Is Algorithmic Discovery Dead?
Or did it never really exist in the first place? I look at what's next for Spotify, and whether it's claims about discovery ring true.
You know the term enshittification right? It was coined a few years ago by a tech journalist and basically means how something (a product, a service) gets shitter over time. It’s an accusation that’s been levelled at Spotify lately - in some ways unfairly, but in others I find, it’s fully deserved.
The enshittification of Google has become clear - search results now showing AI generated answers at the top - that are often incorrect or badly cited - and then there’s ads. Whereas in the past the answer to your search would often be in the top one or two results - something that helped cement Google cement it’s position as the #1 search engine - now you’ve often go to scroll down four or five results to get a reliable answer.
In Spotify’s case they’ve actually added some meaningful updates to the product over time, lyrics are a useful addition, the ability for artists to set an ‘artists pick’ and sell merch likewise, audiobooks a slightly confused addition for a music service and canvas & videos are fairly meaningless, but a pretty gimmick all the same.
My biggest bugbear with Spotify, and something that seems to be more and more prevalent is the radio algorithm, something that seems to be harder and harder to turn off - when I put an album on to listen in full, I want to hear it from Track 1 to the end. That helps me form my critical opinion of it. What I don’t want is something similar to start playing at the end. Sometimes you can tell straight away, but there’s times you can’t. Lately it’s been getting easier to tell as the algorithm seems to be defaulting to well-known songs - billion streaming radio hits for want of a better word. Quite often for me I’m served Wonderwall straight after listening to something 90’s guitar-adjacent, or Close To Me after playing a record from the 80’s. That’s not a bad thing, they’re great songs, but they’re not even the best Oasis or Cure songs. I don’t really need to hear them again, especially if I’m playing something new to me. I like these songs so I rarely skip them, so maybe it’s just the algorithm doing it’s job, but I can’t remember the last time it played me something I wasn’t already familiar with.
For so long Spotify sold itself (to artists and fans) on discovery. “Get your music on here and millions of listeners will discover your music” was the incentive. If Spotify exists for artists (and they spend a lot of time trying to convince them that it does) then it needs to do better - it needs to actively push users towards new music, and here lies the problem. A lot of people don’t use Spotify for discovery and have no desire to. For many users it’s replaced radio, and they want it to play stuff they like, tracks they know. Too many unfamiliar songs and they’ll give up and move to YouTube or Apple Music, or back to radio.
Spotify needs to work out whether it exists for artists or users (the current answer is neither, it exists for shareholders) and tailor it’s offering to them. If it wants to be the all-things-to-all-users, all-you-can-eat musical buffet then fair enough, but it shouldn’t pretend it’s somewhere artists can build a meaningful audience. If it wants to be a genuine place for musical discovery then it needs to accept it won’t have the ginormous market share it currently enjoys. In all aspects Apple Music does a better job at this in my eyes.
I guess it comes down to most people using Spotify in a way different to how you or I would use it, and people like you or I are a niche subsect of users. There’s also the question of how does Spotify know if I’m a mood for discovery, or a mood to just go back and listen to old favourites. I listen to a huge amount of new, or new-to-me music each week, but I’ve generally found it elsewhere, either via friends or journalists or some other way. I never listen to Release Radar and I think I can count on one hand the number of songs I’ve genuinely discovered via Spotify.
This comes across as a bit of an anti-Spotify missive, and that’s not the case. It’s a decent service, and I’ve been a paying user for years. Income from Spotify has indirectly paid my bills for a long time now, and I’m grateful for the way it’s basically killed off music piracy. If they announce a new feature I try and ensure we adopt it straight away. But I feel it’s at a crossroads where it needs to figure things out quickly. It can’t keep adding users at a massive rate and eventually it’s going to change. I want to make sure those changes are for the better, for artists, labels and listeners alike.
So maybe it’s not doing anything all that wrong at all then? It plays it safe, playing me stuff it knows I’ll like, and stuff I won’t skip. But with all the talk of amazing AI and curation wouldn’t it be nice to be challenged once in a while?
