recorded music industry, I see two things:

1. Streaming will be the big gorilla in the room for a long time, but

2. This gorilla’s capacity for big socio-economic disruption is used up

And if true, that raises the question: “Now what?”

The simplicity of that query belies its inordinate weight in the music industry these days. It seems like everyone is trying to find or build an answer, and most of the answers sound similar: Streaming 2.0! Premium tiers! Superfan monetization! This is not your typical “grow the market” kind of energy. It feels like there’s more desperation to it than that, perhaps since none of these proffered solutions has yet to actually materialize. (NFTs, anyone?)

Here’s one reason for the commotion: when Spotify’s new-user growth in developed markets began slowing down a few years ago, it revealed that the streaming revolution left money on the table. A lot of money. Most of us are spending a much smaller portion of our income on listening to music than we did during the physical music era. Do we care less about music today? I seriously doubt it. Can the industry somehow recapture that lost consumer spend? That’s the 21-billion-dollar question.

Before someone says “what about AI,” sorry, but Suno can’t answer this particular question. Generative AI is a cost-reducer on the supply side; it won’t induce more consumer spend on recorded music. In fact, it’ll likely dilute demand, but that’s a whole ’nother story.

We’re gonna party like it’s (about ½ of) 1999

The RIAA used to have this neat-o interactive graph on their website that visualized annual recorded music revenues since 1973, broken out by format. You can still find it on Tableau, though the data only goes to 2024.

I love a graph that paints a story, and this chart tells a familiar one. Just about the time that tape cassettes started to overtake vinyl records in the early ’80s, compact discs arrived and outgrew them both, and the industry expanded like crazy for 10 years. Then Napster and other pirates started picking off CD sales in 2001, but streaming eventually swooped in to “save” the day. (A few days earlier would have been nice, I suppose… and maybe without stealing everyone’s music; but I digress.)

It’s a good story, but the chart isn’t right. Any economist will tell you that when you look at market growth over time, you need to account for inflation, otherwise price increases distort the real trends. So… see that little pair of radio buttons in the top right corner of the chart? It’s an auto-magic inflation adjuster! <click>…

Hmmm. I guess the first half of the story is still roughly the same, but the ending looks quite different. Notice how the streaming lines are now shorter than most of the compact disc lines? In fact, the peak streaming years don’t even beat peak vinyl in 1978. Granted, that was a banner year with both Rumours and Saturday Night Fever, but also: that was 50 years ago.

Another thing that’s clearer in this view is how from 2021-2024 the steep green curve starts to level out. By that time, streaming’s adolescent growth spurt was over.

To bring it more current and global, here’s a graph that compares two peak years of worldwide format-specific revenues, adjusted for inflation: physical music in 1999 and streaming music in 2025.

Now it’s painfully clear: last year’s streaming market was worth barely half of the 1999 physical music market. The streaming revolution left as much as $21 billion per year on the table. Will streaming continue to grow? Certainly. Will it double in size? Hmm <thinking face>. Is anyone going to wait around for that to happen? Not a chance.

This is a story that music business people know, but not many say out loud, probably for fear of angering the streaming gods. Please don’t tease the twenty-billion-pound green gorilla, even if he did turn out to be a little runt compared to his mom.

Streaming 2.0 won’t be streaming

For now, it seems most of the consumer spending “missing” from streaming is still in your pocketbook. It makes sense: for the amount of food you get, an all-you-can-eat buffet is a better deal than ordering dishes from the menu. In consumer decisions, low price usually wins out. Most people will choose the buffet, even if it doesn’t taste quite as good.

But that’s only most people most of the time. Many of us are eager and willing to pay more and/or work harder for a more satisfying music experience. Goldman Sachs thinks we’d pay $4 billion more. The problem is, streaming doesn’t have a menu for that. It’s a low-priced buffet or nothing. All the talk about adding pricier options has so far been just talk.

