Monday, July 06, 2015

Apple Music, Operating System Integration, and Antitrust

by Michael Dorf

The breathless arrival of Apple Music last week occasions this set of reflections on music and technology, with a few comments about antitrust law thrown in. To set things up, I recount my own experience with recorded music technology.

I was born at the tail end of the Baby Boom and so I grew up listening to music on vinyl and on the radio, occasionally supplemented by cassette tapes. I had an 8-track for a brief period but I abandoned it as quickly as everyone else did. Because I'm not a hipster, I happily made the transition to digital. I hung onto my vinyl records through the 1980s as I built my CD collection, and I was an early adopter of mp3 technology, first with a Rio and then with a series of iPods and lately, smartphones. When streaming became a thing, I signed up for Pandora and Rhapsody accounts, which I continue to use extensively. I pay the monthly subscription fees to avoid ads.

I've never been especially committed to any one operating system or platform. For many years I used Wintel desktop machines, first running DOS (yes, I'm old) and later Windows. Even as I recognized the somewhat greater stability of the Mac OS, I stuck with Windows because for a long time businesses (including universities) didn't support Macs. I am now in the process of transitioning all of my devices to Apple, although I still have an android phone that I'll replace with a next-generation iPhone in the fall and one legacy computer running Windows that I share with some family members. I'm switching to all-Apple because of the modest network benefits of a single platform but I'm not an Apple fanboy. I think their products have some advantages and some disadvantages relative to the various competitors in various markets. On the whole, for the uses I mostly have for my various machines, I've found Apple hardware somewhat better.

Now to Apple Music. As is well known, Steve Jobs opposed offering a streaming music product, so Apple starts late, with established streaming services like Spotify, Rdio, Rhapsody, and Pandora already holding lots of users. But that's not at all unusual for Apple, which has on past occasions brought out products that do a better job of doing something that was already being done by others and pretty quickly become the market leader. The iPod and iPhone are good examples.

At launch, it's not obvious that the mix of pros and cons of Apple Music relative to its leading competitors (described, e.g., here and here) makes Apple a clear winner--especially when you factor in the time many users of the rival services have sunk into building playlists, etc. on those services. But the comparative analyses omit or at best downplay what may prove to be Apple Music's best feature, at least within the Apple ecosystem: operating system integration.

Remember back in the 1990s when the federal government brought antitrust charges against Microsoft? One of its claims was that Microsoft was bundling its browser, Internet Explorer (IE), with its operating system, Windows, in order to help IE dislodge the then-dominant browser, Netscape. The charge wasn't that Microsoft was trying to monopolize the browser market; since Microsoft gave away IE for free, it didn't make any money in that market; rather, Microsoft feared that Netscape was developing programs that would run directly on its browser rather than on Windows, and thus would eventually challenge Microsoft's leading position in the operating system market. In other words, Microsoft feared that Netscape would do what Google would later do with its Chromebooks: turn a browser into an operating system. Microsoft was trying to kill Netscape to preserve its dominance in computer operating systems.

The government's theory of antitrust liability against Microsoft in the 90s wouldn't work against Apple for Apple Music because there is no real fear of Apple developing a monopoly in either the mobile OS market (where it trails Android in terms of market share) or in the streaming music service (where, as noted, it is a late entrant). If Apple had a dominant position in the European mobile phone market, then European competititon law would block it from taking unfair advantage of that position to gain advantage in the related market for music streaming. But even after winning some converts in the last year, Apple's iOS still only has about a 20% market share in Europe.

Still, while Apple is in the clear legally, it's worth thinking about operating system integration more broadly. When Microsoft bundled IE with Windows, it claimed that doing so improved the functionality of both. That claim was dubious at the time but today seems obviously right: Given how people use computers, tablets, and mobile phones, a modern operating system needs to be tightly integrated with access to the internet.

Likewise, operating system integration gives Apple Music a potential advantage over stand-alone streaming music apps. I tried a series of experiments to see whether the advantage is real.

On my iPad, I asked Siri to "play 1985 by Bowling for Soup." Within seconds, it opened Apple Music (which is now integrated into iTunes) and did exactly as requested. Next I said "open Rhapsody and play 1985 by Bowling for Soup." The iPad opened the Rhapsody App but didn't play anything, and it showed me two results, both covers of 1985 by other bands.

I wondered whether this was just a Rhapsody problem. I have been less than fully satisfied with Rhapsody for some time, but haven't switched to Spotify or one of the other streaming services because for most of my purposes Rhapsody was good enough. But maybe it isn't. Maybe other services have better access to Siri--which is a big deal if, say, you listen to music while driving or, like me, you have not-so-great-near-field vision without your reading glasses and large fingers that make voice commands very helpful.

I downloaded and installed Spotify on my iPad. I signed up for the 7-day free trial of premium (so that I could request particular songs rather than just shuffle).  Siri didn't do so well with Spotify either, botching the 1985 request as badly as Rhapsody did.

