There was a moment, somewhere around 2019, when the last A&R rep who truly believed in the old way retired or got laid off or just stopped returning calls. The major labels didn't replace them, not really. They replaced them with data analysts and playlist pitching teams and a quiet, desperate faith that the algorithm would find the next star so they wouldn't have to.
They were right. It did. The problem is the algorithm isn't finding what they expected.
Elhé is a French artist who spent years as a professional dancer before she ever released a song. When Spotify placed her in the RADAR program ... which is their artist development initiative for emerging acts, complete with editorial support, marketing budget, and a dedicated playlist with tens of millions of followers ... people kept pointing to her sound as the reason. Atmospheric, minimal, something between bedroom pop and contemporary R&B with a movement quality to it that's hard to name. But the truth is, the sound was only part of it. What RADAR responded to was her strangeness. The fact that she moved between disciplines. That her presence in the space was not reducible to a genre tag or a comp set. She was not easily filed.
The algorithm, it turns out, has developed a taste for exactly that.
What RADAR Is Actually Selecting For
Nobody tells you this but RADAR is not a playlist. It's a bet. Spotify is not just surfacing artists it thinks listeners will enjoy ... it's actively developing artists it thinks it can use to build new listener pockets. The distinction matters enormously if you're an artist trying to get there.
When you pitch editorial or think about algorithmic discovery, most artist development advice still sounds like it's from 2015. Pick a lane. Find your niche. Build a core fanbase in one genre. That advice made sense when the goal was to get a rep at Atlantic or Def Jam to see you perform at the Double Door on Illinois Street and sign you to a four-album deal that would quietly dissolve after the second one didn't move units. That world is gone. And the replacement is not just a digital version of the same machine.
Spotify's recommendation engine ... Autoplay, Radio, Discover Weekly, Release Radar ... is trained to identify novelty within familiarity. It wants artists whose sonic fingerprint creates a bridge between listener clusters that don't typically overlap. A bedroom pop artist who is also a dancer brings a physical vocabulary to songwriting that listeners can feel even if they can't name it. A producer from the South Side who grew up on house but studied jazz theory at DePaul sounds different than either of those things alone. The algorithm doesn't know why it works. But it knows when streams hold and when they skip, when saves accumulate and when they don't, and it is continuously learning that the artists who travel across listener graphs without losing people ... those are the ones worth surfacing.
Genre purity, in this environment, is actually a liability. Not always. Not permanently. But more than people admit.
The Invisible Tax on Specialization
Here is what I keep telling artists I know, the ones grinding through winter at Metro aftershows and building their catalog in a spare bedroom in Logan Square: the skills you think are off-topic are probably your actual differentiator.
If you spent three years doing photography before you committed to music full-time, that is not wasted time. That's a compositional sensibility living in your arrangements. If you danced before you sang, your phrasing is going to land in the body differently than someone who learned music theory first. If you wrote poetry before you wrote lyrics, your line breaks carry weight in a way that pure songwriters often can't fake.
The old industry penalized this. You'd walk into a meeting and someone would say, so what are you, exactly. That question was designed to sort you into a box they already had a marketing strategy for. The answer, if you were honest, was often that you were something between three boxes, and that made them uncomfortable, and so they passed.
The algorithm doesn't ask that question. It just watches what happens when it plays your song after someone who usually listens to something completely different. And if they stay ... if they save it ... if they follow you ... that's signal. That's the thing RADAR is looking for. The artist who opens a door to a listener who didn't know they wanted to walk through it.
Elhé's RADAR placement is not a fluke. It's a data point in a pattern that keeps repeating. Arlo Parks, who came up through spoken word and literature before music, landed editorial support at a scale that pure pop songwriters with slicker productions couldn't match. Remi Wolf, whose whole aesthetic is organized chaos across visual art, fashion, and production, is less an artist than a scene unto herself. The algorithm rewards artists who carry more than one world in their work because those artists are genuinely harder to replace and genuinely better at creating new listener pathways.
The truth is, platform curation has become the new A&R in the most literal sense. Spotify editorial teams now function the way artist development departments used to. They scout. They invest. They shape careers. The difference is they are not looking for artists who fit a format. They are looking for artists who create one.
And that means the game has changed in a direction that is actually more friendly to the weird, the hybrid, the person who can't stop doing three things at once. It rewards the artist who has been stubbornly, expensively themselves despite every piece of advice to simplify and specialize.
Nobody told you that because the people giving advice learned the game before the game changed.
The dancer who becomes a musician. The poet who produces. The photographer who sings. These are not hobbyists dabbling. They are the artists the algorithm is currently writing checks for, in streams, in saves, in editorial support, in the quiet arithmetic of recommendation systems that have learned something the old industry never could quite accept.
Strange is valuable now. Strange is discoverable. Strange is, if you build it right and let it breathe, the most powerful thing you can be.
So if someone asks you what you are exactly ... you don't have to answer them.