Matt Kiser

I’ve been thinking a lot about the rumors that Apple is going to create a Pandora-like streaming radio experience. I don’t think that’s the whole story. It obviously can’t be.

Apple left an umbrella for streaming services, which explains why the Spotify’s, Pandora’s, iHeartRadio’s of the world have been able to creep in on Apple’s turf — there’s a gap. We all see it. Apple’s also maintained a pretty hardline that ownership (.99-cent download) is more important to them than access (subscription plans). It’s important to the music industry, too, as something like 80% of all music sold is sold by iTunes. You can’t just replace download dollars (ok, .99-cents) for streaming tenths of a penny and expect everyone to be ok with it.

So, it makes a lot sense for Apple to narrow this gap and strategically defend its position as the top seller of music, but also rejuvenate their huge install base of 400 million+ installs by creating a hybrid model. I see the standard iTunes Store experience (ownership) more clearly and seamlessly paired-up with a beefier iTunes Match (access to owned content anywhere) — essentially doubling-down on the iTunes Match promise that you can “access your music from all your devices and listen to your entire library, wherever you are.” Coupled with how much iTunes Genius has learned over the years (i.e. listening habits, including time of day, type of music, day of the week, type of device, etc.) and iCloud, you can see a clear convergence happening.

Additionally, we should expect to see a programmed, streaming radio model that uses your library as a guide, but relies on the millions of iTunes Genius data points of user listening history to dictate the music mix (i.e. Genius Mixes on crack). I’ll contend that iTunes Genius is more similar to Songza, which is why I don’t think Pandora is the correct analog. Songza makes more sense from their themed, bite-sized playlists that are easy to consume and navigate. The obvious assumption is that Apple would just create a streaming radio experience and slap their lagging iAds product on it and call it a day. That’s probably going to happen to an extent, but that can’t be the full story either.

I have a hunch that Apple will use and reward users for having a vast library they own (i.e. relying heavily on the music that’s been previously paid for) and then fill in the blanks in the streaming radio programming from the 28 million+ songs in the iTunes library. I believe that the rewards for having a large library of owned music will be realized through less restrictions in the streaming radio space (i.e. no caps on skips or number of artist plays per hour, or amount of “free” listening per month per user, etc.). It’s important to remember that iTunes Match is like renting your own music. Every time a consumer re-downloads or faux-streams their own music, the copyright holder is getting paid. I can’t underscore that enough: for $25 a year you’re putting money into the music industry that was previously lost in exchange for access to your own music. Jeff Price, president of TuneCore, sheds some light on the economics by saying, “Apple keeps 30% of iTunes Match revenues for itself — the same percentage the company keeps from the iTunes and App Stores. The remaining 70% is divided, with 88% going to record labels and 12% going to songwriters. The royalties are split amongst artists based on ‘how many times someone accesses your song’ via iTunes Match and it doesn’t matter if a song is matched or uploaded — the royalty is paid either way.”

Think about it: you’re paying to access music either way. It’s either a subsidized version of your music library paired with the iTunes library of 28 million songs ($25/year + whatever you buy + ad revenue), or it’s Pandora’s 800,000 songs ($36/year -or- just ad revenue) or Spotify’s 15 million songs ($120/year -or- just ad revenue) you pay to access. But in either scenario, the music industry is going to make more money from Apple. And, by Apple keeping people in their ecosystem longer and rewarding them for owning music in the form of subsidized streaming radio, they’re able to sell more music and put even more money in the pockets of the music industry — something none of the other streaming competitors can provide sans potentially Amazon and Google, which we haven’t even talked about. In a nutshell, the pitch is that for something like $25 a year, you can get all your music anywhere at anytime on any device — oh, and one more thing — you can also listen to exceptionally awesome iTunes Genius-powered streaming radio, too. I’d buy.

We like to talk about “social music” as it pertains to music shared with our friends across social networks. Spotify and Rdio’s integration into Facebook being the obvious example of social music. But, that’s not social music. Nor is Facebook’s “Listen with” feature. Even Turntable.fm for all its interestingness is close, but not an actual social experience. 

This is social music:

Imagine a venn diagram. In one circle is music I like. In the other circle is music a friend likes. The overlap is the sweet spot where music crosses over from being an individual experience, to music as a social one. In the vastness of the canon of recorded music, this is the spot at which the two of us intersect — this is where our musical compatibility and relationship begins. That’s powerful knowledge. Now take that concept further and extend it to music recommendations: music you like that your friend might also like. 

Think about it for a second: we don’t actually know what our friends will like. Do you really have a good answer to “what should I listen to?” for every single one of your friends? I’m an outlier and spend an extraordinary amount of time listening to music, and as good as I am at recommending it, my success rate is pitifully low. A “normal” person would have an even lower success rate. 

The ability to make recommendations “smart” and personalized on a person-to-person level is what I love about working at the intersection of music, media, and technology. 

If we could take this concept of social music and apply it to entire friend circles, the results would be incredibly interesting since it would highlight the natural affinity groups centered around particular artists and genres. Further, it would reveal the sub-sub-groups we form when we listen to music communally. Which, in reality, is what being social is about.