NEW YORK (TheStreet) -- The launch of Beats Music Tuesday was a little buggy, but it's a really good app and that alone spells a change in direction in the streaming music sector. While the subscription service still has some problems to iron out, the bugs obvious in the launch appear to have been fixed.
The sector is new -- the roots of free Internet radio service Pandora Media (P - Get Report) go back only to 2000 and Spotify, a free ad-based service with a subscription upgrade, launched a little over five years ago. As a result, there is still a lot of misunderstanding about how the technology works. Yahoo!'s Alyssa Bereznak, in her otherwise excellent review of some streaming services this week, appears to treat Beat Music's underlying algorithm as an unusual feature. It's really not. At its core, it's basically the same technology that powers Spotify and Rdio. The chief difference as Bereznak points out, for discovery, those services seem to do better when they rely on access to users' other social media platforms to inform selections.
I agree with Bereznak that the algorithm used in Beats Music feels like a much different experience. A lot of that has to do with the interface, but some also has to do with the human faces associated with playlists and song recommendations.First, let's be clear: nobody is spinning tracks for you. Just like Spotify and Rdio, Beats choices are run moment to moment by an algorithm that feeds back to you selections based on your history of choices. This is the underlying feature of all the streaming music services. It is simply impossible to do a robust, personalized streaming music service for tens of millions of customers any other way. What Beats Music is able to do is to take some of the monstrous quality of that human-machine relationship and humanize it, drawing from pools of screened playlists, applying cold user data to similar histories of experienced music screeners, to warm up the result. The algorithm may be the same, but the background of screened stuff and prescribed playlists seems to open the landscape somewhat, making it easier to trust that the next song is going to be satisfying. At base, trust is what its all about. Music by definition is an intimate, shared experience and the job of any music service is to recreate that possibility, human-to-human. Pandora has a similarly human-informed selection process, but its location for that human interaction along the production line is different. Pandora emphasizes human input in the database, making sure that musicians and musicologists are involved in the labeling of songs, each involving literally hundreds of different categories. Each song can be selected via any of more than 400 ways. That is what they poetically call the Music Genome Project, and it is at the base of Pandora's business model. With the MGP, Pandora puts humans at the earliest stage of the production line, loading the hyper-individualized conveyors. It also has another department that trails the user experience with data analysis and uses that to tweak the algorithms. With this combination, Pandora's team tries to identify patterns in your selections to recommend choices you'll like better. But the choices themselves are made by machine; the store, as it were, is self-service. Beats Music, to extend the metaphor, puts humans toward the end of the production line, a lot closer to the customer. As the product floods in, celebrity and music specialist playlists screen the possibilities for selection, well ahead of the analysis of the user's likes and dislikes. They are there in the store with you, helping you choose. At least, that's the illusion. The psychological presence of music experts is one of the central elements of the Beats Music app experience, every bit as much as the Genius Bar is central to the Apple Store experience. We have the sense that this team of knowledgeable people making recommendations wants us to have a good experience. I have written before about the need for leadership in the selection process in the streaming sector. If they can appear to provide that effectively, Beats Music only needs to be otherwise technically as good as Spotify or Rdio to create a more positive user experience, giving the service a huge advantage.
BugsThe launch showed the app to be a little buggy, asking users to reenter their initial likes and dislikes over and over again. Sometimes the app locked up completely for a period, stuck in a processing mode. However an update that I downloaded shortly after my first bad experience seems to have fixed those problems. Users trying to sign up for the service also met with a long processing delays, as Beats was apparently unable to keep up with immediate registration demand. Any other problem typically encountered during a registration ("that username is already taken") was aggravated by the lack of response from the overloaded system.
Beats was upfront about the problems users were having in messages on Twitter.
We've been humbled by the outpouring of excitement for Beats Music. We know there's a delay in bringing new users onto the service Beats Music (@beatsmusic) January 22, 2014
The team is working around the clock to resolve this as quickly as possible, while maintaining service for those already using the service Beats Music (@beatsmusic) January 22, 2014Still, I must say my very first encounter with the service and the app -- bugs and all -- was quite positive. The environment is attractive, functional and easy to navigate. I thoroughly enjoyed being recommended such arcane listening choices as Charles Ives and Musique Concrete. Yet it was just as easy to ignore those and play the entire Buena Vista Social Club album (another suggestion based on my initial choices) or follow David Bowie (recommended because I did a search for Kraftwerk).