Streaming rebuilt music distribution the same way Uber rethought taxis, Amazon remade retail, and Netflix reinvented movie rental. The pattern I see in all of these examples is this: once the main thrust of the disruption is fully realized and markets are saturated, further improvements to the core business model are marginal. You may build a very nice penthouse floor, but it’s nothing close to a second building. It seems unlikely that Streaming 2.0, or even 6.0, is going to find all of that $21 billion pot of gold they left behind.

It’s conceivable that the CD market was always an extortion racket, and streaming is simply a twice-as-efficient delivery model for music. I’m sure many of us can remember falling in love with a radio single on first listen and rushing out to buy the CD, only to suffer from deep buyer’s regret after hearing the other 10 tracks. Maybe physical albums were just a price-bundling trick to get us to pay more for music than we really wanted to.

But if streaming is twice as cost-efficient as physical music was, do we truly believe that Spotify and Apple are just generously passing on those massive savings to consumers? Please. Big music is constantly pressuring Spotify and Apple to raise their price, so it’s likely the current cost reflects the maximum value that can be extracted from subscribers under the paid streaming model.

This reinforces the notion that there’s a bunch of unmet demand among music fans out there. The only question is—if not more streaming—what do they want?

Part 2 of this series can be found here.

“We’re beginning to see the outlines emerge of what a post-algorithmic indie music ecosystem could look like.”

A parallel music universe

If you were a young indie head in the 1980s or ’90s, you knew how to get your music fix. Instead of listening to American Top 40 with Casey Kasem, you would tune in to Snap! on KCRW in Santa Monica. Rather than picking out 12 tapes for 1¢ from Columbia House, you’d sign up to have the K Records cassette fanzine snail-mailed to you from Calvin Johnson’s kitchen in Olympia. Or if you were Down Under, you’d subscribe to the Fast Forward “tapezine” from two DJs at Melbourne’s 3RRR radio station.

In London, you could read indie album reviews in GirlFrenzy at an underground comic book store, instead of lusting over the latest boy band in Smash Hits magazine. (Or maybe both, no judgment.) From New Jersey you’d ride the bus to Manhattan to shop at Other Music in the East Village, since you wouldn’t be caught dead at the Sam Goody chain store in the Paramus mall.

In Berlin, you’d see the newest bands at SO36 in Kreuzberg. In Amsterdam you’d go to Paradiso. If you lived in L.A. you’d hit up Spaceland in Silverlake. In New York you’d try to get into CBGB rather than queueing up for Billy Joel at Madison Square Garden. And you wouldn’t do it alone—you’d do all these things together with your friends, each of you spending a sizable chunk of your paycheck from low-wage service jobs to do it.

What “indie” sounds like

Before becoming another node in the Spotify music taxonomy, “indie” was more than a genre tag. It was a decentralized global network of bands, their fans, and not-so-big businesses sharing a collective mission: to discover and support musicians working outside of the major label system (and to look really cool doing it).

Starting in the 1970s with tiny U.K. punk labels, this ecosystem was hand-built by music industry outsiders who craved hearing bands that the charts-driven majors wouldn’t take chances on. Over decades it evolved into a mature industry, eventually becoming the launching pad for whole-genre mainstream breakouts like New York hip-hop in the late ’80s and Seattle grunge in the ’90s.

Significantly, all the elements in the ecosystem—indie labels, indie radio stations, indie publications, indie promoters, indie distributors, indie record stores—operated parallel to, but mostly separate from, the economic engines of the mainstream music business.

Some of the distribution and revenue channels were totally different. College radio stations didn’t run ads, they ran week-long pledge drives to augment their tiny university budgets. Indie album reviews appeared in photocopied zines that would mysteriously show up on a local book shop shelf. Young rappers dropped off stacks of homemade CDs for sale at their nearest indie record store. Sweat equity was the only form of venture capital available.

In those days, “indie” was a vague and confusing label for a music genre, because it didn’t describe a specific sound. It described a genre-spanning cultural industry that was literally independent from Big Music. Indie was a homegrown parallel music universe.