Aslice Musings
I touched on this last week, and I’ve been discussing it with various friends all week. RA did a think piece (of course they did, and then followed it up with a big advertorial for earplugs) which didn’t really touch on the main issues with the service. Shawn Reynaldo put together a much better article in his First Floor newsletter that actually went into the heart of the issue - there’s no appetite from anyone at the top (government, PRO) levels to try and ensure something like this gets made mandatory. I don’t always agree with everything Shawn says but he’s hit the nail on the head with this one. Venues have to pay PRS licence fees, labels have to pay MCPS or face a distribution ban, isn’t it about time DJ’s paid their peers a fair amount too? Most big DJs also produce these days so they’ll probably make back a decent amount compared to what they put in. All it would have taken was one EDC/Tomorrowland headline act to make the service viable, but only a handful of techno DJ’s actually signed up.
In Strict Tempo Book Review
This week I read Street Level Superstar - Will Hodgkinsons account of spending a year with Lawrence. Lawrence is best known for being the driving force behind Felt, but also recorded as Denim, Go-Kart Mozart and currently Mozart Estate. The book was as much a sad read as it was frustrating. Lawrence clearly has his demons (an obsession with pissing, what sounds like undiagnosed OCD) but there’s times you just want to bash his head against a wall and tell him to wake up and stop being so bloody difficult. Success was always just out of reach for Lawrence, but mostly as a result of his own actions. He disbanded Felt (one of the greatest bands of the 80’s) after 10 albums and 10 singles, which in hindsight, if you listen to Denim and the Mozart stuff might not have been the worst decision. He’s obsessed with writing the best pop song of all time, and becoming famous, but he sounds impossible to work with due to the constant self-sabotage. I feel sorry for Lawrence, and recommend the book even if you’re not really familiar with his music. As sad as it is, Lawrence doesn’t come across as a bad guy per se (with a few notable exceptions), and there’s some funny moments, most notably when him and his then partner manage to get a lifetime ban from the branch of HMV I used to work in. I wondered why we never really stocked many Felt records there…
Summertime Smash
That’s it then, the end of summer and into the autumn period leading up to In Strict Tempo’s first birthday. Will there be a party? No, but feel free to send me a cake. Here’s this weeks new music.
Floating Points - Cascade [Ninja Tune]
I haven’t been massively taken with any of the singles so far, so will the album be much better? It’s definitely very 6 Music-friendly, electronica-lite but there’s touches of the old FloPo magic in there. I think I said already but in a more just world he’d be bigger than Fred Again, hopefully this is the album to get him there.
Mika Vainio (ø) - Fermionit [Sahko]
Some of the late Mika Vainio’s last recordings apparently. His death is still felt even seven years later. One of the most important electronic musicians of our times, for sure.
Manuel Gottschung - E2-E4 (Alex Kassian & Mad Professor Remixes) [MG.Art]
About as Balearic as it comes this. Been out for a while now but I kept forgetting to check it out. There’s a Mad Professor remix that does the biz too.
Robin Mackay - By The North Sea [Flatlines]
New spoken word/ambient piece on the Hyperdub sub-label Flatlines. I spent a lot of my childhood summers on the East Coast (Felixstowe specifically) but we ventured further afield to Aldeburgh, Lowestoft and of course Dunwich too, which is the subject of this new work. One for fans of Lovecraft, The Caretaker, Sebald etc
Quiet Husband - Subutex [Drowned by Locals]
A new single from Richie Culver’s noise-techno alias. Fast and ferocious this is proper gear this.
Clark - In Camera [Warp]
Been ages since I thought about Clark, even though a quick scan of his discogs shows he’s been pretty prolific these past few years. This new album is pretty cinematic, not really what I was expecting. My favourite Clark tune is still Growl’s Garden, I think.
Death Is Not The End - Making Records: Home Recordings c. 1890-1920 [Death Is Not The End]
A couple of Bandcamp-only releases now, first up Death Is Not The End’s foray into an archive of Wax Cylinder recordings from the late 1890’s. Not exactly music, but it’s a really interesting listen of a time when home recording was the height of technology.
VA - R.A.R:Deconstructed-Reconstructed [Industrial Coast]
A few months ago Industrial Coast put out a call for people to deconstruct and reconstruct a piece of music from artists associated with the Rock Against Racism movement in the 80’s, in support of a local community centre in their home town of Middlesbrough. Not wanting to shy away from a challenge, I had a crack at breaking down and fucking up an old punk favourite that’s sits on the album alongside a load of better tracks from the likes of Richie Culver and Like Weeds.
That’s all for this week - thanks once again for reading and subscribing!