Spotify on Podcasts and their New Radio AI DJ
Spotify continue to go for your ears
Spotify’s recent Q4 update has resulted in lots of predictions about the future of podcasting, and Spotify’s position in it, partly after CEO Daniel Ek said “I got carried away”. The usually excellent Semafor ran a piece suggesting Spotify’s podcast bet has gone wrong.
James Cridland articulated a different view in Podnews - “If coming from nowhere and being #1 in most countries in less than four years is “going wrong” [where does that leave everyone else]?
I think it’s fair to say there’s been bumps in their podcasting journey, but it would be very difficult to argue that they haven’t made a massive splash and built a business of significant scale, providing the first real competition to Apple’s dominance of the podcast space.
Its intervention in the market has been down to three things. But before we get to that, it’s important to see where they came from. Before their acquisition spree, they had introduced podcasts into their app and had commissioned some of their own shows and made some talent acquisitions (like The Receipts).
The big change, though, was doing three things. The first was building scale in their owned and rep’d shows. Buying Parcast, Gimlet and The Ringer alongside exclusive deals with Joe Rogan, the Obamas and Meghan and Harry, gave them significant downloads and reach. This, combined with their other content initiatives meant they had the scale to drive their second big decision - growing their advertising operation. This is something they grew even further by acquiring Megaphone - which provides ad services to podcasters.
Having Megaphone alongside their other big acquisition - of podcast publishing platform Anchor - meant the ability to build a platform that could provide a great environment for creators.
The latter’s important - instead of continuing to spend loads of money buying content and content companies, you can instead have a selection of services that means creators come to you, and build on your platform.
It’s not a new play. YouTube’s hosting, monetisation and marketing of creators’ video content has built a huge business for Alphabet - one no one has come close to replicating. There’s definitely value in being the place for podcast creators - but I still think Spotify has quite a way to unite their tools, to lose their obsession with Spotify-specific features (polls etc) and to make the subscription journey as smooth as Apple’s.
In summary they now have scale in downloads, a stronger monetisation platform and (potentially) the right creator tools to attract more people to their platform. They also have the number 1 or number 2 podcast app in most territories.
The staff changes, axing of shows, bad integration work of studio acquisitions is, of course, not ideal - but the end result of all that is they are in a very strong place going forward.
Spotify’s AI DJ
But it’s not just podcasts that Spotify have been busy working on. They’re making (another) attempt at ‘taking on radio’. Radio, especially in the US, generates revenues that dwarf podcasting. Daniel Ek has always been focused on breaking broadcast’s stranglehold. Usually unsuccessfully. Whether its content plays like Music + Talk or devices like the Car Thing, Spotify have a history of delivering clunkers. As I’ve talked about before, this is because big tech fails to understand why radio works (and they never seem to hire any people who know!).
Will their latest attempt beat their last lot of cock-ups?
It is, on the surface, a neat idea. Spotify knows what music you like, so they’ve combined this with a ‘DJ’ - an AI voice that talks to you, provides context for what it’s about to play and is generally jolly (well according to the video above).
Ignoring the ‘powered by AI’ strap-line, which is a little band-wagon jumping, I think they still misunderstand what a DJ does. Listeners aren’t crying out for functional interruptions, especially from someone that isn’t real. The DJs that listeners don’t like are the inauthentic ones or those that ‘talk too much’, listener code for ‘I don’t like what they say’. Having an AI make guesses about what I want to hear now, based on what I previously listened to, is perhaps likely to be less successful than they think. All you have to do is look at people taking the mick out of their Spotify Unwrapped, and the reasons it’s wrong - kids, parties, ex-relationship woes - to show your play-log is not how you define your listening today.
The proof of the pudding is in the eating and we’ll see what it’s like IRL! BUT I bet they still don’t have next song markers and intros coded up, so it’ll end up being presenter link followed by a 00.00 start of the song, killing much of the flow. Now sorting that out with AI would be a step forward.
You can catch-up on last week’s Media Podcast where Ann Charles and Steven D. Wright joined to talk about the BBC Chairman, TV taxes and nepo-babies. Plus I spoke to Great British Bake Off boss and creator of The Piano, he unexpectedly did a mea-culpa over the last series of Bake Off, which ended up making the news with pieces in Deadline, Metro and more. You can listen here.
Had this sent to you? Subscribe and get stuff like it, for free, in your inbox:
"Big Tech Fails to understand why radio works..." Radio works? Just my opinion, but I find it the absolute worst way to listen to music that was ever invented - it just happened to be the only option for decades. The DJs are annoying/boring, the advertisements are incessant and they play the same songs over and over and over again each day. I'd love to hear what exactly "works" for radio that Big Tech doesn't understand.
“big tech fails to understand why radio works (and they never seem to hire any people who know!)”. I’d bet if they did hire “people who know”, those people would be out in 6-12 months because they’d got fed up of being told “but we’re Big Tech, and this is the way WE do it” 🙄