The article goes into this, but it's more accurate to say that the aggregator subscription business model saved record labels -- paying a flat but recurring fee

The article goes into this, but it's more accurate to say that the aggregator subscription business model saved record labels -- paying a flat but recurring fee to access a diverse (but ever-changing) library sourced from various rightsholders. This is the same model as Netflix, which licensed all sorts of content when studios were glad anyone was offering, but as they realized they can launch their own streaming services, Netflix found it prudent to produce more and more original content to ward against people who might cancel after waves of popular shows are pulled and put elsewhere. There's not much stopping record labels from pulling a Hulu, except for the fact they already did with Vevo, which was recently killed. Vevo fizzled because it was never clear what value it offered, and its content was also available on YouTube by their own making, in a desperate and bizarre truce against the realities of content discovery. It's like newspapers and their eternal frenemies Facebook and Google, but with more money on the line. The value proposition of aggregators depends on the content they aggregate, but also on matching the content to a receptive audience. Labels may be looking at Spotify and wondering if they can't short-circuit them instead, but when content rightsholders wall themselves off from an aggregator, they must find this receptive audience themselves. And that audience may not be willing to pay the same way: the same content is available through other sources, even if it's just YouTube, or CDs, or digital downloads of the handful of albums they'd miss, not to mention the wide selection of other music from other labels they might warm up to instead. Movies and, in particular, episodic shows are less affordable through legal means, so it's far easier to withdraw them from aggregators and make them exclusives. And unlike music, motion picture is typically consumed once or just a few times, while music is consumed over and over again. Conversely, Spotify might court artists to sidestep record labels, but that only helps for new content. Yet there's other destinations now for emerging artists and future fans thereof, and better platforms for discovery. With subscription business models, retention is always key, so enough palatable content has to be present and no competitor must be overwhelmingly more appealing such that subscribers are driven to cancel. And I don't think the Netflix pivot works nearly as well with music: the whole point of Spotify is to never have to worry about the costs and logistics of acquiring -- legally or illegally -- a personal music library ever again, and to remove the friction and mental gymnastics about picking some music you're in the mood for to play now. The reliance on back catalogs, of nostalgia and emotion of repeated consumption (vs. discovering something new) is too great. If labels start to withdraw, that utility will plummet.