Then came the algorithm.

We were promised jetpacks

If not jetpacks, we were at least promised a delicious, ad-free feast of only songs that we were sure to love. What we got was a giant all-you-can-eat cruise ship buffet filled with stolen mp3s. Of course, the major labels fixed everything by taking Spotify to court—it was only the indie bands and small record labels that couldn’t afford lawyers.

Suddenly beholden to a business model that they didn’t design and never asked for, much of the indie music ecosystem got washed out by streaming. The parts that remained standing were left to cope on their own, their support network now fractured. College radio listenership plunged, along with pledge revenues. Thousands of indie record shops vanished. Respected indie music journalists got displaced by automated discovery playlists. The superstars of the ecosystem, indie musicians, could no longer earn money from their music by selling CDs, while the too-clever “pro-rata” royalties model decided that indie songs were worth about $0.003 per stream.

Some 20+ years later, the streaming tide has begun to ebb, and the edges of the new music landscape that it wrought are now firmly in view. The streaming market’s developed regions are nearly saturated, but streaming still makes way less money than physical music did 25 years ago. Streaming fatigue is real, especially among Gen Z and younger audiences. And it’s become quite clear that, despite vastly expanding the potential listening audience for emerging indie bands, the financial fundamentals of the streaming-buffet business model simply make it incapable of nurturing a truly independent music sector. At least not one that’s as culturally vital as it once was.

But you know what? Fuck that. Nobody puts indie in a corner. Music outsiders are already creating a new independent infrastructure to break out of the boundaries set by the streaming-industrial complex. And this time, we’ve got computers.

Cue the indie ark

This is no ship of nostalgic fools hoping to return to “better days.” Indie never looks back. And to be clear, many of the pillars of the original indie economy weathered the storm and they’re still standing tall, having adapted to the new reality. But if you ask, they’ll tell you: “We need our own thing.”

Fortunately, folks are already stepping up to build it. The new crew is scrappy, determined, and laser-focused on solving the problems of new generations of streaming-native indie fans. It’s early days, but we’re beginning to see the outlines emerge of what a post-algorithmic indie music world could look like.

Community-owned collectives

After streaming washed away the CD market, Bandcamp was manna. It gave indie artists a way to sell music directly to their fans again; i.e. actually earn income. But then the company got sold to a big music licensing firm who immediately laid off half the staff. Bandcamp is still around, for now, but bands who built their livelihood on that platform are justifiably nervous.

So how do you build a future-proof indie digital distribution platform that can’t be sold off to a corporate investor and gutted? If you’re Austin Robey, you invite the entire indie music community to form a collectively owned business cooperative. After years of work to establish a solid legal footing, build an app, and recruit artists, subvert.fm went live in May 2026 with more than 150,000 tracks already in the catalog. Many more bands will join, because it’s free for musicians and labels to become member-owners, which means they also get a vote in how the platform develops. Subvert’s goal for their first year of operations? Just to become the largest artist co-op in world history, with 100,000 members.

There’s still a lot of work ahead to expand Subvert’s features, but it looks like we now have a dependable long-term solution for selling digital downloads, and it’s permanently owned by the community that it serves, not by corporate investor overlords.

Anti-algorithmic music curation

One of the good news/bad news paradoxes of the streaming revolution is that whilst it removed barriers to music distribution for new bands, that also meant flooding the marketplace with new bands. (And that was before Suno. Let’s not talk about Suno.) Taste curators like college radio DJs and indie music journalists were a critical linchpin of the old indie discovery model, and they were among the hardest hit by the streaming takeover.

Trying to fill that role, Cantilever aims to help indie fans sort through the constant cacophony of the new music landscape by giving us a digital version of old-school-style music curation. Similar to the “rotation” concept at indie radio stations, only a limited selection of albums are available on Cantilever, and their tenure is limited to one month. And like the old Pitchfork, they each get a detailed review, thoughtfully written by a serious music journalist. Most importantly, Cantilever is squarely focused on indie music, and only features full albums. They just raised a pre-seed investment round (see global resource pooling below), so let’s see what they do with it.

Street-legal bootleg tapes

It’s not all about software, folks (but mostly it is). UK startup Setmixer invented a machine for making automatic, high-fidelity recordings of concerts straight from a live club’s sound board, so bands could sell their shows to superfans as digital downloads. The great thing about this idea is that it represents a net add to the indie ecosystem. It provides musicians with a brand new income stream from an activity they were doing anyway: playing gigs. Setmixer basically just gave indie musicians a raise.

Founder Pascal de Mul and his team also knew that they needed to deliver recordings that actually sounded good, so they built some very smart software into their hardware to optimize audio from a live sound board and automagically mix it down, master it, and post it online—overnight. As with any live gig, you need a good sound engineer to place your microphones and set the levels properly, but the results are impressive. Go listen to some recordings for free on their streaming page, they sound great.

Promotion without payola

As evidenced by the blowup around Geese hiring a social marketing agency to engineer “organic” virality, the indie community has always had a complicated relationship with promotions and marketing. We’d love to believe that our precious ecosystem is a benevolent meritocracy free from the scourge of manufactured hype. But deep down we know that’s a bit unrealistic.

That being said, Repost Exchange is making a pretty good run at it. Unlike the raft of pay-for-playlisting services that sprang up in the wake of the streaming coup, the currency on Repost Exchange is purely participation. The more you share and comment on other musician’s tracks, the more credits you earn to use amplifying your own songs. No fake TikTok accounts, no bot-stream farms, just musicians sharing music. On the merits of this model—and no doubt the half a million bands already using the service—Soundcloud recently made Repost Exchange an official platform partner, which will allow them to integrate more deeply.

Payout models that don’t suck

Speaking of Soundcloud, they were the first streaming platform to give the finger to the pro-rata payouts model and shift to what they call “fan-powered royalties.” That means if Susie the Soundcloud subscriber only ever listens to your band, you get every dollar of royalties coming out of Susie’s subscription fee. There’s something weirdly logical about that, right Spotify?

That shift rolled out back in 2021, but Soundcloud is still actively evolving their product and business model, in a way that suggests maybe they actually get what indie artists need. Their latest moves include commission-free distribution to other streaming services, buttons for listeners to pay tips to artists, direct-to-fan merch stores, and even on-demand vinyl pressing. (Bandcamp tried selling vinyl too, and gave up, so we’ll see.)

Global resource pooling

The Organization for Recorded Culture and Arts (ORCA) isn’t a startup, but they recently helped fund one. ORCA is an international think tank for indie labels like Sub Pop, Beggars, Domino, Secretly, and more labels from Paris, Madrid, Nashville, L.A., Berlin... well, you get the idea.

In addition to collectively sponsoring detailed research on how to build a stronger indie music ecosystem, several of ORCA’s member companies recently pooled a quarter-million bucks to lead a pre-seed funding round for Cantilever. So I guess indie labels are angel investors now? Oh wait... they always were.

Are we there yet?

There’s no single unicorn startup or magic wand to build a new indie infrastructure overnight. As before, it’ll take a wide range of products and industries, some new and some renewed, to create a different kind of indie music ecosystem that will thrive in the context of major-label streaming dominance.

Organizations like ORCA, subvert.fm, and Repost Exchange signal a new era of global collaboration in the indie music sphere that wasn’t possible in the old scheme. It used to be that ideas like cassette fanzines could only spread slowly and organically from city to city. Now we have fancy electronic mail and video telephones. (Still working on those jetpacks.)

Nonetheless, as before, a new indie system with a shot at real longevity will have to be architected parallel to—not inside—the mainstream music matrix, in order to realize cultural and economic sustainability. That will take some time.

Skipless is developing a new product for indie music that will be announced soon. If you sign up for our mailing list, you’ll be the first to hear about it.