Then I wondered whether Google plays better with third-party streaming apps than Apple does, so I ran the test on my Android device (a Samsung Galaxy S4). It did just as badly as the iPad did with my Rhapsody request, but Google/Android did about as well with Spotify as Siri did with Apple Music.

I ran a series of other tests to see whether this was anomalous. For each of my five conditions--(1) Siri running Apple Music; (2) Siri running Rhapsody; (3) Siri running Spotify; (4) Google/Android running Rhapsody; and (5) Google/Android running Spotify, I next ordered: "play Blinded by the Light by Bruce Springsteen." I then said: "play Blinded by the Light by Manfred Mann" (the more popular but inferior cover). Combinations 1 and 5 (Apple Music on an Apple device and Spotify on an Android device) both worked extremely well. The others did badly.

Finally, I tried Pandora, which under-performed Apple Music on both the iPad and the Galaxy S4. The top level voice command (Siri and Google) were both able to open Pandora to the same station I had last been playing but neither could go to the different station I requested by voice. By contrast, when I asked Siri to "play Exile in Guyville radio" the iPad opened Apple Music to the custom radio station I had created based on the Liz Phair album of that name and began playing songs from it. Google plus Spotify didn't do quite as well on this test. Instead of playing the (parallel) radio station I had created in Spotify, it started playing the Exile in Guyville album.

None of the services did well with classical music. Siri perfectly transcribed my voice request to "play Herbert von Karajan's recording of Beethoven's Third Symphony" but she told me that she couldn't find it on Apple Music, even though in fact the Apple Music library does contain such a recording. I varied the request by asking Siri to "play the Berlin Philharmonic's recording of Beethoven's Third Symphony" but got the same defeatist answer. Things went even worse when I simply asked Siri to "play Beethoven's Third Symphony." Apple Music started to play a Royal Philharmonic Orchestra recording of the third movement of Beethoven's Seventh Symphony. No variation I tried worked. I had similarly dreadful luck with Beethoven via Google plus Spotify (and Rhapsody) on the Galaxy. This experiment leads me to think that Apple, Google and the third-party streaming app makers have invested fewer resources into figuring out how to retrieve classical music relative to other genres. For other genres, it takes some fooling around with phrasing but once you get the hang of it, Siri controls Apple Music quite well. For example, when I asked Siri to "play the Devo cover of I Can't Get No Satisfaction" she was flummoxed, but asking her to "play I Can't Get No Satisfaction by Devo" worked perfectly. I got just about the exact same results using Google and Spotify on the Galaxy.

My little series of experiments suggests to me a number of tentative conclusions:

(1) If I were sticking with an Android, I would switch from Rhapsody to Spotify, notwithstanding the sunk cost in playlists and downloads.

(2) If I'm using an iPad or iPhone, I would want to run Apple Music because it plays better with iOS than do the third-party apps.

(3) If someone were otherwise in equipoise as between Android and iOS, and putting aside pricing issues, Android would have the advantage--not because it does a better job of streaming music. Android plus Spotify is at best equal to, but not better than Apple Music on iOS. But the results of my little series of experiments are consistent with what appears to be a broader pattern. Google deliberately makes Android pull data from third-party apps, whereas Apple simply has its devices run them. A recent NPR story discussed how this pattern fits within the companies' respective approaches to privacy and their business models.  But putting those issues aside, it would seem that you get more with Google: The opportunity to choose from a world of third-party apps that will be able to take full advantage of the Android OS in a way that third-party apps on iOS can't quite match.

Of course, anybody who has been following the tech world for the last three decades-plus knows this is an old story. Apple has always claimed that by tightly integrating its products--including not just the OS and software but hardware too--it can deliver higher quality than an open ecosystem can deliver. The validity of that claim is a giant unresolved and hotly contested question.

For the moment, there's no legal problem with letting the market decide the matter according to consumer preferences. So long as Apple doesn't monopolize the mobile OS market, there's no serious worry that it will use its control of the OS to advantage its own products. Competition from Android will lead Apple to ensure enough access to popular third-party apps on iOS--and in fact things usually go in the other direction, with apps often written first for iOS and then for Android.

In a hypothetical world in which Apple attains market dominance of mobile operating systems, systematically advantaging its own apps over third-party apps still wouldn't violate US antitrust law, so long as doing so wasn't undertaken to maintain its OS monopoly (as per the 1990s charges against Microsoft). But in that hypothetical world, the "abuse" of the monopoly would violate European competition law and it strikes me that, in this respect, European law makes more sense than US law. Fortunately, we need not worry about that problem in this context because, as noted, Apple faces serious competition in the OS market. In this respect, those of us who opt into the Apple ecosystem free ride on the existence of Android to keep Apple (somewhat) honest.

Thus concludes the report of my music experiment. As should be evident, I have taken advantage of the end of the SCOTUS term to indulge my instincts as an amateur scientist.

No comments: